4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.2](#openssl-32)
14 - [OpenSSL 3.1](#openssl-31)
15 - [OpenSSL 3.0](#openssl-30)
16 - [OpenSSL 1.1.1](#openssl-111)
17 - [OpenSSL 1.1.0](#openssl-110)
18 - [OpenSSL 1.0.2](#openssl-102)
19 - [OpenSSL 1.0.1](#openssl-101)
20 - [OpenSSL 1.0.0](#openssl-100)
21 - [OpenSSL 0.9.x](#openssl-09x)
26 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
28 * Added client side support for QUIC
32 * Added SHA256/192 algorithm support.
36 * Provide a new configure option `no-http` that can be used to disable HTTP
41 * TLS round-trip time calculation was added by a Brigham Young University
42 Capstone team partnering with Sandia National Laboratories. A new function
43 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
48 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
49 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
50 option. Use of the "advanced" s_client command command via the "-adv" option
55 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
56 The previous fix for this timing side channel turned out to cause
57 a severe 2-3x performance regression in the typical use case
58 compared to 3.0.7. The new fix uses existing constant time
59 code paths, and restores the previous performance level while
60 fully eliminating all existing timing side channels.
61 The fix was developed by Bernd Edlinger with testing support
66 * Added an "advanced" command mode to s_client. Use this with the "-adv"
67 option. The old "basic" command mode recognises certain letters that must
68 always appear at the start of a line and cannot be escaped. The advanced
69 command mode enables commands to be entered anywhere and there is an
70 escaping mechanism. After starting s_client with "-adv" type "{help}"
71 to show a list of available commands.
75 * Add Raw Public Key (RFC7250) support. Authentication is supported
76 by matching keys against either local policy (TLSA records synthesised
77 from the expected keys) or DANE (TLSA records obtained by the
78 application from DNS). TLSA records will also match the same key in
79 the server certificate, should RPK use not happen to be negotiated.
83 * Added EC_GROUP_to_params which creates an OSSL_PARAM array
84 from a given EC_GROUP.
88 * Added support for pluggable (provider-based) TLS signature algorithms.
89 This enables TLS 1.3 authentication operations with algorithms embedded
90 in providers not included by default in OpenSSL. In combination with
91 the already available pluggable KEM and X.509 support, this enables
92 for example suitable providers to deliver post-quantum or quantum-safe
93 cryptography to OpenSSL users.
97 * Added support for Hybrid Public Key Encryption (HPKE) as defined
98 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
99 Message Layer Security (MLS) and other IETF specifications.
100 HPKE can also be used by other applications that require
101 encrypting "to" an ECDH public key. External APIs are defined in
102 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
106 * Add support for certificate compression (RFC8879), including
107 library support for Brotli and Zstandard compression.
111 * Add the ability to add custom attributes to PKCS12 files. Add a new API
112 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
113 for a user specified callback and optional argument.
114 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
115 added to the existing STACK_OF attrs.
119 * Major refactor of the libssl record layer
123 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
127 * Add a mac salt length option for the pkcs12 command.
131 * Add more SRTP protection profiles from RFC8723 and RFC8269.
135 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
137 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
139 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
140 supported and enabled.
144 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
145 to the list of ciphersuites providing Perfect Forward Secrecy as
146 required by SECLEVEL >= 3.
148 *Dmitry Belyavskiy, Nicola Tuveri*
150 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
151 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
152 SSL_get0_iana_groups() function-like macro, retrieves the list of
153 supported groups sent by the peer.
154 The function SSL_client_hello_get_extension_order() populates
155 a caller-supplied array with the list of extension types present in the
156 ClientHello, in order of appearance.
160 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
161 to make it possible to use empty passphrase strings.
165 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
166 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
167 of 160 bits and above and less than 224 bits were previously accepted by
168 default but are now no longer allowed. By default TLS compression was
169 already disabled in previous OpenSSL versions. At security level 2 it cannot
174 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
179 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
180 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
181 will need to load the legacy crypto provider.
184 * CCM8 cipher suites in TLS have been downgraded to security level zero
185 because they use a short authentication tag which lowers their strength.
189 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
194 * Add X.509 certificate codeSigning purpose and related checks on key usage and
195 extended key usage of the leaf certificate according to the CA/Browser Forum.
199 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
200 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
201 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
202 X.509 version 3 if the certificate information includes X.509 extensions.
206 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
207 such as adding a trace facility for debugging certificate chain building.
211 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
212 in particular supporting requests for central key generation, generalized
213 polling, and various types of genm/genp exchanges defined in CMP Updates.
217 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
218 like correcting the TLS and proxy support and adding tracing for debugging.
222 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
226 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
227 if a certificate to be added is already present.
228 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
229 in their `certs` argument and no longer throw an error for them.
233 * Fixed and extended `util/check-format.pl` for checking adherence to the
234 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
235 The checks are meanwhile more complete and yield fewer false positives.
239 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
240 sending and receiving multiple messages in a single call. An implementation
241 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
245 * Support for loading root certificates from the Windows certificate store
246 has been added. The support is in the form of a store which recognises the
247 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
248 arguments. This store is built by default and can be disabled using the new
249 compile-time option `no-winstore`. This store is not currently used by
250 default and must be loaded explicitly using the above store URI. It is
251 expected to be loaded by default in the future.
255 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
256 kernel versions that support KTLS have a known bug in CCM processing. That
257 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
258 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
263 * Zerocopy KTLS sendfile() support on Linux.
267 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
268 decryption as a protection against Bleichenbacher-like attacks.
269 The RSA decryption API will now return a randomly generated deterministic
270 message instead of an error in case it detects an error when checking
271 padding during PKCS#1 v1.5 decryption. This is a general protection against
272 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
274 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
275 on the RSA decryption context.
279 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
280 basic thread pool implementation for select platforms.
287 ### Changes between 3.1.1 and 3.1.2 [xx XXX xxxx]
289 * When building with the `enable-fips` option and using the resulting
290 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
291 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
292 not operate with truncated digests (FIPS 140-3 IG G.R).
296 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
298 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
299 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
301 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
302 numeric text form. For gigantic sub-identifiers, this would take a very
303 long time, the time complexity being O(n^2) where n is the size of that
304 sub-identifier. ([CVE-2023-2650])
306 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
307 IDENTIFIER to canonical numeric text form if the size of that OBJECT
308 IDENTIFIER is 586 bytes or less, and fail otherwise.
310 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
311 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
312 most 128 sub-identifiers, and that the maximum value that each sub-
313 identifier may have is 2^32-1 (4294967295 decimal).
315 For each byte of every sub-identifier, only the 7 lower bits are part of
316 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
317 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
322 * Multiple algorithm implementation fixes for ARM BE platforms.
326 * Added a -pedantic option to fipsinstall that adjusts the various
327 settings to ensure strict FIPS compliance rather than backwards
332 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
333 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
334 trigger a crash of an application using AES-XTS decryption if the memory
335 just after the buffer being decrypted is not mapped.
336 Thanks to Anton Romanov (Amazon) for discovering the issue.
341 * Add FIPS provider configuration option to disallow the use of
342 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
343 The option '-no_drbg_truncated_digests' can optionally be
344 supplied to 'openssl fipsinstall'.
348 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
349 that it does not enable policy checking. Thanks to David Benjamin for
350 discovering this issue.
355 * Fixed an issue where invalid certificate policies in leaf certificates are
356 silently ignored by OpenSSL and other certificate policy checks are skipped
357 for that certificate. A malicious CA could use this to deliberately assert
358 invalid certificate policies in order to circumvent policy checking on the
359 certificate altogether.
364 * Limited the number of nodes created in a policy tree to mitigate
365 against CVE-2023-0464. The default limit is set to 1000 nodes, which
366 should be sufficient for most installations. If required, the limit
367 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
368 time define to a desired maximum number of nodes or zero to allow
374 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
376 * Add FIPS provider configuration option to enforce the
377 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
378 The option '-ems-check' can optionally be supplied to
379 'openssl fipsinstall'.
383 * The FIPS provider includes a few non-approved algorithms for
384 backward compatibility purposes and the "fips=yes" property query
385 must be used for all algorithm fetches to ensure FIPS compliance.
387 The algorithms that are included but not approved are Triple DES ECB,
388 Triple DES CBC and EdDSA.
392 * Added support for KMAC in KBKDF.
396 * RNDR and RNDRRS support in provider functions to provide
397 random number generation for Arm CPUs (aarch64).
401 * s_client and s_server apps now explicitly say when the TLS version
402 does not include the renegotiation mechanism. This avoids confusion
403 between that scenario versus when the TLS version includes secure
404 renegotiation but the peer lacks support for it.
408 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
410 *Tomasz Kantecki, Andrey Matyukov*
412 * The various OBJ_* functions have been made thread safe.
416 * Parallel dual-prime 1536/2048-bit modular exponentiation for
417 AVX512_IFMA capable processors.
419 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
421 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
422 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
423 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
424 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
425 `OPENSSL_NO_DEPRECATED_3_1`.
427 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
428 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
429 definitions for these functions regardless of whether
430 `OPENSSL_NO_DEPRECATED_3_1` is defined.
432 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
433 functions regardless of whether they are using them. It is recommended that
434 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
438 * When generating safe-prime DH parameters set the recommended private key
439 length equivalent to minimum key lengths as in RFC 7919.
443 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
444 maximum size that is smaller or equal to the digest length to comply with
445 FIPS 186-4 section 5. This is implemented by a new option
446 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
447 `rsa_pss_saltlen` parameter, which is now the default. Signature
448 verification is not affected by this change and continues to work as before.
455 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
456 listed here are only a brief description.
457 The migration guide contains more detailed information related to new features,
458 breaking changes, and mappings for the large list of deprecated functions.
460 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
462 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
464 * Fixed NULL dereference during PKCS7 data verification.
466 A NULL pointer can be dereferenced when signatures are being
467 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
468 algorithm used for the signature is known to the OpenSSL library but
469 the implementation of the hash algorithm is not available the digest
470 initialization will fail. There is a missing check for the return
471 value from the initialization function which later leads to invalid
472 usage of the digest API most likely leading to a crash.
475 PKCS7 data is processed by the SMIME library calls and also by the
476 time stamp (TS) library calls. The TLS implementation in OpenSSL does
477 not call these functions however third party applications would be
478 affected if they call these functions to verify signatures on untrusted
483 * Fixed X.400 address type confusion in X.509 GeneralName.
485 There is a type confusion vulnerability relating to X.400 address processing
486 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
487 but the public structure definition for GENERAL_NAME incorrectly specified
488 the type of the x400Address field as ASN1_TYPE. This field is subsequently
489 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
492 When CRL checking is enabled (i.e. the application sets the
493 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
494 pass arbitrary pointers to a memcmp call, enabling them to read memory
495 contents or enact a denial of service.
500 * Fixed NULL dereference validating DSA public key.
502 An invalid pointer dereference on read can be triggered when an
503 application tries to check a malformed DSA public key by the
504 EVP_PKEY_public_check() function. This will most likely lead
505 to an application crash. This function can be called on public
506 keys supplied from untrusted sources which could allow an attacker
507 to cause a denial of service attack.
509 The TLS implementation in OpenSSL does not call this function
510 but applications might call the function if there are additional
511 security requirements imposed by standards such as FIPS 140-3.
514 *Shane Lontis, Tomáš Mráz*
516 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
518 An invalid pointer dereference on read can be triggered when an
519 application tries to load malformed PKCS7 data with the
520 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
522 The result of the dereference is an application crash which could
523 lead to a denial of service attack. The TLS implementation in OpenSSL
524 does not call this function however third party applications might
525 call these functions on untrusted data.
530 * Fixed Use-after-free following BIO_new_NDEF.
532 The public API function BIO_new_NDEF is a helper function used for
533 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
534 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
535 be called directly by end user applications.
537 The function receives a BIO from the caller, prepends a new BIO_f_asn1
538 filter BIO onto the front of it to form a BIO chain, and then returns
539 the new head of the BIO chain to the caller. Under certain conditions,
540 for example if a CMS recipient public key is invalid, the new filter BIO
541 is freed and the function returns a NULL result indicating a failure.
542 However, in this case, the BIO chain is not properly cleaned up and the
543 BIO passed by the caller still retains internal pointers to the previously
544 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
545 then a use-after-free will occur. This will most likely result in a crash.
548 *Viktor Dukhovni, Matt Caswell*
550 * Fixed Double free after calling PEM_read_bio_ex.
552 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
553 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
554 data. If the function succeeds then the "name_out", "header" and "data"
555 arguments are populated with pointers to buffers containing the relevant
556 decoded data. The caller is responsible for freeing those buffers. It is
557 possible to construct a PEM file that results in 0 bytes of payload data.
558 In this case PEM_read_bio_ex() will return a failure code but will populate
559 the header argument with a pointer to a buffer that has already been freed.
560 If the caller also frees this buffer then a double free will occur. This
561 will most likely lead to a crash.
563 The functions PEM_read_bio() and PEM_read() are simple wrappers around
564 PEM_read_bio_ex() and therefore these functions are also directly affected.
566 These functions are also called indirectly by a number of other OpenSSL
567 functions including PEM_X509_INFO_read_bio_ex() and
568 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
569 internal uses of these functions are not vulnerable because the caller does
570 not free the header argument if PEM_read_bio_ex() returns a failure code.
573 *Kurt Roeckx, Matt Caswell*
575 * Fixed Timing Oracle in RSA Decryption.
577 A timing based side channel exists in the OpenSSL RSA Decryption
578 implementation which could be sufficient to recover a plaintext across
579 a network in a Bleichenbacher style attack. To achieve a successful
580 decryption an attacker would have to be able to send a very large number
581 of trial messages for decryption. The vulnerability affects all RSA padding
582 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
585 *Dmitry Belyavsky, Hubert Kario*
587 * Fixed X.509 Name Constraints Read Buffer Overflow.
589 A read buffer overrun can be triggered in X.509 certificate verification,
590 specifically in name constraint checking. The read buffer overrun might
591 result in a crash which could lead to a denial of service attack.
592 In a TLS client, this can be triggered by connecting to a malicious
593 server. In a TLS server, this can be triggered if the server requests
594 client authentication and a malicious client connects.
599 * Fixed X.509 Policy Constraints Double Locking security issue.
601 If an X.509 certificate contains a malformed policy constraint and
602 policy processing is enabled, then a write lock will be taken twice
603 recursively. On some operating systems (most widely: Windows) this
604 results in a denial of service when the affected process hangs. Policy
605 processing being enabled on a publicly facing server is not considered
606 to be a common setup.
611 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
612 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
613 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
614 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
615 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
616 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
617 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
618 for legacy EC and SM2 keys is also changed similarly to honor the
619 equivalent conversion format flag as specified in the underlying
620 `EC_KEY` object being exported to a provider, when this function is
621 called through `EVP_PKEY_export()`.
625 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
627 * Fixed two buffer overflows in punycode decoding functions.
629 A buffer overrun can be triggered in X.509 certificate verification,
630 specifically in name constraint checking. Note that this occurs after
631 certificate chain signature verification and requires either a CA to
632 have signed the malicious certificate or for the application to continue
633 certificate verification despite failure to construct a path to a trusted
636 In a TLS client, this can be triggered by connecting to a malicious
637 server. In a TLS server, this can be triggered if the server requests
638 client authentication and a malicious client connects.
640 An attacker can craft a malicious email address to overflow
641 an arbitrary number of bytes containing the `.` character (decimal 46)
642 on the stack. This buffer overflow could result in a crash (causing a
646 An attacker can craft a malicious email address to overflow four
647 attacker-controlled bytes on the stack. This buffer overflow could
648 result in a crash (causing a denial of service) or potentially remote code
649 execution depending on stack layout for any given platform/compiler.
654 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
655 parameters in OpenSSL code.
656 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
657 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
658 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
659 Using these invalid names may cause algorithms to use slower methods
660 that ignore the CRT parameters.
664 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
669 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
670 data to be signed before signing the certificate.
674 * Added RIPEMD160 to the default provider.
678 * Ensured that the key share group sent or accepted for the key exchange
679 is allowed for the protocol version.
683 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
685 * OpenSSL supports creating a custom cipher via the legacy
686 EVP_CIPHER_meth_new() function and associated function calls. This function
687 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
688 to use the new provider mechanism in order to implement custom ciphers.
690 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
691 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
692 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
693 and decryption initialisation functions). Instead of using the custom cipher
694 directly it incorrectly tries to fetch an equivalent cipher from the
695 available providers. An equivalent cipher is found based on the NID passed to
696 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
697 given cipher. However it is possible for an application to incorrectly pass
698 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
699 is used in this way the OpenSSL encryption/decryption initialisation function
700 will match the NULL cipher as being equivalent and will fetch this from the
701 available providers. This will succeed if the default provider has been
702 loaded (or if a third party provider has been loaded that offers this
703 cipher). Using the NULL cipher means that the plaintext is emitted as the
706 Applications are only affected by this issue if they call
707 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
708 encryption/decryption initialisation function. Applications that only use
709 SSL/TLS are not impacted by this issue.
714 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
719 * Fixed the linux-mips64 Configure target which was missing the
720 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
725 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
730 * Correctly handle a retransmitted ClientHello in DTLS
734 * Fixed detection of ktls support in cross-compile environment on Linux
738 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
743 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
744 report correct results in some cases
748 * Fix UWP builds by defining VirtualLock
752 * For known safe primes use the minimum key length according to RFC 7919.
753 Longer private key sizes unnecessarily raise the cycles needed to compute the
754 shared secret without any increase of the real security. This fixes a
755 regression from 1.1.1 where these shorter keys were generated for the known
760 * Added the loongarch64 target
764 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
765 only passed to the FIPS provider and not to the default or legacy provider.
769 * Fixed reported performance degradation on aarch64. Restored the
770 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
771 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
772 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
773 The new algorithm is still used for 32 bit targets.
777 * Added a missing header for memcmp that caused compilation failure on some
782 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
784 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
785 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
786 This issue makes the RSA implementation with 2048 bit private keys
787 incorrect on such machines and memory corruption will happen during
788 the computation. As a consequence of the memory corruption an attacker
789 may be able to trigger a remote code execution on the machine performing
792 SSL/TLS servers or other servers using 2048 bit RSA private keys running
793 on machines supporting AVX512IFMA instructions of the X86_64 architecture
794 are affected by this issue.
799 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
800 implementation would not encrypt the entirety of the data under some
801 circumstances. This could reveal sixteen bytes of data that was
802 preexisting in the memory that wasn't written. In the special case of
803 "in place" encryption, sixteen bytes of the plaintext would be revealed.
805 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
806 they are both unaffected.
809 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
811 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
813 * In addition to the c_rehash shell command injection identified in
814 CVE-2022-1292, further bugs where the c_rehash script does not
815 properly sanitise shell metacharacters to prevent command injection have been
818 When the CVE-2022-1292 was fixed it was not discovered that there
819 are other places in the script where the file names of certificates
820 being hashed were possibly passed to a command executed through the shell.
822 This script is distributed by some operating systems in a manner where
823 it is automatically executed. On such operating systems, an attacker
824 could execute arbitrary commands with the privileges of the script.
826 Use of the c_rehash script is considered obsolete and should be replaced
827 by the OpenSSL rehash command line tool.
830 *Daniel Fiala, Tomáš Mráz*
832 * Case insensitive string comparison no longer uses locales. It has instead
833 been directly implemented.
837 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
839 * Case insensitive string comparison is reimplemented via new locale-agnostic
840 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
841 comparison. The previous implementation had problems when the Turkish locale
846 * Fixed a bug in the c_rehash script which was not properly sanitising shell
847 metacharacters to prevent command injection. This script is distributed by
848 some operating systems in a manner where it is automatically executed. On
849 such operating systems, an attacker could execute arbitrary commands with the
850 privileges of the script.
852 Use of the c_rehash script is considered obsolete and should be replaced
853 by the OpenSSL rehash command line tool.
858 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
859 certificate on an OCSP response. The bug caused the function in the case
860 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
861 response (meaning a successful verification) even in the case where the
862 response signing certificate fails to verify.
864 It is anticipated that most users of `OCSP_basic_verify` will not use the
865 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
866 a negative value (indicating a fatal error) in the case of a certificate
867 verification failure. The normal expected return value in this case would be
870 This issue also impacts the command line OpenSSL "ocsp" application. When
871 verifying an ocsp response with the "-no_cert_checks" option the command line
872 application will report that the verification is successful even though it
873 has in fact failed. In this case the incorrect successful response will also
874 be accompanied by error messages showing the failure and contradicting the
875 apparently successful result.
880 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
881 AAD data as the MAC key. This made the MAC key trivially predictable.
883 An attacker could exploit this issue by performing a man-in-the-middle attack
884 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
885 that the modified data would still pass the MAC integrity check.
887 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
888 endpoint will always be rejected by the recipient and the connection will
889 fail at that point. Many application protocols require data to be sent from
890 the client to the server first. Therefore, in such a case, only an OpenSSL
891 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
893 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
894 sent in both directions. In this case both clients and servers could be
895 affected, regardless of the application protocol.
897 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
898 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
899 the handshake when using this ciphersuite.
901 The confidentiality of data is not impacted by this issue, i.e. an attacker
902 cannot decrypt data that has been encrypted using this ciphersuite - they can
905 In order for this attack to work both endpoints must legitimately negotiate
906 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
907 OpenSSL 3.0, and is not available within the default provider or the default
908 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
909 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
910 following must have occurred:
912 1) OpenSSL must have been compiled with the (non-default) compile time option
913 enable-weak-ssl-ciphers
915 2) OpenSSL must have had the legacy provider explicitly loaded (either
916 through application code or via configuration)
918 3) The ciphersuite must have been explicitly added to the ciphersuite list
920 4) The libssl security level must have been set to 0 (default is 1)
922 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
924 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
925 others that both endpoints have in common
930 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
931 occupied by the removed hash table entries.
933 This function is used when decoding certificates or keys. If a long lived
934 process periodically decodes certificates or keys its memory usage will
935 expand without bounds and the process might be terminated by the operating
936 system causing a denial of service. Also traversing the empty hash table
937 entries will take increasingly more time.
939 Typically such long lived processes might be TLS clients or TLS servers
940 configured to accept client certificate authentication.
943 *Hugo Landau, Aliaksei Levin*
945 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
946 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
947 statistics are no longer supported. For compatibility, these statistics are
948 still listed in the output but are now always reported as zero.
952 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
954 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
955 for non-prime moduli.
957 Internally this function is used when parsing certificates that contain
958 elliptic curve public keys in compressed form or explicit elliptic curve
959 parameters with a base point encoded in compressed form.
961 It is possible to trigger the infinite loop by crafting a certificate that
962 has invalid explicit curve parameters.
964 Since certificate parsing happens prior to verification of the certificate
965 signature, any process that parses an externally supplied certificate may thus
966 be subject to a denial of service attack. The infinite loop can also be
967 reached when parsing crafted private keys as they can contain explicit
968 elliptic curve parameters.
970 Thus vulnerable situations include:
972 - TLS clients consuming server certificates
973 - TLS servers consuming client certificates
974 - Hosting providers taking certificates or private keys from customers
975 - Certificate authorities parsing certification requests from subscribers
976 - Anything else which parses ASN.1 elliptic curve parameters
978 Also any other applications that use the BN_mod_sqrt() where the attacker
979 can control the parameter values are vulnerable to this DoS issue.
984 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
985 to the list of ciphersuites providing Perfect Forward Secrecy as
986 required by SECLEVEL >= 3.
988 *Dmitry Belyavskiy, Nicola Tuveri*
990 * Made the AES constant time code for no-asm configurations
991 optional due to the resulting 95% performance degradation.
992 The AES constant time code can be enabled, for no assembly
993 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
997 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1002 * The negative return value handling of the certificate verification callback
1003 was reverted. The replacement is to set the verification retry state with
1004 the SSL_set_retry_verify() function.
1008 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1010 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1011 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1012 verify a certificate supplied by a server. That function may return a
1013 negative return value to indicate an internal error (for example out of
1014 memory). Such a negative return value is mishandled by OpenSSL and will cause
1015 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1016 success and a subsequent call to SSL_get_error() to return the value
1017 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1018 returned by OpenSSL if the application has previously called
1019 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1020 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1021 totally unexpected and applications may not behave correctly as a result. The
1022 exact behaviour will depend on the application but it could result in
1023 crashes, infinite loops or other similar incorrect responses.
1025 This issue is made more serious in combination with a separate bug in OpenSSL
1026 3.0 that will cause X509_verify_cert() to indicate an internal error when
1027 processing a certificate chain. This will occur where a certificate does not
1028 include the Subject Alternative Name extension but where a Certificate
1029 Authority has enforced name constraints. This issue can occur even with valid
1035 * Corrected a few file name and file reference bugs in the build,
1036 installation and setup scripts, which lead to installation verification
1037 failures. Slightly enhanced the installation verification script.
1041 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1046 * Fixed PVK encoder to properly query for the passphrase.
1050 * Multiple fixes in the OSSL_HTTP API functions.
1054 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1055 OSSL_PARAM_INTEGER data type and return error on negative numbers
1056 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1057 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1061 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1065 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1069 * Multiple threading fixes.
1073 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1077 * Allow fetching an operation from the provider that owns an unexportable key
1078 as a fallback if that is still allowed by the property query.
1082 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1084 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1089 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1090 S390X capability vector to zero. This simplifies testing of different code
1091 paths on S390X architecture.
1095 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1096 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1097 SP 800-38D". The communication will fail at this point.
1101 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1102 confidential in EC_GROUP data.
1106 * The byte order mark (BOM) character is ignored if encountered at the
1107 beginning of a PEM-formatted file.
1111 * Added CMS support for the Russian GOST algorithms.
1115 * Due to move of the implementation of cryptographic operations
1116 to the providers, validation of various operation parameters can
1117 be postponed until the actual operation is executed where previously
1118 it happened immediately when an operation parameter was set.
1120 For example when setting an unsupported curve with
1121 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1122 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1124 *OpenSSL team members and many third party contributors*
1126 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1127 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1128 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1129 instead to retrieve these algorithms from a provider.
1133 * On build targets where the multilib postfix is set in the build
1134 configuration the libdir directory was changing based on whether
1135 the lib directory with the multilib postfix exists on the system
1136 or not. This unpredictable behavior was removed and eventual
1137 multilib postfix is now always added to the default libdir. Use
1138 `--libdir=lib` to override the libdir if adding the postfix is
1143 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1144 no longer interoperable with OpenSSL 1.1.1.
1148 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1149 function codes, this function can only cause problems for calling
1154 * Add a configurable flag to output date formats as ISO 8601. Does not
1155 change the default date format.
1159 * Version of MSVC earlier than 1300 could get link warnings, which could
1160 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1161 Support for this flag has been removed.
1165 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1166 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1167 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1168 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1169 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1173 * The signatures of the functions to get and set options on SSL and
1174 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1175 Some source code changes may be required.
1179 * The public definitions of conf_method_st and conf_st have been
1180 deprecated. They will be made opaque in a future release.
1182 *Rich Salz and Tomáš Mráz*
1184 * Client-initiated renegotiation is disabled by default. To allow it, use
1185 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1186 flag, or the "ClientRenegotiation" config parameter as appropriate.
1190 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1191 or modify relative pathname inclusion.
1195 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1196 validated. Please consult the README-FIPS and
1197 README-PROVIDERS files, as well as the migration guide.
1199 *OpenSSL team members and many third party contributors*
1201 * For the key types DH and DHX the allowed settable parameters are now different.
1205 * The openssl commands that read keys, certificates, and CRLs now
1206 automatically detect the PEM or DER format of the input files.
1208 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1210 * Added enhanced PKCS#12 APIs which accept a library context.
1214 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1218 * Added support for Kernel TLS (KTLS).
1220 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1222 * Support for RFC 5746 secure renegotiation is now required by default for
1223 SSL or TLS connections to succeed.
1227 * The signature of the `copy` functional parameter of the
1228 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1229 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1230 the signature of the `pub_decode` functional parameter of the
1231 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1232 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1236 * The error return values from some control calls (ctrl) have changed.
1240 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1244 * Many functions in the EVP_ namespace that are getters of values from
1245 implementations or contexts were renamed to include get or get0 in their
1246 names. Old names are provided as macro aliases for compatibility and
1251 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1252 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1253 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1258 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1261 * The output from the command line applications may have minor
1266 * The output from numerous "printing" may have minor changes.
1270 * Windows thread synchronization uses read/write primitives (SRWLock) when
1271 supported by the OS, otherwise CriticalSection continues to be used.
1275 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1276 work on read only BIO source/sinks that do not support these functions.
1277 This allows piping or redirection of a file BIO using stdin to be buffered
1278 into memory. This is used internally in OSSL_DECODER_from_bio().
1282 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1283 this function would return one of the values OSSL_STORE_INFO_NAME,
1284 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1285 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1286 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1287 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1288 using this function should be amended to handle the changed return value.
1292 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1293 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1294 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1295 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1296 contains more than one certificate identifier: This means that all
1297 certificates referenced there MUST be part of the validation chain.
1301 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1302 RC5, DESX and DES have been moved to the legacy provider.
1306 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1307 RIPEMD-160 have been moved to the legacy provider.
1311 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1316 * The deprecated functions EVP_PKEY_get0_RSA(),
1317 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1318 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1319 well as the similarly named "get1" functions behave differently in
1324 * A number of functions handling low-level keys or engines were deprecated
1325 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1326 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1327 EVP_PKEY_get0_siphash().
1331 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1332 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1333 will need to load the legacy crypto provider. This includes these PBE
1334 algorithms which use this KDF:
1335 - NID_pbeWithMD2AndDES_CBC
1336 - NID_pbeWithMD5AndDES_CBC
1337 - NID_pbeWithSHA1AndRC2_CBC
1338 - NID_pbeWithMD2AndRC2_CBC
1339 - NID_pbeWithMD5AndRC2_CBC
1340 - NID_pbeWithSHA1AndDES_CBC
1344 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1345 BIO_debug_callback() functions.
1349 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1350 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1354 * The RAND_METHOD APIs have been deprecated.
1358 * The SRP APIs have been deprecated.
1362 * Add a compile time option to prevent the caching of provider fetched
1363 algorithms. This is enabled by including the no-cached-fetch option
1364 at configuration time.
1368 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1369 count of PKCS12_DEFAULT_ITER.
1371 *Tomáš Mráz and Sahana Prasad*
1373 * The openssl speed command does not use low-level API calls anymore.
1377 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1380 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1382 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1386 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1387 providers may supply their own group implementations (using either the "key
1388 exchange" or the "key encapsulation" methods) which will automatically be
1389 detected and used by libssl.
1391 *Matt Caswell, Nicola Tuveri*
1393 * The undocumented function X509_certificate_type() has been deprecated;
1397 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1401 * Removed RSA padding mode for SSLv23 (which was only used for
1402 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1403 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1408 * Deprecated the obsolete X9.31 RSA key generation related functions.
1410 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1411 is not allowed to return a value > 1, this is no more taken as failure.
1413 *Viktor Dukhovni and David von Oheimb*
1415 * Deprecated the obsolete X9.31 RSA key generation related functions
1416 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1417 BN_X931_generate_prime_ex().
1421 * The default key generation method for the regular 2-prime RSA keys was
1422 changed to the FIPS 186-4 B.3.6 method.
1426 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1430 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1434 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1435 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1437 *Rich Salz, Richard Levitte, and David von Oheimb*
1439 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1443 * Deprecated `OCSP_parse_url()`.
1447 * Validation of SM2 keys has been separated from the validation of regular EC
1452 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1453 switches: a validation failure triggers an early exit, returning a failure
1454 exit status to the parent process.
1458 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1459 to ignore unknown ciphers.
1463 * The `-cipher-commands` and `-digest-commands` options
1464 of the command line utility `list` have been deprecated.
1465 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1469 * Added convenience functions for generating asymmetric key pairs:
1470 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1471 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1475 * All of the low-level EC_KEY functions have been deprecated.
1477 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1479 * Deprecated all the libcrypto and libssl error string loading
1484 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1485 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1490 * The `-crypt` option to the `passwd` command line tool has been removed.
1494 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1499 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1503 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1504 EVP_PKEY_get1_tls_encodedpoint().
1508 * The security callback, which can be customised by application code, supports
1509 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1510 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1514 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1515 interface. Their functionality remains unchanged.
1519 * Added new option for 'openssl list', '-providers', which will display the
1520 list of loaded providers, their names, version and status. It optionally
1521 displays their gettable parameters.
1525 * Removed EVP_PKEY_set_alias_type().
1529 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1530 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1534 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1535 parameters are still checked for type safety at compile time via helper
1540 * Remove the RAND_DRBG API
1542 *Paul Dale and Matthias St. Pierre*
1544 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1545 as well as actual hostnames.
1549 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1550 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1551 conversely, silently ignore DTLS protocol version bounds when configuring
1552 TLS-based contexts. The commands can be repeated to set bounds of both
1553 types. The same applies with the corresponding "min_protocol" and
1554 "max_protocol" command-line switches, in case some application uses both TLS
1557 SSL_CTX instances that are created for a fixed protocol version (e.g.
1558 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1559 attempts to apply bounds to these protocol versions would result in an
1560 error. Now only the "version-flexible" SSL_CTX instances are subject to
1561 limits in configuration files in command-line options.
1565 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1570 * Reworked the recorded ERR codes to make better space for system errors.
1571 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1572 given code is a system error (true) or an OpenSSL error (false).
1576 * Reworked the test perl framework to better allow parallel testing.
1578 *Nicola Tuveri and David von Oheimb*
1580 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1581 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1585 * 'Configure' has been changed to figure out the configuration target if
1586 none is given on the command line. Consequently, the 'config' script is
1587 now only a mere wrapper. All documentation is changed to only mention
1590 *Rich Salz and Richard Levitte*
1592 * Added a library context `OSSL_LIB_CTX` that applications as well as
1593 other libraries can use to form a separate context within which
1594 libcrypto operations are performed.
1598 * Added various `_ex` functions to the OpenSSL API that support using
1599 a non-default `OSSL_LIB_CTX`.
1603 * Handshake now fails if Extended Master Secret extension is dropped
1608 * Dropped interactive mode from the `openssl` program.
1612 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1614 *David von Oheimb and Shane Lontis*
1616 * Deprecated `EC_METHOD_get_field_type()`.
1620 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1621 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1622 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1626 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1630 * Add CAdES-BES signature verification support, mostly derived
1631 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1633 *Filipe Raimundo da Silva*
1635 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1639 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1640 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1644 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1648 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1649 EC_KEY_precompute_mult().
1653 * Deprecated EC_POINTs_mul().
1657 * Removed FIPS_mode() and FIPS_mode_set().
1661 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1665 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1666 EC_POINT_get_Jprojective_coordinates_GFp().
1670 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1671 arrays to be more easily constructed via a series of utility functions.
1672 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1673 the various push functions and finally convert to a passable OSSL_PARAM
1674 array using OSSL_PARAM_BLD_to_param().
1678 * The security strength of SHA1 and MD5 based signatures in TLS has been
1683 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1684 contain a provider side internal key.
1688 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1692 * Project text documents not yet having a proper file name extension
1693 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1694 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1698 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1699 have been converted to Markdown with the goal to produce documents
1700 which not only look pretty when viewed online in the browser, but
1701 remain well readable inside a plain text editor.
1703 To achieve this goal, a 'minimalistic' Markdown style has been applied
1704 which avoids formatting elements that interfere too much with the
1705 reading flow in the text file. For example, it
1707 * avoids [ATX headings][] and uses [setext headings][] instead
1708 (which works for `<h1>` and `<h2>` headings only).
1709 * avoids [inline links][] and uses [reference links][] instead.
1710 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1712 [ATX headings]: https://github.github.com/gfm/#atx-headings
1713 [setext headings]: https://github.github.com/gfm/#setext-headings
1714 [inline links]: https://github.github.com/gfm/#inline-link
1715 [reference links]: https://github.github.com/gfm/#reference-link
1716 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1717 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1719 *Matthias St. Pierre*
1721 * The test suite is changed to preserve results of each test recipe.
1722 A new directory test-runs/ with subdirectories named like the
1723 test recipes are created in the build tree for this purpose.
1727 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1728 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1729 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1731 *David von Oheimb, Martin Peylo*
1733 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1734 It supports arbitrary request and response content types, GET redirection,
1735 TLS, connections via HTTP(S) proxies, connections and exchange via
1736 user-defined BIOs (allowing implicit connections), persistent connections,
1737 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1738 The legacy OCSP-focused (and only partly documented) API
1739 is retained for backward compatibility, while most of it is deprecated.
1743 * Added `util/check-format.pl`, a tool for checking adherence to the
1744 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1745 The checks performed are incomplete and yield some false positives.
1746 Still the tool should be useful for detecting most typical glitches.
1750 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1751 If domain name resolution yields multiple IP addresses all of them are tried
1752 after `connect()` failures.
1756 * All of the low-level RSA functions have been deprecated.
1760 * X509 certificates signed using SHA1 are no longer allowed at security
1765 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1766 modified to use PKEY APIs. These commands are now in maintenance mode
1767 and no new features will be added to them.
1771 * The command line utility rsautl has been deprecated.
1775 * The command line utilities genrsa and rsa have been modified to use PKEY
1776 APIs. They now write PKCS#8 keys by default. These commands are now in
1777 maintenance mode and no new features will be added to them.
1781 * All of the low-level DH functions have been deprecated.
1783 *Paul Dale and Matt Caswell*
1785 * All of the low-level DSA functions have been deprecated.
1789 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1790 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1794 * Deprecated low-level ECDH and ECDSA functions.
1798 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1802 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1803 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1804 a new formulation to include all the things it can be used for,
1805 as well as words of caution.
1809 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1813 * All of the low-level HMAC functions have been deprecated.
1815 *Paul Dale and David von Oheimb*
1817 * Over two thousand fixes were made to the documentation, including:
1818 - Common options (such as -rand/-writerand, TLS version control, etc)
1819 were refactored and point to newly-enhanced descriptions in openssl.pod.
1820 - Added style conformance for all options (with help from Richard Levitte),
1821 documented all reported missing options, added a CI build to check
1822 that all options are documented and that no unimplemented options
1824 - Documented some internals, such as all use of environment variables.
1825 - Addressed all internal broken L<> references.
1829 * All of the low-level CMAC functions have been deprecated.
1833 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1834 functions have been deprecated.
1836 *Paul Dale and David von Oheimb*
1838 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1839 set of functions. The documentation mentioned negative values for some
1840 errors, but this was never the case, so the mention of negative values
1843 Code that followed the documentation and thereby check with something
1844 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1848 * All of the low-level cipher functions have been deprecated.
1850 *Matt Caswell and Paul Dale*
1852 * Removed include/openssl/opensslconf.h.in and replaced it with
1853 include/openssl/configuration.h.in, which differs in not including
1854 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1855 was added to include both.
1857 This allows internal hacks where one might need to modify the set
1858 of configured macros, for example this if deprecated symbols are
1859 still supposed to be available internally:
1861 #include <openssl/configuration.h>
1863 #undef OPENSSL_NO_DEPRECATED
1864 #define OPENSSL_SUPPRESS_DEPRECATED
1866 #include <openssl/macros.h>
1868 This should not be used by applications that use the exported
1869 symbols, as that will lead to linking errors.
1873 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1874 used in exponentiation with 512-bit moduli. No EC algorithms are
1875 affected. Analysis suggests that attacks against 2-prime RSA1024,
1876 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1877 difficult to perform and are not believed likely. Attacks against DH512
1878 are considered just feasible. However, for an attack the target would
1879 have to re-use the DH512 private key, which is not recommended anyway.
1880 Also applications directly using the low-level API BN_mod_exp may be
1881 affected if they use BN_FLG_CONSTTIME.
1886 * Most memory-debug features have been deprecated, and the functionality
1887 replaced with no-ops.
1891 * Added documentation for the STACK API.
1895 * Introduced a new method type and API, OSSL_ENCODER, to represent
1896 generic encoders. These do the same sort of job that PEM writers
1897 and d2i functions do, but with support for methods supplied by
1898 providers, and the possibility for providers to support other
1903 * Introduced a new method type and API, OSSL_DECODER, to represent
1904 generic decoders. These do the same sort of job that PEM readers
1905 and i2d functions do, but with support for methods supplied by
1906 providers, and the possibility for providers to support other
1911 * Added a .pragma directive to the syntax of configuration files, to
1912 allow varying behavior in a supported and predictable manner.
1913 Currently added pragma:
1917 This allows dollar signs to be a keyword character unless it's
1918 followed by a opening brace or parenthesis. This is useful for
1919 platforms where dollar signs are commonly used in names, such as
1920 volume names and system directory names on VMS.
1924 * Added functionality to create an EVP_PKEY from user data.
1928 * Change the interpretation of the '--api' configuration option to
1929 mean that this is a desired API compatibility level with no
1930 further meaning. The previous interpretation, that this would
1931 also mean to remove all deprecated symbols up to and including
1932 the given version, no requires that 'no-deprecated' is also used
1933 in the configuration.
1935 When building applications, the desired API compatibility level
1936 can be set with the OPENSSL_API_COMPAT macro like before. For
1937 API compatibility version below 3.0, the old style numerical
1938 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1939 For version 3.0 and on, the value is expected to be the decimal
1940 value calculated from the major and minor version like this:
1942 MAJOR * 10000 + MINOR * 100
1946 -DOPENSSL_API_COMPAT=30000 For 3.0
1947 -DOPENSSL_API_COMPAT=30200 For 3.2
1949 To hide declarations that are deprecated up to and including the
1950 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1951 given when building the application as well.
1955 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1956 access to certificate and CRL stores via URIs and OSSL_STORE
1959 This adds the following functions:
1961 - X509_LOOKUP_store()
1962 - X509_STORE_load_file()
1963 - X509_STORE_load_path()
1964 - X509_STORE_load_store()
1965 - SSL_add_store_cert_subjects_to_stack()
1966 - SSL_CTX_set_default_verify_store()
1967 - SSL_CTX_load_verify_file()
1968 - SSL_CTX_load_verify_dir()
1969 - SSL_CTX_load_verify_store()
1973 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1974 The presence of this system service is determined at run-time.
1978 * Added functionality to create an EVP_PKEY context based on data
1979 for methods from providers. This takes an algorithm name and a
1980 property query string and simply stores them, with the intent
1981 that any operation that uses this context will use those strings
1982 to fetch the needed methods implicitly, thereby making the port
1983 of application written for pre-3.0 OpenSSL easier.
1987 * The undocumented function NCONF_WIN32() has been deprecated; for
1988 conversion details see the HISTORY section of doc/man5/config.pod
1992 * Introduced the new functions EVP_DigestSignInit_ex() and
1993 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1994 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1995 pages for further details.
1999 * Over two thousand fixes were made to the documentation, including:
2000 adding missing command flags, better style conformance, documentation
2003 *Rich Salz, Richard Levitte*
2005 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2006 X25519, X448, Ed25519 and Ed448.
2010 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2015 * Deprecated the public definition of `ERR_STATE` as well as the function
2016 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2021 * Added ERR functionality to give callers access to the stored function
2022 names that have replaced the older function code based functions.
2024 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2025 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2026 ERR_peek_error_all() and ERR_peek_last_error_all().
2028 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2029 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2030 ERR_func_error_string().
2034 * Extended testing to be verbose for failing tests only. The make variables
2035 VERBOSE_FAILURE or VF can be used to enable this:
2037 $ make VF=1 test # Unix
2038 $ mms /macro=(VF=1) test ! OpenVMS
2039 $ nmake VF=1 test # Windows
2043 * Added the `-copy_extensions` option to the `x509` command for use with
2044 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2045 all extensions in the request are copied to the certificate or vice versa.
2047 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2049 * Added the `-copy_extensions` option to the `req` command for use with
2050 `-x509`. When given with the `copy` or `copyall` argument,
2051 all extensions in the certification request are copied to the certificate.
2055 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2056 they generate are by default RFC 5280 compliant in the following sense:
2057 There is a subjectKeyIdentifier extension with a hash value of the public key
2058 and for not self-signed certs there is an authorityKeyIdentifier extension
2059 with a keyIdentifier field or issuer information identifying the signing key.
2060 This is done unless some configuration overrides the new default behavior,
2061 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2065 * Added several checks to `X509_verify_cert()` according to requirements in
2066 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2067 (which may be done by using the CLI option `-x509_strict`):
2068 * The basicConstraints of CA certificates must be marked critical.
2069 * CA certificates must explicitly include the keyUsage extension.
2070 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2071 * The issuer name of any certificate must not be empty.
2072 * The subject name of CA certs, certs with keyUsage crlSign,
2073 and certs without subjectAlternativeName must not be empty.
2074 * If a subjectAlternativeName extension is given it must not be empty.
2075 * The signatureAlgorithm field and the cert signature must be consistent.
2076 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2077 must not be marked critical.
2078 * The authorityKeyIdentifier must be given for X.509v3 certs
2079 unless they are self-signed.
2080 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2084 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2085 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2089 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2090 used even when parsing explicit parameters, when loading a encoded key
2091 or calling `EC_GROUP_new_from_ecpkparameters()`/
2092 `EC_GROUP_new_from_ecparameters()`.
2093 This prevents bypass of security hardening and performance gains,
2094 especially for curves with specialized EC_METHODs.
2095 By default, if a key encoded with explicit parameters is loaded and later
2096 encoded, the output is still encoded with explicit parameters, even if
2097 internally a "named" EC_GROUP is used for computation.
2101 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2102 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2103 NULL. After this change, only the cofactor parameter can be NULL. It also
2104 does some minimal sanity checks on the passed order.
2109 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2110 An attack is simple, if the first CMS_recipientInfo is valid but the
2111 second CMS_recipientInfo is chosen ciphertext. If the second
2112 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2113 encryption key will be replaced by garbage, and the message cannot be
2114 decoded, but if the RSA decryption fails, the correct encryption key is
2115 used and the recipient will not notice the attack.
2116 As a work around for this potential attack the length of the decrypted
2117 key must be equal to the cipher default key length, in case the
2118 certificate is not given and all recipientInfo are tried out.
2119 The old behaviour can be re-enabled in the CMS code by setting the
2120 CMS_DEBUG_DECRYPT flag.
2124 * Early start up entropy quality from the DEVRANDOM seed source has been
2125 improved for older Linux systems. The RAND subsystem will wait for
2126 /dev/random to be producing output before seeding from /dev/urandom.
2127 The seeded state is stored for future library initialisations using
2128 a system global shared memory segment. The shared memory identifier
2129 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2130 the desired value. The default identifier is 114.
2134 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2135 when primes for RSA keys are computed.
2136 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2137 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2138 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2139 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2140 This avoids possible fingerprinting of newly generated RSA modules.
2144 * Correct the extended master secret constant on EBCDIC systems. Without this
2145 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2146 negotiate EMS will fail. Unfortunately this also means that TLS connections
2147 between EBCDIC systems with this fix, and EBCDIC systems without this
2148 fix will fail if they negotiate EMS.
2152 * Changed the library initialisation so that the config file is now loaded
2153 by default. This was already the case for libssl. It now occurs for both
2154 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2155 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2159 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2160 where the former acts as a replacement for `ERR_put_error()`, and the
2161 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2162 `ERR_raise_data()` adds more flexibility by taking a format string and
2163 an arbitrary number of arguments following it, to be processed with
2168 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2169 to check if a named provider is loaded and available. When called, it
2170 will also activate all fallback providers if such are still present.
2174 * Enforce a minimum DH modulus size of 512 bits.
2178 * Changed DH parameters to generate the order q subgroup instead of 2q.
2179 Previously generated DH parameters are still accepted by DH_check
2180 but DH_generate_key works around that by clearing bit 0 of the
2181 private key for those. This avoids leaking bit 0 of the private key.
2185 * Significantly reduce secure memory usage by the randomness pools.
2189 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2194 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2195 algorithms. An implementation of a key exchange algorithm can be obtained
2196 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2197 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2198 the older EVP_PKEY_derive_init() function. See the man pages for the new
2199 functions for further details.
2203 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2207 * Removed the function names from error messages and deprecated the
2212 * Removed NextStep support and the macro OPENSSL_UNISTD
2216 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2217 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2218 Also removed "export var as function" capability; we do not export
2219 variables, only functions.
2223 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2224 an error and 1 indicating success. In previous versions of OpenSSL this
2225 was a void type. If a key was set longer than the maximum possible this
2230 * Support SM2 signing and verification schemes with X509 certificate.
2234 * Use SHA256 as the default digest for TS query in the `ts` app.
2238 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2242 * Default cipher lists/suites are now available via a function, the
2243 #defines are deprecated.
2247 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2248 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2249 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2253 * Join the directories crypto/x509 and crypto/x509v3
2257 * Added command 'openssl kdf' that uses the EVP_KDF API.
2261 * Added command 'openssl mac' that uses the EVP_MAC API.
2265 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2266 as default directories. Also added the command 'openssl info'
2267 for scripting purposes.
2271 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2276 * Add prediction resistance to the DRBG reseeding process.
2280 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2281 mandated by IEEE Std 1619-2018.
2285 * Added newline escaping functionality to a filename when using openssl dgst.
2286 This output format is to replicate the output format found in the `*sum`
2287 checksum programs. This aims to preserve backward compatibility.
2289 *Matt Eaton, Richard Levitte, and Paul Dale*
2291 * Removed the heartbeat message in DTLS feature, as it has very
2292 little usage and doesn't seem to fulfill a valuable purpose.
2293 The configuration option is now deprecated.
2297 * Changed the output of 'openssl {digestname} < file' to display the
2298 digest name in its output.
2302 * Added a new generic trace API which provides support for enabling
2303 instrumentation through trace output.
2305 *Richard Levitte & Matthias St. Pierre*
2307 * Added build tests for C++. These are generated files that only do one
2308 thing, to include one public OpenSSL head file each. This tests that
2309 the public header files can be usefully included in a C++ application.
2311 This test isn't enabled by default. It can be enabled with the option
2312 'enable-buildtest-c++'.
2316 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2320 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2324 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2328 * Added KMAC to EVP_MAC.
2332 * Added property based algorithm implementation selection framework to
2337 * Added SCA hardening for modular field inversion in EC_GROUP through
2338 a new dedicated field_inv() pointer in EC_METHOD.
2339 This also addresses a leakage affecting conversions from projective
2340 to affine coordinates.
2342 *Billy Bob Brumley, Nicola Tuveri*
2344 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2345 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2346 those algorithms that were already supported through the EVP_PKEY API
2347 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2348 and scrypt are now wrappers that call EVP_KDF.
2352 * Build devcrypto engine as a dynamic engine.
2354 *Eneas U de Queiroz*
2356 * Add keyed BLAKE2 to EVP_MAC.
2360 * Fix a bug in the computation of the endpoint-pair shared secret used
2361 by DTLS over SCTP. This breaks interoperability with older versions
2362 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2363 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2364 interoperability with such broken implementations. However, enabling
2365 this switch breaks interoperability with correct implementations.
2367 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2368 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2372 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2376 * Changed the license to the Apache License v2.0.
2380 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2382 - Major releases (indicated by incrementing the MAJOR release number)
2383 may introduce incompatible API/ABI changes.
2384 - Minor releases (indicated by incrementing the MINOR release number)
2385 may introduce new features but retain API/ABI compatibility.
2386 - Patch releases (indicated by incrementing the PATCH number)
2387 are intended for bug fixes and other improvements of existing
2388 features only (like improving performance or adding documentation)
2389 and retain API/ABI compatibility.
2393 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2397 * Remove the 'dist' target and add a tarball building script. The
2398 'dist' target has fallen out of use, and it shouldn't be
2399 necessary to configure just to create a source distribution.
2403 * Recreate the OS390-Unix config target. It no longer relies on a
2404 special script like it did for OpenSSL pre-1.1.0.
2408 * Instead of having the source directories listed in Configure, add
2409 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2414 * Add GMAC to EVP_MAC.
2418 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2422 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2423 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2424 to facilitate the continued use of MACs through raw private keys in
2425 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2429 * Deprecate ECDH_KDF_X9_62().
2433 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2434 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2435 are retained for backwards compatibility.
2439 * AES-XTS mode now enforces that its two keys are different to mitigate
2440 the attacked described in "Efficient Instantiations of Tweakable
2441 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2442 Details of this attack can be obtained from:
2443 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2447 * Rename the object files, i.e. give them other names than in previous
2448 versions. Their names now include the name of the final product, as
2449 well as its type mnemonic (bin, lib, shlib).
2453 * Added new option for 'openssl list', '-objects', which will display the
2454 list of built in objects, i.e. OIDs with names.
2458 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2459 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2464 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2465 improves application performance by removing data copies and providing
2466 applications with zero-copy system calls such as sendfile and splice.
2470 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2474 * `PKCS12_parse` now maintains the order of the parsed certificates
2475 when outputting them via `*ca` (rather than reversing it).
2479 * Deprecated pthread fork support methods.
2483 * Added support for FFDHE key exchange in TLS 1.3.
2487 * Added a new concept for OpenSSL plugability: providers. This
2488 functionality is designed to replace the ENGINE API and ENGINE
2489 implementations, and to be much more dynamic, allowing provider
2490 authors to introduce new algorithms among other things, as long as
2491 there's an API that supports the algorithm type.
2493 With this concept comes a new core API for interaction between
2494 libcrypto and provider implementations. Public libcrypto functions
2495 that want to use providers do so through this core API.
2497 The main documentation for this core API is found in
2498 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2499 refer to other manuals describing the API specific for supported
2500 algorithm types (also called operations).
2507 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2509 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2511 * Avoid loading of a dynamic engine twice.
2515 * Prioritise DANE TLSA issuer certs over peer certs
2519 * Fixed random API for MacOS prior to 10.12
2521 These MacOS versions don't support the CommonCrypto APIs
2525 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2527 * Fixed an SM2 Decryption Buffer Overflow.
2529 In order to decrypt SM2 encrypted data an application is expected to
2530 call the API function EVP_PKEY_decrypt(). Typically an application will
2531 call this function twice. The first time, on entry, the "out" parameter
2532 can be NULL and, on exit, the "outlen" parameter is populated with the
2533 buffer size required to hold the decrypted plaintext. The application
2534 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2535 again, but this time passing a non-NULL value for the "out" parameter.
2537 A bug in the implementation of the SM2 decryption code means that the
2538 calculation of the buffer size required to hold the plaintext returned
2539 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2540 size required by the second call. This can lead to a buffer overflow
2541 when EVP_PKEY_decrypt() is called by the application a second time with
2542 a buffer that is too small.
2544 A malicious attacker who is able present SM2 content for decryption to
2545 an application could cause attacker chosen data to overflow the buffer
2546 by up to a maximum of 62 bytes altering the contents of other data held
2547 after the buffer, possibly changing application behaviour or causing
2548 the application to crash. The location of the buffer is application
2549 dependent but is typically heap allocated.
2554 * Fixed various read buffer overruns processing ASN.1 strings
2556 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2557 structure which contains a buffer holding the string data and a field
2558 holding the buffer length. This contrasts with normal C strings which
2559 are represented as a buffer for the string data which is terminated
2560 with a NUL (0) byte.
2562 Although not a strict requirement, ASN.1 strings that are parsed using
2563 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2564 well as any string whose value has been set with the ASN1_STRING_set()
2565 function will additionally NUL terminate the byte array in the
2566 ASN1_STRING structure.
2568 However, it is possible for applications to directly construct valid
2569 ASN1_STRING structures which do not NUL terminate the byte array by
2570 directly setting the "data" and "length" fields in the ASN1_STRING
2571 array. This can also happen by using the ASN1_STRING_set0() function.
2573 Numerous OpenSSL functions that print ASN.1 data have been found to
2574 assume that the ASN1_STRING byte array will be NUL terminated, even
2575 though this is not guaranteed for strings that have been directly
2576 constructed. Where an application requests an ASN.1 structure to be
2577 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2578 been directly constructed by the application without NUL terminating
2579 the "data" field, then a read buffer overrun can occur.
2581 The same thing can also occur during name constraints processing
2582 of certificates (for example if a certificate has been directly
2583 constructed by the application instead of loading it via the OpenSSL
2584 parsing functions, and the certificate contains non NUL terminated
2585 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2586 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2588 If a malicious actor can cause an application to directly construct an
2589 ASN1_STRING and then process it through one of the affected OpenSSL
2590 functions then this issue could be hit. This might result in a crash
2591 (causing a Denial of Service attack). It could also result in the
2592 disclosure of private memory contents (such as private keys, or
2593 sensitive plaintext).
2598 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2600 * Fixed a problem with verifying a certificate chain when using the
2601 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2602 the certificates present in a certificate chain. It is not set by default.
2604 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2605 the chain that have explicitly encoded elliptic curve parameters was added
2606 as an additional strict check.
2608 An error in the implementation of this check meant that the result of a
2609 previous check to confirm that certificates in the chain are valid CA
2610 certificates was overwritten. This effectively bypasses the check
2611 that non-CA certificates must not be able to issue other certificates.
2613 If a "purpose" has been configured then there is a subsequent opportunity
2614 for checks that the certificate is a valid CA. All of the named "purpose"
2615 values implemented in libcrypto perform this check. Therefore, where
2616 a purpose is set the certificate chain will still be rejected even when the
2617 strict flag has been used. A purpose is set by default in libssl client and
2618 server certificate verification routines, but it can be overridden or
2619 removed by an application.
2621 In order to be affected, an application must explicitly set the
2622 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2623 for the certificate verification or, in the case of TLS client or server
2624 applications, override the default purpose.
2629 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2630 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2631 renegotiation ClientHello omits the signature_algorithms extension (where it
2632 was present in the initial ClientHello), but includes a
2633 signature_algorithms_cert extension then a NULL pointer dereference will
2634 result, leading to a crash and a denial of service attack.
2636 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2637 (which is the default configuration). OpenSSL TLS clients are not impacted by
2641 *Peter Kästle and Samuel Sapalski*
2643 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2645 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2646 create a unique hash value based on the issuer and serial number data
2647 contained within an X509 certificate. However, it was failing to correctly
2648 handle any errors that may occur while parsing the issuer field (which might
2649 occur if the issuer field is maliciously constructed). This may subsequently
2650 result in a NULL pointer deref and a crash leading to a potential denial of
2656 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2657 padding mode to correctly check for rollback attacks. This is considered a
2658 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2663 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2664 functions. Previously they could overflow the output length argument in some
2665 cases where the input length is close to the maximum permissible length for
2666 an integer on the platform. In such cases the return value from the function
2667 call would be 1 (indicating success), but the output length value would be
2668 negative. This could cause applications to behave incorrectly or crash.
2673 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2674 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2675 could be exploited in a side channel attack to recover the password. Since
2676 the attack is local host only this is outside of the current OpenSSL
2677 threat model and therefore no CVE is assigned.
2679 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2684 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2686 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2687 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2688 If an attacker can control both items being compared then this could lead
2689 to a possible denial of service attack. OpenSSL itself uses the
2690 GENERAL_NAME_cmp function for two purposes:
2691 1) Comparing CRL distribution point names between an available CRL and a
2692 CRL distribution point embedded in an X509 certificate
2693 2) When verifying that a timestamp response token signer matches the
2694 timestamp authority name (exposed via the API functions
2695 TS_RESP_verify_response and TS_RESP_verify_token)
2700 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2702 * Certificates with explicit curve parameters are now disallowed in
2703 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2707 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2708 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2709 conversely, silently ignore DTLS protocol version bounds when configuring
2710 TLS-based contexts. The commands can be repeated to set bounds of both
2711 types. The same applies with the corresponding "min_protocol" and
2712 "max_protocol" command-line switches, in case some application uses both TLS
2715 SSL_CTX instances that are created for a fixed protocol version (e.g.
2716 TLSv1_server_method()) also silently ignore version bounds. Previously
2717 attempts to apply bounds to these protocol versions would result in an
2718 error. Now only the "version-flexible" SSL_CTX instances are subject to
2719 limits in configuration files in command-line options.
2723 * Handshake now fails if Extended Master Secret extension is dropped
2728 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2730 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2732 * Fixed segmentation fault in SSL_check_chain()
2733 Server or client applications that call the SSL_check_chain() function
2734 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2735 dereference as a result of incorrect handling of the
2736 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2737 or unrecognised signature algorithm is received from the peer. This could
2738 be exploited by a malicious peer in a Denial of Service attack.
2743 * Added AES consttime code for no-asm configurations
2744 an optional constant time support for AES was added
2745 when building openssl for no-asm.
2746 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2747 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2748 At this time this feature is by default disabled.
2749 It will be enabled by default in 3.0.
2753 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2755 * Revert the change of EOF detection while reading in libssl to avoid
2756 regressions in applications depending on the current way of reporting
2757 the EOF. As the existing method is not fully accurate the change to
2758 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2759 branch and will be present in the 3.0 release.
2763 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2764 when primes for RSA keys are computed.
2765 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2766 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2767 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2768 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2769 This avoids possible fingerprinting of newly generated RSA modules.
2773 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2775 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2776 while reading in libssl then we would report an error back to the
2777 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2778 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2779 therefore give a hint as to what went wrong.
2783 * Check that ed25519 and ed448 are allowed by the security level. Previously
2784 signature algorithms not using an MD were not being checked that they were
2785 allowed by the security level.
2789 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2790 was not quite right. The behaviour was not consistent between resumption
2791 and normal handshakes, and also not quite consistent with historical
2792 behaviour. The behaviour in various scenarios has been clarified and
2793 it has been updated to make it match historical behaviour as closely as
2798 * *[VMS only]* The header files that the VMS compilers include automatically,
2799 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2800 that the C++ compiler doesn't understand. This is a shortcoming in the
2801 compiler, but can be worked around with `__cplusplus` guards.
2803 C++ applications that use OpenSSL libraries must be compiled using the
2804 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2805 functions. Otherwise, only functions with symbols of less than 31
2806 characters can be used, as the linker will not be able to successfully
2807 resolve symbols with longer names.
2811 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2812 The presence of this system service is determined at run-time.
2816 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2821 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2823 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2824 number generator (RNG). This was intended to include protection in the
2825 event of a fork() system call in order to ensure that the parent and child
2826 processes did not share the same RNG state. However, this protection was not
2827 being used in the default case.
2829 A partial mitigation for this issue is that the output from a high
2830 precision timer is mixed into the RNG state so the likelihood of a parent
2831 and child process sharing state is significantly reduced.
2833 If an application already calls OPENSSL_init_crypto() explicitly using
2834 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2837 *Matthias St. Pierre*
2839 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2840 used even when parsing explicit parameters, when loading a encoded key
2841 or calling `EC_GROUP_new_from_ecpkparameters()`/
2842 `EC_GROUP_new_from_ecparameters()`.
2843 This prevents bypass of security hardening and performance gains,
2844 especially for curves with specialized EC_METHODs.
2845 By default, if a key encoded with explicit parameters is loaded and later
2846 encoded, the output is still encoded with explicit parameters, even if
2847 internally a "named" EC_GROUP is used for computation.
2851 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2852 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2853 NULL. After this change, only the cofactor parameter can be NULL. It also
2854 does some minimal sanity checks on the passed order.
2859 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2860 An attack is simple, if the first CMS_recipientInfo is valid but the
2861 second CMS_recipientInfo is chosen ciphertext. If the second
2862 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2863 encryption key will be replaced by garbage, and the message cannot be
2864 decoded, but if the RSA decryption fails, the correct encryption key is
2865 used and the recipient will not notice the attack.
2866 As a work around for this potential attack the length of the decrypted
2867 key must be equal to the cipher default key length, in case the
2868 certificate is not given and all recipientInfo are tried out.
2869 The old behaviour can be re-enabled in the CMS code by setting the
2870 CMS_DEBUG_DECRYPT flag.
2875 * Early start up entropy quality from the DEVRANDOM seed source has been
2876 improved for older Linux systems. The RAND subsystem will wait for
2877 /dev/random to be producing output before seeding from /dev/urandom.
2878 The seeded state is stored for future library initialisations using
2879 a system global shared memory segment. The shared memory identifier
2880 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2881 the desired value. The default identifier is 114.
2885 * Correct the extended master secret constant on EBCDIC systems. Without this
2886 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2887 negotiate EMS will fail. Unfortunately this also means that TLS connections
2888 between EBCDIC systems with this fix, and EBCDIC systems without this
2889 fix will fail if they negotiate EMS.
2893 * Use Windows installation paths in the mingw builds
2895 Mingw isn't a POSIX environment per se, which means that Windows
2896 paths should be used for installation.
2901 * Changed DH_check to accept parameters with order q and 2q subgroups.
2902 With order 2q subgroups the bit 0 of the private key is not secret
2903 but DH_generate_key works around that by clearing bit 0 of the
2904 private key for those. This avoids leaking bit 0 of the private key.
2908 * Significantly reduce secure memory usage by the randomness pools.
2912 * Revert the DEVRANDOM_WAIT feature for Linux systems
2914 The DEVRANDOM_WAIT feature added a select() call to wait for the
2915 /dev/random device to become readable before reading from the
2916 /dev/urandom device.
2918 It turned out that this change had negative side effects on
2919 performance which were not acceptable. After some discussion it
2920 was decided to revert this feature and leave it up to the OS
2921 resp. the platform maintainer to ensure a proper initialization
2922 during early boot time.
2924 *Matthias St. Pierre*
2926 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2928 * Add build tests for C++. These are generated files that only do one
2929 thing, to include one public OpenSSL head file each. This tests that
2930 the public header files can be usefully included in a C++ application.
2932 This test isn't enabled by default. It can be enabled with the option
2933 'enable-buildtest-c++'.
2937 * Enable SHA3 pre-hashing for ECDSA and DSA.
2941 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2942 This changes the size when using the `genpkey` command when no size is given.
2943 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2944 generation commands to use 2048 bits by default.
2948 * Reorganize the manual pages to consistently have RETURN VALUES,
2949 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2950 util/fix-doc-nits accordingly.
2952 *Paul Yang, Joshua Lock*
2954 * Add the missing accessor EVP_PKEY_get0_engine()
2958 * Have commands like `s_client` and `s_server` output the signature scheme
2959 along with other cipher suite parameters when debugging.
2963 * Make OPENSSL_config() error agnostic again.
2967 * Do the error handling in RSA decryption constant time.
2971 * Prevent over long nonces in ChaCha20-Poly1305.
2973 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2974 for every encryption operation. RFC 7539 specifies that the nonce value
2975 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2976 and front pads the nonce with 0 bytes if it is less than 12
2977 bytes. However it also incorrectly allows a nonce to be set of up to 16
2978 bytes. In this case only the last 12 bytes are significant and any
2979 additional leading bytes are ignored.
2981 It is a requirement of using this cipher that nonce values are
2982 unique. Messages encrypted using a reused nonce value are susceptible to
2983 serious confidentiality and integrity attacks. If an application changes
2984 the default nonce length to be longer than 12 bytes and then makes a
2985 change to the leading bytes of the nonce expecting the new value to be a
2986 new unique nonce then such an application could inadvertently encrypt
2987 messages with a reused nonce.
2989 Additionally the ignored bytes in a long nonce are not covered by the
2990 integrity guarantee of this cipher. Any application that relies on the
2991 integrity of these ignored leading bytes of a long nonce may be further
2992 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2993 is safe because no such use sets such a long nonce value. However user
2994 applications that use this cipher directly and set a non-default nonce
2995 length to be longer than 12 bytes may be vulnerable.
2997 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3003 * Add DEVRANDOM_WAIT feature for Linux systems
3005 On older Linux systems where the getrandom() system call is not available,
3006 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3007 Contrary to getrandom(), the /dev/urandom device will not block during
3008 early boot when the kernel CSPRNG has not been seeded yet.
3010 To mitigate this known weakness, use select() to wait for /dev/random to
3011 become readable before reading from /dev/urandom.
3013 * Ensure that SM2 only uses SM3 as digest algorithm
3017 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3019 * Change the info callback signals for the start and end of a post-handshake
3020 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3021 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3022 confused by this and assume that a TLSv1.2 renegotiation has started. This
3023 can break KeyUpdate handling. Instead we no longer signal the start and end
3024 of a post handshake message exchange (although the messages themselves are
3025 still signalled). This could break some applications that were expecting
3026 the old signals. However without this KeyUpdate is not usable for many
3031 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3033 * Timing vulnerability in DSA signature generation
3035 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3036 timing side channel attack. An attacker could use variations in the signing
3037 algorithm to recover the private key.
3039 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3044 * Timing vulnerability in ECDSA signature generation
3046 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3047 timing side channel attack. An attacker could use variations in the signing
3048 algorithm to recover the private key.
3050 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3055 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3056 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3057 of two gigabytes and the error handling improved.
3059 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3060 categorized as a normal bug, not a security issue, because the DRBG reseeds
3061 automatically and is fully functional even without additional randomness
3062 provided by the application.
3064 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3066 * Add a new ClientHello callback. Provides a callback interface that gives
3067 the application the ability to adjust the nascent SSL object at the
3068 earliest stage of ClientHello processing, immediately after extensions have
3069 been collected but before they have been processed. In particular, this
3070 callback can adjust the supported TLS versions in response to the contents
3075 * Add SM2 base algorithm support.
3079 * s390x assembly pack: add (improved) hardware-support for the following
3080 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3081 aes-cfb/cfb8, aes-ecb.
3085 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3086 parameter is no longer accepted, as it leads to a corrupt table. NULL
3087 pem_str is reserved for alias entries only.
3091 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3092 step for prime curves. The new implementation is based on formulae from
3093 differential addition-and-doubling in homogeneous projective coordinates
3094 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3095 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3096 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3097 to work in projective coordinates.
3099 *Billy Bob Brumley, Nicola Tuveri*
3101 * Change generating and checking of primes so that the error rate of not
3102 being prime depends on the intended use based on the size of the input.
3103 For larger primes this will result in more rounds of Miller-Rabin.
3104 The maximal error rate for primes with more than 1080 bits is lowered
3107 *Kurt Roeckx, Annie Yousar*
3109 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3113 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3114 moving between systems, and to avoid confusion when a Windows build is
3115 done with mingw vs with MSVC. For POSIX installs, there's still a
3116 symlink or copy named 'tsget' to avoid that confusion as well.
3120 * Revert blinding in ECDSA sign and instead make problematic addition
3121 length-invariant. Switch even to fixed-length Montgomery multiplication.
3125 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3126 step for binary curves. The new implementation is based on formulae from
3127 differential addition-and-doubling in mixed Lopez-Dahab projective
3128 coordinates, modified to independently blind the operands.
3130 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3132 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3133 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3134 EC_METHODs to implement their own specialized "ladder step", to take
3135 advantage of more favorable coordinate systems or more efficient
3136 differential addition-and-doubling algorithms.
3138 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3140 * Modified the random device based seed sources to keep the relevant
3141 file descriptors open rather than reopening them on each access.
3142 This allows such sources to operate in a chroot() jail without
3143 the associated device nodes being available. This behaviour can be
3144 controlled using RAND_keep_random_devices_open().
3148 * Numerous side-channel attack mitigations have been applied. This may have
3149 performance impacts for some algorithms for the benefit of improved
3150 security. Specific changes are noted in this change log by their respective
3155 * AIX shared library support overhaul. Switch to AIX "natural" way of
3156 handling shared libraries, which means collecting shared objects of
3157 different versions and bitnesses in one common archive. This allows to
3158 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3159 doesn't affect the way 3rd party applications are linked, only how
3160 multi-version installation is managed.
3164 * Make ec_group_do_inverse_ord() more robust and available to other
3165 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3166 mitigations are applied to the fallback BN_mod_inverse().
3167 When using this function rather than BN_mod_inverse() directly, new
3168 EC cryptosystem implementations are then safer-by-default.
3172 * Add coordinate blinding for EC_POINT and implement projective
3173 coordinate blinding for generic prime curves as a countermeasure to
3174 chosen point SCA attacks.
3176 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3178 * Add blinding to ECDSA and DSA signatures to protect against side channel
3179 attacks discovered by Keegan Ryan (NCC Group).
3183 * Enforce checking in the `pkeyutl` command to ensure that the input
3184 length does not exceed the maximum supported digest length when performing
3185 a sign, verify or verifyrecover operation.
3189 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3190 I/O in combination with something like select() or poll() will hang. This
3191 can be turned off again using SSL_CTX_clear_mode().
3192 Many applications do not properly handle non-application data records, and
3193 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3194 around the problems in those applications, but can also break some.
3195 It's recommended to read the manpages about SSL_read(), SSL_write(),
3196 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3197 SSL_CTX_set_read_ahead() again.
3201 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3202 now allow empty (zero character) pass phrases.
3206 * Apply blinding to binary field modular inversion and remove patent
3207 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3211 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3212 binary and prime elliptic curves.
3216 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3217 constant time fixed point multiplication.
3221 * Revise elliptic curve scalar multiplication with timing attack
3222 defenses: ec_wNAF_mul redirects to a constant time implementation
3223 when computing fixed point and variable point multiplication (which
3224 in OpenSSL are mostly used with secret scalars in keygen, sign,
3225 ECDH derive operations).
3226 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3229 * Updated CONTRIBUTING
3233 * Updated DRBG / RAND to request nonce and additional low entropy
3234 randomness from the system.
3236 *Matthias St. Pierre*
3238 * Updated 'openssl rehash' to use OpenSSL consistent default.
3242 * Moved the load of the ssl_conf module to libcrypto, which helps
3243 loading engines that libssl uses before libssl is initialised.
3247 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3251 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3253 *Ingo Schwarze, Rich Salz*
3255 * Added output of accepting IP address and port for 'openssl s_server'
3259 * Added a new API for TLSv1.3 ciphersuites:
3260 SSL_CTX_set_ciphersuites()
3261 SSL_set_ciphersuites()
3265 * Memory allocation failures consistently add an error to the error
3270 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3271 in libcrypto when run as setuid/setgid.
3275 * Load any config file by default when libssl is used.
3279 * Added new public header file <openssl/rand_drbg.h> and documentation
3280 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3282 *Matthias St. Pierre*
3284 * QNX support removed (cannot find contributors to get their approval
3285 for the license change).
3289 * TLSv1.3 replay protection for early data has been implemented. See the
3290 SSL_read_early_data() man page for further details.
3294 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3295 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3296 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3297 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3298 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3299 configuration has been separated out. See the ciphers man page or the
3300 SSL_CTX_set_ciphersuites() man page for more information.
3304 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3305 in responder mode now supports the new "-multi" option, which
3306 spawns the specified number of child processes to handle OCSP
3307 requests. The "-timeout" option now also limits the OCSP
3308 responder's patience to wait to receive the full client request
3309 on a newly accepted connection. Child processes are respawned
3310 as needed, and the CA index file is automatically reloaded
3311 when changed. This makes it possible to run the "ocsp" responder
3312 as a long-running service, making the OpenSSL CA somewhat more
3313 feature-complete. In this mode, most diagnostic messages logged
3314 after entering the event loop are logged via syslog(3) rather than
3319 * Added support for X448 and Ed448. Heavily based on original work by
3324 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3325 objects loaded. This adds the functions OSSL_STORE_expect() and
3326 OSSL_STORE_find() as well as needed tools to construct searches and
3327 get the search data out of them.
3331 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3332 version of OpenSSL should review their configuration settings to ensure
3333 that they are still appropriate for TLSv1.3. For further information see:
3334 <https://wiki.openssl.org/index.php/TLS1.3>
3338 * Grand redesign of the OpenSSL random generator
3340 The default RAND method now utilizes an AES-CTR DRBG according to
3341 NIST standard SP 800-90Ar1. The new random generator is essentially
3342 a port of the default random generator from the OpenSSL FIPS 2.0
3343 object module. It is a hybrid deterministic random bit generator
3344 using an AES-CTR bit stream and which seeds and reseeds itself
3345 automatically using trusted system entropy sources.
3347 Some of its new features are:
3348 - Support for multiple DRBG instances with seed chaining.
3349 - The default RAND method makes use of a DRBG.
3350 - There is a public and private DRBG instance.
3351 - The DRBG instances are fork-safe.
3352 - Keep all global DRBG instances on the secure heap if it is enabled.
3353 - The public and private DRBG instance are per thread for lock free
3356 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3358 * Changed Configure so it only says what it does and doesn't dump
3359 so much data. Instead, ./configdata.pm should be used as a script
3360 to display all sorts of configuration data.
3364 * Added processing of "make variables" to Configure.
3368 * Added SHA512/224 and SHA512/256 algorithm support.
3372 * The last traces of Netware support, first removed in 1.1.0, have
3377 * Get rid of Makefile.shared, and in the process, make the processing
3378 of certain files (rc.obj, or the .def/.map/.opt files produced from
3379 the ordinal files) more visible and hopefully easier to trace and
3380 debug (or make silent).
3384 * Make it possible to have environment variable assignments as
3385 arguments to config / Configure.
3389 * Add multi-prime RSA (RFC 8017) support.
3393 * Add SM3 implemented according to GB/T 32905-2016
3394 *Jack Lloyd <jack.lloyd@ribose.com>,*
3395 *Ronald Tse <ronald.tse@ribose.com>,*
3396 *Erick Borsboom <erick.borsboom@ribose.com>*
3398 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3399 as documented in RFC6066.
3400 Based on a patch from Tomasz Moń
3402 *Filipe Raimundo da Silva*
3404 * Add SM4 implemented according to GB/T 32907-2016.
3405 *Jack Lloyd <jack.lloyd@ribose.com>,*
3406 *Ronald Tse <ronald.tse@ribose.com>,*
3407 *Erick Borsboom <erick.borsboom@ribose.com>*
3409 * Reimplement -newreq-nodes and ERR_error_string_n; the
3410 original author does not agree with the license change.
3414 * Add ARIA AEAD TLS support.
3418 * Some macro definitions to support VS6 have been removed. Visual
3419 Studio 6 has not worked since 1.1.0
3423 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3424 without clearing the errors.
3428 * Add "atfork" functions. If building on a system that without
3429 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3430 requirements. The RAND facility now uses/requires this.
3438 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3439 not possible to disable entirely. However, it's still possible to
3440 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3443 To disable, configure with 'no-ui-console'. 'no-ui' is still
3444 possible to use as an alias. Check at compile time with the
3445 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3446 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3450 * Add a STORE module, which implements a uniform and URI based reader of
3451 stores that can contain keys, certificates, CRLs and numerous other
3452 objects. The main API is loosely based on a few stdio functions,
3453 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3454 OSSL_STORE_error and OSSL_STORE_close.
3455 The implementation uses backends called "loaders" to implement arbitrary
3456 URI schemes. There is one built in "loader" for the 'file' scheme.
3460 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3461 then adjusted to work on FreeBSD 8.4 as well.
3462 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3463 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3467 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3468 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3469 error code calls like this:
3471 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3473 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3474 that can be encoded in C. For the foreseeable future, this will only
3477 *Richard Levitte and Tim Hudson*
3479 * Removed BSD cryptodev engine.
3483 * Add a build target 'build_all_generated', to build all generated files
3484 and only that. This can be used to prepare everything that requires
3485 things like perl for a system that lacks perl and then move everything
3486 to that system and do the rest of the build there.
3490 * In the UI interface, make it possible to duplicate the user data. This
3491 can be used by engines that need to retain the data for a longer time
3492 than just the call where this user data is passed.
3496 * Ignore the '-named_curve auto' value for compatibility of applications
3499 *Tomáš Mráz <tmraz@fedoraproject.org>*
3501 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3502 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3503 alerts across multiple records (some of which could be empty). In practice
3504 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3505 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3506 support this at all. Supporting it adds significant complexity to the
3507 record layer, and its removal is unlikely to cause interoperability
3512 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3513 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3514 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3519 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3520 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3522 *Richard Levitte, Andy Polyakov*
3524 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3529 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3530 platform rather than 'mingw'.
3534 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3535 success if they are asked to add an object which already exists
3536 in the store. This change cascades to other functions which load
3537 certificates and CRLs.
3541 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3542 facilitate stack unwinding even from assembly subroutines.
3546 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3547 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3551 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3552 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3553 which is the minimum version we support.
3557 * Certificate time validation (X509_cmp_time) enforces stricter
3558 compliance with RFC 5280. Fractional seconds and timezone offsets
3559 are no longer allowed.
3563 * Add support for ARIA
3567 * s_client will now send the Server Name Indication (SNI) extension by
3568 default unless the new "-noservername" option is used. The server name is
3569 based on the host provided to the "-connect" option unless overridden by
3570 using "-servername".
3574 * Add support for SipHash
3578 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3579 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3580 prevent issues where no progress is being made and the peer continually
3581 sends unrecognised record types, using up resources processing them.
3585 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3586 using the algorithm defined in
3587 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3591 * Heartbeat support has been removed; the ABI is changed for now.
3593 *Richard Levitte, Rich Salz*
3595 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3599 * The RSA "null" method, which was partially supported to avoid patent
3600 issues, has been replaced to always returns NULL.
3607 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3609 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3610 used even when parsing explicit parameters, when loading a encoded key
3611 or calling `EC_GROUP_new_from_ecpkparameters()`/
3612 `EC_GROUP_new_from_ecparameters()`.
3613 This prevents bypass of security hardening and performance gains,
3614 especially for curves with specialized EC_METHODs.
3615 By default, if a key encoded with explicit parameters is loaded and later
3616 encoded, the output is still encoded with explicit parameters, even if
3617 internally a "named" EC_GROUP is used for computation.
3621 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3622 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3623 NULL. After this change, only the cofactor parameter can be NULL. It also
3624 does some minimal sanity checks on the passed order.
3629 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3630 An attack is simple, if the first CMS_recipientInfo is valid but the
3631 second CMS_recipientInfo is chosen ciphertext. If the second
3632 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3633 encryption key will be replaced by garbage, and the message cannot be
3634 decoded, but if the RSA decryption fails, the correct encryption key is
3635 used and the recipient will not notice the attack.
3636 As a work around for this potential attack the length of the decrypted
3637 key must be equal to the cipher default key length, in case the
3638 certificate is not given and all recipientInfo are tried out.
3639 The old behaviour can be re-enabled in the CMS code by setting the
3640 CMS_DEBUG_DECRYPT flag.
3645 * Use Windows installation paths in the mingw builds
3647 Mingw isn't a POSIX environment per se, which means that Windows
3648 paths should be used for installation.
3653 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3655 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3656 This changes the size when using the `genpkey` command when no size is given.
3657 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3658 generation commands to use 2048 bits by default.
3662 * Prevent over long nonces in ChaCha20-Poly1305.
3664 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3665 for every encryption operation. RFC 7539 specifies that the nonce value
3666 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3667 and front pads the nonce with 0 bytes if it is less than 12
3668 bytes. However it also incorrectly allows a nonce to be set of up to 16
3669 bytes. In this case only the last 12 bytes are significant and any
3670 additional leading bytes are ignored.
3672 It is a requirement of using this cipher that nonce values are
3673 unique. Messages encrypted using a reused nonce value are susceptible to
3674 serious confidentiality and integrity attacks. If an application changes
3675 the default nonce length to be longer than 12 bytes and then makes a
3676 change to the leading bytes of the nonce expecting the new value to be a
3677 new unique nonce then such an application could inadvertently encrypt
3678 messages with a reused nonce.
3680 Additionally the ignored bytes in a long nonce are not covered by the
3681 integrity guarantee of this cipher. Any application that relies on the
3682 integrity of these ignored leading bytes of a long nonce may be further
3683 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3684 is safe because no such use sets such a long nonce value. However user
3685 applications that use this cipher directly and set a non-default nonce
3686 length to be longer than 12 bytes may be vulnerable.
3688 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3694 * Added SCA hardening for modular field inversion in EC_GROUP through
3695 a new dedicated field_inv() pointer in EC_METHOD.
3696 This also addresses a leakage affecting conversions from projective
3697 to affine coordinates.
3699 *Billy Bob Brumley, Nicola Tuveri*
3701 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3702 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3706 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3710 * Remove the 'dist' target and add a tarball building script. The
3711 'dist' target has fallen out of use, and it shouldn't be
3712 necessary to configure just to create a source distribution.
3716 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3718 * Timing vulnerability in DSA signature generation
3720 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3721 timing side channel attack. An attacker could use variations in the signing
3722 algorithm to recover the private key.
3724 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3729 * Timing vulnerability in ECDSA signature generation
3731 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3732 timing side channel attack. An attacker could use variations in the signing
3733 algorithm to recover the private key.
3735 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3740 * Add coordinate blinding for EC_POINT and implement projective
3741 coordinate blinding for generic prime curves as a countermeasure to
3742 chosen point SCA attacks.
3744 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3746 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3748 * Client DoS due to large DH parameter
3750 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3751 malicious server can send a very large prime value to the client. This will
3752 cause the client to spend an unreasonably long period of time generating a
3753 key for this prime resulting in a hang until the client has finished. This
3754 could be exploited in a Denial Of Service attack.
3756 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3761 * Cache timing vulnerability in RSA Key Generation
3763 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3764 a cache timing side channel attack. An attacker with sufficient access to
3765 mount cache timing attacks during the RSA key generation process could
3766 recover the private key.
3768 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3769 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3774 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3775 parameter is no longer accepted, as it leads to a corrupt table. NULL
3776 pem_str is reserved for alias entries only.
3780 * Revert blinding in ECDSA sign and instead make problematic addition
3781 length-invariant. Switch even to fixed-length Montgomery multiplication.
3785 * Change generating and checking of primes so that the error rate of not
3786 being prime depends on the intended use based on the size of the input.
3787 For larger primes this will result in more rounds of Miller-Rabin.
3788 The maximal error rate for primes with more than 1080 bits is lowered
3791 *Kurt Roeckx, Annie Yousar*
3793 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3797 * Add blinding to ECDSA and DSA signatures to protect against side channel
3798 attacks discovered by Keegan Ryan (NCC Group).
3802 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3803 now allow empty (zero character) pass phrases.
3807 * Certificate time validation (X509_cmp_time) enforces stricter
3808 compliance with RFC 5280. Fractional seconds and timezone offsets
3809 are no longer allowed.
3813 * Fixed a text canonicalisation bug in CMS
3815 Where a CMS detached signature is used with text content the text goes
3816 through a canonicalisation process first prior to signing or verifying a
3817 signature. This process strips trailing space at the end of lines, converts
3818 line terminators to CRLF and removes additional trailing line terminators
3819 at the end of a file. A bug in the canonicalisation process meant that
3820 some characters, such as form-feed, were incorrectly treated as whitespace
3821 and removed. This is contrary to the specification (RFC5485). This fix
3822 could mean that detached text data signed with an earlier version of
3823 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3824 signed with a fixed OpenSSL may fail to verify with an earlier version of
3825 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3826 and use the "-binary" flag (for the "cms" command line application) or set
3827 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3831 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3833 * Constructed ASN.1 types with a recursive definition could exceed the stack
3835 Constructed ASN.1 types with a recursive definition (such as can be found
3836 in PKCS7) could eventually exceed the stack given malicious input with
3837 excessive recursion. This could result in a Denial Of Service attack. There
3838 are no such structures used within SSL/TLS that come from untrusted sources
3839 so this is considered safe.
3841 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3847 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3849 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3850 effectively reduced to only comparing the least significant bit of each
3851 byte. This allows an attacker to forge messages that would be considered as
3852 authenticated in an amount of tries lower than that guaranteed by the
3853 security claims of the scheme. The module can only be compiled by the
3854 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3856 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3862 * Add a build target 'build_all_generated', to build all generated files
3863 and only that. This can be used to prepare everything that requires
3864 things like perl for a system that lacks perl and then move everything
3865 to that system and do the rest of the build there.
3869 * Backport SSL_OP_NO_RENGOTIATION
3871 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3872 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3873 changes this is no longer possible in 1.1.0. Therefore, the new
3874 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3875 1.1.0 to provide equivalent functionality.
3877 Note that if an application built against 1.1.0h headers (or above) is run
3878 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3879 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3883 * Removed the OS390-Unix config target. It relied on a script that doesn't
3888 * rsaz_1024_mul_avx2 overflow bug on x86_64
3890 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3891 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3892 Analysis suggests that attacks against RSA and DSA as a result of this
3893 defect would be very difficult to perform and are not believed likely.
3894 Attacks against DH1024 are considered just feasible, because most of the
3895 work necessary to deduce information about a private key may be performed
3896 offline. The amount of resources required for such an attack would be
3897 significant. However, for an attack on TLS to be meaningful, the server
3898 would have to share the DH1024 private key among multiple clients, which is
3899 no longer an option since CVE-2016-0701.
3901 This only affects processors that support the AVX2 but not ADX extensions
3902 like Intel Haswell (4th generation).
3904 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3905 was originally found via the OSS-Fuzz project.
3910 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3912 * bn_sqrx8x_internal carry bug on x86_64
3914 There is a carry propagating bug in the x86_64 Montgomery squaring
3915 procedure. No EC algorithms are affected. Analysis suggests that attacks
3916 against RSA and DSA as a result of this defect would be very difficult to
3917 perform and are not believed likely. Attacks against DH are considered just
3918 feasible (although very difficult) because most of the work necessary to
3919 deduce information about a private key may be performed offline. The amount
3920 of resources required for such an attack would be very significant and
3921 likely only accessible to a limited number of attackers. An attacker would
3922 additionally need online access to an unpatched system using the target
3923 private key in a scenario with persistent DH parameters and a private
3924 key that is shared between multiple clients.
3926 This only affects processors that support the BMI1, BMI2 and ADX extensions
3927 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3929 This issue was reported to OpenSSL by the OSS-Fuzz project.
3934 * Malformed X.509 IPAddressFamily could cause OOB read
3936 If an X.509 certificate has a malformed IPAddressFamily extension,
3937 OpenSSL could do a one-byte buffer overread. The most likely result
3938 would be an erroneous display of the certificate in text format.
3940 This issue was reported to OpenSSL by the OSS-Fuzz project.
3945 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3947 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3948 platform rather than 'mingw'.
3952 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3953 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3954 which is the minimum version we support.
3958 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3960 * Encrypt-Then-Mac renegotiation crash
3962 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3963 negotiated where it was not in the original handshake (or vice-versa) then
3964 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3965 and servers are affected.
3967 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3972 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3974 * Truncated packet could crash via OOB read
3976 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3977 cipher is being used, then a truncated packet can cause that host to
3978 perform an out-of-bounds read, usually resulting in a crash.
3980 This issue was reported to OpenSSL by Robert Święcki of Google.
3985 * Bad (EC)DHE parameters cause a client crash
3987 If a malicious server supplies bad parameters for a DHE or ECDHE key
3988 exchange then this can result in the client attempting to dereference a
3989 NULL pointer leading to a client crash. This could be exploited in a Denial
3992 This issue was reported to OpenSSL by Guido Vranken.
3997 * BN_mod_exp may produce incorrect results on x86_64
3999 There is a carry propagating bug in the x86_64 Montgomery squaring
4000 procedure. No EC algorithms are affected. Analysis suggests that attacks
4001 against RSA and DSA as a result of this defect would be very difficult to
4002 perform and are not believed likely. Attacks against DH are considered just
4003 feasible (although very difficult) because most of the work necessary to
4004 deduce information about a private key may be performed offline. The amount
4005 of resources required for such an attack would be very significant and
4006 likely only accessible to a limited number of attackers. An attacker would
4007 additionally need online access to an unpatched system using the target
4008 private key in a scenario with persistent DH parameters and a private
4009 key that is shared between multiple clients. For example this can occur by
4010 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4011 similar to CVE-2015-3193 but must be treated as a separate problem.
4013 This issue was reported to OpenSSL by the OSS-Fuzz project.
4018 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4020 * ChaCha20/Poly1305 heap-buffer-overflow
4022 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4023 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4024 crash. This issue is not considered to be exploitable beyond a DoS.
4026 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4031 * CMS Null dereference
4033 Applications parsing invalid CMS structures can crash with a NULL pointer
4034 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4035 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4036 structure callback if an attempt is made to free certain invalid encodings.
4037 Only CHOICE structures using a callback which do not handle NULL value are
4040 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4045 * Montgomery multiplication may produce incorrect results
4047 There is a carry propagating bug in the Broadwell-specific Montgomery
4048 multiplication procedure that handles input lengths divisible by, but
4049 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4050 and DH private keys are impossible. This is because the subroutine in
4051 question is not used in operations with the private key itself and an input
4052 of the attacker's direct choice. Otherwise the bug can manifest itself as
4053 transient authentication and key negotiation failures or reproducible
4054 erroneous outcome of public-key operations with specially crafted input.
4055 Among EC algorithms only Brainpool P-512 curves are affected and one
4056 presumably can attack ECDH key negotiation. Impact was not analyzed in
4057 detail, because pre-requisites for attack are considered unlikely. Namely
4058 multiple clients have to choose the curve in question and the server has to
4059 share the private key among them, neither of which is default behaviour.
4060 Even then only clients that chose the curve will be affected.
4062 This issue was publicly reported as transient failures and was not
4063 initially recognized as a security issue. Thanks to Richard Morgan for
4064 providing reproducible case.
4069 * Removed automatic addition of RPATH in shared libraries and executables,
4070 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4074 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4076 * Fix Use After Free for large message sizes
4078 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4079 message larger than approx 16k is received then the underlying buffer to
4080 store the incoming message is reallocated and moved. Unfortunately a
4081 dangling pointer to the old location is left which results in an attempt to
4082 write to the previously freed location. This is likely to result in a
4083 crash, however it could potentially lead to execution of arbitrary code.
4085 This issue only affects OpenSSL 1.1.0a.
4087 This issue was reported to OpenSSL by Robert Święcki.
4092 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4094 * OCSP Status Request extension unbounded memory growth
4096 A malicious client can send an excessively large OCSP Status Request
4097 extension. If that client continually requests renegotiation, sending a
4098 large OCSP Status Request extension each time, then there will be unbounded
4099 memory growth on the server. This will eventually lead to a Denial Of
4100 Service attack through memory exhaustion. Servers with a default
4101 configuration are vulnerable even if they do not support OCSP. Builds using
4102 the "no-ocsp" build time option are not affected.
4104 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4109 * SSL_peek() hang on empty record
4111 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4112 sends an empty record. This could be exploited by a malicious peer in a
4113 Denial Of Service attack.
4115 This issue was reported to OpenSSL by Alex Gaynor.
4120 * Excessive allocation of memory in tls_get_message_header() and
4121 dtls1_preprocess_fragment()
4123 A (D)TLS message includes 3 bytes for its length in the header for the
4124 message. This would allow for messages up to 16Mb in length. Messages of
4125 this length are excessive and OpenSSL includes a check to ensure that a
4126 peer is sending reasonably sized messages in order to avoid too much memory
4127 being consumed to service a connection. A flaw in the logic of version
4128 1.1.0 means that memory for the message is allocated too early, prior to
4129 the excessive message length check. Due to way memory is allocated in
4130 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4131 to service a connection. This could lead to a Denial of Service through
4132 memory exhaustion. However, the excessive message length check still takes
4133 place, and this would cause the connection to immediately fail. Assuming
4134 that the application calls SSL_free() on the failed connection in a timely
4135 manner then the 21Mb of allocated memory will then be immediately freed
4136 again. Therefore, the excessive memory allocation will be transitory in
4137 nature. This then means that there is only a security impact if:
4139 1) The application does not call SSL_free() in a timely manner in the event
4140 that the connection fails
4142 2) The application is working in a constrained environment where there is
4143 very little free memory
4145 3) The attacker initiates multiple connection attempts such that there are
4146 multiple connections in a state where memory has been allocated for the
4147 connection; SSL_free() has not yet been called; and there is insufficient
4148 memory to service the multiple requests.
4150 Except in the instance of (1) above any Denial Of Service is likely to be
4151 transitory because as soon as the connection fails the memory is
4152 subsequently freed again in the SSL_free() call. However there is an
4153 increased risk during this period of application crashes due to the lack of
4154 memory - which would then mean a more serious Denial of Service.
4156 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4157 (CVE-2016-6307 and CVE-2016-6308)
4161 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4162 had to be removed. Primary reason is that vendor assembler can't
4163 assemble our modules with -KPIC flag. As result it, assembly
4164 support, was not even available as option. But its lack means
4165 lack of side-channel resistant code, which is incompatible with
4166 security by todays standards. Fortunately gcc is readily available
4167 prepackaged option, which we firmly point at...
4171 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4173 * Windows command-line tool supports UTF-8 opt-in option for arguments
4174 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4175 (to any value) allows Windows user to access PKCS#12 file generated
4176 with Windows CryptoAPI and protected with non-ASCII password, as well
4177 as files generated under UTF-8 locale on Linux also protected with
4182 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4183 have been disabled by default and removed from DEFAULT, just like RC4.
4184 See the RC4 item below to re-enable both.
4188 * The method for finding the storage location for the Windows RAND seed file
4189 has changed. First we check %RANDFILE%. If that is not set then we check
4190 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4191 all else fails we fall back to C:\.
4195 * The EVP_EncryptUpdate() function has had its return type changed from void
4196 to int. A return of 0 indicates and error while a return of 1 indicates
4201 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4202 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4203 off the constant time implementation for RSA, DSA and DH have been made
4204 no-ops and deprecated.
4208 * Windows RAND implementation was simplified to only get entropy by
4209 calling CryptGenRandom(). Various other RAND-related tickets
4212 *Joseph Wylie Yandle, Rich Salz*
4214 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4215 and `OPENSSL_LH_`, respectively. The old names are available
4216 with API compatibility. They new names are now completely documented.
4220 * Unify TYPE_up_ref(obj) methods signature.
4221 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4222 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4223 int (instead of void) like all others TYPE_up_ref() methods.
4224 So now these methods also check the return value of CRYPTO_atomic_add(),
4225 and the validity of object reference counter.
4227 *fdasilvayy@gmail.com*
4229 * With Windows Visual Studio builds, the .pdb files are installed
4230 alongside the installed libraries and executables. For a static
4231 library installation, ossl_static.pdb is the associate compiler
4232 generated .pdb file to be used when linking programs.
4236 * Remove openssl.spec. Packaging files belong with the packagers.
4240 * Automatic Darwin/OSX configuration has had a refresh, it will now
4241 recognise x86_64 architectures automatically. You can still decide
4242 to build for a different bitness with the environment variable
4243 KERNEL_BITS (can be 32 or 64), for example:
4245 KERNEL_BITS=32 ./config
4249 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4250 256 bit AES and HMAC with SHA256.
4254 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4258 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4262 * To enable users to have their own config files and build file templates,
4263 Configure looks in the directory indicated by the environment variable
4264 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4265 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4266 name and is used as is.
4270 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4271 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4272 X509_CERT_FILE_CTX was removed.
4276 * "shared" builds are now the default. To create only static libraries use
4277 the "no-shared" Configure option.
4281 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4282 All of these option have not worked for some while and are fundamental
4287 * Make various cleanup routines no-ops and mark them as deprecated. Most
4288 global cleanup functions are no longer required because they are handled
4289 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4290 Explicitly de-initing can cause problems (e.g. where a library that uses
4291 OpenSSL de-inits, but an application is still using it). The affected
4292 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4293 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4294 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4295 COMP_zlib_cleanup().
4299 * --strict-warnings no longer enables runtime debugging options
4300 such as REF_DEBUG. Instead, debug options are automatically
4301 enabled with '--debug' builds.
4303 *Andy Polyakov, Emilia Käsper*
4305 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4306 have been moved out of the public header files. New functions for managing
4307 these have been added.
4311 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4312 objects have been moved out of the public header files. New
4313 functions for managing these have been added.
4317 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4318 have been moved out of the public header files. New functions for managing
4319 these have been added.
4323 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4324 moved out of the public header files. New functions for managing these
4329 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4333 * Removed the mk1mf build scripts.
4337 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4338 it is always safe to #include a header now.
4342 * Removed the aged BC-32 config and all its supporting scripts
4346 * Removed support for Ultrix, Netware, and OS/2.
4350 * Add support for HKDF.
4352 *Alessandro Ghedini*
4354 * Add support for blake2b and blake2s
4358 * Added support for "pipelining". Ciphers that have the
4359 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4360 encryptions/decryptions simultaneously. There are currently no built-in
4361 ciphers with this property but the expectation is that engines will be able
4362 to offer it to significantly improve throughput. Support has been extended
4363 into libssl so that multiple records for a single connection can be
4364 processed in one go (for >=TLS 1.1).
4368 * Added the AFALG engine. This is an async capable engine which is able to
4369 offload work to the Linux kernel. In this initial version it only supports
4370 AES128-CBC. The kernel must be version 4.1.0 or greater.
4374 * OpenSSL now uses a new threading API. It is no longer necessary to
4375 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4376 are two supported threading models: pthreads and windows threads. It is
4377 also possible to configure OpenSSL at compile time for "no-threads". The
4378 old threading API should no longer be used. The functions have been
4379 replaced with "no-op" compatibility macros.
4381 *Alessandro Ghedini, Matt Caswell*
4383 * Modify behavior of ALPN to invoke callback after SNI/servername
4384 callback, such that updates to the SSL_CTX affect ALPN.
4388 * Add SSL_CIPHER queries for authentication and key-exchange.
4392 * Changes to the DEFAULT cipherlist:
4393 - Prefer (EC)DHE handshakes over plain RSA.
4394 - Prefer AEAD ciphers over legacy ciphers.
4395 - Prefer ECDSA over RSA when both certificates are available.
4396 - Prefer TLSv1.2 ciphers/PRF.
4397 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4402 * Change the ECC default curve list to be this, in order: x25519,
4403 secp256r1, secp521r1, secp384r1.
4407 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4408 disabled by default. They can be re-enabled using the
4409 enable-weak-ssl-ciphers option to Configure.
4413 * If the server has ALPN configured, but supports no protocols that the
4414 client advertises, send a fatal "no_application_protocol" alert.
4415 This behaviour is SHALL in RFC 7301, though it isn't universally
4416 implemented by other servers.
4420 * Add X25519 support.
4421 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4422 for public and private key encoding using the format documented in
4423 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4424 key generation and key derivation.
4426 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4431 * Deprecate SRP_VBASE_get_by_user.
4432 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4433 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4434 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4435 seed, even if the seed is configured.
4437 Users should use SRP_VBASE_get1_by_user instead. Note that in
4438 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4439 also that even though configuring the SRP seed attempts to hide
4440 invalid usernames by continuing the handshake with fake
4441 credentials, this behaviour is not constant time and no strong
4442 guarantees are made that the handshake is indistinguishable from
4443 that of a valid user.
4447 * Configuration change; it's now possible to build dynamic engines
4448 without having to build shared libraries and vice versa. This
4449 only applies to the engines in `engines/`, those in `crypto/engine/`
4450 will always be built into libcrypto (i.e. "static").
4452 Building dynamic engines is enabled by default; to disable, use
4453 the configuration option "disable-dynamic-engine".
4455 The only requirements for building dynamic engines are the
4456 presence of the DSO module and building with position independent
4457 code, so they will also automatically be disabled if configuring
4458 with "disable-dso" or "disable-pic".
4460 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4461 are also taken away from openssl/opensslconf.h, as they are
4466 * Configuration change; if there is a known flag to compile
4467 position independent code, it will always be applied on the
4468 libcrypto and libssl object files, and never on the application
4469 object files. This means other libraries that use routines from
4470 libcrypto / libssl can be made into shared libraries regardless
4471 of how OpenSSL was configured.
4473 If this isn't desirable, the configuration options "disable-pic"
4474 or "no-pic" can be used to disable the use of PIC. This will
4475 also disable building shared libraries and dynamic engines.
4479 * Removed JPAKE code. It was experimental and has no wide use.
4483 * The INSTALL_PREFIX Makefile variable has been renamed to
4484 DESTDIR. That makes for less confusion on what this variable
4485 is for. Also, the configuration option --install_prefix is
4490 * Heartbeat for TLS has been removed and is disabled by default
4491 for DTLS; configure with enable-heartbeats. Code that uses the
4492 old #define's might need to be updated.
4494 *Emilia Käsper, Rich Salz*
4496 * Rename REF_CHECK to REF_DEBUG.
4500 * New "unified" build system
4502 The "unified" build system is aimed to be a common system for all
4503 platforms we support. With it comes new support for VMS.
4505 This system builds supports building in a different directory tree
4506 than the source tree. It produces one Makefile (for unix family
4507 or lookalikes), or one descrip.mms (for VMS).
4509 The source of information to make the Makefile / descrip.mms is
4510 small files called 'build.info', holding the necessary
4511 information for each directory with source to compile, and a
4512 template in Configurations, like unix-Makefile.tmpl or
4515 With this change, the library names were also renamed on Windows
4516 and on VMS. They now have names that are closer to the standard
4517 on Unix, and include the major version number, and in certain
4518 cases, the architecture they are built for. See "Notes on shared
4519 libraries" in INSTALL.
4521 We rely heavily on the perl module Text::Template.
4525 * Added support for auto-initialisation and de-initialisation of the library.
4526 OpenSSL no longer requires explicit init or deinit routines to be called,
4527 except in certain circumstances. See the OPENSSL_init_crypto() and
4528 OPENSSL_init_ssl() man pages for further information.
4532 * The arguments to the DTLSv1_listen function have changed. Specifically the
4533 "peer" argument is now expected to be a BIO_ADDR object.
4535 * Rewrite of BIO networking library. The BIO library lacked consistent
4536 support of IPv6, and adding it required some more extensive
4537 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4538 which hold all types of addresses and chains of address information.
4539 It also introduces a new API, with functions like BIO_socket,
4540 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4541 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4542 have been adapted accordingly.
4546 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4551 * CRIME protection: disable compression by default, even if OpenSSL is
4552 compiled with zlib enabled. Applications can still enable compression
4553 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4554 using the SSL_CONF library to configure compression.
4558 * The signature of the session callback configured with
4559 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4560 was explicitly marked as `const unsigned char*` instead of
4565 * Always DPURIFY. Remove the use of uninitialized memory in the
4566 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4570 * Removed many obsolete configuration items, including
4571 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4572 MD2_CHAR, MD2_INT, MD2_LONG
4574 IDEA_SHORT, IDEA_LONG
4575 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4577 *Rich Salz, with advice from Andy Polyakov*
4579 * Many BN internals have been moved to an internal header file.
4581 *Rich Salz with help from Andy Polyakov*
4583 * Configuration and writing out the results from it has changed.
4584 Files such as Makefile include/openssl/opensslconf.h and are now
4585 produced through general templates, such as Makefile.in and
4586 crypto/opensslconf.h.in and some help from the perl module
4589 Also, the center of configuration information is no longer
4590 Makefile. Instead, Configure produces a perl module in
4591 configdata.pm which holds most of the config data (in the hash
4592 table %config), the target data that comes from the target
4593 configuration in one of the `Configurations/*.conf` files (in
4598 * To clarify their intended purposes, the Configure options
4599 --prefix and --openssldir change their semantics, and become more
4600 straightforward and less interdependent.
4602 --prefix shall be used exclusively to give the location INSTALLTOP
4603 where programs, scripts, libraries, include files and manuals are
4604 going to be installed. The default is now /usr/local.
4606 --openssldir shall be used exclusively to give the default
4607 location OPENSSLDIR where certificates, private keys, CRLs are
4608 managed. This is also where the default openssl.cnf gets
4610 If the directory given with this option is a relative path, the
4611 values of both the --prefix value and the --openssldir value will
4612 be combined to become OPENSSLDIR.
4613 The default for --openssldir is INSTALLTOP/ssl.
4615 Anyone who uses --openssldir to specify where OpenSSL is to be
4616 installed MUST change to use --prefix instead.
4620 * The GOST engine was out of date and therefore it has been removed. An up
4621 to date GOST engine is now being maintained in an external repository.
4622 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4623 support for GOST ciphersuites (these are only activated if a GOST engine
4628 * EGD is no longer supported by default; use enable-egd when
4631 *Ben Kaduk and Rich Salz*
4633 * The distribution now has Makefile.in files, which are used to
4634 create Makefile's when Configure is run. *Configure must be run
4635 before trying to build now.*
4639 * The return value for SSL_CIPHER_description() for error conditions
4644 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4646 Obtaining and performing DNSSEC validation of TLSA records is
4647 the application's responsibility. The application provides
4648 the TLSA records of its choice to OpenSSL, and these are then
4649 used to authenticate the peer.
4651 The TLSA records need not even come from DNS. They can, for
4652 example, be used to implement local end-entity certificate or
4653 trust-anchor "pinning", where the "pin" data takes the form
4654 of TLSA records, which can augment or replace verification
4655 based on the usual WebPKI public certification authorities.
4659 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4660 continues to support deprecated interfaces in default builds.
4661 However, applications are strongly advised to compile their
4662 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4663 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4664 or the 1.1.0 releases.
4666 In environments in which all applications have been ported to
4667 not use any deprecated interfaces OpenSSL's Configure script
4668 should be used with the --api=1.1.0 option to entirely remove
4669 support for the deprecated features from the library and
4670 unconditionally disable them in the installed headers.
4671 Essentially the same effect can be achieved with the "no-deprecated"
4672 argument to Configure, except that this will always restrict
4673 the build to just the latest API, rather than a fixed API
4676 As applications are ported to future revisions of the API,
4677 they should update their compile-time OPENSSL_API_COMPAT define
4678 accordingly, but in most cases should be able to continue to
4679 compile with later releases.
4681 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4682 0x10000000L and 0x00908000L, respectively. However those
4683 versions did not support the OPENSSL_API_COMPAT feature, and
4684 so applications are not typically tested for explicit support
4685 of just the undeprecated features of either release.
4689 * Add support for setting the minimum and maximum supported protocol.
4690 It can bet set via the SSL_set_min_proto_version() and
4691 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4692 MaxProtocol. It's recommended to use the new APIs to disable
4693 protocols instead of disabling individual protocols using
4694 SSL_set_options() or SSL_CONF's Protocol. This change also
4695 removes support for disabling TLS 1.2 in the OpenSSL TLS
4696 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4700 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4704 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4705 and integrates ECDSA and ECDH functionality into EC. Implementations can
4706 now redirect key generation and no longer need to convert to or from
4709 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4710 include the ec.h header file instead.
4714 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4715 ciphers who are no longer supported and drops support the ephemeral RSA key
4716 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4720 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4721 opaque. For HMAC_CTX, the following constructors and destructors
4724 HMAC_CTX *HMAC_CTX_new(void);
4725 void HMAC_CTX_free(HMAC_CTX *ctx);
4727 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4728 destroy such methods has been added. See EVP_MD_meth_new(3) and
4729 EVP_CIPHER_meth_new(3) for documentation.
4732 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4733 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4734 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4735 an already created structure.
4736 2) For consistency with the majority of our object creators and
4737 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4738 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4739 for deprecated builds.
4743 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4744 cryptographic operations to be performed asynchronously as long as an
4745 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4746 further details. Libssl has also had this capability integrated with the
4747 introduction of the new mode SSL_MODE_ASYNC and associated error
4748 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4749 pages. This work was developed in partnership with Intel Corp.
4753 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4754 always enabled now. If you want to disable the support you should
4755 exclude it using the list of supported ciphers. This also means that the
4756 "-no_ecdhe" option has been removed from s_server.
4760 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4761 SSL_{CTX_}set1_curves() which can set a list.
4765 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4766 curve you want to support using SSL_{CTX_}set1_curves().
4770 * State machine rewrite. The state machine code has been significantly
4771 refactored in order to remove much duplication of code and solve issues
4772 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4773 further details). This change does have some associated API changes.
4774 Notably the SSL_state() function has been removed and replaced by
4775 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4776 SSL_set_state() has been removed altogether. The previous handshake states
4777 defined in ssl.h and ssl3.h have also been removed.
4781 * All instances of the string "ssleay" in the public API were replaced
4782 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4783 Some error codes related to internal RSA_eay API's were renamed.
4787 * The demo files in crypto/threads were moved to demo/threads.
4791 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4794 *Matt Caswell, Rich Salz*
4796 * New ASN.1 embed macro.
4798 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4799 structure is not allocated: it is part of the parent. That is instead of
4807 This reduces memory fragmentation and make it impossible to accidentally
4808 set a mandatory field to NULL.
4810 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4811 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4812 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4817 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4821 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4822 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4823 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4824 DES and RC4 ciphersuites.
4828 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4829 This changes the decoding behaviour for some invalid messages,
4830 though the change is mostly in the more lenient direction, and
4831 legacy behaviour is preserved as much as possible.
4835 * Fix no-stdio build.
4836 *David Woodhouse <David.Woodhouse@intel.com> and also*
4837 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4839 * New testing framework
4840 The testing framework has been largely rewritten and is now using
4841 perl and the perl modules Test::Harness and an extended variant of
4842 Test::More called OpenSSL::Test to do its work. All test scripts in
4843 test/ have been rewritten into test recipes, and all direct calls to
4844 executables in test/Makefile have become individual recipes using the
4845 simplified testing OpenSSL::Test::Simple.
4847 For documentation on our testing modules, do:
4849 perldoc test/testlib/OpenSSL/Test/Simple.pm
4850 perldoc test/testlib/OpenSSL/Test.pm
4854 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4855 are used; the latter aborts on memory leaks (usually checked on exit).
4856 Some undocumented "set malloc, etc., hooks" functions were removed
4857 and others were changed. All are now documented.
4861 * In DSA_generate_parameters_ex, if the provided seed is too short,
4864 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4866 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4867 from RFC4279, RFC4785, RFC5487, RFC5489.
4869 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4870 original RSA_PSK patch.
4874 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4875 era flag was never set throughout the codebase (only read). Also removed
4876 SSL3_FLAGS_POP_BUFFER which was only used if
4877 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4881 * Changed the default name options in the "ca", "crl", "req" and "x509"
4882 to be "oneline" instead of "compat".
4886 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4887 not aware of clients that still exhibit this bug, and the workaround
4888 hasn't been working properly for a while.
4892 * The return type of BIO_number_read() and BIO_number_written() as well as
4893 the corresponding num_read and num_write members in the BIO structure has
4894 changed from unsigned long to uint64_t. On platforms where an unsigned
4895 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4900 * Given the pervasive nature of TLS extensions it is inadvisable to run
4901 OpenSSL without support for them. It also means that maintaining
4902 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4903 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4907 * Removed support for the two export grade static DH ciphersuites
4908 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4909 were newly added (along with a number of other static DH ciphersuites) to
4910 1.0.2. However the two export ones have *never* worked since they were
4911 introduced. It seems strange in any case to be adding new export
4912 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4916 * Version negotiation has been rewritten. In particular SSLv23_method(),
4917 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4918 and turned into macros which simply call the new preferred function names
4919 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4920 should use the new names instead. Also as part of this change the ssl23.h
4921 header file has been removed.
4925 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4926 code and the associated standard is no longer considered fit-for-purpose.
4930 * RT2547 was closed. When generating a private key, try to make the
4931 output file readable only by the owner. This behavior change might
4932 be noticeable when interacting with other software.
4934 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4939 * Added HTTP GET support to the ocsp command.
4943 * Changed default digest for the dgst and enc commands from MD5 to
4948 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4952 * Added support for TLS extended master secret from
4953 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4954 initial patch which was a great help during development.
4958 * All libssl internal structures have been removed from the public header
4959 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4960 now redundant). Users should not attempt to access internal structures
4961 directly. Instead they should use the provided API functions.
4965 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4966 Access to deprecated functions can be re-enabled by running config with
4967 "enable-deprecated". In addition applications wishing to use deprecated
4968 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4969 will, by default, disable some transitive includes that previously existed
4970 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4974 * Added support for OCB mode. OpenSSL has been granted a patent license
4975 compatible with the OpenSSL license for use of OCB. Details are available
4976 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4977 for OCB can be removed by calling config with no-ocb.
4981 * SSLv2 support has been removed. It still supports receiving an SSLv2
4982 compatible client hello.
4986 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4987 done while fixing the error code for the key-too-small case.
4989 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4991 * CA.sh has been removed; use CA.pl instead.
4995 * Removed old DES API.
4999 * Remove various unsupported platforms:
5005 Sinix/ReliantUNIX RM400
5010 16-bit platforms such as WIN16
5014 * Clean up OPENSSL_NO_xxx #define's
5015 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5016 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5017 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5018 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5019 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5020 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5021 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5022 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5023 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5024 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5028 * Cleaned up dead code
5029 Remove all but one '#ifdef undef' which is to be looked at.
5033 * Clean up calling of xxx_free routines.
5034 Just like free(), fix most of the xxx_free routines to accept
5035 NULL. Remove the non-null checks from callers. Save much code.
5039 * Add secure heap for storage of private keys (when possible).
5040 Add BIO_s_secmem(), CBIGNUM, etc.
5041 Contributed by Akamai Technologies under our Corporate CLA.
5045 * Experimental support for a new, fast, unbiased prime candidate generator,
5046 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5048 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5050 * New output format NSS in the sess_id command line tool. This allows
5051 exporting the session id and the master key in NSS keylog format.
5053 *Martin Kaiser <martin@kaiser.cx>*
5055 * Harmonize version and its documentation. -f flag is used to display
5058 *mancha <mancha1@zoho.com>*
5060 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5061 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5063 *mancha <mancha1@zoho.com>*
5065 * Fix some double frees. These are not thought to be exploitable.
5067 *mancha <mancha1@zoho.com>*
5069 * A missing bounds check in the handling of the TLS heartbeat extension
5070 can be used to reveal up to 64k of memory to a connected client or
5073 Thanks for Neel Mehta of Google Security for discovering this bug and to
5074 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5075 preparing the fix ([CVE-2014-0160])
5077 *Adam Langley, Bodo Moeller*
5079 * Fix for the attack described in the paper "Recovering OpenSSL
5080 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5081 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5082 <http://eprint.iacr.org/2014/140>
5084 Thanks to Yuval Yarom and Naomi Benger for discovering this
5085 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5087 *Yuval Yarom and Naomi Benger*
5089 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5090 this fixes a limitation in previous versions of OpenSSL.
5094 * Experimental encrypt-then-mac support.
5096 Experimental support for encrypt then mac from
5097 draft-gutmann-tls-encrypt-then-mac-02.txt
5099 To enable it set the appropriate extension number (0x42 for the test
5100 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5102 For non-compliant peers (i.e. just about everything) this should have no
5105 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5109 * Add EVP support for key wrapping algorithms, to avoid problems with
5110 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5111 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5112 algorithms and include tests cases.
5116 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5121 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5122 MGF1 digest and OAEP label.
5126 * Make openssl verify return errors.
5128 *Chris Palmer <palmer@google.com> and Ben Laurie*
5130 * New function ASN1_TIME_diff to calculate the difference between two
5131 ASN1_TIME structures or one structure and the current time.
5135 * Update fips_test_suite to support multiple command line options. New
5136 test to induce all self test errors in sequence and check expected
5141 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5142 sign or verify all in one operation.
5146 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5147 test programs and fips_test_suite. Includes functionality to parse
5148 the minimal script output of fipsalgest.pl directly.
5152 * Add authorisation parameter to FIPS_module_mode_set().
5156 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5160 * Use separate DRBG fields for internal and external flags. New function
5161 FIPS_drbg_health_check() to perform on demand health checking. Add
5162 generation tests to fips_test_suite with reduced health check interval to
5163 demonstrate periodic health checking. Add "nodh" option to
5164 fips_test_suite to skip very slow DH test.
5168 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5173 * More extensive health check for DRBG checking many more failure modes.
5174 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5175 combination: call this in fips_test_suite.
5179 * Add support for canonical generation of DSA parameter 'g'. See
5182 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5183 POST to handle HMAC cases.
5187 * Add functions FIPS_module_version() and FIPS_module_version_text()
5188 to return numerical and string versions of the FIPS module number.
5192 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5193 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5194 outside the validated module in the FIPS capable OpenSSL.
5198 * Minor change to DRBG entropy callback semantics. In some cases
5199 there is no multiple of the block length between min_len and
5200 max_len. Allow the callback to return more than max_len bytes
5201 of entropy but discard any extra: it is the callback's responsibility
5202 to ensure that the extra data discarded does not impact the
5203 requested amount of entropy.
5207 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5208 information in FIPS186-3, SP800-57 and SP800-131A.
5212 * CCM support via EVP. Interface is very similar to GCM case except we
5213 must supply all data in one chunk (i.e. no update, final) and the
5214 message length must be supplied if AAD is used. Add algorithm test
5219 * Initial version of POST overhaul. Add POST callback to allow the status
5220 of POST to be monitored and/or failures induced. Modify fips_test_suite
5221 to use callback. Always run all selftests even if one fails.
5225 * XTS support including algorithm test driver in the fips_gcmtest program.
5226 Note: this does increase the maximum key length from 32 to 64 bytes but
5227 there should be no binary compatibility issues as existing applications
5228 will never use XTS mode.
5232 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5233 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5234 performs algorithm blocking for unapproved PRNG types. Also do not
5235 set PRNG type in FIPS_mode_set(): leave this to the application.
5236 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5237 the standard OpenSSL PRNG: set additional data to a date time vector.
5241 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5242 This shouldn't present any incompatibility problems because applications
5243 shouldn't be using these directly and any that are will need to rethink
5244 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5248 * Extensive self tests and health checking required by SP800-90 DRBG.
5249 Remove strength parameter from FIPS_drbg_instantiate and always
5250 instantiate at maximum supported strength.
5254 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5258 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5262 * New function DH_compute_key_padded() to compute a DH key and pad with
5263 leading zeroes if needed: this complies with SP800-56A et al.
5267 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5268 anything, incomplete, subject to change and largely untested at present.
5272 * Modify fipscanisteronly build option to only build the necessary object
5273 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5277 * Add experimental option FIPSSYMS to give all symbols in
5278 fipscanister.o and FIPS or fips prefix. This will avoid
5279 conflicts with future versions of OpenSSL. Add perl script
5280 util/fipsas.pl to preprocess assembly language source files
5281 and rename any affected symbols.
5285 * Add selftest checks and algorithm block of non-fips algorithms in
5286 FIPS mode. Remove DES2 from selftests.
5290 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5291 return internal method without any ENGINE dependencies. Add new
5292 tiny fips sign and verify functions.
5296 * New build option no-ec2m to disable characteristic 2 code.
5300 * New build option "fipscanisteronly". This only builds fipscanister.o
5301 and (currently) associated fips utilities. Uses the file Makefile.fips
5302 instead of Makefile.org as the prototype.
5306 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5307 Update fips_gcmtest to use IV generator.
5311 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5312 setting output buffer to NULL. The `*Final` function must be
5313 called although it will not retrieve any additional data. The tag
5314 can be set or retrieved with a ctrl. The IV length is by default 12
5315 bytes (96 bits) but can be set to an alternative value. If the IV
5316 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5321 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5322 underlying do_cipher function handles all cipher semantics itself
5323 including padding and finalisation. This is useful if (for example)
5324 an ENGINE cipher handles block padding itself. The behaviour of
5325 do_cipher is subtly changed if this flag is set: the return value
5326 is the number of characters written to the output buffer (zero is
5327 no longer an error code) or a negative error code. Also if the
5328 input buffer is NULL and length 0 finalisation should be performed.
5332 * If a candidate issuer certificate is already part of the constructed
5333 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5337 * Improve forward-security support: add functions
5339 void SSL_CTX_set_not_resumable_session_callback(
5340 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5341 void SSL_set_not_resumable_session_callback(
5342 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5344 for use by SSL/TLS servers; the callback function will be called whenever a
5345 new session is created, and gets to decide whether the session may be
5346 cached to make it resumable (return 0) or not (return 1). (As by the
5347 SSL/TLS protocol specifications, the session_id sent by the server will be
5348 empty to indicate that the session is not resumable; also, the server will
5349 not generate RFC 4507 (RFC 5077) session tickets.)
5351 A simple reasonable callback implementation is to return is_forward_secure.
5352 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5353 by the SSL/TLS server library, indicating whether it can provide forward
5356 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5358 * New -verify_name option in command line utilities to set verification
5363 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5364 Add CMAC pkey methods.
5368 * Experimental renegotiation in s_server -www mode. If the client
5369 browses /reneg connection is renegotiated. If /renegcert it is
5370 renegotiated requesting a certificate.
5374 * Add an "external" session cache for debugging purposes to s_server. This
5375 should help trace issues which normally are only apparent in deployed
5376 multi-process servers.
5380 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5381 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5382 BIO_set_cipher() and some obscure PEM functions were changed so they
5383 can now return an error. The RAND changes required a change to the
5384 RAND_METHOD structure.
5388 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5389 a gcc attribute to warn if the result of a function is ignored. This
5390 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5391 whose return value is often ignored.
5395 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5396 These allow SCTs (signed certificate timestamps) to be requested and
5397 validated when establishing a connection.
5399 *Rob Percival <robpercival@google.com>*
5404 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5406 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5407 used even when parsing explicit parameters, when loading a encoded key
5408 or calling `EC_GROUP_new_from_ecpkparameters()`/
5409 `EC_GROUP_new_from_ecparameters()`.
5410 This prevents bypass of security hardening and performance gains,
5411 especially for curves with specialized EC_METHODs.
5412 By default, if a key encoded with explicit parameters is loaded and later
5413 encoded, the output is still encoded with explicit parameters, even if
5414 internally a "named" EC_GROUP is used for computation.
5418 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5419 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5420 NULL. After this change, only the cofactor parameter can be NULL. It also
5421 does some minimal sanity checks on the passed order.
5426 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5427 An attack is simple, if the first CMS_recipientInfo is valid but the
5428 second CMS_recipientInfo is chosen ciphertext. If the second
5429 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5430 encryption key will be replaced by garbage, and the message cannot be
5431 decoded, but if the RSA decryption fails, the correct encryption key is
5432 used and the recipient will not notice the attack.
5433 As a work around for this potential attack the length of the decrypted
5434 key must be equal to the cipher default key length, in case the
5435 certificate is not given and all recipientInfo are tried out.
5436 The old behaviour can be re-enabled in the CMS code by setting the
5437 CMS_DEBUG_DECRYPT flag.
5442 * Document issue with installation paths in diverse Windows builds
5444 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5445 binaries and run-time config file.
5450 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5452 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5453 This changes the size when using the `genpkey` command when no size is given.
5454 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5455 generation commands to use 2048 bits by default.
5459 * Add FIPS support for Android Arm 64-bit
5461 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5462 Module in Version 2.0.10. For some reason, the corresponding target
5463 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5464 built with FIPS support on Android Arm 64-bit. This omission has been
5467 *Matthias St. Pierre*
5469 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5471 * 0-byte record padding oracle
5473 If an application encounters a fatal protocol error and then calls
5474 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5475 then OpenSSL can respond differently to the calling application if a 0 byte
5476 record is received with invalid padding compared to if a 0 byte record is
5477 received with an invalid MAC. If the application then behaves differently
5478 based on that in a way that is detectable to the remote peer, then this
5479 amounts to a padding oracle that could be used to decrypt data.
5481 In order for this to be exploitable "non-stitched" ciphersuites must be in
5482 use. Stitched ciphersuites are optimised implementations of certain
5483 commonly used ciphersuites. Also the application must call SSL_shutdown()
5484 twice even if a protocol error has occurred (applications should not do
5485 this but some do anyway).
5487 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5488 Aviram, with additional investigation by Steven Collison and Andrew
5489 Hourselt. It was reported to OpenSSL on 10th December 2018.
5494 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5498 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5500 * Microarchitecture timing vulnerability in ECC scalar multiplication
5502 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5503 shown to be vulnerable to a microarchitecture timing side channel attack.
5504 An attacker with sufficient access to mount local timing attacks during
5505 ECDSA signature generation could recover the private key.
5507 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5508 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5514 * Timing vulnerability in DSA signature generation
5516 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5517 timing side channel attack. An attacker could use variations in the signing
5518 algorithm to recover the private key.
5520 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5525 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5526 Module, accidentally introduced while backporting security fixes from the
5527 development branch and hindering the use of ECC in FIPS mode.
5531 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5533 * Client DoS due to large DH parameter
5535 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5536 malicious server can send a very large prime value to the client. This will
5537 cause the client to spend an unreasonably long period of time generating a
5538 key for this prime resulting in a hang until the client has finished. This
5539 could be exploited in a Denial Of Service attack.
5541 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5546 * Cache timing vulnerability in RSA Key Generation
5548 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5549 a cache timing side channel attack. An attacker with sufficient access to
5550 mount cache timing attacks during the RSA key generation process could
5551 recover the private key.
5553 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5554 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5559 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5560 parameter is no longer accepted, as it leads to a corrupt table. NULL
5561 pem_str is reserved for alias entries only.
5565 * Revert blinding in ECDSA sign and instead make problematic addition
5566 length-invariant. Switch even to fixed-length Montgomery multiplication.
5570 * Change generating and checking of primes so that the error rate of not
5571 being prime depends on the intended use based on the size of the input.
5572 For larger primes this will result in more rounds of Miller-Rabin.
5573 The maximal error rate for primes with more than 1080 bits is lowered
5576 *Kurt Roeckx, Annie Yousar*
5578 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5582 * Add blinding to ECDSA and DSA signatures to protect against side channel
5583 attacks discovered by Keegan Ryan (NCC Group).
5587 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5588 now allow empty (zero character) pass phrases.
5592 * Certificate time validation (X509_cmp_time) enforces stricter
5593 compliance with RFC 5280. Fractional seconds and timezone offsets
5594 are no longer allowed.
5598 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5600 * Constructed ASN.1 types with a recursive definition could exceed the stack
5602 Constructed ASN.1 types with a recursive definition (such as can be found
5603 in PKCS7) could eventually exceed the stack given malicious input with
5604 excessive recursion. This could result in a Denial Of Service attack. There
5605 are no such structures used within SSL/TLS that come from untrusted sources
5606 so this is considered safe.
5608 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5614 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5616 * Read/write after SSL object in error state
5618 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5619 mechanism. The intent was that if a fatal error occurred during a handshake
5620 then OpenSSL would move into the error state and would immediately fail if
5621 you attempted to continue the handshake. This works as designed for the
5622 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5623 SSL_connect()), however due to a bug it does not work correctly if
5624 SSL_read() or SSL_write() is called directly. In that scenario, if the
5625 handshake fails then a fatal error will be returned in the initial function
5626 call. If SSL_read()/SSL_write() is subsequently called by the application
5627 for the same SSL object then it will succeed and the data is passed without
5628 being decrypted/encrypted directly from the SSL/TLS record layer.
5630 In order to exploit this issue an application bug would have to be present
5631 that resulted in a call to SSL_read()/SSL_write() being issued after having
5632 already received a fatal error.
5634 This issue was reported to OpenSSL by David Benjamin (Google).
5639 * rsaz_1024_mul_avx2 overflow bug on x86_64
5641 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5642 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5643 Analysis suggests that attacks against RSA and DSA as a result of this
5644 defect would be very difficult to perform and are not believed likely.
5645 Attacks against DH1024 are considered just feasible, because most of the
5646 work necessary to deduce information about a private key may be performed
5647 offline. The amount of resources required for such an attack would be
5648 significant. However, for an attack on TLS to be meaningful, the server
5649 would have to share the DH1024 private key among multiple clients, which is
5650 no longer an option since CVE-2016-0701.
5652 This only affects processors that support the AVX2 but not ADX extensions
5653 like Intel Haswell (4th generation).
5655 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5656 was originally found via the OSS-Fuzz project.
5661 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5663 * bn_sqrx8x_internal carry bug on x86_64
5665 There is a carry propagating bug in the x86_64 Montgomery squaring
5666 procedure. No EC algorithms are affected. Analysis suggests that attacks
5667 against RSA and DSA as a result of this defect would be very difficult to
5668 perform and are not believed likely. Attacks against DH are considered just
5669 feasible (although very difficult) because most of the work necessary to
5670 deduce information about a private key may be performed offline. The amount
5671 of resources required for such an attack would be very significant and
5672 likely only accessible to a limited number of attackers. An attacker would
5673 additionally need online access to an unpatched system using the target
5674 private key in a scenario with persistent DH parameters and a private
5675 key that is shared between multiple clients.
5677 This only affects processors that support the BMI1, BMI2 and ADX extensions
5678 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5680 This issue was reported to OpenSSL by the OSS-Fuzz project.
5685 * Malformed X.509 IPAddressFamily could cause OOB read
5687 If an X.509 certificate has a malformed IPAddressFamily extension,
5688 OpenSSL could do a one-byte buffer overread. The most likely result
5689 would be an erroneous display of the certificate in text format.
5691 This issue was reported to OpenSSL by the OSS-Fuzz project.
5695 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5697 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5698 platform rather than 'mingw'.
5702 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5704 * Truncated packet could crash via OOB read
5706 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5707 cipher is being used, then a truncated packet can cause that host to
5708 perform an out-of-bounds read, usually resulting in a crash.
5710 This issue was reported to OpenSSL by Robert Święcki of Google.
5715 * BN_mod_exp may produce incorrect results on x86_64
5717 There is a carry propagating bug in the x86_64 Montgomery squaring
5718 procedure. No EC algorithms are affected. Analysis suggests that attacks
5719 against RSA and DSA as a result of this defect would be very difficult to
5720 perform and are not believed likely. Attacks against DH are considered just
5721 feasible (although very difficult) because most of the work necessary to
5722 deduce information about a private key may be performed offline. The amount
5723 of resources required for such an attack would be very significant and
5724 likely only accessible to a limited number of attackers. An attacker would
5725 additionally need online access to an unpatched system using the target
5726 private key in a scenario with persistent DH parameters and a private
5727 key that is shared between multiple clients. For example this can occur by
5728 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5729 similar to CVE-2015-3193 but must be treated as a separate problem.
5731 This issue was reported to OpenSSL by the OSS-Fuzz project.
5736 * Montgomery multiplication may produce incorrect results
5738 There is a carry propagating bug in the Broadwell-specific Montgomery
5739 multiplication procedure that handles input lengths divisible by, but
5740 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5741 and DH private keys are impossible. This is because the subroutine in
5742 question is not used in operations with the private key itself and an input
5743 of the attacker's direct choice. Otherwise the bug can manifest itself as
5744 transient authentication and key negotiation failures or reproducible
5745 erroneous outcome of public-key operations with specially crafted input.
5746 Among EC algorithms only Brainpool P-512 curves are affected and one
5747 presumably can attack ECDH key negotiation. Impact was not analyzed in
5748 detail, because pre-requisites for attack are considered unlikely. Namely
5749 multiple clients have to choose the curve in question and the server has to
5750 share the private key among them, neither of which is default behaviour.
5751 Even then only clients that chose the curve will be affected.
5753 This issue was publicly reported as transient failures and was not
5754 initially recognized as a security issue. Thanks to Richard Morgan for
5755 providing reproducible case.
5760 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5761 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5762 prevent issues where no progress is being made and the peer continually
5763 sends unrecognised record types, using up resources processing them.
5767 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5769 * Missing CRL sanity check
5771 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5772 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5773 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5775 This issue only affects the OpenSSL 1.0.2i
5780 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5782 * OCSP Status Request extension unbounded memory growth
5784 A malicious client can send an excessively large OCSP Status Request
5785 extension. If that client continually requests renegotiation, sending a
5786 large OCSP Status Request extension each time, then there will be unbounded
5787 memory growth on the server. This will eventually lead to a Denial Of
5788 Service attack through memory exhaustion. Servers with a default
5789 configuration are vulnerable even if they do not support OCSP. Builds using
5790 the "no-ocsp" build time option are not affected.
5792 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5797 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5800 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5806 * OOB write in MDC2_Update()
5808 An overflow can occur in MDC2_Update() either if called directly or
5809 through the EVP_DigestUpdate() function using MDC2. If an attacker
5810 is able to supply very large amounts of input data after a previous
5811 call to EVP_EncryptUpdate() with a partial block then a length check
5812 can overflow resulting in a heap corruption.
5814 The amount of data needed is comparable to SIZE_MAX which is impractical
5817 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5822 * Malformed SHA512 ticket DoS
5824 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5825 DoS attack where a malformed ticket will result in an OOB read which will
5828 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5829 a custom server callback and ticket lookup mechanism.
5831 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5836 * OOB write in BN_bn2dec()
5838 The function BN_bn2dec() does not check the return value of BN_div_word().
5839 This can cause an OOB write if an application uses this function with an
5840 overly large BIGNUM. This could be a problem if an overly large certificate
5841 or CRL is printed out from an untrusted source. TLS is not affected because
5842 record limits will reject an oversized certificate before it is parsed.
5844 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5849 * OOB read in TS_OBJ_print_bio()
5851 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5852 the total length the OID text representation would use and not the amount
5853 of data written. This will result in OOB reads when large OIDs are
5856 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5861 * Pointer arithmetic undefined behaviour
5863 Avoid some undefined pointer arithmetic
5865 A common idiom in the codebase is to check limits in the following manner:
5868 Where "p" points to some malloc'd data of SIZE bytes and
5871 "len" here could be from some externally supplied data (e.g. from a TLS
5874 The rules of C pointer arithmetic are such that "p + len" is only well
5875 defined where len <= SIZE. Therefore the above idiom is actually
5876 undefined behaviour.
5878 For example this could cause problems if some malloc implementation
5879 provides an address for "p" such that "p + len" actually overflows for
5880 values of len that are too big and therefore p + len < limit.
5882 This issue was reported to OpenSSL by Guido Vranken
5887 * Constant time flag not preserved in DSA signing
5889 Operations in the DSA signing algorithm should run in constant time in
5890 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5891 implementation means that a non-constant time codepath is followed for
5892 certain operations. This has been demonstrated through a cache-timing
5893 attack to be sufficient for an attacker to recover the private DSA key.
5895 This issue was reported by César Pereida (Aalto University), Billy Brumley
5896 (Tampere University of Technology), and Yuval Yarom (The University of
5897 Adelaide and NICTA).
5902 * DTLS buffered message DoS
5904 In a DTLS connection where handshake messages are delivered out-of-order
5905 those messages that OpenSSL is not yet ready to process will be buffered
5906 for later use. Under certain circumstances, a flaw in the logic means that
5907 those messages do not get removed from the buffer even though the handshake
5908 has been completed. An attacker could force up to approx. 15 messages to
5909 remain in the buffer when they are no longer required. These messages will
5910 be cleared when the DTLS connection is closed. The default maximum size for
5911 a message is 100k. Therefore, the attacker could force an additional 1500k
5912 to be consumed per connection. By opening many simultaneous connections an
5913 attacker could cause a DoS attack through memory exhaustion.
5915 This issue was reported to OpenSSL by Quan Luo.
5920 * DTLS replay protection DoS
5922 A flaw in the DTLS replay attack protection mechanism means that records
5923 that arrive for future epochs update the replay protection "window" before
5924 the MAC for the record has been validated. This could be exploited by an
5925 attacker by sending a record for the next epoch (which does not have to
5926 decrypt or have a valid MAC), with a very large sequence number. This means
5927 that all subsequent legitimate packets are dropped causing a denial of
5928 service for a specific DTLS connection.
5930 This issue was reported to OpenSSL by the OCAP audit team.
5935 * Certificate message OOB reads
5937 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5938 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5939 theoretical DoS risk but this has not been observed in practice on common
5942 The messages affected are client certificate, client certificate request
5943 and server certificate. As a result the attack can only be performed
5944 against a client or a server which enables client authentication.
5946 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5951 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5953 * Prevent padding oracle in AES-NI CBC MAC check
5955 A MITM attacker can use a padding oracle attack to decrypt traffic
5956 when the connection uses an AES CBC cipher and the server support
5959 This issue was introduced as part of the fix for Lucky 13 padding
5960 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5961 constant time by making sure that always the same bytes are read and
5962 compared against either the MAC or padding bytes. But it no longer
5963 checked that there was enough data to have both the MAC and padding
5966 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5970 * Fix EVP_EncodeUpdate overflow
5972 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5973 Base64 encoding of binary data. If an attacker is able to supply very large
5974 amounts of input data then a length check can overflow resulting in a heap
5977 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5978 the `PEM_write_bio*` family of functions. These are mainly used within the
5979 OpenSSL command line applications, so any application which processes data
5980 from an untrusted source and outputs it as a PEM file should be considered
5981 vulnerable to this issue. User applications that call these APIs directly
5982 with large amounts of untrusted data may also be vulnerable.
5984 This issue was reported by Guido Vranken.
5989 * Fix EVP_EncryptUpdate overflow
5991 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5992 is able to supply very large amounts of input data after a previous call to
5993 EVP_EncryptUpdate() with a partial block then a length check can overflow
5994 resulting in a heap corruption. Following an analysis of all OpenSSL
5995 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5996 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5997 the first called function after an EVP_EncryptInit(), and therefore that
5998 specific call must be safe. The second form is where the length passed to
5999 EVP_EncryptUpdate() can be seen from the code to be some small value and
6000 therefore there is no possibility of an overflow. Since all instances are
6001 one of these two forms, it is believed that there can be no overflows in
6002 internal code due to this problem. It should be noted that
6003 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6004 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6005 of these calls have also been analysed too and it is believed there are no
6006 instances in internal usage where an overflow could occur.
6008 This issue was reported by Guido Vranken.
6013 * Prevent ASN.1 BIO excessive memory allocation
6015 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6016 a short invalid encoding can cause allocation of large amounts of memory
6017 potentially consuming excessive resources or exhausting memory.
6019 Any application parsing untrusted data through d2i BIO functions is
6020 affected. The memory based functions such as d2i_X509() are *not* affected.
6021 Since the memory based functions are used by the TLS library, TLS
6022 applications are not affected.
6024 This issue was reported by Brian Carpenter.
6031 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6032 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6033 in arbitrary stack data being returned in the buffer.
6035 This issue was reported by Guido Vranken.
6040 * Modify behavior of ALPN to invoke callback after SNI/servername
6041 callback, such that updates to the SSL_CTX affect ALPN.
6045 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6050 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6051 methods are enabled and ssl2 is disabled the methods return NULL.
6055 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6057 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6058 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6059 provide any "EXPORT" or "LOW" strength ciphers.
6063 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6064 is by default disabled at build-time. Builds that are not configured with
6065 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6066 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6067 will need to explicitly call either of:
6069 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6071 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6073 as appropriate. Even if either of those is used, or the application
6074 explicitly uses the version-specific SSLv2_method() or its client and
6075 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6076 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6077 ciphers, and SSLv2 56-bit DES are no longer available.
6082 * Fix a double-free in DSA code
6084 A double free bug was discovered when OpenSSL parses malformed DSA private
6085 keys and could lead to a DoS attack or memory corruption for applications
6086 that receive DSA private keys from untrusted sources. This scenario is
6089 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6095 * Disable SRP fake user seed to address a server memory leak.
6097 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6099 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6100 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6101 was changed to ignore the "fake user" SRP seed, even if the seed
6104 Users should use SRP_VBASE_get1_by_user instead. Note that in
6105 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6106 also that even though configuring the SRP seed attempts to hide
6107 invalid usernames by continuing the handshake with fake
6108 credentials, this behaviour is not constant time and no strong
6109 guarantees are made that the handshake is indistinguishable from
6110 that of a valid user.
6115 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6117 In the BN_hex2bn function the number of hex digits is calculated using an
6118 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6119 large values of `i` this can result in `bn_expand` not allocating any
6120 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6121 field as NULL leading to a subsequent NULL ptr deref. For very large values
6122 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6123 In this case memory is allocated to the internal BIGNUM data field, but it
6124 is insufficiently sized leading to heap corruption. A similar issue exists
6125 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6126 is ever called by user applications with very large untrusted hex/dec data.
6127 This is anticipated to be a rare occurrence.
6129 All OpenSSL internal usage of these functions use data that is not expected
6130 to be untrusted, e.g. config file data or application command line
6131 arguments. If user developed applications generate config file data based
6132 on untrusted data then it is possible that this could also lead to security
6133 consequences. This is also anticipated to be rare.
6135 This issue was reported to OpenSSL by Guido Vranken.
6140 * Fix memory issues in `BIO_*printf` functions
6142 The internal `fmtstr` function used in processing a "%s" format string in
6143 the `BIO_*printf` functions could overflow while calculating the length of a
6144 string and cause an OOB read when printing very long strings.
6146 Additionally the internal `doapr_outch` function can attempt to write to an
6147 OOB memory location (at an offset from the NULL pointer) in the event of a
6148 memory allocation failure. In 1.0.2 and below this could be caused where
6149 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6150 could be in processing a very long "%s" format string. Memory leaks can
6153 The first issue may mask the second issue dependent on compiler behaviour.
6154 These problems could enable attacks where large amounts of untrusted data
6155 is passed to the `BIO_*printf` functions. If applications use these functions
6156 in this way then they could be vulnerable. OpenSSL itself uses these
6157 functions when printing out human-readable dumps of ASN.1 data. Therefore
6158 applications that print this data could be vulnerable if the data is from
6159 untrusted sources. OpenSSL command line applications could also be
6160 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6161 as command line arguments.
6163 Libssl is not considered directly vulnerable. Additionally certificates etc
6164 received via remote connections via libssl are also unlikely to be able to
6165 trigger these issues because of message size limits enforced within libssl.
6167 This issue was reported to OpenSSL Guido Vranken.
6172 * Side channel attack on modular exponentiation
6174 A side-channel attack was found which makes use of cache-bank conflicts on
6175 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6176 of RSA keys. The ability to exploit this issue is limited as it relies on
6177 an attacker who has control of code in a thread running on the same
6178 hyper-threaded core as the victim thread which is performing decryptions.
6180 This issue was reported to OpenSSL by Yuval Yarom, The University of
6181 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6182 Nadia Heninger, University of Pennsylvania with more information at
6183 <http://cachebleed.info>.
6188 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6189 if no keysize is specified with default_bits. This fixes an
6190 omission in an earlier change that changed all RSA/DSA key generation
6191 commands to use 2048 bits by default.
6195 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6197 * DH small subgroups
6199 Historically OpenSSL only ever generated DH parameters based on "safe"
6200 primes. More recently (in version 1.0.2) support was provided for
6201 generating X9.42 style parameter files such as those required for RFC 5114
6202 support. The primes used in such files may not be "safe". Where an
6203 application is using DH configured with parameters based on primes that are
6204 not "safe" then an attacker could use this fact to find a peer's private
6205 DH exponent. This attack requires that the attacker complete multiple
6206 handshakes in which the peer uses the same private DH exponent. For example
6207 this could be used to discover a TLS server's private DH exponent if it's
6208 reusing the private DH exponent or it's using a static DH ciphersuite.
6210 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6211 TLS. It is not on by default. If the option is not set then the server
6212 reuses the same private DH exponent for the life of the server process and
6213 would be vulnerable to this attack. It is believed that many popular
6214 applications do set this option and would therefore not be at risk.
6216 The fix for this issue adds an additional check where a "q" parameter is
6217 available (as is the case in X9.42 based parameters). This detects the
6218 only known attack, and is the only possible defense for static DH
6219 ciphersuites. This could have some performance impact.
6221 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6222 default and cannot be disabled. This could have some performance impact.
6224 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6229 * SSLv2 doesn't block disabled ciphers
6231 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6232 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6233 been disabled, provided that the SSLv2 protocol was not also disabled via
6236 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6237 and Sebastian Schinzel.
6242 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6244 * BN_mod_exp may produce incorrect results on x86_64
6246 There is a carry propagating bug in the x86_64 Montgomery squaring
6247 procedure. No EC algorithms are affected. Analysis suggests that attacks
6248 against RSA and DSA as a result of this defect would be very difficult to
6249 perform and are not believed likely. Attacks against DH are considered just
6250 feasible (although very difficult) because most of the work necessary to
6251 deduce information about a private key may be performed offline. The amount
6252 of resources required for such an attack would be very significant and
6253 likely only accessible to a limited number of attackers. An attacker would
6254 additionally need online access to an unpatched system using the target
6255 private key in a scenario with persistent DH parameters and a private
6256 key that is shared between multiple clients. For example this can occur by
6257 default in OpenSSL DHE based SSL/TLS ciphersuites.
6259 This issue was reported to OpenSSL by Hanno Böck.
6264 * Certificate verify crash with missing PSS parameter
6266 The signature verification routines will crash with a NULL pointer
6267 dereference if presented with an ASN.1 signature using the RSA PSS
6268 algorithm and absent mask generation function parameter. Since these
6269 routines are used to verify certificate signature algorithms this can be
6270 used to crash any certificate verification operation and exploited in a
6271 DoS attack. Any application which performs certificate verification is
6272 vulnerable including OpenSSL clients and servers which enable client
6275 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6280 * X509_ATTRIBUTE memory leak
6282 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6283 memory. This structure is used by the PKCS#7 and CMS routines so any
6284 application which reads PKCS#7 or CMS data from untrusted sources is
6285 affected. SSL/TLS is not affected.
6287 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6293 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6294 This changes the decoding behaviour for some invalid messages,
6295 though the change is mostly in the more lenient direction, and
6296 legacy behaviour is preserved as much as possible.
6300 * In DSA_generate_parameters_ex, if the provided seed is too short,
6303 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6305 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6307 * Alternate chains certificate forgery
6309 During certificate verification, OpenSSL will attempt to find an
6310 alternative certificate chain if the first attempt to build such a chain
6311 fails. An error in the implementation of this logic can mean that an
6312 attacker could cause certain checks on untrusted certificates to be
6313 bypassed, such as the CA flag, enabling them to use a valid leaf
6314 certificate to act as a CA and "issue" an invalid certificate.
6316 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6321 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6323 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6324 incompatibility in the handling of HMAC. The previous ABI has now been
6329 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6331 * Malformed ECParameters causes infinite loop
6333 When processing an ECParameters structure OpenSSL enters an infinite loop
6334 if the curve specified is over a specially malformed binary polynomial
6337 This can be used to perform denial of service against any
6338 system which processes public keys, certificate requests or
6339 certificates. This includes TLS clients and TLS servers with
6340 client authentication enabled.
6342 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6347 * Exploitable out-of-bounds read in X509_cmp_time
6349 X509_cmp_time does not properly check the length of the ASN1_TIME
6350 string and can read a few bytes out of bounds. In addition,
6351 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6354 An attacker can use this to craft malformed certificates and CRLs of
6355 various sizes and potentially cause a segmentation fault, resulting in
6356 a DoS on applications that verify certificates or CRLs. TLS clients
6357 that verify CRLs are affected. TLS clients and servers with client
6358 authentication enabled may be affected if they use custom verification
6361 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6362 independently by Hanno Böck.
6367 * PKCS7 crash with missing EnvelopedContent
6369 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6370 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6371 with missing content and trigger a NULL pointer dereference on parsing.
6373 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6374 structures from untrusted sources are affected. OpenSSL clients and
6375 servers are not affected.
6377 This issue was reported to OpenSSL by Michal Zalewski (Google).
6382 * CMS verify infinite loop with unknown hash function
6384 When verifying a signedData message the CMS code can enter an infinite loop
6385 if presented with an unknown hash function OID. This can be used to perform
6386 denial of service against any system which verifies signedData messages using
6388 This issue was reported to OpenSSL by Johannes Bauer.
6393 * Race condition handling NewSessionTicket
6395 If a NewSessionTicket is received by a multi-threaded client when attempting to
6396 reuse a previous ticket then a race condition can occur potentially leading to
6397 a double free of the ticket data.
6402 * Only support 256-bit or stronger elliptic curves with the
6403 'ecdh_auto' setting (server) or by default (client). Of supported
6404 curves, prefer P-256 (both).
6408 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6410 * ClientHello sigalgs DoS fix
6412 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6413 invalid signature algorithms extension a NULL pointer dereference will
6414 occur. This can be exploited in a DoS attack against the server.
6416 This issue was was reported to OpenSSL by David Ramos of Stanford
6420 *Stephen Henson and Matt Caswell*
6422 * Multiblock corrupted pointer fix
6424 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6425 feature only applies on 64 bit x86 architecture platforms that support AES
6426 NI instructions. A defect in the implementation of "multiblock" can cause
6427 OpenSSL's internal write buffer to become incorrectly set to NULL when
6428 using non-blocking IO. Typically, when the user application is using a
6429 socket BIO for writing, this will only result in a failed connection.
6430 However if some other BIO is used then it is likely that a segmentation
6431 fault will be triggered, thus enabling a potential DoS attack.
6433 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6438 * Segmentation fault in DTLSv1_listen fix
6440 The DTLSv1_listen function is intended to be stateless and processes the
6441 initial ClientHello from many peers. It is common for user code to loop
6442 over the call to DTLSv1_listen until a valid ClientHello is received with
6443 an associated cookie. A defect in the implementation of DTLSv1_listen means
6444 that state is preserved in the SSL object from one invocation to the next
6445 that can lead to a segmentation fault. Errors processing the initial
6446 ClientHello can trigger this scenario. An example of such an error could be
6447 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6450 This issue was reported to OpenSSL by Per Allansson.
6455 * Segmentation fault in ASN1_TYPE_cmp fix
6457 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6458 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6459 certificate signature algorithm consistency this can be used to crash any
6460 certificate verification operation and exploited in a DoS attack. Any
6461 application which performs certificate verification is vulnerable including
6462 OpenSSL clients and servers which enable client authentication.
6467 * Segmentation fault for invalid PSS parameters fix
6469 The signature verification routines will crash with a NULL pointer
6470 dereference if presented with an ASN.1 signature using the RSA PSS
6471 algorithm and invalid parameters. Since these routines are used to verify
6472 certificate signature algorithms this can be used to crash any
6473 certificate verification operation and exploited in a DoS attack. Any
6474 application which performs certificate verification is vulnerable including
6475 OpenSSL clients and servers which enable client authentication.
6477 This issue was was reported to OpenSSL by Brian Carpenter.
6482 * ASN.1 structure reuse memory corruption fix
6484 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6485 memory corruption via an invalid write. Such reuse is and has been
6486 strongly discouraged and is believed to be rare.
6488 Applications that parse structures containing CHOICE or ANY DEFINED BY
6489 components may be affected. Certificate parsing (d2i_X509 and related
6490 functions) are however not affected. OpenSSL clients and servers are
6496 * PKCS7 NULL pointer dereferences fix
6498 The PKCS#7 parsing code does not handle missing outer ContentInfo
6499 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6500 missing content and trigger a NULL pointer dereference on parsing.
6502 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6503 otherwise parse PKCS#7 structures from untrusted sources are
6504 affected. OpenSSL clients and servers are not affected.
6506 This issue was reported to OpenSSL by Michal Zalewski (Google).
6511 * DoS via reachable assert in SSLv2 servers fix
6513 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6514 servers that both support SSLv2 and enable export cipher suites by sending
6515 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6517 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6518 (OpenSSL development team).
6523 * Empty CKE with client auth and DHE fix
6525 If client auth is used then a server can seg fault in the event of a DHE
6526 ciphersuite being selected and a zero length ClientKeyExchange message
6527 being sent by the client. This could be exploited in a DoS attack.
6532 * Handshake with unseeded PRNG fix
6534 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6535 with an unseeded PRNG. The conditions are:
6536 - The client is on a platform where the PRNG has not been seeded
6537 automatically, and the user has not seeded manually
6538 - A protocol specific client method version has been used (i.e. not
6539 SSL_client_methodv23)
6540 - A ciphersuite is used that does not require additional random data from
6541 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6543 If the handshake succeeds then the client random that has been used will
6544 have been generated from a PRNG with insufficient entropy and therefore the
6545 output may be predictable.
6547 For example using the following command with an unseeded openssl will
6548 succeed on an unpatched platform:
6550 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6555 * Use After Free following d2i_ECPrivatekey error fix
6557 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6558 could cause a use after free condition. This, in turn, could cause a double
6559 free in several private key parsing functions (such as d2i_PrivateKey
6560 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6561 for applications that receive EC private keys from untrusted
6562 sources. This scenario is considered rare.
6564 This issue was discovered by the BoringSSL project and fixed in their
6570 * X509_to_X509_REQ NULL pointer deref fix
6572 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6573 the certificate key is invalid. This function is rarely used in practice.
6575 This issue was discovered by Brian Carpenter.
6580 * Removed the export ciphers from the DEFAULT ciphers
6584 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6586 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6587 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6588 So far those who have to target multiple platforms would compromise
6589 and argue that binary targeting say ARMv5 would still execute on
6590 ARMv8. "Universal" build resolves this compromise by providing
6591 near-optimal performance even on newer platforms.
6595 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6596 (other platforms pending).
6598 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6600 * Add support for the SignedCertificateTimestampList certificate and
6601 OCSP response extensions from RFC6962.
6605 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6606 for corner cases. (Certain input points at infinity could lead to
6607 bogus results, with non-infinity inputs mapped to infinity too.)
6611 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6612 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6613 common cases are optimized and there still is room for further
6614 improvements. Vector Permutation AES for Altivec is also added.
6618 * Add support for little-endian ppc64 Linux target.
6620 *Marcelo Cerri (IBM)*
6622 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6623 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6624 are optimized and there still is room for further improvements.
6625 Both 32- and 64-bit modes are supported.
6627 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6629 * Improved ARMv7 NEON support.
6633 * Support for SPARC Architecture 2011 crypto extensions, first
6634 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6635 SHA256/512, MD5, GHASH and modular exponentiation.
6637 *Andy Polyakov, David Miller*
6639 * Accelerated modular exponentiation for Intel processors, a.k.a.
6642 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6644 * Support for new and upcoming Intel processors, including AVX2,
6645 BMI and SHA ISA extensions. This includes additional "stitched"
6646 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6649 This work was sponsored by Intel Corp.
6653 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6654 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6655 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6659 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6660 this fixes a limitation in previous versions of OpenSSL.
6664 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6665 MGF1 digest and OAEP label.
6669 * Add EVP support for key wrapping algorithms, to avoid problems with
6670 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6671 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6672 algorithms and include tests cases.
6676 * Add functions to allocate and set the fields of an ECDSA_METHOD
6679 *Douglas E. Engert, Steve Henson*
6681 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6682 difference in days and seconds between two tm or ASN1_TIME structures.
6686 * Add -rev test option to s_server to just reverse order of characters
6687 received by client and send back to server. Also prints an abbreviated
6688 summary of the connection parameters.
6692 * New option -brief for s_client and s_server to print out a brief summary
6693 of connection parameters.
6697 * Add callbacks for arbitrary TLS extensions.
6699 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6701 * New option -crl_download in several openssl utilities to download CRLs
6702 from CRLDP extension in certificates.
6706 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6710 * New function X509_CRL_diff to generate a delta CRL from the difference
6711 of two full CRLs. Add support to "crl" utility.
6715 * New functions to set lookup_crls function and to retrieve
6716 X509_STORE from X509_STORE_CTX.
6720 * Print out deprecated issuer and subject unique ID fields in
6725 * Extend OCSP I/O functions so they can be used for simple general purpose
6726 HTTP as well as OCSP. New wrapper function which can be used to download
6727 CRLs using the OCSP API.
6731 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6735 * `SSL_CONF*` functions. These provide a common framework for application
6736 configuration using configuration files or command lines.
6740 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6741 message callback and prints the results. Needs compile time option
6742 "enable-ssl-trace". New options to s_client and s_server to enable
6747 * New ctrl and macro to retrieve supported points extensions.
6748 Print out extension in s_server and s_client.
6752 * New functions to retrieve certificate signature and signature
6757 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6762 * New Suite B modes for TLS code. These use and enforce the requirements
6763 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6764 only use Suite B curves. The Suite B modes can be set by using the
6765 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6769 * New chain verification flags for Suite B levels of security. Check
6770 algorithms are acceptable when flags are set in X509_verify_cert.
6774 * Make tls1_check_chain return a set of flags indicating checks passed
6775 by a certificate chain. Add additional tests to handle client
6776 certificates: checks for matching certificate type and issuer name
6781 * If an attempt is made to use a signature algorithm not in the peer
6782 preference list abort the handshake. If client has no suitable
6783 signature algorithms in response to a certificate request do not
6784 use the certificate.
6788 * If server EC tmp key is not in client preference list abort handshake.
6792 * Add support for certificate stores in CERT structure. This makes it
6793 possible to have different stores per SSL structure or one store in
6794 the parent SSL_CTX. Include distinct stores for certificate chain
6795 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6796 to build and store a certificate chain in CERT structure: returning
6797 an error if the chain cannot be built: this will allow applications
6798 to test if a chain is correctly configured.
6800 Note: if the CERT based stores are not set then the parent SSL_CTX
6801 store is used to retain compatibility with existing behaviour.
6805 * New function ssl_set_client_disabled to set a ciphersuite disabled
6806 mask based on the current session, check mask when sending client
6807 hello and checking the requested ciphersuite.
6811 * New ctrls to retrieve and set certificate types in a certificate
6812 request message. Print out received values in s_client. If certificate
6813 types is not set with custom values set sensible values based on
6814 supported signature algorithms.
6818 * Support for distinct client and server supported signature algorithms.
6822 * Add certificate callback. If set this is called whenever a certificate
6823 is required by client or server. An application can decide which
6824 certificate chain to present based on arbitrary criteria: for example
6825 supported signature algorithms. Add very simple example to s_server.
6826 This fixes many of the problems and restrictions of the existing client
6827 certificate callback: for example you can now clear an existing
6828 certificate and specify the whole chain.
6832 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6833 the certificate can be used for (if anything). Set valid_flags field
6834 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6835 to have similar checks in it.
6837 Add new "cert_flags" field to CERT structure and include a "strict mode".
6838 This enforces some TLS certificate requirements (such as only permitting
6839 certificate signature algorithms contained in the supported algorithms
6840 extension) which some implementations ignore: this option should be used
6841 with caution as it could cause interoperability issues.
6845 * Update and tidy signature algorithm extension processing. Work out
6846 shared signature algorithms based on preferences and peer algorithms
6847 and print them out in s_client and s_server. Abort handshake if no
6848 shared signature algorithms.
6852 * Add new functions to allow customised supported signature algorithms
6853 for SSL and SSL_CTX structures. Add options to s_client and s_server
6858 * New function SSL_certs_clear() to delete all references to certificates
6859 from an SSL structure. Before this once a certificate had been added
6860 it couldn't be removed.
6864 * Integrate hostname, email address and IP address checking with certificate
6865 verification. New verify options supporting checking in openssl utility.
6869 * Fixes and wildcard matching support to hostname and email checking
6870 functions. Add manual page.
6872 *Florian Weimer (Red Hat Product Security Team)*
6874 * New functions to check a hostname email or IP address against a
6875 certificate. Add options x509 utility to print results of checks against
6880 * Fix OCSP checking.
6882 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6884 * Initial experimental support for explicitly trusted non-root CAs.
6885 OpenSSL still tries to build a complete chain to a root but if an
6886 intermediate CA has a trust setting included that is used. The first
6887 setting is used: whether to trust (e.g., -addtrust option to the x509
6892 * Add -trusted_first option which attempts to find certificates in the
6893 trusted store even if an untrusted chain is also supplied.
6897 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6898 platform support for Linux and Android.
6902 * Support for linux-x32, ILP32 environment in x86_64 framework.
6906 * Experimental multi-implementation support for FIPS capable OpenSSL.
6907 When in FIPS mode the approved implementations are used as normal,
6908 when not in FIPS mode the internal unapproved versions are used instead.
6909 This means that the FIPS capable OpenSSL isn't forced to use the
6910 (often lower performance) FIPS implementations outside FIPS mode.
6914 * Transparently support X9.42 DH parameters when calling
6915 PEM_read_bio_DHparameters. This means existing applications can handle
6916 the new parameter format automatically.
6920 * Initial experimental support for X9.42 DH parameter format: mainly
6921 to support use of 'q' parameter for RFC5114 parameters.
6925 * Add DH parameters from RFC5114 including test data to dhtest.
6929 * Support for automatic EC temporary key parameter selection. If enabled
6930 the most preferred EC parameters are automatically used instead of
6931 hardcoded fixed parameters. Now a server just has to call:
6932 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6933 support ECDH and use the most appropriate parameters.
6937 * Enhance and tidy EC curve and point format TLS extension code. Use
6938 static structures instead of allocation if default values are used.
6939 New ctrls to set curves we wish to support and to retrieve shared curves.
6940 Print out shared curves in s_server. New options to s_server and s_client
6941 to set list of supported curves.
6945 * New ctrls to retrieve supported signature algorithms and
6946 supported curve values as an array of NIDs. Extend openssl utility
6947 to print out received values.
6951 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6952 between NIDs and the more common NIST names such as "P-256". Enhance
6953 ecparam utility and ECC method to recognise the NIST names for curves.
6957 * Enhance SSL/TLS certificate chain handling to support different
6958 chains for each certificate instead of one chain in the parent SSL_CTX.
6962 * Support for fixed DH ciphersuite client authentication: where both
6963 server and client use DH certificates with common parameters.
6967 * Support for fixed DH ciphersuites: those requiring DH server
6972 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6974 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6975 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6976 X509_CINF_get_signature were reverted post internal team review.
6981 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6983 * OCSP Status Request extension unbounded memory growth
6985 A malicious client can send an excessively large OCSP Status Request
6986 extension. If that client continually requests renegotiation, sending a
6987 large OCSP Status Request extension each time, then there will be unbounded
6988 memory growth on the server. This will eventually lead to a Denial Of
6989 Service attack through memory exhaustion. Servers with a default
6990 configuration are vulnerable even if they do not support OCSP. Builds using
6991 the "no-ocsp" build time option are not affected.
6993 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6998 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7001 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7007 * OOB write in MDC2_Update()
7009 An overflow can occur in MDC2_Update() either if called directly or
7010 through the EVP_DigestUpdate() function using MDC2. If an attacker
7011 is able to supply very large amounts of input data after a previous
7012 call to EVP_EncryptUpdate() with a partial block then a length check
7013 can overflow resulting in a heap corruption.
7015 The amount of data needed is comparable to SIZE_MAX which is impractical
7018 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7023 * Malformed SHA512 ticket DoS
7025 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7026 DoS attack where a malformed ticket will result in an OOB read which will
7029 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7030 a custom server callback and ticket lookup mechanism.
7032 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7037 * OOB write in BN_bn2dec()
7039 The function BN_bn2dec() does not check the return value of BN_div_word().
7040 This can cause an OOB write if an application uses this function with an
7041 overly large BIGNUM. This could be a problem if an overly large certificate
7042 or CRL is printed out from an untrusted source. TLS is not affected because
7043 record limits will reject an oversized certificate before it is parsed.
7045 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7050 * OOB read in TS_OBJ_print_bio()
7052 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7053 the total length the OID text representation would use and not the amount
7054 of data written. This will result in OOB reads when large OIDs are
7057 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7062 * Pointer arithmetic undefined behaviour
7064 Avoid some undefined pointer arithmetic
7066 A common idiom in the codebase is to check limits in the following manner:
7069 Where "p" points to some malloc'd data of SIZE bytes and
7072 "len" here could be from some externally supplied data (e.g. from a TLS
7075 The rules of C pointer arithmetic are such that "p + len" is only well
7076 defined where len <= SIZE. Therefore, the above idiom is actually
7077 undefined behaviour.
7079 For example this could cause problems if some malloc implementation
7080 provides an address for "p" such that "p + len" actually overflows for
7081 values of len that are too big and therefore p + len < limit.
7083 This issue was reported to OpenSSL by Guido Vranken
7088 * Constant time flag not preserved in DSA signing
7090 Operations in the DSA signing algorithm should run in constant time in
7091 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7092 implementation means that a non-constant time codepath is followed for
7093 certain operations. This has been demonstrated through a cache-timing
7094 attack to be sufficient for an attacker to recover the private DSA key.
7096 This issue was reported by César Pereida (Aalto University), Billy Brumley
7097 (Tampere University of Technology), and Yuval Yarom (The University of
7098 Adelaide and NICTA).
7103 * DTLS buffered message DoS
7105 In a DTLS connection where handshake messages are delivered out-of-order
7106 those messages that OpenSSL is not yet ready to process will be buffered
7107 for later use. Under certain circumstances, a flaw in the logic means that
7108 those messages do not get removed from the buffer even though the handshake
7109 has been completed. An attacker could force up to approx. 15 messages to
7110 remain in the buffer when they are no longer required. These messages will
7111 be cleared when the DTLS connection is closed. The default maximum size for
7112 a message is 100k. Therefore, the attacker could force an additional 1500k
7113 to be consumed per connection. By opening many simultaneous connections an
7114 attacker could cause a DoS attack through memory exhaustion.
7116 This issue was reported to OpenSSL by Quan Luo.
7121 * DTLS replay protection DoS
7123 A flaw in the DTLS replay attack protection mechanism means that records
7124 that arrive for future epochs update the replay protection "window" before
7125 the MAC for the record has been validated. This could be exploited by an
7126 attacker by sending a record for the next epoch (which does not have to
7127 decrypt or have a valid MAC), with a very large sequence number. This means
7128 that all subsequent legitimate packets are dropped causing a denial of
7129 service for a specific DTLS connection.
7131 This issue was reported to OpenSSL by the OCAP audit team.
7136 * Certificate message OOB reads
7138 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7139 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7140 theoretical DoS risk but this has not been observed in practice on common
7143 The messages affected are client certificate, client certificate request
7144 and server certificate. As a result the attack can only be performed
7145 against a client or a server which enables client authentication.
7147 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7152 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7154 * Prevent padding oracle in AES-NI CBC MAC check
7156 A MITM attacker can use a padding oracle attack to decrypt traffic
7157 when the connection uses an AES CBC cipher and the server support
7160 This issue was introduced as part of the fix for Lucky 13 padding
7161 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7162 constant time by making sure that always the same bytes are read and
7163 compared against either the MAC or padding bytes. But it no longer
7164 checked that there was enough data to have both the MAC and padding
7167 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7172 * Fix EVP_EncodeUpdate overflow
7174 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7175 Base64 encoding of binary data. If an attacker is able to supply very large
7176 amounts of input data then a length check can overflow resulting in a heap
7179 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7180 the `PEM_write_bio*` family of functions. These are mainly used within the
7181 OpenSSL command line applications, so any application which processes data
7182 from an untrusted source and outputs it as a PEM file should be considered
7183 vulnerable to this issue. User applications that call these APIs directly
7184 with large amounts of untrusted data may also be vulnerable.
7186 This issue was reported by Guido Vranken.
7191 * Fix EVP_EncryptUpdate overflow
7193 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7194 is able to supply very large amounts of input data after a previous call to
7195 EVP_EncryptUpdate() with a partial block then a length check can overflow
7196 resulting in a heap corruption. Following an analysis of all OpenSSL
7197 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7198 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7199 the first called function after an EVP_EncryptInit(), and therefore that
7200 specific call must be safe. The second form is where the length passed to
7201 EVP_EncryptUpdate() can be seen from the code to be some small value and
7202 therefore there is no possibility of an overflow. Since all instances are
7203 one of these two forms, it is believed that there can be no overflows in
7204 internal code due to this problem. It should be noted that
7205 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7206 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7207 of these calls have also been analysed too and it is believed there are no
7208 instances in internal usage where an overflow could occur.
7210 This issue was reported by Guido Vranken.
7215 * Prevent ASN.1 BIO excessive memory allocation
7217 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7218 a short invalid encoding can casuse allocation of large amounts of memory
7219 potentially consuming excessive resources or exhausting memory.
7221 Any application parsing untrusted data through d2i BIO functions is
7222 affected. The memory based functions such as d2i_X509() are *not* affected.
7223 Since the memory based functions are used by the TLS library, TLS
7224 applications are not affected.
7226 This issue was reported by Brian Carpenter.
7233 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7234 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7235 in arbitrary stack data being returned in the buffer.
7237 This issue was reported by Guido Vranken.
7242 * Modify behavior of ALPN to invoke callback after SNI/servername
7243 callback, such that updates to the SSL_CTX affect ALPN.
7247 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7252 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7253 methods are enabled and ssl2 is disabled the methods return NULL.
7257 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7259 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7260 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7261 provide any "EXPORT" or "LOW" strength ciphers.
7265 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7266 is by default disabled at build-time. Builds that are not configured with
7267 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7268 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7269 will need to explicitly call either of:
7271 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7273 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7275 as appropriate. Even if either of those is used, or the application
7276 explicitly uses the version-specific SSLv2_method() or its client and
7277 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7278 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7279 ciphers, and SSLv2 56-bit DES are no longer available.
7284 * Fix a double-free in DSA code
7286 A double free bug was discovered when OpenSSL parses malformed DSA private
7287 keys and could lead to a DoS attack or memory corruption for applications
7288 that receive DSA private keys from untrusted sources. This scenario is
7291 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7297 * Disable SRP fake user seed to address a server memory leak.
7299 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7301 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7302 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7303 was changed to ignore the "fake user" SRP seed, even if the seed
7306 Users should use SRP_VBASE_get1_by_user instead. Note that in
7307 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7308 also that even though configuring the SRP seed attempts to hide
7309 invalid usernames by continuing the handshake with fake
7310 credentials, this behaviour is not constant time and no strong
7311 guarantees are made that the handshake is indistinguishable from
7312 that of a valid user.
7317 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7319 In the BN_hex2bn function the number of hex digits is calculated using an
7320 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7321 large values of `i` this can result in `bn_expand` not allocating any
7322 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7323 field as NULL leading to a subsequent NULL ptr deref. For very large values
7324 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7325 In this case memory is allocated to the internal BIGNUM data field, but it
7326 is insufficiently sized leading to heap corruption. A similar issue exists
7327 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7328 is ever called by user applications with very large untrusted hex/dec data.
7329 This is anticipated to be a rare occurrence.
7331 All OpenSSL internal usage of these functions use data that is not expected
7332 to be untrusted, e.g. config file data or application command line
7333 arguments. If user developed applications generate config file data based
7334 on untrusted data then it is possible that this could also lead to security
7335 consequences. This is also anticipated to be rare.
7337 This issue was reported to OpenSSL by Guido Vranken.
7342 * Fix memory issues in `BIO_*printf` functions
7344 The internal `fmtstr` function used in processing a "%s" format string in
7345 the `BIO_*printf` functions could overflow while calculating the length of a
7346 string and cause an OOB read when printing very long strings.
7348 Additionally the internal `doapr_outch` function can attempt to write to an
7349 OOB memory location (at an offset from the NULL pointer) in the event of a
7350 memory allocation failure. In 1.0.2 and below this could be caused where
7351 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7352 could be in processing a very long "%s" format string. Memory leaks can
7355 The first issue may mask the second issue dependent on compiler behaviour.
7356 These problems could enable attacks where large amounts of untrusted data
7357 is passed to the `BIO_*printf` functions. If applications use these functions
7358 in this way then they could be vulnerable. OpenSSL itself uses these
7359 functions when printing out human-readable dumps of ASN.1 data. Therefore
7360 applications that print this data could be vulnerable if the data is from
7361 untrusted sources. OpenSSL command line applications could also be
7362 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7363 as command line arguments.
7365 Libssl is not considered directly vulnerable. Additionally certificates etc
7366 received via remote connections via libssl are also unlikely to be able to
7367 trigger these issues because of message size limits enforced within libssl.
7369 This issue was reported to OpenSSL Guido Vranken.
7374 * Side channel attack on modular exponentiation
7376 A side-channel attack was found which makes use of cache-bank conflicts on
7377 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7378 of RSA keys. The ability to exploit this issue is limited as it relies on
7379 an attacker who has control of code in a thread running on the same
7380 hyper-threaded core as the victim thread which is performing decryptions.
7382 This issue was reported to OpenSSL by Yuval Yarom, The University of
7383 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7384 Nadia Heninger, University of Pennsylvania with more information at
7385 <http://cachebleed.info>.
7390 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7391 if no keysize is specified with default_bits. This fixes an
7392 omission in an earlier change that changed all RSA/DSA key generation
7393 commands to use 2048 bits by default.
7397 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7399 * Protection for DH small subgroup attacks
7401 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7402 switched on by default and cannot be disabled. This could have some
7407 * SSLv2 doesn't block disabled ciphers
7409 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7410 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7411 been disabled, provided that the SSLv2 protocol was not also disabled via
7414 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7415 and Sebastian Schinzel.
7420 * Reject DH handshakes with parameters shorter than 1024 bits.
7424 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7426 * Certificate verify crash with missing PSS parameter
7428 The signature verification routines will crash with a NULL pointer
7429 dereference if presented with an ASN.1 signature using the RSA PSS
7430 algorithm and absent mask generation function parameter. Since these
7431 routines are used to verify certificate signature algorithms this can be
7432 used to crash any certificate verification operation and exploited in a
7433 DoS attack. Any application which performs certificate verification is
7434 vulnerable including OpenSSL clients and servers which enable client
7437 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7442 * X509_ATTRIBUTE memory leak
7444 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7445 memory. This structure is used by the PKCS#7 and CMS routines so any
7446 application which reads PKCS#7 or CMS data from untrusted sources is
7447 affected. SSL/TLS is not affected.
7449 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7455 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7456 This changes the decoding behaviour for some invalid messages,
7457 though the change is mostly in the more lenient direction, and
7458 legacy behaviour is preserved as much as possible.
7462 * In DSA_generate_parameters_ex, if the provided seed is too short,
7463 use a random seed, as already documented.
7465 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7467 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7469 * Alternate chains certificate forgery
7471 During certificate verification, OpenSSL will attempt to find an
7472 alternative certificate chain if the first attempt to build such a chain
7473 fails. An error in the implementation of this logic can mean that an
7474 attacker could cause certain checks on untrusted certificates to be
7475 bypassed, such as the CA flag, enabling them to use a valid leaf
7476 certificate to act as a CA and "issue" an invalid certificate.
7478 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7484 * Race condition handling PSK identify hint
7486 If PSK identity hints are received by a multi-threaded client then
7487 the values are wrongly updated in the parent SSL_CTX structure. This can
7488 result in a race condition potentially leading to a double free of the
7494 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7496 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7497 incompatibility in the handling of HMAC. The previous ABI has now been
7500 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7502 * Malformed ECParameters causes infinite loop
7504 When processing an ECParameters structure OpenSSL enters an infinite loop
7505 if the curve specified is over a specially malformed binary polynomial
7508 This can be used to perform denial of service against any
7509 system which processes public keys, certificate requests or
7510 certificates. This includes TLS clients and TLS servers with
7511 client authentication enabled.
7513 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7518 * Exploitable out-of-bounds read in X509_cmp_time
7520 X509_cmp_time does not properly check the length of the ASN1_TIME
7521 string and can read a few bytes out of bounds. In addition,
7522 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7525 An attacker can use this to craft malformed certificates and CRLs of
7526 various sizes and potentially cause a segmentation fault, resulting in
7527 a DoS on applications that verify certificates or CRLs. TLS clients
7528 that verify CRLs are affected. TLS clients and servers with client
7529 authentication enabled may be affected if they use custom verification
7532 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7533 independently by Hanno Böck.
7538 * PKCS7 crash with missing EnvelopedContent
7540 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7541 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7542 with missing content and trigger a NULL pointer dereference on parsing.
7544 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7545 structures from untrusted sources are affected. OpenSSL clients and
7546 servers are not affected.
7548 This issue was reported to OpenSSL by Michal Zalewski (Google).
7553 * CMS verify infinite loop with unknown hash function
7555 When verifying a signedData message the CMS code can enter an infinite loop
7556 if presented with an unknown hash function OID. This can be used to perform
7557 denial of service against any system which verifies signedData messages using
7559 This issue was reported to OpenSSL by Johannes Bauer.
7564 * Race condition handling NewSessionTicket
7566 If a NewSessionTicket is received by a multi-threaded client when attempting to
7567 reuse a previous ticket then a race condition can occur potentially leading to
7568 a double free of the ticket data.
7573 * Reject DH handshakes with parameters shorter than 768 bits.
7575 *Kurt Roeckx and Emilia Kasper*
7577 * dhparam: generate 2048-bit parameters by default.
7579 *Kurt Roeckx and Emilia Kasper*
7581 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7583 * Segmentation fault in ASN1_TYPE_cmp fix
7585 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7586 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7587 certificate signature algorithm consistency this can be used to crash any
7588 certificate verification operation and exploited in a DoS attack. Any
7589 application which performs certificate verification is vulnerable including
7590 OpenSSL clients and servers which enable client authentication.
7595 * ASN.1 structure reuse memory corruption fix
7597 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7598 memory corruption via an invalid write. Such reuse is and has been
7599 strongly discouraged and is believed to be rare.
7601 Applications that parse structures containing CHOICE or ANY DEFINED BY
7602 components may be affected. Certificate parsing (d2i_X509 and related
7603 functions) are however not affected. OpenSSL clients and servers are
7609 * PKCS7 NULL pointer dereferences fix
7611 The PKCS#7 parsing code does not handle missing outer ContentInfo
7612 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7613 missing content and trigger a NULL pointer dereference on parsing.
7615 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7616 otherwise parse PKCS#7 structures from untrusted sources are
7617 affected. OpenSSL clients and servers are not affected.
7619 This issue was reported to OpenSSL by Michal Zalewski (Google).
7624 * DoS via reachable assert in SSLv2 servers fix
7626 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7627 servers that both support SSLv2 and enable export cipher suites by sending
7628 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7630 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7631 (OpenSSL development team).
7636 * Use After Free following d2i_ECPrivatekey error fix
7638 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7639 could cause a use after free condition. This, in turn, could cause a double
7640 free in several private key parsing functions (such as d2i_PrivateKey
7641 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7642 for applications that receive EC private keys from untrusted
7643 sources. This scenario is considered rare.
7645 This issue was discovered by the BoringSSL project and fixed in their
7651 * X509_to_X509_REQ NULL pointer deref fix
7653 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7654 the certificate key is invalid. This function is rarely used in practice.
7656 This issue was discovered by Brian Carpenter.
7661 * Removed the export ciphers from the DEFAULT ciphers
7665 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7667 * Build fixes for the Windows and OpenVMS platforms
7669 *Matt Caswell and Richard Levitte*
7671 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7673 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7674 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7675 dereference. This could lead to a Denial Of Service attack. Thanks to
7676 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7681 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7682 dtls1_buffer_record function under certain conditions. In particular this
7683 could occur if an attacker sent repeated DTLS records with the same
7684 sequence number but for the next epoch. The memory leak could be exploited
7685 by an attacker in a Denial of Service attack through memory exhaustion.
7686 Thanks to Chris Mueller for reporting this issue.
7691 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7692 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7693 method would be set to NULL which could later result in a NULL pointer
7694 dereference. Thanks to Frank Schmirler for reporting this issue.
7699 * Abort handshake if server key exchange message is omitted for ephemeral
7702 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7703 reporting this issue.
7708 * Remove non-export ephemeral RSA code on client and server. This code
7709 violated the TLS standard by allowing the use of temporary RSA keys in
7710 non-export ciphersuites and could be used by a server to effectively
7711 downgrade the RSA key length used to a value smaller than the server
7712 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7713 INRIA or reporting this issue.
7718 * Fixed issue where DH client certificates are accepted without verification.
7719 An OpenSSL server will accept a DH certificate for client authentication
7720 without the certificate verify message. This effectively allows a client to
7721 authenticate without the use of a private key. This only affects servers
7722 which trust a client certificate authority which issues certificates
7723 containing DH keys: these are extremely rare and hardly ever encountered.
7724 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7730 * Ensure that the session ID context of an SSL is updated when its
7731 SSL_CTX is updated via SSL_set_SSL_CTX.
7733 The session ID context is typically set from the parent SSL_CTX,
7734 and can vary with the CTX.
7738 * Fix various certificate fingerprint issues.
7740 By using non-DER or invalid encodings outside the signed portion of a
7741 certificate the fingerprint can be changed without breaking the signature.
7742 Although no details of the signed portion of the certificate can be changed
7743 this can cause problems with some applications: e.g. those using the
7744 certificate fingerprint for blacklists.
7746 1. Reject signatures with non zero unused bits.
7748 If the BIT STRING containing the signature has non zero unused bits reject
7749 the signature. All current signature algorithms require zero unused bits.
7751 2. Check certificate algorithm consistency.
7753 Check the AlgorithmIdentifier inside TBS matches the one in the
7754 certificate signature. NB: this will result in signature failure
7755 errors for some broken certificates.
7757 Thanks to Konrad Kraszewski from Google for reporting this issue.
7759 3. Check DSA/ECDSA signatures use DER.
7761 Re-encode DSA/ECDSA signatures and compare with the original received
7762 signature. Return an error if there is a mismatch.
7764 This will reject various cases including garbage after signature
7765 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7766 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7767 (negative or with leading zeroes).
7769 Further analysis was conducted and fixes were developed by Stephen Henson
7770 of the OpenSSL core team.
7776 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7777 results on some platforms, including x86_64. This bug occurs at random
7778 with a very low probability, and is not known to be exploitable in any
7779 way, though its exact impact is difficult to determine. Thanks to Pieter
7780 Wuille (Blockstream) who reported this issue and also suggested an initial
7781 fix. Further analysis was conducted by the OpenSSL development team and
7782 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7783 the OpenSSL core team.
7788 * Do not resume sessions on the server if the negotiated protocol
7789 version does not match the session's version. Resuming with a different
7790 version, while not strictly forbidden by the RFC, is of questionable
7791 sanity and breaks all known clients.
7793 *David Benjamin, Emilia Käsper*
7795 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7796 early CCS messages during renegotiation. (Note that because
7797 renegotiation is encrypted, this early CCS was not exploitable.)
7801 * Tighten client-side session ticket handling during renegotiation:
7802 ensure that the client only accepts a session ticket if the server sends
7803 the extension anew in the ServerHello. Previously, a TLS client would
7804 reuse the old extension state and thus accept a session ticket if one was
7805 announced in the initial ServerHello.
7807 Similarly, ensure that the client requires a session ticket if one
7808 was advertised in the ServerHello. Previously, a TLS client would
7809 ignore a missing NewSessionTicket message.
7813 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7817 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7818 sends a carefully crafted handshake message, to cause OpenSSL to fail
7819 to free up to 64k of memory causing a memory leak. This could be
7820 exploited in a Denial Of Service attack. This issue affects OpenSSL
7821 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7822 whether SRTP is used or configured. Implementations of OpenSSL that
7823 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7825 The fix was developed by the OpenSSL team.
7830 * Session Ticket Memory Leak.
7832 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7833 integrity of that ticket is first verified. In the event of a session
7834 ticket integrity check failing, OpenSSL will fail to free memory
7835 causing a memory leak. By sending a large number of invalid session
7836 tickets an attacker could exploit this issue in a Denial Of Service
7842 * Build option no-ssl3 is incomplete.
7844 When OpenSSL is configured with "no-ssl3" as a build option, servers
7845 could accept and complete an SSL 3.0 handshake, and clients could be
7846 configured to send them.
7849 *Akamai and the OpenSSL team*
7851 * Add support for TLS_FALLBACK_SCSV.
7852 Client applications doing fallback retries should call
7853 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7856 *Adam Langley, Bodo Moeller*
7858 * Add additional DigestInfo checks.
7860 Re-encode DigestInto in DER and check against the original when
7861 verifying RSA signature: this will reject any improperly encoded
7862 DigestInfo structures.
7864 Note: this is a precautionary measure and no attacks are currently known.
7868 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7870 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7871 SRP code can be overrun an internal buffer. Add sanity check that
7872 g, A, B < N to SRP code.
7874 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7875 Group for discovering this issue.
7880 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7881 TLS 1.0 instead of higher protocol versions when the ClientHello message
7882 is badly fragmented. This allows a man-in-the-middle attacker to force a
7883 downgrade to TLS 1.0 even if both the server and the client support a
7884 higher protocol version, by modifying the client's TLS records.
7886 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7887 researching this issue.
7892 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7893 to a denial of service attack. A malicious server can crash the client
7894 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7895 ciphersuite and sending carefully crafted handshake messages.
7897 Thanks to Felix Gröbert (Google) for discovering and researching this
7903 * By sending carefully crafted DTLS packets an attacker could cause openssl
7904 to leak memory. This can be exploited through a Denial of Service attack.
7905 Thanks to Adam Langley for discovering and researching this issue.
7910 * An attacker can force openssl to consume large amounts of memory whilst
7911 processing DTLS handshake messages. This can be exploited through a
7912 Denial of Service attack.
7913 Thanks to Adam Langley for discovering and researching this issue.
7918 * An attacker can force an error condition which causes openssl to crash
7919 whilst processing DTLS packets due to memory being freed twice. This
7920 can be exploited through a Denial of Service attack.
7921 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7927 * If a multithreaded client connects to a malicious server using a resumed
7928 session and the server sends an ec point format extension it could write
7929 up to 255 bytes to freed memory.
7931 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7937 * A malicious server can crash an OpenSSL client with a null pointer
7938 dereference (read) by specifying an SRP ciphersuite even though it was not
7939 properly negotiated with the client. This can be exploited through a
7940 Denial of Service attack.
7942 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7943 discovering and researching this issue.
7948 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7949 X509_name_oneline, X509_name_print_ex et al. to leak some information
7950 from the stack. Applications may be affected if they echo pretty printing
7951 output to the attacker.
7953 Thanks to Ivan Fratric (Google) for discovering this issue.
7956 *Emilia Käsper, and Steve Henson*
7958 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7959 for corner cases. (Certain input points at infinity could lead to
7960 bogus results, with non-infinity inputs mapped to infinity too.)
7964 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7966 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7967 handshake can force the use of weak keying material in OpenSSL
7968 SSL/TLS clients and servers.
7970 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7971 researching this issue. ([CVE-2014-0224])
7973 *KIKUCHI Masashi, Steve Henson*
7975 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7976 OpenSSL DTLS client the code can be made to recurse eventually crashing
7979 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7982 *Imre Rad, Steve Henson*
7984 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7985 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7986 client or server. This is potentially exploitable to run arbitrary
7987 code on a vulnerable client or server.
7989 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7991 *Jüri Aedla, Steve Henson*
7993 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7994 are subject to a denial of service attack.
7996 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7997 this issue. ([CVE-2014-3470])
7999 *Felix Gröbert, Ivan Fratric, Steve Henson*
8001 * Harmonize version and its documentation. -f flag is used to display
8004 *mancha <mancha1@zoho.com>*
8006 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8007 in i2d_ECPrivateKey.
8009 *mancha <mancha1@zoho.com>*
8011 * Fix some double frees. These are not thought to be exploitable.
8013 *mancha <mancha1@zoho.com>*
8015 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8017 * A missing bounds check in the handling of the TLS heartbeat extension
8018 can be used to reveal up to 64k of memory to a connected client or
8021 Thanks for Neel Mehta of Google Security for discovering this bug and to
8022 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8023 preparing the fix ([CVE-2014-0160])
8025 *Adam Langley, Bodo Moeller*
8027 * Fix for the attack described in the paper "Recovering OpenSSL
8028 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8029 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8030 <http://eprint.iacr.org/2014/140>
8032 Thanks to Yuval Yarom and Naomi Benger for discovering this
8033 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8035 *Yuval Yarom and Naomi Benger*
8037 * TLS pad extension: draft-agl-tls-padding-03
8039 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8040 TLS client Hello record length value would otherwise be > 255 and
8041 less that 512 pad with a dummy extension containing zeroes so it
8042 is at least 512 bytes long.
8044 *Adam Langley, Steve Henson*
8046 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8048 * Fix for TLS record tampering bug. A carefully crafted invalid
8049 handshake could crash OpenSSL with a NULL pointer exception.
8050 Thanks to Anton Johansson for reporting this issues.
8053 * Keep original DTLS digest and encryption contexts in retransmission
8054 structures so we can use the previous session parameters if they need
8055 to be resent. ([CVE-2013-6450])
8059 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8060 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8061 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8062 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8063 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8064 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8066 *Rob Stradling, Adam Langley*
8068 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8070 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8071 supporting platforms or when small records were transferred.
8073 *Andy Polyakov, Steve Henson*
8075 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8077 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8079 This addresses the flaw in CBC record processing discovered by
8080 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8081 at: <http://www.isg.rhul.ac.uk/tls/>
8083 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8084 Security Group at Royal Holloway, University of London
8085 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8086 Emilia Käsper for the initial patch.
8089 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8091 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8092 ciphersuites which can be exploited in a denial of service attack.
8093 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8094 and detecting this bug and to Wolfgang Ettlinger
8095 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8100 * Return an error when checking OCSP signatures when key is NULL.
8101 This fixes a DoS attack. ([CVE-2013-0166])
8105 * Make openssl verify return errors.
8107 *Chris Palmer <palmer@google.com> and Ben Laurie*
8109 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8110 the right response is stapled. Also change SSL_get_certificate()
8111 so it returns the certificate actually sent.
8112 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8114 *Rob Stradling <rob.stradling@comodo.com>*
8116 * Fix possible deadlock when decoding public keys.
8120 * Don't use TLS 1.0 record version number in initial client hello
8125 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8127 * Sanity check record length before skipping explicit IV in TLS
8128 1.2, 1.1 and DTLS to fix DoS attack.
8130 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8131 fuzzing as a service testing platform.
8136 * Initialise tkeylen properly when encrypting CMS messages.
8137 Thanks to Solar Designer of Openwall for reporting this issue.
8141 * In FIPS mode don't try to use composite ciphers as they are not
8146 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8148 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8149 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8150 mean any application compiled against OpenSSL 1.0.0 headers setting
8151 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8152 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8153 0x10000000L Any application which was previously compiled against
8154 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8155 will need to be recompiled as a result. Letting be results in
8156 inability to disable specifically TLS 1.1 and in client context,
8157 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8161 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8162 disable just protocol X, but all protocols above X *if* there are
8163 protocols *below* X still enabled. In more practical terms it means
8164 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8165 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8166 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8171 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8173 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8174 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8175 in CRYPTO_realloc_clean.
8177 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8178 issue and to Adam Langley <agl@chromium.org> for fixing it.
8181 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8183 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8187 * Workarounds for some broken servers that "hang" if a client hello
8188 record length exceeds 255 bytes.
8190 1. Do not use record version number > TLS 1.0 in initial client
8191 hello: some (but not all) hanging servers will now work.
8192 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8193 the number of ciphers sent in the client hello. This should be
8194 set to an even number, such as 50, for example by passing:
8195 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8196 Most broken servers should now work.
8197 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8198 TLS 1.2 client support entirely.
8202 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8206 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8208 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8209 STRING form instead of a DigestInfo.
8213 * The format used for MDC2 RSA signatures is inconsistent between EVP
8214 and the RSA_sign/RSA_verify functions. This was made more apparent when
8215 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8216 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8217 the correct format in RSA_verify so both forms transparently work.
8221 * Some servers which support TLS 1.0 can choke if we initially indicate
8222 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8223 encrypted premaster secret. As a workaround use the maximum permitted
8224 client version in client hello, this should keep such servers happy
8225 and still work with previous versions of OpenSSL.
8229 * Add support for TLS/DTLS heartbeats.
8231 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8233 * Add support for SCTP.
8235 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8237 * Improved PRNG seeding for VOS.
8239 *Paul Green <Paul.Green@stratus.com>*
8241 * Extensive assembler packs updates, most notably:
8243 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8244 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8245 - x86_64: bit-sliced AES implementation;
8246 - ARM: NEON support, contemporary platforms optimizations;
8247 - s390x: z196 support;
8248 - `*`: GHASH and GF(2^m) multiplication implementations;
8252 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8253 (removal of unnecessary code)
8255 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8257 * Add TLS key material exporter from RFC 5705.
8261 * Add DTLS-SRTP negotiation from RFC 5764.
8265 * Add Next Protocol Negotiation,
8266 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8267 disabled with a no-npn flag to config or Configure. Code donated
8270 *Adam Langley <agl@google.com> and Ben Laurie*
8272 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8273 NIST-P256, NIST-P521, with constant-time single point multiplication on
8274 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8275 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8276 Code made available under Apache License version 2.0.
8278 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8279 line to include this in your build of OpenSSL, and run "make depend" (or
8280 "make update"). This enables the following EC_METHODs:
8282 EC_GFp_nistp224_method()
8283 EC_GFp_nistp256_method()
8284 EC_GFp_nistp521_method()
8286 EC_GROUP_new_by_curve_name() will automatically use these (while
8287 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8290 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8292 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8293 all platforms. Move ssize_t definition from e_os.h to the public
8294 header file e_os2.h as it now appears in public header file cms.h
8298 * New -sigopt option to the ca, req and x509 utilities. Additional
8299 signature parameters can be passed using this option and in
8304 * Add RSA PSS signing function. This will generate and set the
8305 appropriate AlgorithmIdentifiers for PSS based on those in the
8306 corresponding EVP_MD_CTX structure. No application support yet.
8310 * Support for companion algorithm specific ASN1 signing routines.
8311 New function ASN1_item_sign_ctx() signs a pre-initialised
8312 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8313 the appropriate parameters.
8317 * Add new algorithm specific ASN1 verification initialisation function
8318 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8319 handling will be the same no matter what EVP_PKEY_METHOD is used.
8320 Add a PSS handler to support verification of PSS signatures: checked
8321 against a number of sample certificates.
8325 * Add signature printing for PSS. Add PSS OIDs.
8327 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8329 * Add algorithm specific signature printing. An individual ASN1 method
8330 can now print out signatures instead of the standard hex dump.
8332 More complex signatures (e.g. PSS) can print out more meaningful
8333 information. Include DSA version that prints out the signature
8338 * Password based recipient info support for CMS library: implementing
8343 * Split password based encryption into PBES2 and PBKDF2 functions. This
8344 neatly separates the code into cipher and PBE sections and is required
8345 for some algorithms that split PBES2 into separate pieces (such as
8346 password based CMS).
8350 * Session-handling fixes:
8351 - Fix handling of connections that are resuming with a session ID,
8352 but also support Session Tickets.
8353 - Fix a bug that suppressed issuing of a new ticket if the client
8354 presented a ticket with an expired session.
8355 - Try to set the ticket lifetime hint to something reasonable.
8356 - Make tickets shorter by excluding irrelevant information.
8357 - On the client side, don't ignore renewed tickets.
8359 *Adam Langley, Bodo Moeller (Google)*
8361 * Fix PSK session representation.
8365 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8367 This work was sponsored by Intel.
8371 * Add GCM support to TLS library. Some custom code is needed to split
8372 the IV between the fixed (from PRF) and explicit (from TLS record)
8373 portions. This adds all GCM ciphersuites supported by RFC5288 and
8374 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8375 add a special AESGCM string for GCM only.
8379 * Expand range of ctrls for AES GCM. Permit setting invocation
8380 field on decrypt and retrieval of invocation field only on encrypt.
8384 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8385 As required by RFC5289 these ciphersuites cannot be used if for
8386 versions of TLS earlier than 1.2.
8390 * For FIPS capable OpenSSL interpret a NULL default public key method
8391 as unset and return the appropriate default but do *not* set the default.
8392 This means we can return the appropriate method in applications that
8393 switch between FIPS and non-FIPS modes.
8397 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8398 ENGINE is used then we cannot handle that in the FIPS module so we
8399 keep original code iff non-FIPS operations are allowed.
8403 * Add -attime option to openssl utilities.
8405 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8407 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8411 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8412 FIPS EC methods unconditionally for now.
8416 * New build option no-ec2m to disable characteristic 2 code.
8420 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8421 all cases can be covered as some introduce binary incompatibilities.
8425 * Redirect RSA operations to FIPS module including keygen,
8426 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8430 * Add similar low-level API blocking to ciphers.
8434 * low-level digest APIs are not approved in FIPS mode: any attempt
8435 to use these will cause a fatal error. Applications that *really* want
8436 to use them can use the `private_*` version instead.
8440 * Redirect cipher operations to FIPS module for FIPS builds.
8444 * Redirect digest operations to FIPS module for FIPS builds.
8448 * Update build system to add "fips" flag which will link in fipscanister.o
8449 for static and shared library builds embedding a signature if needed.
8453 * Output TLS supported curves in preference order instead of numerical
8454 order. This is currently hardcoded for the highest order curves first.
8455 This should be configurable so applications can judge speed vs strength.
8459 * Add TLS v1.2 server support for client authentication.
8463 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8468 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8469 FIPS modules versions.
8473 * Add TLS v1.2 client side support for client authentication. Keep cache
8474 of handshake records longer as we don't know the hash algorithm to use
8475 until after the certificate request message is received.
8479 * Initial TLS v1.2 client support. Add a default signature algorithms
8480 extension including all the algorithms we support. Parse new signature
8481 format in client key exchange. Relax some ECC signing restrictions for
8482 TLS v1.2 as indicated in RFC5246.
8486 * Add server support for TLS v1.2 signature algorithms extension. Switch
8487 to new signature format when needed using client digest preference.
8488 All server ciphersuites should now work correctly in TLS v1.2. No client
8489 support yet and no support for client certificates.
8493 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8494 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8495 ciphersuites. At present only RSA key exchange ciphersuites work with
8496 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8497 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8498 and version checking.
8502 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8503 with this defined it will not be affected by any changes to ssl internal
8504 structures. Add several utility functions to allow openssl application
8505 to work with OPENSSL_NO_SSL_INTERN defined.
8509 * A long standing patch to add support for SRP from EdelWeb (Peter
8510 Sylvester and Christophe Renou) was integrated.
8511 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8512 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8515 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8519 * Permit abbreviated handshakes when renegotiating using the function
8520 SSL_renegotiate_abbreviated().
8522 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8524 * Add call to ENGINE_register_all_complete() to
8525 ENGINE_load_builtin_engines(), so some implementations get used
8526 automatically instead of needing explicit application support.
8530 * Add support for TLS key exporter as described in RFC5705.
8532 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8534 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8535 a few changes are required:
8537 Add SSL_OP_NO_TLSv1_1 flag.
8538 Add TLSv1_1 methods.
8539 Update version checking logic to handle version 1.1.
8540 Add explicit IV handling (ported from DTLS code).
8541 Add command line options to s_client/s_server.
8548 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8550 * X509_ATTRIBUTE memory leak
8552 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8553 memory. This structure is used by the PKCS#7 and CMS routines so any
8554 application which reads PKCS#7 or CMS data from untrusted sources is
8555 affected. SSL/TLS is not affected.
8557 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8563 * Race condition handling PSK identify hint
8565 If PSK identity hints are received by a multi-threaded client then
8566 the values are wrongly updated in the parent SSL_CTX structure. This can
8567 result in a race condition potentially leading to a double free of the
8573 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8575 * Malformed ECParameters causes infinite loop
8577 When processing an ECParameters structure OpenSSL enters an infinite loop
8578 if the curve specified is over a specially malformed binary polynomial
8581 This can be used to perform denial of service against any
8582 system which processes public keys, certificate requests or
8583 certificates. This includes TLS clients and TLS servers with
8584 client authentication enabled.
8586 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8591 * Exploitable out-of-bounds read in X509_cmp_time
8593 X509_cmp_time does not properly check the length of the ASN1_TIME
8594 string and can read a few bytes out of bounds. In addition,
8595 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8598 An attacker can use this to craft malformed certificates and CRLs of
8599 various sizes and potentially cause a segmentation fault, resulting in
8600 a DoS on applications that verify certificates or CRLs. TLS clients
8601 that verify CRLs are affected. TLS clients and servers with client
8602 authentication enabled may be affected if they use custom verification
8605 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8606 independently by Hanno Böck.
8611 * PKCS7 crash with missing EnvelopedContent
8613 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8614 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8615 with missing content and trigger a NULL pointer dereference on parsing.
8617 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8618 structures from untrusted sources are affected. OpenSSL clients and
8619 servers are not affected.
8621 This issue was reported to OpenSSL by Michal Zalewski (Google).
8626 * CMS verify infinite loop with unknown hash function
8628 When verifying a signedData message the CMS code can enter an infinite loop
8629 if presented with an unknown hash function OID. This can be used to perform
8630 denial of service against any system which verifies signedData messages using
8632 This issue was reported to OpenSSL by Johannes Bauer.
8637 * Race condition handling NewSessionTicket
8639 If a NewSessionTicket is received by a multi-threaded client when attempting to
8640 reuse a previous ticket then a race condition can occur potentially leading to
8641 a double free of the ticket data.
8646 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8648 * Segmentation fault in ASN1_TYPE_cmp fix
8650 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8651 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8652 certificate signature algorithm consistency this can be used to crash any
8653 certificate verification operation and exploited in a DoS attack. Any
8654 application which performs certificate verification is vulnerable including
8655 OpenSSL clients and servers which enable client authentication.
8660 * ASN.1 structure reuse memory corruption fix
8662 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8663 memory corruption via an invalid write. Such reuse is and has been
8664 strongly discouraged and is believed to be rare.
8666 Applications that parse structures containing CHOICE or ANY DEFINED BY
8667 components may be affected. Certificate parsing (d2i_X509 and related
8668 functions) are however not affected. OpenSSL clients and servers are
8674 * PKCS7 NULL pointer dereferences fix
8676 The PKCS#7 parsing code does not handle missing outer ContentInfo
8677 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8678 missing content and trigger a NULL pointer dereference on parsing.
8680 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8681 otherwise parse PKCS#7 structures from untrusted sources are
8682 affected. OpenSSL clients and servers are not affected.
8684 This issue was reported to OpenSSL by Michal Zalewski (Google).
8689 * DoS via reachable assert in SSLv2 servers fix
8691 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8692 servers that both support SSLv2 and enable export cipher suites by sending
8693 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8695 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8696 (OpenSSL development team).
8701 * Use After Free following d2i_ECPrivatekey error fix
8703 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8704 could cause a use after free condition. This, in turn, could cause a double
8705 free in several private key parsing functions (such as d2i_PrivateKey
8706 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8707 for applications that receive EC private keys from untrusted
8708 sources. This scenario is considered rare.
8710 This issue was discovered by the BoringSSL project and fixed in their
8716 * X509_to_X509_REQ NULL pointer deref fix
8718 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8719 the certificate key is invalid. This function is rarely used in practice.
8721 This issue was discovered by Brian Carpenter.
8726 * Removed the export ciphers from the DEFAULT ciphers
8730 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8732 * Build fixes for the Windows and OpenVMS platforms
8734 *Matt Caswell and Richard Levitte*
8736 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8738 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8739 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8740 dereference. This could lead to a Denial Of Service attack. Thanks to
8741 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8746 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8747 dtls1_buffer_record function under certain conditions. In particular this
8748 could occur if an attacker sent repeated DTLS records with the same
8749 sequence number but for the next epoch. The memory leak could be exploited
8750 by an attacker in a Denial of Service attack through memory exhaustion.
8751 Thanks to Chris Mueller for reporting this issue.
8756 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8757 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8758 method would be set to NULL which could later result in a NULL pointer
8759 dereference. Thanks to Frank Schmirler for reporting this issue.
8764 * Abort handshake if server key exchange message is omitted for ephemeral
8767 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8768 reporting this issue.
8773 * Remove non-export ephemeral RSA code on client and server. This code
8774 violated the TLS standard by allowing the use of temporary RSA keys in
8775 non-export ciphersuites and could be used by a server to effectively
8776 downgrade the RSA key length used to a value smaller than the server
8777 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8778 INRIA or reporting this issue.
8783 * Fixed issue where DH client certificates are accepted without verification.
8784 An OpenSSL server will accept a DH certificate for client authentication
8785 without the certificate verify message. This effectively allows a client to
8786 authenticate without the use of a private key. This only affects servers
8787 which trust a client certificate authority which issues certificates
8788 containing DH keys: these are extremely rare and hardly ever encountered.
8789 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8795 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8796 results on some platforms, including x86_64. This bug occurs at random
8797 with a very low probability, and is not known to be exploitable in any
8798 way, though its exact impact is difficult to determine. Thanks to Pieter
8799 Wuille (Blockstream) who reported this issue and also suggested an initial
8800 fix. Further analysis was conducted by the OpenSSL development team and
8801 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8802 the OpenSSL core team.
8807 * Fix various certificate fingerprint issues.
8809 By using non-DER or invalid encodings outside the signed portion of a
8810 certificate the fingerprint can be changed without breaking the signature.
8811 Although no details of the signed portion of the certificate can be changed
8812 this can cause problems with some applications: e.g. those using the
8813 certificate fingerprint for blacklists.
8815 1. Reject signatures with non zero unused bits.
8817 If the BIT STRING containing the signature has non zero unused bits reject
8818 the signature. All current signature algorithms require zero unused bits.
8820 2. Check certificate algorithm consistency.
8822 Check the AlgorithmIdentifier inside TBS matches the one in the
8823 certificate signature. NB: this will result in signature failure
8824 errors for some broken certificates.
8826 Thanks to Konrad Kraszewski from Google for reporting this issue.
8828 3. Check DSA/ECDSA signatures use DER.
8830 Re-encode DSA/ECDSA signatures and compare with the original received
8831 signature. Return an error if there is a mismatch.
8833 This will reject various cases including garbage after signature
8834 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8835 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8836 (negative or with leading zeroes).
8838 Further analysis was conducted and fixes were developed by Stephen Henson
8839 of the OpenSSL core team.
8845 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8847 * Session Ticket Memory Leak.
8849 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8850 integrity of that ticket is first verified. In the event of a session
8851 ticket integrity check failing, OpenSSL will fail to free memory
8852 causing a memory leak. By sending a large number of invalid session
8853 tickets an attacker could exploit this issue in a Denial Of Service
8859 * Build option no-ssl3 is incomplete.
8861 When OpenSSL is configured with "no-ssl3" as a build option, servers
8862 could accept and complete an SSL 3.0 handshake, and clients could be
8863 configured to send them.
8866 *Akamai and the OpenSSL team*
8868 * Add support for TLS_FALLBACK_SCSV.
8869 Client applications doing fallback retries should call
8870 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8873 *Adam Langley, Bodo Moeller*
8875 * Add additional DigestInfo checks.
8877 Re-encode DigestInto in DER and check against the original when
8878 verifying RSA signature: this will reject any improperly encoded
8879 DigestInfo structures.
8881 Note: this is a precautionary measure and no attacks are currently known.
8885 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8887 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8888 to a denial of service attack. A malicious server can crash the client
8889 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8890 ciphersuite and sending carefully crafted handshake messages.
8892 Thanks to Felix Gröbert (Google) for discovering and researching this
8898 * By sending carefully crafted DTLS packets an attacker could cause openssl
8899 to leak memory. This can be exploited through a Denial of Service attack.
8900 Thanks to Adam Langley for discovering and researching this issue.
8905 * An attacker can force openssl to consume large amounts of memory whilst
8906 processing DTLS handshake messages. This can be exploited through a
8907 Denial of Service attack.
8908 Thanks to Adam Langley for discovering and researching this issue.
8913 * An attacker can force an error condition which causes openssl to crash
8914 whilst processing DTLS packets due to memory being freed twice. This
8915 can be exploited through a Denial of Service attack.
8916 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8922 * If a multithreaded client connects to a malicious server using a resumed
8923 session and the server sends an ec point format extension it could write
8924 up to 255 bytes to freed memory.
8926 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8932 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8933 X509_name_oneline, X509_name_print_ex et al. to leak some information
8934 from the stack. Applications may be affected if they echo pretty printing
8935 output to the attacker.
8937 Thanks to Ivan Fratric (Google) for discovering this issue.
8940 *Emilia Käsper, and Steve Henson*
8942 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8943 for corner cases. (Certain input points at infinity could lead to
8944 bogus results, with non-infinity inputs mapped to infinity too.)
8948 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8950 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8951 handshake can force the use of weak keying material in OpenSSL
8952 SSL/TLS clients and servers.
8954 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8955 researching this issue. ([CVE-2014-0224])
8957 *KIKUCHI Masashi, Steve Henson*
8959 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8960 OpenSSL DTLS client the code can be made to recurse eventually crashing
8963 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8966 *Imre Rad, Steve Henson*
8968 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8969 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8970 client or server. This is potentially exploitable to run arbitrary
8971 code on a vulnerable client or server.
8973 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8975 *Jüri Aedla, Steve Henson*
8977 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8978 are subject to a denial of service attack.
8980 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8981 this issue. ([CVE-2014-3470])
8983 *Felix Gröbert, Ivan Fratric, Steve Henson*
8985 * Harmonize version and its documentation. -f flag is used to display
8988 *mancha <mancha1@zoho.com>*
8990 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8991 in i2d_ECPrivateKey.
8993 *mancha <mancha1@zoho.com>*
8995 * Fix some double frees. These are not thought to be exploitable.
8997 *mancha <mancha1@zoho.com>*
8999 * Fix for the attack described in the paper "Recovering OpenSSL
9000 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9001 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9002 <http://eprint.iacr.org/2014/140>
9004 Thanks to Yuval Yarom and Naomi Benger for discovering this
9005 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9007 *Yuval Yarom and Naomi Benger*
9009 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9011 * Keep original DTLS digest and encryption contexts in retransmission
9012 structures so we can use the previous session parameters if they need
9013 to be resent. ([CVE-2013-6450])
9017 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9018 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9019 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9020 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9021 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9022 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9024 *Rob Stradling, Adam Langley*
9026 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9028 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9030 This addresses the flaw in CBC record processing discovered by
9031 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9032 at: <http://www.isg.rhul.ac.uk/tls/>
9034 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9035 Security Group at Royal Holloway, University of London
9036 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9037 Emilia Käsper for the initial patch.
9040 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9042 * Return an error when checking OCSP signatures when key is NULL.
9043 This fixes a DoS attack. ([CVE-2013-0166])
9047 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9048 the right response is stapled. Also change SSL_get_certificate()
9049 so it returns the certificate actually sent.
9050 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9051 (This is a backport)
9053 *Rob Stradling <rob.stradling@comodo.com>*
9055 * Fix possible deadlock when decoding public keys.
9059 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9061 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9064 * Sanity check record length before skipping explicit IV in DTLS
9067 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9068 fuzzing as a service testing platform.
9073 * Initialise tkeylen properly when encrypting CMS messages.
9074 Thanks to Solar Designer of Openwall for reporting this issue.
9078 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9080 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9081 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9082 in CRYPTO_realloc_clean.
9084 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9085 issue and to Adam Langley <agl@chromium.org> for fixing it.
9088 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9090 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9092 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9093 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9094 content decryption and always return the same error. Note: this attack
9095 needs on average 2^20 messages so it only affects automated senders. The
9096 old behaviour can be re-enabled in the CMS code by setting the
9097 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9098 an MMA defence is not necessary.
9099 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9100 this issue. ([CVE-2012-0884])
9104 * Fix CVE-2011-4619: make sure we really are receiving a
9105 client hello before rejecting multiple SGC restarts. Thanks to
9106 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9110 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9112 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9113 Thanks to Antonio Martin, Enterprise Secure Access Research and
9114 Development, Cisco Systems, Inc. for discovering this bug and
9115 preparing a fix. ([CVE-2012-0050])
9119 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9121 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9122 of the Vaudenay padding oracle attack on CBC mode encryption
9123 which enables an efficient plaintext recovery attack against
9124 the OpenSSL implementation of DTLS. Their attack exploits timing
9125 differences arising during decryption processing. A research
9126 paper describing this attack can be found at:
9127 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9128 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9129 Security Group at Royal Holloway, University of London
9130 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9131 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9132 for preparing the fix. ([CVE-2011-4108])
9134 *Robin Seggelmann, Michael Tuexen*
9136 * Clear bytes used for block padding of SSL 3.0 records.
9139 *Adam Langley (Google)*
9141 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9142 Kadianakis <desnacked@gmail.com> for discovering this issue and
9143 Adam Langley for preparing the fix. ([CVE-2011-4619])
9145 *Adam Langley (Google)*
9147 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9149 *Andrey Kulikov <amdeich@gmail.com>*
9151 * Prevent malformed RFC3779 data triggering an assertion failure.
9152 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9153 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9155 *Rob Austein <sra@hactrn.net>*
9157 * Improved PRNG seeding for VOS.
9159 *Paul Green <Paul.Green@stratus.com>*
9161 * Fix ssl_ciph.c set-up race.
9163 *Adam Langley (Google)*
9165 * Fix spurious failures in ecdsatest.c.
9167 *Emilia Käsper (Google)*
9169 * Fix the BIO_f_buffer() implementation (which was mixing different
9170 interpretations of the `..._len` fields).
9172 *Adam Langley (Google)*
9174 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9175 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9176 threads won't reuse the same blinding coefficients.
9178 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9179 lock to call BN_BLINDING_invert_ex, and avoids one use of
9180 BN_BLINDING_update for each BN_BLINDING structure (previously,
9181 the last update always remained unused).
9183 *Emilia Käsper (Google)*
9185 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9187 *Bob Buckholz (Google)*
9189 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9191 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9192 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9194 *Kaspar Brand <ossl@velox.ch>*
9196 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9197 for multi-threaded use of ECDH. ([CVE-2011-3210])
9199 *Adam Langley (Google)*
9201 * Fix x509_name_ex_d2i memory leak on bad inputs.
9205 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9206 signature public key algorithm by using OID xref utilities instead.
9207 Before this you could only use some ECC ciphersuites with SHA1 only.
9211 * Add protection against ECDSA timing attacks as mentioned in the paper
9212 by Billy Bob Brumley and Nicola Tuveri, see:
9213 <http://eprint.iacr.org/2011/232.pdf>
9215 *Billy Bob Brumley and Nicola Tuveri*
9217 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9219 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9221 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9223 * Fix bug in string printing code: if *any* escaping is enabled we must
9224 escape the escape character (backslash) or the resulting string is
9229 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9231 * Disable code workaround for ancient and obsolete Netscape browsers
9232 and servers: an attacker can use it in a ciphersuite downgrade attack.
9233 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9237 * Fixed J-PAKE implementation error, originally discovered by
9238 Sebastien Martini, further info and confirmation from Stefan
9239 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9243 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9245 * Fix extension code to avoid race conditions which can result in a buffer
9246 overrun vulnerability: resumed sessions must not be modified as they can
9247 be shared by multiple threads. CVE-2010-3864
9251 * Fix WIN32 build system to correctly link an ENGINE directory into
9256 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9258 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9261 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9263 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9265 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9266 context. The operation can be customised via the ctrl mechanism in
9267 case ENGINEs want to include additional functionality.
9271 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9275 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9276 output hashes compatible with older versions of OpenSSL.
9278 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9280 * Fix compression algorithm handling: if resuming a session use the
9281 compression algorithm of the resumed session instead of determining
9282 it from client hello again. Don't allow server to change algorithm.
9286 * Add load_crls() function to commands tidying load_certs() too. Add option
9287 to verify utility to allow additional CRLs to be included.
9291 * Update OCSP request code to permit adding custom headers to the request:
9292 some responders need this.
9296 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9299 *Julia Lawall <julia@diku.dk>*
9301 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9302 needlessly dereferenced structures, used obsolete functions and
9303 didn't handle all updated verify codes correctly.
9307 * Disable MD2 in the default configuration.
9311 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9312 indicate the initial BIO being pushed or popped. This makes it possible
9313 to determine whether the BIO is the one explicitly called or as a result
9314 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9315 it handles reference counts correctly and doesn't zero out the I/O bio
9316 when it is not being explicitly popped. WARNING: applications which
9317 included workarounds for the old buggy behaviour will need to be modified
9318 or they could free up already freed BIOs.
9322 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9323 renaming to all platforms (within the 0.9.8 branch, this was
9324 done conditionally on Netware platforms to avoid a name clash).
9326 *Guenter <lists@gknw.net>*
9328 * Add ECDHE and PSK support to DTLS.
9330 *Michael Tuexen <tuexen@fh-muenster.de>*
9332 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9337 * Add "missing" function EVP_MD_flags() (without this the only way to
9338 retrieve a digest flags is by accessing the structure directly. Update
9339 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9340 or cipher is registered as in the "from" argument. Print out all
9341 registered digests in the dgst usage message instead of manually
9342 attempting to work them out.
9346 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9347 this allows the use of compression and extensions. Change default cipher
9348 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9349 by default unless an application cipher string requests it.
9353 * Alter match criteria in PKCS12_parse(). It used to try to use local
9354 key ids to find matching certificates and keys but some PKCS#12 files
9355 don't follow the (somewhat unwritten) rules and this strategy fails.
9356 Now just gather all certificates together and the first private key
9357 then look for the first certificate that matches the key.
9361 * Support use of registered digest and cipher names for dgst and cipher
9362 commands instead of having to add each one as a special case. So now
9369 openssl dgst -sha256 foo
9371 and this works for ENGINE based algorithms too.
9375 * Update Gost ENGINE to support parameter files.
9377 *Victor B. Wagner <vitus@cryptocom.ru>*
9379 * Support GeneralizedTime in ca utility.
9381 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9383 * Enhance the hash format used for certificate directory links. The new
9384 form uses the canonical encoding (meaning equivalent names will work
9385 even if they aren't identical) and uses SHA1 instead of MD5. This form
9386 is incompatible with the older format and as a result c_rehash should
9387 be used to rebuild symbolic links.
9391 * Make PKCS#8 the default write format for private keys, replacing the
9392 traditional format. This form is standardised, more secure and doesn't
9393 include an implicit MD5 dependency.
9397 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9398 committed to OpenSSL should pass this lot as a minimum.
9402 * Add session ticket override functionality for use by EAP-FAST.
9404 *Jouni Malinen <j@w1.fi>*
9406 * Modify HMAC functions to return a value. Since these can be implemented
9407 in an ENGINE errors can occur.
9411 * Type-checked OBJ_bsearch_ex.
9415 * Type-checked OBJ_bsearch. Also some constification necessitated
9416 by type-checking. Still to come: TXT_DB, bsearch(?),
9417 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9422 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9423 seconds to a tm structure directly, instead of going through OS
9424 specific date routines. This avoids any issues with OS routines such
9425 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9426 and X509_time_adj_ex() to cover the extended range. The existing
9427 X509_time_adj() is still usable and will no longer have any date issues.
9431 * Delta CRL support. New use deltas option which will attempt to locate
9432 and search any appropriate delta CRLs available.
9434 This work was sponsored by Google.
9438 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9439 code and add additional score elements. Validate alternate CRL paths
9440 as part of the CRL checking and indicate a new error "CRL path validation
9441 error" in this case. Applications wanting additional details can use
9442 the verify callback and check the new "parent" field. If this is not
9443 NULL CRL path validation is taking place. Existing applications won't
9444 see this because it requires extended CRL support which is off by
9447 This work was sponsored by Google.
9451 * Support for freshest CRL extension.
9453 This work was sponsored by Google.
9457 * Initial indirect CRL support. Currently only supported in the CRLs
9458 passed directly and not via lookup. Process certificate issuer
9459 CRL entry extension and lookup CRL entries by bother issuer name
9460 and serial number. Check and process CRL issuer entry in IDP extension.
9462 This work was sponsored by Google.
9466 * Add support for distinct certificate and CRL paths. The CRL issuer
9467 certificate is validated separately in this case. Only enabled if
9468 an extended CRL support flag is set: this flag will enable additional
9469 CRL functionality in future.
9471 This work was sponsored by Google.
9475 * Add support for policy mappings extension.
9477 This work was sponsored by Google.
9481 * Fixes to pathlength constraint, self issued certificate handling,
9482 policy processing to align with RFC3280 and PKITS tests.
9484 This work was sponsored by Google.
9488 * Support for name constraints certificate extension. DN, email, DNS
9489 and URI types are currently supported.
9491 This work was sponsored by Google.
9495 * To cater for systems that provide a pointer-based thread ID rather
9496 than numeric, deprecate the current numeric thread ID mechanism and
9497 replace it with a structure and associated callback type. This
9498 mechanism allows a numeric "hash" to be extracted from a thread ID in
9499 either case, and on platforms where pointers are larger than 'long',
9500 mixing is done to help ensure the numeric 'hash' is usable even if it
9501 can't be guaranteed unique. The default mechanism is to use "&errno"
9502 as a pointer-based thread ID to distinguish between threads.
9504 Applications that want to provide their own thread IDs should now use
9505 CRYPTO_THREADID_set_callback() to register a callback that will call
9506 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9508 Note that ERR_remove_state() is now deprecated, because it is tied
9509 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9510 to free the current thread's error state should be replaced by
9511 ERR_remove_thread_state(NULL).
9513 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9514 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9515 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9516 application was previously providing a numeric thread callback that
9517 was inappropriate for distinguishing threads, then uniqueness might
9518 have been obtained with &errno that happened immediately in the
9519 intermediate development versions of OpenSSL; this is no longer the
9520 case, the numeric thread callback will now override the automatic use
9523 *Geoff Thorpe, with help from Bodo Moeller*
9525 * Initial support for different CRL issuing certificates. This covers a
9526 simple case where the self issued certificates in the chain exist and
9527 the real CRL issuer is higher in the existing chain.
9529 This work was sponsored by Google.
9533 * Removed effectively defunct crypto/store from the build.
9537 * Revamp of STACK to provide stronger type-checking. Still to come:
9538 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9539 ASN1_STRING, CONF_VALUE.
9543 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9544 RAM on SSL connections. This option can save about 34k per idle SSL.
9548 * Revamp of LHASH to provide stronger type-checking. Still to come:
9549 STACK, TXT_DB, bsearch, qsort.
9553 * Initial support for Cryptographic Message Syntax (aka CMS) based
9554 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9555 support for data, signedData, compressedData, digestedData and
9556 encryptedData, envelopedData types included. Scripts to check against
9557 RFC4134 examples draft and interop and consistency checks of many
9558 content types and variants.
9562 * Add options to enc utility to support use of zlib compression BIO.
9566 * Extend mk1mf to support importing of options and assembly language
9567 files from Configure script, currently only included in VC-WIN32.
9568 The assembly language rules can now optionally generate the source
9569 files from the associated perl scripts.
9573 * Implement remaining functionality needed to support GOST ciphersuites.
9574 Interop testing has been performed using CryptoPro implementations.
9576 *Victor B. Wagner <vitus@cryptocom.ru>*
9578 * s390x assembler pack.
9582 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9587 * Implement Opaque PRF Input TLS extension as specified in
9588 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9589 official specification yet and no extension type assignment by
9590 IANA exists, this extension (for now) will have to be explicitly
9591 enabled when building OpenSSL by providing the extension number
9592 to use. For example, specify an option
9594 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9596 to the "config" or "Configure" script to enable the extension,
9597 assuming extension number 0x9527 (which is a completely arbitrary
9598 and unofficial assignment based on the MD5 hash of the Internet
9599 Draft). Note that by doing so, you potentially lose
9600 interoperability with other TLS implementations since these might
9601 be using the same extension number for other purposes.
9603 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9604 opaque PRF input value to use in the handshake. This will create
9605 an internal copy of the length-'len' string at 'src', and will
9606 return non-zero for success.
9608 To get more control and flexibility, provide a callback function
9611 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9612 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9616 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9619 Callback function 'cb' will be called in handshakes, and is
9620 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9621 Argument 'arg' is for application purposes (the value as given to
9622 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9623 be provided to the callback function). The callback function
9624 has to return non-zero to report success: usually 1 to use opaque
9625 PRF input just if possible, or 2 to enforce use of the opaque PRF
9626 input. In the latter case, the library will abort the handshake
9627 if opaque PRF input is not successfully negotiated.
9629 Arguments 'peerinput' and 'len' given to the callback function
9630 will always be NULL and 0 in the case of a client. A server will
9631 see the client's opaque PRF input through these variables if
9632 available (NULL and 0 otherwise). Note that if the server
9633 provides an opaque PRF input, the length must be the same as the
9634 length of the client's opaque PRF input.
9636 Note that the callback function will only be called when creating
9637 a new session (session resumption can resume whatever was
9638 previously negotiated), and will not be called in SSL 2.0
9639 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9640 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9641 for applications that need to enforce opaque PRF input.
9645 * Update ssl code to support digests other than SHA1+MD5 for handshake
9648 *Victor B. Wagner <vitus@cryptocom.ru>*
9650 * Add RFC4507 support to OpenSSL. This includes the corrections in
9651 RFC4507bis. The encrypted ticket format is an encrypted encoded
9652 SSL_SESSION structure, that way new session features are automatically
9655 If a client application caches session in an SSL_SESSION structure
9656 support is transparent because tickets are now stored in the encoded
9659 The SSL_CTX structure automatically generates keys for ticket
9660 protection in servers so again support should be possible
9661 with no application modification.
9663 If a client or server wishes to disable RFC4507 support then the option
9664 SSL_OP_NO_TICKET can be set.
9666 Add a TLS extension debugging callback to allow the contents of any client
9667 or server extensions to be examined.
9669 This work was sponsored by Google.
9673 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9674 OpenSSL should now compile cleanly on gcc 4.2
9676 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9678 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9679 support including streaming MAC support: this is required for GOST
9680 ciphersuite support.
9682 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9684 * Add option -stream to use PKCS#7 streaming in smime utility. New
9685 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9686 to output in BER and PEM format.
9690 * Experimental support for use of HMAC via EVP_PKEY interface. This
9691 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9692 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9693 ENGINE support for HMAC keys which are unextractable. New -mac and
9694 -macopt options to dgst utility.
9698 * New option -sigopt to dgst utility. Update dgst to use
9699 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9700 alternative signing parameters such as X9.31 or PSS in the dgst
9705 * Change ssl_cipher_apply_rule(), the internal function that does
9706 the work each time a ciphersuite string requests enabling
9707 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9708 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9709 the order of disabled ciphersuites such that those ciphersuites
9710 that most recently went from enabled to disabled not only stay
9711 in order with respect to each other, but also have higher priority
9712 than other disabled ciphersuites the next time ciphersuites are
9715 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9716 the same ciphersuites as with "HIGH" alone, but in a specific
9717 order where the PSK ciphersuites come first (since they are the
9718 most recently disabled ciphersuites when "HIGH" is parsed).
9720 Also, change ssl_create_cipher_list() (using this new
9721 functionality) such that between otherwise identical
9722 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9727 * Change ssl_create_cipher_list() so that it automatically
9728 arranges the ciphersuites in reasonable order before starting
9729 to process the rule string. Thus, the definition for "DEFAULT"
9730 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9731 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9732 This makes it much easier to arrive at a reasonable default order
9733 in applications for which anonymous ciphers are OK (meaning
9734 that you can't actually use DEFAULT).
9736 *Bodo Moeller; suggested by Victor Duchovni*
9738 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9739 processing) into multiple integers instead of setting
9740 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9741 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9742 (These masks as well as the individual bit definitions are hidden
9743 away into the non-exported interface ssl/ssl_locl.h, so this
9744 change to the definition of the SSL_CIPHER structure shouldn't
9745 affect applications.) This give us more bits for each of these
9746 categories, so there is no longer a need to coagulate AES128 and
9747 AES256 into a single algorithm bit, and to coagulate Camellia128
9748 and Camellia256 into a single algorithm bit, which has led to all
9751 Thus, among other things, the kludge introduced in 0.9.7m and
9752 0.9.8e for masking out AES256 independently of AES128 or masking
9753 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9755 With the change, we also introduce new ciphersuite aliases that
9756 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9761 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9762 Use the leftmost N bytes of the signature input if the input is
9763 larger than the prime q (with N being the size in bytes of q).
9767 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9768 it yet and it is largely untested.
9772 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9776 * Initial incomplete changes to avoid need for function casts in OpenSSL
9777 some compilers (gcc 4.2 and later) reject their use. Safestack is
9778 reimplemented. Update ASN1 to avoid use of legacy functions.
9782 * Win32/64 targets are linked with Winsock2.
9786 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9787 to external functions. This can be used to increase CRL handling
9788 efficiency especially when CRLs are very large by (for example) storing
9789 the CRL revoked certificates in a database.
9793 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9794 new CRLs added to a directory can be used. New command line option
9795 -verify_return_error to s_client and s_server. This causes real errors
9796 to be returned by the verify callback instead of carrying on no matter
9797 what. This reflects the way a "real world" verify callback would behave.
9801 * GOST engine, supporting several GOST algorithms and public key formats.
9802 Kindly donated by Cryptocom.
9806 * Partial support for Issuing Distribution Point CRL extension. CRLs
9807 partitioned by DP are handled but no indirect CRL or reason partitioning
9808 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9809 selected via a scoring technique which handles IDP and AKID in CRLs.
9813 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9814 will ultimately be used for all verify operations: this will remove the
9815 X509_STORE dependency on certificate verification and allow alternative
9816 lookup methods. X509_STORE based implementations of these two callbacks.
9820 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9821 Modify get_crl() to find a valid (unexpired) CRL if possible.
9825 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9826 this would be called X509_CRL_cmp() but that name is already used by
9827 a function that just compares CRL issuer names. Cache several CRL
9828 extensions in X509_CRL structure and cache CRLDP in X509.
9832 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9833 this maps equivalent X509_NAME structures into a consistent structure.
9834 Name comparison can then be performed rapidly using memcmp().
9838 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9843 * Allow digests to supply their own micalg string for S/MIME type using
9844 the ctrl EVP_MD_CTRL_MICALG.
9848 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9849 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9850 ctrl. It can then customise the structure before and/or after signing
9855 * New function OBJ_add_sigid() to allow application defined signature OIDs
9856 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9857 to free up any added signature OIDs.
9861 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9862 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9863 digest and cipher tables. New options added to openssl utility:
9864 list-message-digest-algorithms and list-cipher-algorithms.
9868 * Change the array representation of binary polynomials: the list
9869 of degrees of non-zero coefficients is now terminated with -1.
9870 Previously it was terminated with 0, which was also part of the
9871 value; thus, the array representation was not applicable to
9872 polynomials where t^0 has coefficient zero. This change makes
9873 the array representation useful in a more general context.
9877 * Various modifications and fixes to SSL/TLS cipher string
9878 handling. For ECC, the code now distinguishes between fixed ECDH
9879 with RSA certificates on the one hand and with ECDSA certificates
9880 on the other hand, since these are separate ciphersuites. The
9881 unused code for Fortezza ciphersuites has been removed.
9883 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9884 (not "ECDHE"). For consistency with the code for DH
9885 certificates, use of ECDH certificates is now considered ECDH
9886 authentication, not RSA or ECDSA authentication (the latter is
9887 merely the CA's signing algorithm and not actively used in the
9890 The temporary ciphersuite alias "ECCdraft" is no longer
9891 available, and ECC ciphersuites are no longer excluded from "ALL"
9892 and "DEFAULT". The following aliases now exist for RFC 4492
9893 ciphersuites, most of these by analogy with the DH case:
9895 kECDHr - ECDH cert, signed with RSA
9896 kECDHe - ECDH cert, signed with ECDSA
9897 kECDH - ECDH cert (signed with either RSA or ECDSA)
9898 kEECDH - ephemeral ECDH
9899 ECDH - ECDH cert or ephemeral ECDH
9905 AECDH - anonymous ECDH
9906 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9910 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9911 Use correct micalg parameters depending on digest(s) in signed message.
9915 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9916 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9920 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9921 an engine to register a method. Add ENGINE lookups for methods and
9922 functional reference processing.
9926 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9927 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9932 * New -resign option to smime utility. This adds one or more signers
9933 to an existing PKCS#7 signedData structure. Also -md option to use an
9934 alternative message digest algorithm for signing.
9938 * Tidy up PKCS#7 routines and add new functions to make it easier to
9939 create PKCS7 structures containing multiple signers. Update smime
9940 application to support multiple signers.
9944 * New -macalg option to pkcs12 utility to allow setting of an alternative
9949 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9950 Reorganize PBE internals to lookup from a static table using NIDs,
9951 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9952 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9953 PRF which will be automatically used with PBES2.
9957 * Replace the algorithm specific calls to generate keys in "req" with the
9962 * Update PKCS#7 enveloped data routines to use new API. This is now
9963 supported by any public key method supporting the encrypt operation. A
9964 ctrl is added to allow the public key algorithm to examine or modify
9965 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9970 * Add a ctrl to asn1 method to allow a public key algorithm to express
9971 a default digest type to use. In most cases this will be SHA1 but some
9972 algorithms (such as GOST) need to specify an alternative digest. The
9973 return value indicates how strong the preference is 1 means optional and
9974 2 is mandatory (that is it is the only supported type). Modify
9975 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9976 use the default md. Update openssl utilities to use the default digest
9977 type for signing if it is not explicitly indicated.
9981 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9982 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9983 signing method from the key type. This effectively removes the link
9984 between digests and public key types.
9988 * Add an OID cross reference table and utility functions. Its purpose is to
9989 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9990 rsaEncryption. This will allow some of the algorithm specific hackery
9991 needed to use the correct OID to be removed.
9995 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9996 structures for PKCS7_sign(). They are now set up by the relevant public
10001 * Add provisional EC pkey method with support for ECDSA and ECDH.
10005 * Add support for key derivation (agreement) in the API, DH method and
10010 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10011 public and private key formats. As a side effect these add additional
10012 command line functionality not previously available: DSA signatures can be
10013 generated and verified using pkeyutl and DH key support and generation in
10020 *Oliver Tappe <zooey@hirschkaefer.de>*
10022 * New make target "install_html_docs" installs HTML renditions of the
10025 *Oliver Tappe <zooey@hirschkaefer.de>*
10027 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10028 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10029 support key and parameter generation and add initial key generation
10030 functionality for RSA.
10034 * Add functions for main EVP_PKEY_method operations. The undocumented
10035 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10036 `EVP_PKEY_{encrypt,decrypt}_old`.
10040 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10041 key API, doesn't do much yet.
10045 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10046 public key algorithms. New option to openssl utility:
10047 "list-public-key-algorithms" to print out info.
10051 * Implement the Supported Elliptic Curves Extension for
10052 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10056 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10057 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10061 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10062 utilities such as rsa, dsa, dsaparam etc except they process any key
10067 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10068 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10069 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10074 * Initial support for pluggable public key ASN1.
10075 De-spaghettify the public key ASN1 handling. Move public and private
10076 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10077 algorithm specific handling to a single module within the relevant
10078 algorithm directory. Add functions to allow (near) opaque processing
10079 of public and private key structures.
10083 * Implement the Supported Point Formats Extension for
10084 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10088 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10089 for the psk identity [hint] and the psk callback functions to the
10090 SSL_SESSION, SSL and SSL_CTX structure.
10093 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10097 SSL_CTX_use_psk_identity_hint
10098 SSL_get_psk_identity_hint
10099 SSL_get_psk_identity
10100 SSL_use_psk_identity_hint
10102 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10104 * Add RFC 3161 compliant time stamp request creation, response generation
10105 and response verification functionality.
10107 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10109 * Add initial support for TLS extensions, specifically for the server_name
10110 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10111 have new members for a hostname. The SSL data structure has an
10112 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10113 stored in that context to allow for session resumption, even after the
10114 SSL has been switched to a new SSL_CTX in reaction to a client's
10115 server_name extension.
10117 New functions (subject to change):
10119 SSL_get_servername()
10120 SSL_get_servername_type()
10123 New CTRL codes and macros (subject to change):
10125 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10126 - SSL_CTX_set_tlsext_servername_callback()
10127 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10128 - SSL_CTX_set_tlsext_servername_arg()
10129 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10131 openssl s_client has a new '-servername ...' option.
10133 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10134 '-key2 ...', '-servername_fatal' (subject to change). This allows
10135 testing the HostName extension for a specific single hostname ('-cert'
10136 and '-key' remain fallbacks for handshakes without HostName
10137 negotiation). If the unrecognized_name alert has to be sent, this by
10138 default is a warning; it becomes fatal with the '-servername_fatal'
10141 *Peter Sylvester, Remy Allais, Christophe Renou*
10143 * Whirlpool hash implementation is added.
10147 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10148 bn(64,32). Because of instruction set limitations it doesn't have
10149 any negative impact on performance. This was done mostly in order
10150 to make it possible to share assembler modules, such as bn_mul_mont
10151 implementations, between 32- and 64-bit builds without hassle.
10155 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10156 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10161 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10162 dedicated Montgomery multiplication procedure, is introduced.
10163 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10164 "64-bit" performance on certain 32-bit targets.
10168 * New option SSL_OP_NO_COMP to disable use of compression selectively
10169 in SSL structures. New SSL ctrl to set maximum send fragment size.
10170 Save memory by setting the I/O buffer sizes dynamically instead of
10171 using the maximum available value.
10175 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10176 in addition to the text details.
10180 * Very, very preliminary EXPERIMENTAL support for printing of general
10181 ASN1 structures. This currently produces rather ugly output and doesn't
10182 handle several customised structures at all.
10186 * Integrated support for PVK file format and some related formats such
10187 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10188 these in the 'rsa' and 'dsa' utilities.
10192 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10196 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10197 place for the (very old) "NETSCAPE" format certificates which are now
10198 handled using new ASN1 code equivalents.
10202 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10203 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10204 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10208 * Modify CRL distribution points extension code to print out previously
10209 unsupported fields. Enhance extension setting code to allow setting of
10214 * Add print and set support for Issuing Distribution Point CRL extension.
10218 * Change 'Configure' script to enable Camellia by default.
10225 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10227 * When rejecting SSL/TLS records due to an incorrect version number, never
10228 update s->server with a new major version number. As of
10229 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10230 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10231 the previous behavior could result in a read attempt at NULL when
10232 receiving specific incorrect SSL/TLS records once record payload
10233 protection is active. ([CVE-2010-0740])
10235 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10237 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10238 could be crashed if the relevant tables were not present (e.g. chrooted).
10240 *Tomas Hoger <thoger@redhat.com>*
10242 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10244 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10246 *Martin Olsson, Neel Mehta*
10248 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10249 accommodate for stack sorting, always a write lock!).
10253 * On some versions of WIN32 Heap32Next is very slow. This can cause
10254 excessive delays in the RAND_poll(): over a minute. As a workaround
10255 include a time check in the inner Heap32Next loop too.
10259 * The code that handled flushing of data in SSL/TLS originally used the
10260 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10261 the problem outlined in PR#1949. The fix suggested there however can
10262 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10263 of Apache). So instead simplify the code to flush unconditionally.
10264 This should be fine since flushing with no data to flush is a no op.
10268 * Handle TLS versions 2.0 and later properly and correctly use the
10269 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10270 off ancient servers have a habit of sticking around for a while...
10274 * Modify compression code so it frees up structures without using the
10275 ex_data callbacks. This works around a problem where some applications
10276 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10277 restarting) then use compression (e.g. SSL with compression) later.
10278 This results in significant per-connection memory leaks and
10279 has caused some security issues including CVE-2008-1678 and
10284 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10285 change when encrypting or decrypting.
10289 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10290 connect and renegotiate with servers which do not support RI.
10291 Until RI is more widely deployed this option is enabled by default.
10295 * Add "missing" ssl ctrls to clear options and mode.
10299 * If client attempts to renegotiate and doesn't support RI respond with
10300 a no_renegotiation alert as required by RFC5746. Some renegotiating
10301 TLS clients will continue a connection gracefully when they receive
10302 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10303 waiting for a server hello which it will never receive. Now we treat a
10304 received no_renegotiation alert as a fatal error. This is because
10305 applications requesting a renegotiation might well expect it to succeed
10306 and would have no code in place to handle the server denying it so the
10307 only safe thing to do is to terminate the connection.
10311 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10312 peer supports secure renegotiation and 0 otherwise. Print out peer
10313 renegotiation support in s_client/s_server.
10317 * Replace the highly broken and deprecated SPKAC certification method with
10318 the updated NID creation version. This should correctly handle UTF8.
10322 * Implement RFC5746. Re-enable renegotiation but require the extension
10323 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10324 turns out to be a bad idea. It has been replaced by
10325 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10326 SSL_CTX_set_options(). This is really not recommended unless you
10327 know what you are doing.
10329 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10331 * Fixes to stateless session resumption handling. Use initial_ctx when
10332 issuing and attempting to decrypt tickets in case it has changed during
10333 servername handling. Use a non-zero length session ID when attempting
10334 stateless session resumption: this makes it possible to determine if
10335 a resumption has occurred immediately after receiving server hello
10336 (several places in OpenSSL subtly assume this) instead of later in
10341 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10342 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10343 fixes for a few places where the return code is not checked
10346 *Julia Lawall <julia@diku.dk>*
10348 * Add --strict-warnings option to Configure script to include devteam
10349 warnings in other configurations.
10353 * Add support for --libdir option and LIBDIR variable in makefiles. This
10354 makes it possible to install openssl libraries in locations which
10355 have names other than "lib", for example "/usr/lib64" which some
10358 *Steve Henson, based on patch from Jeremy Utley*
10360 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10361 X690 8.9.12 and can produce some misleading textual output of OIDs.
10363 *Steve Henson, reported by Dan Kaminsky*
10365 * Delete MD2 from algorithm tables. This follows the recommendation in
10366 several standards that it is not used in new applications due to
10367 several cryptographic weaknesses. For binary compatibility reasons
10368 the MD2 API is still compiled in by default.
10372 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10377 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10378 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10381 *Guenter <lists@gknw.net>*
10383 * Fix the server certificate chain building code to use X509_verify_cert(),
10384 it used to have an ad-hoc builder which was unable to cope with anything
10385 other than a simple chain.
10387 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10389 * Don't check self signed certificate signatures in X509_verify_cert()
10390 by default (a flag can override this): it just wastes time without
10391 adding any security. As a useful side effect self signed root CAs
10392 with non-FIPS digests are now usable in FIPS mode.
10396 * In dtls1_process_out_of_seq_message() the check if the current message
10397 is already buffered was missing. For every new message was memory
10398 allocated, allowing an attacker to perform an denial of service attack
10399 with sending out of seq handshake messages until there is no memory
10400 left. Additionally every future message was buffered, even if the
10401 sequence number made no sense and would be part of another handshake.
10402 So only messages with sequence numbers less than 10 in advance will be
10403 buffered. ([CVE-2009-1378])
10405 *Robin Seggelmann, discovered by Daniel Mentz*
10407 * Records are buffered if they arrive with a future epoch to be
10408 processed after finishing the corresponding handshake. There is
10409 currently no limitation to this buffer allowing an attacker to perform
10410 a DOS attack with sending records with future epochs until there is no
10411 memory left. This patch adds the pqueue_size() function to determine
10412 the size of a buffer and limits the record buffer to 100 entries.
10415 *Robin Seggelmann, discovered by Daniel Mentz*
10417 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10418 parent structure is freed. ([CVE-2009-1379])
10422 * Handle non-blocking I/O properly in SSL_shutdown() call.
10424 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10426 * Add `2.5.4.*` OIDs
10428 *Ilya O. <vrghost@gmail.com>*
10430 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10432 * Disable renegotiation completely - this fixes a severe security
10433 problem ([CVE-2009-3555]) at the cost of breaking all
10434 renegotiation. Renegotiation can be re-enabled by setting
10435 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10436 run-time. This is really not recommended unless you know what
10441 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10443 * Don't set val to NULL when freeing up structures, it is freed up by
10444 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10445 zeroing past the valid field. ([CVE-2009-0789])
10447 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10449 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10450 checked correctly. This would allow some invalid signed attributes to
10451 appear to verify correctly. ([CVE-2009-0591])
10453 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10455 * Reject UniversalString and BMPString types with invalid lengths. This
10456 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10457 a legal length. ([CVE-2009-0590])
10461 * Set S/MIME signing as the default purpose rather than setting it
10462 unconditionally. This allows applications to override it at the store
10467 * Permit restricted recursion of ASN1 strings. This is needed in practice
10468 to handle some structures.
10472 * Improve efficiency of mem_gets: don't search whole buffer each time
10475 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10477 * New -hex option for openssl rand.
10481 * Print out UTF8String and NumericString when parsing ASN1.
10485 * Support NumericString type for name components.
10489 * Allow CC in the environment to override the automatically chosen
10490 compiler. Note that nothing is done to ensure flags work with the
10495 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10497 * Properly check EVP_VerifyFinal() and similar return values
10500 *Ben Laurie, Bodo Moeller, Google Security Team*
10502 * Enable TLS extensions by default.
10506 * Allow the CHIL engine to be loaded, whether the application is
10507 multithreaded or not. (This does not release the developer from the
10508 obligation to set up the dynamic locking callbacks.)
10510 *Sander Temme <sander@temme.net>*
10512 * Use correct exit code if there is an error in dgst command.
10514 *Steve Henson; problem pointed out by Roland Dirlewanger*
10516 * Tweak Configure so that you need to say "experimental-jpake" to enable
10517 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10521 * Add experimental JPAKE support, including demo authentication in
10522 s_client and s_server.
10526 * Set the comparison function in v3_addr_canonize().
10528 *Rob Austein <sra@hactrn.net>*
10530 * Add support for XMPP STARTTLS in s_client.
10532 *Philip Paeps <philip@freebsd.org>*
10534 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10535 to ensure that even with this option, only ciphersuites in the
10536 server's preference list will be accepted. (Note that the option
10537 applies only when resuming a session, so the earlier behavior was
10538 just about the algorithm choice for symmetric cryptography.)
10542 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10544 * Fix NULL pointer dereference if a DTLS server received
10545 ChangeCipherSpec as first record ([CVE-2009-1386]).
10549 * Fix a state transition in s3_srvr.c and d1_srvr.c
10550 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10552 *Nagendra Modadugu*
10554 * The fix in 0.9.8c that supposedly got rid of unsafe
10555 double-checked locking was incomplete for RSA blinding,
10556 addressing just one layer of what turns out to have been
10557 doubly unsafe triple-checked locking.
10559 So now fix this for real by retiring the MONT_HELPER macro
10560 in crypto/rsa/rsa_eay.c.
10562 *Bodo Moeller; problem pointed out by Marius Schilder*
10564 * Various precautionary measures:
10566 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10568 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10569 (NB: This would require knowledge of the secret session ticket key
10570 to exploit, in which case you'd be SOL either way.)
10572 - Change bn_nist.c so that it will properly handle input BIGNUMs
10573 outside the expected range.
10575 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10578 *Neel Mehta, Bodo Moeller*
10580 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10581 the load fails. Useful for distros.
10583 *Ben Laurie and the FreeBSD team*
10585 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10589 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10593 * Expand ENGINE to support engine supplied SSL client certificate functions.
10595 This work was sponsored by Logica.
10599 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10600 keystores. Support for SSL/TLS client authentication too.
10601 Not compiled unless enable-capieng specified to Configure.
10603 This work was sponsored by Logica.
10607 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10608 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10609 attribute creation routines such as certificate requests and PKCS#12
10614 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10616 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10617 handshake which could lead to a client crash as found using the
10618 Codenomicon TLS test suite ([CVE-2008-1672])
10620 *Steve Henson, Mark Cox*
10622 * Fix double free in TLS server name extensions which could lead to
10623 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10627 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10629 Clear the error queue to ensure that error entries left from
10630 older function calls do not interfere with the correct operation.
10632 *Lutz Jaenicke, Erik de Castro Lopo*
10634 * Remove root CA certificates of commercial CAs:
10636 The OpenSSL project does not recommend any specific CA and does not
10637 have any policy with respect to including or excluding any CA.
10638 Therefore, it does not make any sense to ship an arbitrary selection
10639 of root CA certificates with the OpenSSL software.
10643 * RSA OAEP patches to fix two separate invalid memory reads.
10644 The first one involves inputs when 'lzero' is greater than
10645 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10646 before the beginning of from). The second one involves inputs where
10647 the 'db' section contains nothing but zeroes (there is a one-byte
10648 invalid read after the end of 'db').
10650 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10652 * Partial backport from 0.9.9-dev:
10654 Introduce bn_mul_mont (dedicated Montgomery multiplication
10655 procedure) as a candidate for BIGNUM assembler implementation.
10656 While 0.9.9-dev uses assembler for various architectures, only
10657 x86_64 is available by default here in the 0.9.8 branch, and
10658 32-bit x86 is available through a compile-time setting.
10660 To try the 32-bit x86 assembler implementation, use Configure
10661 option "enable-montasm" (which exists only for this backport).
10663 As "enable-montasm" for 32-bit x86 disclaims code stability
10664 anyway, in this constellation we activate additional code
10665 backported from 0.9.9-dev for further performance improvements,
10666 namely BN_from_montgomery_word. (To enable this otherwise,
10667 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10669 *Andy Polyakov (backport partially by Bodo Moeller)*
10671 * Add TLS session ticket callback. This allows an application to set
10672 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10673 values. This is useful for key rollover for example where several key
10674 sets may exist with different names.
10678 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10679 This was broken until now in 0.9.8 releases, such that the only way
10680 a registered ENGINE could be used (assuming it initialises
10681 successfully on the host) was to explicitly set it as the default
10682 for the relevant algorithms. This is in contradiction with 0.9.7
10683 behaviour and the documentation. With this fix, when an ENGINE is
10684 registered into a given algorithm's table of implementations, the
10685 'uptodate' flag is reset so that auto-discovery will be used next
10686 time a new context for that algorithm attempts to select an
10689 *Ian Lister (tweaked by Geoff Thorpe)*
10691 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10692 implementation in the following ways:
10694 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10697 Lack of BER streaming support means one pass streaming processing is
10698 only supported if data is detached: setting the streaming flag is
10699 ignored for embedded content.
10701 CMS support is disabled by default and must be explicitly enabled
10702 with the enable-cms configuration option.
10706 * Update the GMP engine glue to do direct copies between BIGNUM and
10707 mpz_t when openssl and GMP use the same limb size. Otherwise the
10708 existing "conversion via a text string export" trick is still used.
10710 *Paul Sheer <paulsheer@gmail.com>*
10712 * Zlib compression BIO. This is a filter BIO which compressed and
10713 uncompresses any data passed through it.
10717 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10718 RFC3394 compatible AES key wrapping.
10722 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10723 sets string data without copying. X509_ALGOR_set0() and
10724 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10725 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10726 from an X509_ATTRIBUTE structure optionally checking it occurs only
10727 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10732 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10733 to get the expected BN_FLG_CONSTTIME behavior.
10735 *Bodo Moeller (Google)*
10739 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10740 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10741 - added some more tests to do_tests.pl
10742 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10743 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10744 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10745 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10746 - various changes to netware.pl to enable gcc-cross builds on Win32
10748 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10749 - various changes to fix missing prototype warnings
10750 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10751 - added AES, WHIRLPOOL and CPUID assembler code to build files
10752 - added missing AES assembler make rules to mk1mf.pl
10753 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10755 *Guenter Knauf <eflash@gmx.net>*
10757 * Implement certificate status request TLS extension defined in RFC3546.
10758 A client can set the appropriate parameters and receive the encoded
10759 OCSP response via a callback. A server can query the supplied parameters
10760 and set the encoded OCSP response in the callback. Add simplified examples
10761 to s_client and s_server.
10765 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10767 * Fix various bugs:
10768 + Binary incompatibility of ssl_ctx_st structure
10769 + DTLS interoperation with non-compliant servers
10770 + Don't call get_session_cb() without proposed session
10771 + Fix ia64 assembler code
10773 *Andy Polyakov, Steve Henson*
10775 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10777 * DTLS Handshake overhaul. There were longstanding issues with
10778 OpenSSL DTLS implementation, which were making it impossible for
10779 RFC 4347 compliant client to communicate with OpenSSL server.
10780 Unfortunately just fixing these incompatibilities would "cut off"
10781 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10782 server keeps tolerating non RFC compliant syntax. The opposite is
10783 not true, 0.9.8f client can not communicate with earlier server.
10784 This update even addresses CVE-2007-4995.
10788 * Changes to avoid need for function casts in OpenSSL: some compilers
10789 (gcc 4.2 and later) reject their use.
10790 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10793 * Add RFC4507 support to OpenSSL. This includes the corrections in
10794 RFC4507bis. The encrypted ticket format is an encrypted encoded
10795 SSL_SESSION structure, that way new session features are automatically
10798 If a client application caches session in an SSL_SESSION structure
10799 support is transparent because tickets are now stored in the encoded
10802 The SSL_CTX structure automatically generates keys for ticket
10803 protection in servers so again support should be possible
10804 with no application modification.
10806 If a client or server wishes to disable RFC4507 support then the option
10807 SSL_OP_NO_TICKET can be set.
10809 Add a TLS extension debugging callback to allow the contents of any client
10810 or server extensions to be examined.
10812 This work was sponsored by Google.
10816 * Add initial support for TLS extensions, specifically for the server_name
10817 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10818 have new members for a hostname. The SSL data structure has an
10819 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10820 stored in that context to allow for session resumption, even after the
10821 SSL has been switched to a new SSL_CTX in reaction to a client's
10822 server_name extension.
10824 New functions (subject to change):
10826 SSL_get_servername()
10827 SSL_get_servername_type()
10830 New CTRL codes and macros (subject to change):
10832 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10833 - SSL_CTX_set_tlsext_servername_callback()
10834 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10835 - SSL_CTX_set_tlsext_servername_arg()
10836 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10838 openssl s_client has a new '-servername ...' option.
10840 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10841 '-key2 ...', '-servername_fatal' (subject to change). This allows
10842 testing the HostName extension for a specific single hostname ('-cert'
10843 and '-key' remain fallbacks for handshakes without HostName
10844 negotiation). If the unrecognized_name alert has to be sent, this by
10845 default is a warning; it becomes fatal with the '-servername_fatal'
10848 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10850 * Add AES and SSE2 assembly language support to VC++ build.
10854 * Mitigate attack on final subtraction in Montgomery reduction.
10858 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10859 (which previously caused an internal error).
10863 * Squeeze another 10% out of IGE mode when in != out.
10867 * AES IGE mode speedup.
10869 *Dean Gaudet (Google)*
10871 * Add the Korean symmetric 128-bit cipher SEED (see
10872 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10873 add SEED ciphersuites from RFC 4162:
10875 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10876 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10877 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10878 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10880 To minimize changes between patchlevels in the OpenSSL 0.9.8
10881 series, SEED remains excluded from compilation unless OpenSSL
10882 is configured with 'enable-seed'.
10884 *KISA, Bodo Moeller*
10886 * Mitigate branch prediction attacks, which can be practical if a
10887 single processor is shared, allowing a spy process to extract
10888 information. For detailed background information, see
10889 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10890 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10891 and Necessary Software Countermeasures"). The core of the change
10892 are new versions BN_div_no_branch() and
10893 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10894 respectively, which are slower, but avoid the security-relevant
10895 conditional branches. These are automatically called by BN_div()
10896 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10897 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10898 remove a conditional branch.
10900 BN_FLG_CONSTTIME is the new name for the previous
10901 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10902 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10903 in the exponent causes BN_mod_exp_mont() to use the alternative
10904 implementation in BN_mod_exp_mont_consttime().) The old name
10905 remains as a deprecated alias.
10907 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10908 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10909 constant-time implementations for more than just exponentiation.
10910 Here too the old name is kept as a deprecated alias.
10912 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10913 the BN_BLINDING structure gets an independent copy of the
10914 modulus. This means that the previous `BIGNUM *m` argument to
10915 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10916 essentially becomes `const BIGNUM *m`, although we can't actually
10917 change this in the header file before 0.9.9. It allows
10918 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10919 enable BN_FLG_CONSTTIME.
10921 *Matthew D Wood (Intel Corp)*
10923 * In the SSL/TLS server implementation, be strict about session ID
10924 context matching (which matters if an application uses a single
10925 external cache for different purposes). Previously,
10926 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10927 set. This did ensure strict client verification, but meant that,
10928 with applications using a single external cache for quite
10929 different requirements, clients could circumvent ciphersuite
10930 restrictions for a given session ID context by starting a session
10931 in a different context.
10935 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10936 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10937 authentication-only ciphersuites.
10941 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10942 not complete and could lead to a possible single byte overflow
10943 ([CVE-2007-5135]) [Ben Laurie]
10945 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10947 * Since AES128 and AES256 (and similarly Camellia128 and
10948 Camellia256) share a single mask bit in the logic of
10949 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10950 kludge to work properly if AES128 is available and AES256 isn't
10951 (or if Camellia128 is available and Camellia256 isn't).
10955 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10956 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10957 When a point or a seed is encoded in a BIT STRING, we need to
10958 prevent the removal of trailing zero bits to get the proper DER
10959 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10960 of a NamedBitList, for which trailing 0 bits need to be removed.)
10964 * Have SSL/TLS server implementation tolerate "mismatched" record
10965 protocol version while receiving ClientHello even if the
10966 ClientHello is fragmented. (The server can't insist on the
10967 particular protocol version it has chosen before the ServerHello
10968 message has informed the client about his choice.)
10972 * Add RFC 3779 support.
10974 *Rob Austein for ARIN, Ben Laurie*
10976 * Load error codes if they are not already present instead of using a
10977 static variable. This allows them to be cleanly unloaded and reloaded.
10978 Improve header file function name parsing.
10982 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10983 or CAPABILITY handshake as required by RFCs.
10985 *Goetz Babin-Ebell*
10987 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10989 * Introduce limits to prevent malicious keys being able to
10990 cause a denial of service. ([CVE-2006-2940])
10992 *Steve Henson, Bodo Moeller*
10994 * Fix ASN.1 parsing of certain invalid structures that can result
10995 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10997 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10998 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11000 * Fix SSL client code which could crash if connecting to a
11001 malicious SSLv2 server. ([CVE-2006-4343])
11003 *Tavis Ormandy and Will Drewry, Google Security Team*
11005 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11006 match only those. Before that, "AES256-SHA" would be interpreted
11007 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11008 the same strength classification in 0.9.7h) as we currently only
11009 have a single AES bit in the ciphersuite description bitmap.
11010 That change, however, also applied to ciphersuite strings such as
11011 "RC4-MD5" that intentionally matched multiple ciphersuites --
11012 namely, SSL 2.0 ciphersuites in addition to the more common ones
11013 from SSL 3.0/TLS 1.0.
11015 So we change the selection algorithm again: Naming an explicit
11016 ciphersuite selects this one ciphersuite, and any other similar
11017 ciphersuite (same bitmap) from *other* protocol versions.
11018 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11019 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11021 Since SSL 2.0 does not have any ciphersuites for which the
11022 128/256 bit distinction would be relevant, this works for now.
11023 The proper fix will be to use different bits for AES128 and
11024 AES256, which would have avoided the problems from the beginning;
11025 however, bits are scarce, so we can only do this in a new release
11026 (not just a patchlevel) when we can change the SSL_CIPHER
11027 definition to split the single 'unsigned long mask' bitmap into
11028 multiple values to extend the available space.
11032 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11034 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11035 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11037 * Add AES IGE and biIGE modes.
11041 * Change the Unix randomness entropy gathering to use poll() when
11042 possible instead of select(), since the latter has some
11043 undesirable limitations.
11045 *Darryl Miles via Richard Levitte and Bodo Moeller*
11047 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11048 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11049 cannot be implicitly activated as part of, e.g., the "AES" alias.
11050 However, please upgrade to OpenSSL 0.9.9[-dev] for
11051 non-experimental use of the ECC ciphersuites to get TLS extension
11052 support, which is required for curve and point format negotiation
11053 to avoid potential handshake problems.
11057 * Disable rogue ciphersuites:
11059 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11060 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11061 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11063 The latter two were purportedly from
11064 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11067 Also deactivate the remaining ciphersuites from
11068 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11069 unofficial, and the ID has long expired.
11073 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11074 dual-core machines) and other potential thread-safety issues.
11078 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11079 versions), which is now available for royalty-free use
11080 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11081 Also, add Camellia TLS ciphersuites from RFC 4132.
11083 To minimize changes between patchlevels in the OpenSSL 0.9.8
11084 series, Camellia remains excluded from compilation unless OpenSSL
11085 is configured with 'enable-camellia'.
11089 * Disable the padding bug check when compression is in use. The padding
11090 bug check assumes the first packet is of even length, this is not
11091 necessarily true if compression is enabled and can result in false
11092 positives causing handshake failure. The actual bug test is ancient
11093 code so it is hoped that implementations will either have fixed it by
11094 now or any which still have the bug do not support compression.
11098 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11100 * When applying a cipher rule check to see if string match is an explicit
11101 cipher suite and only match that one cipher suite if it is.
11105 * Link in manifests for VC++ if needed.
11107 *Austin Ziegler <halostatue@gmail.com>*
11109 * Update support for ECC-based TLS ciphersuites according to
11110 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11111 TLS extensions, which are supported starting with the 0.9.9
11112 branch, not in the OpenSSL 0.9.8 branch).
11116 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11117 opaque EVP_CIPHER_CTX handling.
11121 * Fixes and enhancements to zlib compression code. We now only use
11122 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11123 to conform with the standards mentioned here:
11124 <http://www.zlib.net/DLL_FAQ.txt>
11125 Static zlib linking now works on Windows and the new --with-zlib-include
11126 --with-zlib-lib options to Configure can be used to supply the location
11127 of the headers and library. Gracefully handle case where zlib library
11132 * Several fixes and enhancements to the OID generation code. The old code
11133 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11134 handle numbers larger than ULONG_MAX, truncated printing and had a
11135 non standard OBJ_obj2txt() behaviour.
11139 * Add support for building of engines under engine/ as shared libraries
11140 under VC++ build system.
11144 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11145 Hopefully, we will not see any false combination of paths any more.
11149 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11151 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11152 (part of SSL_OP_ALL). This option used to disable the
11153 countermeasure against man-in-the-middle protocol-version
11154 rollback in the SSL 2.0 server implementation, which is a bad
11155 idea. ([CVE-2005-2969])
11157 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11158 for Information Security, National Institute of Advanced Industrial
11159 Science and Technology [AIST], Japan)*
11161 * Add two function to clear and return the verify parameter flags.
11165 * Keep cipherlists sorted in the source instead of sorting them at
11166 runtime, thus removing the need for a lock.
11170 * Avoid some small subgroup attacks in Diffie-Hellman.
11172 *Nick Mathewson and Ben Laurie*
11174 * Add functions for well-known primes.
11178 * Extended Windows CE support.
11180 *Satoshi Nakamura and Andy Polyakov*
11182 * Initialize SSL_METHOD structures at compile time instead of during
11183 runtime, thus removing the need for a lock.
11187 * Make PKCS7_decrypt() work even if no certificate is supplied by
11188 attempting to decrypt each encrypted key in turn. Add support to
11193 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11195 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11198 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11202 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11203 key into the same file any more.
11207 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11211 * Add -utf8 command line and config file option to 'ca'.
11213 *Stefan <stf@udoma.org*
11215 * Removed the macro des_crypt(), as it seems to conflict with some
11216 libraries. Use DES_crypt().
11220 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11221 involves renaming the source and generated shared-libs for
11222 both. The engines will accept the corrected or legacy ids
11223 ('ncipher' and '4758_cca' respectively) when binding. NB,
11224 this only applies when building 'shared'.
11226 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11228 * Add attribute functions to EVP_PKEY structure. Modify
11229 PKCS12_create() to recognize a CSP name attribute and
11230 use it. Make -CSP option work again in pkcs12 utility.
11234 * Add new functionality to the bn blinding code:
11235 - automatic re-creation of the BN_BLINDING parameters after
11236 a fixed number of uses (currently 32)
11237 - add new function for parameter creation
11238 - introduce flags to control the update behaviour of the
11239 BN_BLINDING parameters
11240 - hide BN_BLINDING structure
11241 Add a second BN_BLINDING slot to the RSA structure to improve
11242 performance when a single RSA object is shared among several
11247 * Add support for DTLS.
11249 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11251 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11252 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11256 * Remove buggy and incomplete DH cert support from
11257 ssl/ssl_rsa.c and ssl/s3_both.c
11261 * Use SHA-1 instead of MD5 as the default digest algorithm for
11262 the `apps/openssl` commands.
11266 * Compile clean with "-Wall -Wmissing-prototypes
11267 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11268 DEBUG_SAFESTACK must also be set.
11272 * Change ./Configure so that certain algorithms can be disabled by default.
11273 The new counterpiece to "no-xxx" is "enable-xxx".
11275 The patented RC5 and MDC2 algorithms will now be disabled unless
11276 "enable-rc5" and "enable-mdc2", respectively, are specified.
11278 (IDEA remains enabled despite being patented. This is because IDEA
11279 is frequently required for interoperability, and there is no license
11280 fee for non-commercial use. As before, "no-idea" can be used to
11281 avoid this algorithm.)
11285 * Add processing of proxy certificates (see RFC 3820). This work was
11286 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11287 EGEE (Enabling Grids for E-science in Europe).
11291 * RC4 performance overhaul on modern architectures/implementations, such
11292 as Intel P4, IA-64 and AMD64.
11296 * New utility extract-section.pl. This can be used specify an alternative
11297 section number in a pod file instead of having to treat each file as
11298 a separate case in Makefile. This can be done by adding two lines to the
11301 =for comment openssl_section:XXX
11303 The blank line is mandatory.
11307 * New arguments -certform, -keyform and -pass for s_client and s_server
11308 to allow alternative format key and certificate files and passphrase
11313 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11314 update associated structures and add various utility functions.
11316 Add new policy related verify parameters, include policy checking in
11317 standard verify code. Enhance 'smime' application with extra parameters
11318 to support policy checking and print out.
11322 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11323 Nehemiah processors. These extensions support AES encryption in hardware
11324 as well as RNG (though RNG support is currently disabled).
11326 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11328 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11332 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11334 *Andy Polyakov and a number of other people*
11336 * Improved PowerPC platform support. Most notably BIGNUM assembler
11337 implementation contributed by IBM.
11339 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11341 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11342 exponent rather than 'unsigned long'. There is a corresponding change to
11343 the new 'rsa_keygen' element of the RSA_METHOD structure.
11345 *Jelte Jansen, Geoff Thorpe*
11347 * Functionality for creating the initial serial number file is now
11348 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11350 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11351 number file to 1, which is bound to cause problems. To avoid
11352 the problems while respecting compatibility between different 0.9.7
11353 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11354 CA.pl for serial number initialization. With the new release 0.9.8,
11355 we can fix the problem directly in the 'ca' utility.)
11359 * Reduced header interdependencies by declaring more opaque objects in
11360 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11361 give fewer recursive includes, which could break lazy source code - so
11362 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11363 developers should define this symbol when building and using openssl to
11364 ensure they track the recommended behaviour, interfaces, [etc], but
11365 backwards-compatible behaviour prevails when this isn't defined.
11369 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11373 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11374 This will generate a random key of the appropriate length based on the
11375 cipher context. The EVP_CIPHER can provide its own random key generation
11376 routine to support keys of a specific form. This is used in the des and
11377 3des routines to generate a key of the correct parity. Update S/MIME
11378 code to use new functions and hence generate correct parity DES keys.
11379 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11380 valid (weak or incorrect parity).
11384 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11385 as looking them up. This is useful when the verified structure may contain
11386 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11387 present unless the new PKCS7_NO_CRL flag is asserted.
11391 * Extend ASN1 oid configuration module. It now additionally accepts the
11394 shortName = some long name, 1.2.3.4
11398 * Reimplemented the BN_CTX implementation. There is now no more static
11399 limitation on the number of variables it can handle nor the depth of the
11400 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11401 information can now expand as required, and rather than having a single
11402 static array of bignums, BN_CTX now uses a linked-list of such arrays
11403 allowing it to expand on demand whilst maintaining the usefulness of
11404 BN_CTX's "bundling".
11408 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11409 to allow all RSA operations to function using a single BN_CTX.
11413 * Preliminary support for certificate policy evaluation and checking. This
11414 is initially intended to pass the tests outlined in "Conformance Testing
11415 of Relying Party Client Certificate Path Processing Logic" v1.07.
11419 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11420 remained unused and not that useful. A variety of other little bignum
11421 tweaks and fixes have also been made continuing on from the audit (see
11426 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11427 associated ASN1, EVP and SSL functions and old ASN1 macros.
11431 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11432 and this should never fail. So the return value from the use of
11433 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11434 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11438 * BN_CTX_get() should return zero-valued bignums, providing the same
11439 initialised value as BN_new().
11441 *Geoff Thorpe, suggested by Ulf Möller*
11443 * Support for inhibitAnyPolicy certificate extension.
11447 * An audit of the BIGNUM code is underway, for which debugging code is
11448 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11449 is considered valid when processing BIGNUMs, and causes execution to
11450 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11451 further steps are taken to deliberately pollute unused data in BIGNUM
11452 structures to try and expose faulty code further on. For now, openssl will
11453 (in its default mode of operation) continue to tolerate the inconsistent
11454 forms that it has tolerated in the past, but authors and packagers should
11455 consider trying openssl and their own applications when compiled with
11456 these debugging symbols defined. It will help highlight potential bugs in
11457 their own code, and will improve the test coverage for OpenSSL itself. At
11458 some point, these tighter rules will become openssl's default to improve
11459 maintainability, though the assert()s and other overheads will remain only
11460 in debugging configurations. See bn.h for more details.
11462 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11464 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11465 that can only be obtained through BN_CTX_new() (which implicitly
11466 initialises it). The presence of this function only made it possible
11467 to overwrite an existing structure (and cause memory leaks).
11471 * Because of the callback-based approach for implementing LHASH as a
11472 template type, lh_insert() adds opaque objects to hash-tables and
11473 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11474 to clean up those corresponding objects before destroying the hash table
11475 (and losing the object pointers). So some over-zealous constifications in
11476 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11477 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11478 prototyped to have "const" restrictions on the object pointers they are
11479 given (and so aren't required to cast them away any more).
11483 * The tmdiff.h API was so ugly and minimal that our own timing utility
11484 (speed) prefers to use its own implementation. The two implementations
11485 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11486 its object type properly exposed (MS_TM) instead of casting to/from
11487 `char *`. This may still change yet if someone realises MS_TM and
11489 aren't necessarily the greatest nomenclatures - but this is what was used
11490 internally to the implementation so I've used that for now.
11494 * Ensure that deprecated functions do not get compiled when
11495 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11496 the self-tests were still using deprecated key-generation functions so
11497 these have been updated also.
11501 * Reorganise PKCS#7 code to separate the digest location functionality
11502 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11503 New function PKCS7_set_digest() to set the digest type for PKCS#7
11504 digestedData type. Add additional code to correctly generate the
11505 digestedData type and add support for this type in PKCS7 initialization
11510 * New function PKCS7_set0_type_other() this initializes a PKCS7
11511 structure of type "other".
11515 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11516 sure the loop does correctly stop and breaking ("division by zero")
11517 modulus operations are not performed. The (pre-generated) prime
11518 table crypto/bn/bn_prime.h was already correct, but it could not be
11519 re-generated on some platforms because of the "division by zero"
11520 situation in the script.
11522 *Ralf S. Engelschall*
11524 * Update support for ECC-based TLS ciphersuites according to
11525 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11526 SHA-1 now is only used for "small" curves (where the
11527 representation of a field element takes up to 24 bytes); for
11528 larger curves, the field element resulting from ECDH is directly
11529 used as premaster secret.
11531 *Douglas Stebila (Sun Microsystems Laboratories)*
11533 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11534 curve secp160r1 to the tests.
11536 *Douglas Stebila (Sun Microsystems Laboratories)*
11538 * Add the possibility to load symbols globally with DSO.
11540 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11542 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11543 control of the error stack.
11547 * Add support for STORE in ENGINE.
11551 * Add the STORE type. The intention is to provide a common interface
11552 to certificate and key stores, be they simple file-based stores, or
11553 HSM-type store, or LDAP stores, or...
11554 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11558 * Add a generic structure called OPENSSL_ITEM. This can be used to
11559 pass a list of arguments to any function as well as provide a way
11560 for a function to pass data back to the caller.
11564 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11565 works like BUF_strdup() but can be used to duplicate a portion of
11566 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11571 * Add the function sk_find_ex() which works like sk_find(), but will
11572 return an index to an element even if an exact match couldn't be
11573 found. The index is guaranteed to point at the element where the
11574 searched-for key would be inserted to preserve sorting order.
11578 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11579 takes an extra flags argument for optional functionality. Currently,
11580 the following flags are defined:
11582 OBJ_BSEARCH_VALUE_ON_NOMATCH
11583 This one gets OBJ_bsearch_ex() to return a pointer to the first
11584 element where the comparing function returns a negative or zero
11587 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11588 This one gets OBJ_bsearch_ex() to return a pointer to the first
11589 element where the comparing function returns zero. This is useful
11590 if there are more than one element where the comparing function
11595 * Make it possible to create self-signed certificates with 'openssl ca'
11596 in such a way that the self-signed certificate becomes part of the
11597 CA database and uses the same mechanisms for serial number generation
11598 as all other certificate signing. The new flag '-selfsign' enables
11599 this functionality. Adapt CA.sh and CA.pl.in.
11603 * Add functionality to check the public key of a certificate request
11604 against a given private. This is useful to check that a certificate
11605 request can be signed by that key (self-signing).
11609 * Make it possible to have multiple active certificates with the same
11610 subject in the CA index file. This is done only if the keyword
11611 'unique_subject' is set to 'no' in the main CA section (default
11612 if 'CA_default') of the configuration file. The value is saved
11613 with the database itself in a separate index attribute file,
11614 named like the index file with '.attr' appended to the name.
11618 * Generate multi-valued AVAs using '+' notation in config files for
11623 * Support for nameConstraints certificate extension.
11627 * Support for policyConstraints certificate extension.
11631 * Support for policyMappings certificate extension.
11635 * Make sure the default DSA_METHOD implementation only uses its
11636 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11637 and change its own handlers to be NULL so as to remove unnecessary
11638 indirection. This lets alternative implementations fallback to the
11639 default implementation more easily.
11643 * Support for directoryName in GeneralName related extensions
11648 * Make it possible to link applications using Makefile.shared.
11649 Make that possible even when linking against static libraries!
11653 * Support for single pass processing for S/MIME signing. This now
11654 means that S/MIME signing can be done from a pipe, in addition
11655 cleartext signing (multipart/signed type) is effectively streaming
11656 and the signed data does not need to be all held in memory.
11658 This is done with a new flag PKCS7_STREAM. When this flag is set
11659 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11660 is done after the data is output (and digests calculated) in
11661 SMIME_write_PKCS7().
11665 * Add full support for -rpath/-R, both in shared libraries and
11666 applications, at least on the platforms where it's known how
11671 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11672 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11673 will now compute a table of multiples of the generator that
11674 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11675 faster (notably in the case of a single point multiplication,
11676 scalar * generator).
11678 *Nils Larsch, Bodo Moeller*
11680 * IPv6 support for certificate extensions. The various extensions
11681 which use the IP:a.b.c.d can now take IPv6 addresses using the
11682 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11687 * Added an ENGINE that implements RSA by performing private key
11688 exponentiations with the GMP library. The conversions to and from
11689 GMP's mpz_t format aren't optimised nor are any montgomery forms
11690 cached, and on x86 it appears OpenSSL's own performance has caught up.
11691 However there are likely to be other architectures where GMP could
11692 provide a boost. This ENGINE is not built in by default, but it can be
11693 specified at Configure time and should be accompanied by the necessary
11694 linker additions, eg;
11695 ./config -DOPENSSL_USE_GMP -lgmp
11699 * "openssl engine" will not display ENGINE/DSO load failure errors when
11700 testing availability of engines with "-t" - the old behaviour is
11701 produced by increasing the feature's verbosity with "-tt".
11705 * ECDSA routines: under certain error conditions uninitialized BN objects
11706 could be freed. Solution: make sure initialization is performed early
11707 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11712 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11713 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11714 software implementations. For DSA and DH, parameter generation can
11715 also be overridden by providing the appropriate method callbacks.
11719 * Change the "progress" mechanism used in key-generation and
11720 primality testing to functions that take a new BN_GENCB pointer in
11721 place of callback/argument pairs. The new API functions have `_ex`
11722 postfixes and the older functions are reimplemented as wrappers for
11723 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11724 declarations of the old functions to help (graceful) attempts to
11725 migrate to the new functions. Also, the new key-generation API
11726 functions operate on a caller-supplied key-structure and return
11727 success/failure rather than returning a key or NULL - this is to
11728 help make "keygen" another member function of RSA_METHOD etc.
11730 Example for using the new callback interface:
11732 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11733 void *my_arg = ...;
11736 BN_GENCB_set(&my_cb, my_callback, my_arg);
11738 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11739 /* For the meaning of a, b in calls to my_callback(), see the
11740 * documentation of the function that calls the callback.
11741 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11742 * my_callback should return 1 if it wants BN_is_prime_ex()
11743 * to continue, or 0 to stop.
11748 * Change the ZLIB compression method to be stateful, and make it
11749 available to TLS with the number defined in
11750 draft-ietf-tls-compression-04.txt.
11754 * Add the ASN.1 structures and functions for CertificatePair, which
11755 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11757 CertificatePair ::= SEQUENCE {
11758 forward [0] Certificate OPTIONAL,
11759 reverse [1] Certificate OPTIONAL,
11760 -- at least one of the pair shall be present -- }
11762 Also implement the PEM functions to read and write certificate
11763 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11765 This needed to be defined, mostly for the sake of the LDAP
11766 attribute crossCertificatePair, but may prove useful elsewhere as
11771 * Make it possible to inhibit symlinking of shared libraries in
11772 Makefile.shared, for Cygwin's sake.
11776 * Extend the BIGNUM API by creating a function
11777 void BN_set_negative(BIGNUM *a, int neg);
11778 and a macro that behave like
11779 int BN_is_negative(const BIGNUM *a);
11781 to avoid the need to access 'a->neg' directly in applications.
11785 * Implement fast modular reduction for pseudo-Mersenne primes
11786 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11787 EC_GROUP_new_curve_GFp() will now automatically use this
11790 *Nils Larsch <nla@trustcenter.de>*
11792 * Add new lock type (CRYPTO_LOCK_BN).
11796 * Change the ENGINE framework to automatically load engines
11797 dynamically from specific directories unless they could be
11798 found to already be built in or loaded. Move all the
11799 current engines except for the cryptodev one to a new
11800 directory engines/.
11801 The engines in engines/ are built as shared libraries if
11802 the "shared" options was given to ./Configure or ./config.
11803 Otherwise, they are inserted in libcrypto.a.
11804 /usr/local/ssl/engines is the default directory for dynamic
11805 engines, but that can be overridden at configure time through
11806 the usual use of --prefix and/or --openssldir, and at run
11807 time with the environment variable OPENSSL_ENGINES.
11809 *Geoff Thorpe and Richard Levitte*
11811 * Add Makefile.shared, a helper makefile to build shared
11812 libraries. Adapt Makefile.org.
11816 * Add version info to Win32 DLLs.
11818 *Peter 'Luna' Runestig" <peter@runestig.com>*
11820 * Add new 'medium level' PKCS#12 API. Certificates and keys
11821 can be added using this API to created arbitrary PKCS#12
11822 files while avoiding the low-level API.
11824 New options to PKCS12_create(), key or cert can be NULL and
11825 will then be omitted from the output file. The encryption
11826 algorithm NIDs can be set to -1 for no encryption, the mac
11827 iteration count can be set to 0 to omit the mac.
11829 Enhance pkcs12 utility by making the -nokeys and -nocerts
11830 options work when creating a PKCS#12 file. New option -nomac
11831 to omit the mac, NONE can be set for an encryption algorithm.
11832 New code is modified to use the enhanced PKCS12_create()
11833 instead of the low-level API.
11837 * Extend ASN1 encoder to support indefinite length constructed
11838 encoding. This can output sequences tags and octet strings in
11839 this form. Modify pk7_asn1.c to support indefinite length
11840 encoding. This is experimental and needs additional code to
11841 be useful, such as an ASN1 bio and some enhanced streaming
11844 Extend template encode functionality so that tagging is passed
11845 down to the template encoder.
11849 * Let 'openssl req' fail if an argument to '-newkey' is not
11850 recognized instead of using RSA as a default.
11854 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11855 As these are not official, they are not included in "ALL";
11856 the "ECCdraft" ciphersuite group alias can be used to select them.
11858 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11860 * Add ECDH engine support.
11862 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11864 * Add ECDH in new directory crypto/ecdh/.
11866 *Douglas Stebila (Sun Microsystems Laboratories)*
11868 * Let BN_rand_range() abort with an error after 100 iterations
11869 without success (which indicates a broken PRNG).
11873 * Change BN_mod_sqrt() so that it verifies that the input value
11874 is really the square of the return value. (Previously,
11875 BN_mod_sqrt would show GIGO behaviour.)
11879 * Add named elliptic curves over binary fields from X9.62, SECG,
11880 and WAP/WTLS; add OIDs that were still missing.
11882 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11884 * Extend the EC library for elliptic curves over binary fields
11885 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11888 EC_GF2m_simple_method
11892 EC_GROUP_new_curve_GF2m
11893 EC_GROUP_set_curve_GF2m
11894 EC_GROUP_get_curve_GF2m
11895 EC_POINT_set_affine_coordinates_GF2m
11896 EC_POINT_get_affine_coordinates_GF2m
11897 EC_POINT_set_compressed_coordinates_GF2m
11899 Point compression for binary fields is disabled by default for
11900 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11903 As binary polynomials are represented as BIGNUMs, various members
11904 of the EC_GROUP and EC_POINT data structures can be shared
11905 between the implementations for prime fields and binary fields;
11906 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11907 are essentially identical to their `..._GFp` counterparts.
11908 (For simplicity, the `..._GFp` prefix has been dropped from
11909 various internal method names.)
11911 An internal 'field_div' method (similar to 'field_mul' and
11912 'field_sqr') has been added; this is used only for binary fields.
11914 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11916 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11917 through methods ('mul', 'precompute_mult').
11919 The generic implementations (now internally called 'ec_wNAF_mul'
11920 and 'ec_wNAF_precomputed_mult') remain the default if these
11921 methods are undefined.
11923 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11925 * New function EC_GROUP_get_degree, which is defined through
11926 EC_METHOD. For curves over prime fields, this returns the bit
11927 length of the modulus.
11929 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11931 * New functions EC_GROUP_dup, EC_POINT_dup.
11932 (These simply call ..._new and ..._copy).
11934 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11936 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11937 Polynomials are represented as BIGNUMs (where the sign bit is not
11938 used) in the following functions [macros]:
11941 BN_GF2m_sub [= BN_GF2m_add]
11942 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11943 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11944 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11946 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11947 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11948 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11949 BN_GF2m_cmp [= BN_ucmp]
11951 (Note that only the 'mod' functions are actually for fields GF(2^m).
11952 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11954 For some functions, an the irreducible polynomial defining a
11955 field can be given as an 'unsigned int[]' with strictly
11956 decreasing elements giving the indices of those bits that are set;
11957 i.e., p[] represents the polynomial
11958 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11960 p[0] > p[1] > ... > p[k] = 0.
11961 This applies to the following functions:
11964 BN_GF2m_mod_mul_arr
11965 BN_GF2m_mod_sqr_arr
11966 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11967 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11968 BN_GF2m_mod_exp_arr
11969 BN_GF2m_mod_sqrt_arr
11970 BN_GF2m_mod_solve_quad_arr
11974 Conversion can be performed by the following functions:
11979 bntest.c has additional tests for binary polynomial arithmetic.
11981 Two implementations for BN_GF2m_mod_div() are available.
11982 The default algorithm simply uses BN_GF2m_mod_inv() and
11983 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11984 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11985 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11987 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11989 * Add new error code 'ERR_R_DISABLED' that can be used when some
11990 functionality is disabled at compile-time.
11992 *Douglas Stebila <douglas.stebila@sun.com>*
11994 * Change default behaviour of 'openssl asn1parse' so that more
11995 information is visible when viewing, e.g., a certificate:
11997 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11998 mode the content of non-printable OCTET STRINGs is output in a
11999 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12000 avoid the appearance of a printable string.
12002 *Nils Larsch <nla@trustcenter.de>*
12004 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12006 EC_GROUP_set_asn1_flag()
12007 EC_GROUP_get_asn1_flag()
12008 EC_GROUP_set_point_conversion_form()
12009 EC_GROUP_get_point_conversion_form()
12010 These control ASN1 encoding details:
12011 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12012 has been set to OPENSSL_EC_NAMED_CURVE.
12013 - Points are encoded in uncompressed form by default; options for
12014 asn1_for are as for point2oct, namely
12015 POINT_CONVERSION_COMPRESSED
12016 POINT_CONVERSION_UNCOMPRESSED
12017 POINT_CONVERSION_HYBRID
12019 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12021 EC_GROUP_set_seed()
12022 EC_GROUP_get0_seed()
12023 EC_GROUP_get_seed_len()
12024 This is used only for ASN1 purposes (so far).
12026 *Nils Larsch <nla@trustcenter.de>*
12028 * Add 'field_type' member to EC_METHOD, which holds the NID
12029 of the appropriate field type OID. The new function
12030 EC_METHOD_get_field_type() returns this value.
12032 *Nils Larsch <nla@trustcenter.de>*
12035 EC_POINT_point2bn()
12036 EC_POINT_bn2point()
12037 EC_POINT_point2hex()
12038 EC_POINT_hex2point()
12039 providing useful interfaces to EC_POINT_point2oct() and
12040 EC_POINT_oct2point().
12042 *Nils Larsch <nla@trustcenter.de>*
12044 * Change internals of the EC library so that the functions
12045 EC_GROUP_set_generator()
12046 EC_GROUP_get_generator()
12047 EC_GROUP_get_order()
12048 EC_GROUP_get_cofactor()
12049 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12050 to methods, which would lead to unnecessary code duplication when
12051 adding different types of curves.
12053 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12055 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12056 arithmetic, and such that modified wNAFs are generated
12057 (which avoid length expansion in many cases).
12061 * Add a function EC_GROUP_check_discriminant() (defined via
12062 EC_METHOD) that verifies that the curve discriminant is non-zero.
12064 Add a function EC_GROUP_check() that makes some sanity tests
12065 on a EC_GROUP, its generator and order. This includes
12066 EC_GROUP_check_discriminant().
12068 *Nils Larsch <nla@trustcenter.de>*
12070 * Add ECDSA in new directory crypto/ecdsa/.
12072 Add applications 'openssl ecparam' and 'openssl ecdsa'
12073 (these are based on 'openssl dsaparam' and 'openssl dsa').
12075 ECDSA support is also included in various other files across the
12076 library. Most notably,
12077 - 'openssl req' now has a '-newkey ecdsa:file' option;
12078 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12079 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12080 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12081 them suitable for ECDSA where domain parameters must be
12082 extracted before the specific public key;
12083 - ECDSA engine support has been added.
12085 *Nils Larsch <nla@trustcenter.de>*
12087 * Include some named elliptic curves, and add OIDs from X9.62,
12088 SECG, and WAP/WTLS. Each curve can be obtained from the new
12090 EC_GROUP_new_by_curve_name(),
12091 and the list of available named curves can be obtained with
12092 EC_get_builtin_curves().
12093 Also add a 'curve_name' member to EC_GROUP objects, which can be
12095 EC_GROUP_set_curve_name()
12096 EC_GROUP_get_curve_name()
12098 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12100 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12101 was actually never needed) and in BN_mul(). The removal in BN_mul()
12102 required a small change in bn_mul_part_recursive() and the addition
12103 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12104 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12105 bn_sub_words() and bn_add_words() except they take arrays with
12110 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12112 * Cleanse PEM buffers before freeing them since they may contain
12115 *Benjamin Bennett <ben@psc.edu>*
12117 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12118 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12119 authentication-only ciphersuites.
12123 * Since AES128 and AES256 share a single mask bit in the logic of
12124 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12125 kludge to work properly if AES128 is available and AES256 isn't.
12129 * Expand security boundary to match 1.1.1 module.
12133 * Remove redundant features: hash file source, editing of test vectors
12134 modify fipsld to use external fips_premain.c signature.
12138 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12139 run algorithm test programs.
12143 * Make algorithm test programs more tolerant of whitespace.
12147 * Have SSL/TLS server implementation tolerate "mismatched" record
12148 protocol version while receiving ClientHello even if the
12149 ClientHello is fragmented. (The server can't insist on the
12150 particular protocol version it has chosen before the ServerHello
12151 message has informed the client about his choice.)
12155 * Load error codes if they are not already present instead of using a
12156 static variable. This allows them to be cleanly unloaded and reloaded.
12160 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12162 * Introduce limits to prevent malicious keys being able to
12163 cause a denial of service. ([CVE-2006-2940])
12165 *Steve Henson, Bodo Moeller*
12167 * Fix ASN.1 parsing of certain invalid structures that can result
12168 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12170 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12171 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12173 * Fix SSL client code which could crash if connecting to a
12174 malicious SSLv2 server. ([CVE-2006-4343])
12176 *Tavis Ormandy and Will Drewry, Google Security Team*
12178 * Change ciphersuite string processing so that an explicit
12179 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12180 will no longer include "AES128-SHA"), and any other similar
12181 ciphersuite (same bitmap) from *other* protocol versions (so that
12182 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12183 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12184 changes from 0.9.8b and 0.9.8d.
12188 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12190 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12191 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12193 * Change the Unix randomness entropy gathering to use poll() when
12194 possible instead of select(), since the latter has some
12195 undesirable limitations.
12197 *Darryl Miles via Richard Levitte and Bodo Moeller*
12199 * Disable rogue ciphersuites:
12201 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12202 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12203 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12205 The latter two were purportedly from
12206 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12209 Also deactivate the remaining ciphersuites from
12210 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12211 unofficial, and the ID has long expired.
12215 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12216 dual-core machines) and other potential thread-safety issues.
12220 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12222 * Adapt fipsld and the build system to link against the validated FIPS
12223 module in FIPS mode.
12227 * Fixes for VC++ 2005 build under Windows.
12231 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12232 from a Windows bash shell such as MSYS. It is autodetected from the
12233 "config" script when run from a VC++ environment. Modify standard VC++
12234 build to use fipscanister.o from the GNU make build.
12238 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12240 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12241 The value now differs depending on if you build for FIPS or not.
12242 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12243 safely run with a non-FIPSed libcrypto, as it may crash because of
12244 the difference induced by this change.
12248 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12250 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12251 (part of SSL_OP_ALL). This option used to disable the
12252 countermeasure against man-in-the-middle protocol-version
12253 rollback in the SSL 2.0 server implementation, which is a bad
12254 idea. ([CVE-2005-2969])
12256 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12257 for Information Security, National Institute of Advanced Industrial
12258 Science and Technology [AIST, Japan)]*
12260 * Minimal support for X9.31 signatures and PSS padding modes. This is
12261 mainly for FIPS compliance and not fully integrated at this stage.
12265 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12266 the exponentiation using a fixed-length exponent. (Otherwise,
12267 the information leaked through timing could expose the secret key
12268 after many signatures; cf. Bleichenbacher's attack on DSA with
12273 * Make a new fixed-window mod_exp implementation the default for
12274 RSA, DSA, and DH private-key operations so that the sequence of
12275 squares and multiplies and the memory access pattern are
12276 independent of the particular secret key. This will mitigate
12277 cache-timing and potential related attacks.
12279 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12280 and this is automatically used by BN_mod_exp_mont() if the new flag
12281 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12282 will use this BN flag for private exponents unless the flag
12283 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12284 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12286 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12288 * Change the client implementation for SSLv23_method() and
12289 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12290 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12291 (Previously, the SSL 2.0 backwards compatible Client Hello
12292 message format would be used even with SSL_OP_NO_SSLv2.)
12296 * Add support for smime-type MIME parameter in S/MIME messages which some
12301 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12302 a threadsafe manner. Modify rsa code to use new function and add calls
12303 to dsa and dh code (which had race conditions before).
12307 * Include the fixed error library code in the C error file definitions
12308 instead of fixing them up at runtime. This keeps the error code
12309 structures constant.
12313 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12315 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12318 * Fixes for newer kerberos headers. NB: the casts are needed because
12319 the 'length' field is signed on one version and unsigned on another
12320 with no (?) obvious way to tell the difference, without these VC++
12321 complains. Also the "definition" of FAR (blank) is no longer included
12322 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12323 some needed definitions.
12327 * Undo Cygwin change.
12331 * Added support for proxy certificates according to RFC 3820.
12332 Because they may be a security thread to unaware applications,
12333 they must be explicitly allowed in run-time. See
12334 docs/HOWTO/proxy_certificates.txt for further information.
12338 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12340 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12341 server and client random values. Previously
12342 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12343 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12345 This change has negligible security impact because:
12347 1. Server and client random values still have 24 bytes of pseudo random
12350 2. Server and client random values are sent in the clear in the initial
12353 3. The master secret is derived using the premaster secret (48 bytes in
12354 size for static RSA ciphersuites) as well as client server and random
12357 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12360 *Stephen Henson, reported by UK NISCC*
12362 * Use Windows randomness collection on Cygwin.
12366 * Fix hang in EGD/PRNGD query when communication socket is closed
12367 prematurely by EGD/PRNGD.
12369 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12371 * Prompt for pass phrases when appropriate for PKCS12 input format.
12375 * Back-port of selected performance improvements from development
12376 branch, as well as improved support for PowerPC platforms.
12380 * Add lots of checks for memory allocation failure, error codes to indicate
12381 failure and freeing up memory if a failure occurs.
12383 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12385 * Add new -passin argument to dgst.
12389 * Perform some character comparisons of different types in X509_NAME_cmp:
12390 this is needed for some certificates that re-encode DNs into UTF8Strings
12391 (in violation of RFC3280) and can't or won't issue name rollover
12396 * Make an explicit check during certificate validation to see that
12397 the CA setting in each certificate on the chain is correct. As a
12398 side effect always do the following basic checks on extensions,
12399 not just when there's an associated purpose to the check:
12401 - if there is an unhandled critical extension (unless the user
12402 has chosen to ignore this fault)
12403 - if the path length has been exceeded (if one is set at all)
12404 - that certain extensions fit the associated purpose (if one has
12409 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12411 * Avoid a race condition when CRLs are checked in a multi threaded
12412 environment. This would happen due to the reordering of the revoked
12413 entries during signature checking and serial number lookup. Now the
12414 encoding is cached and the serial number sort performed under a lock.
12415 Add new STACK function sk_is_sorted().
12419 * Add Delta CRL to the extension code.
12423 * Various fixes to s3_pkt.c so alerts are sent properly.
12425 *David Holmes <d.holmes@f5.com>*
12427 * Reduce the chances of duplicate issuer name and serial numbers (in
12428 violation of RFC3280) using the OpenSSL certificate creation utilities.
12429 This is done by creating a random 64 bit value for the initial serial
12430 number when a serial number file is created or when a self signed
12431 certificate is created using 'openssl req -x509'. The initial serial
12432 number file is created using 'openssl x509 -next_serial' in CA.pl
12433 rather than being initialized to 1.
12437 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12439 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12440 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12442 *Joe Orton, Steve Henson*
12444 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12447 *Joe Orton, Steve Henson*
12449 * Make it possible to have multiple active certificates with the same
12450 subject in the CA index file. This is done only if the keyword
12451 'unique_subject' is set to 'no' in the main CA section (default
12452 if 'CA_default') of the configuration file. The value is saved
12453 with the database itself in a separate index attribute file,
12454 named like the index file with '.attr' appended to the name.
12458 * X509 verify fixes. Disable broken certificate workarounds when
12459 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12460 keyUsage extension present. Don't accept CRLs with unhandled critical
12461 extensions: since verify currently doesn't process CRL extensions this
12462 rejects a CRL with *any* critical extensions. Add new verify error codes
12467 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12468 A clarification of RFC2560 will require the use of OCTET STRINGs and
12469 some implementations cannot handle the current raw format. Since OpenSSL
12470 copies and compares OCSP nonces as opaque blobs without any attempt at
12471 parsing them this should not create any compatibility issues.
12475 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12476 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12477 this HMAC (and other) operations are several times slower than OpenSSL
12482 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12484 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12486 * Use the correct content when signing type "other".
12490 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12492 * Fix various bugs revealed by running the NISCC test suite:
12494 Stop out of bounds reads in the ASN1 code when presented with
12495 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12497 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12499 If verify callback ignores invalid public key errors don't try to check
12500 certificate signature with the NULL public key.
12504 * New -ignore_err option in ocsp application to stop the server
12505 exiting on the first error in a request.
12509 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12510 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12515 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12516 extra data after the compression methods not only for TLS 1.0
12517 but also for SSL 3.0 (as required by the specification).
12519 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12521 * Change X509_certificate_type() to mark the key as exported/exportable
12522 when it's 512 *bits* long, not 512 bytes.
12526 * Change AES_cbc_encrypt() so it outputs exact multiple of
12527 blocks during encryption.
12531 * Various fixes to base64 BIO and non blocking I/O. On write
12532 flushes were not handled properly if the BIO retried. On read
12533 data was not being buffered properly and had various logic bugs.
12534 This also affects blocking I/O when the data being decoded is a
12539 * Various S/MIME bugfixes and compatibility changes:
12540 output correct application/pkcs7 MIME type if
12541 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12542 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12543 of files as .eml work). Correctly handle very long lines in MIME
12548 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12550 * Countermeasure against the Klima-Pokorny-Rosa extension of
12551 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12552 a protocol version number mismatch like a decryption error
12553 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12557 * Turn on RSA blinding by default in the default implementation
12558 to avoid a timing attack. Applications that don't want it can call
12559 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12560 They would be ill-advised to do so in most cases.
12562 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12564 * Change RSA blinding code so that it works when the PRNG is not
12565 seeded (in this case, the secret RSA exponent is abused as
12566 an unpredictable seed -- if it is not unpredictable, there
12567 is no point in blinding anyway). Make RSA blinding thread-safe
12568 by remembering the creator's thread ID in rsa->blinding and
12569 having all other threads use local one-time blinding factors
12570 (this requires more computation than sharing rsa->blinding, but
12571 avoids excessive locking; and if an RSA object is not shared
12572 between threads, blinding will still be very fast).
12576 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12577 ENGINE as defaults for all supported algorithms irrespective of
12578 the 'flags' parameter. 'flags' is now honoured, so applications
12579 should make sure they are passing it correctly.
12583 * Target "mingw" now allows native Windows code to be generated in
12584 the Cygwin environment as well as with the MinGW compiler.
12588 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12590 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12591 via timing by performing a MAC computation even if incorrect
12592 block cipher padding has been found. This is a countermeasure
12593 against active attacks where the attacker has to distinguish
12594 between bad padding and a MAC verification error. ([CVE-2003-0078])
12596 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12597 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12598 Martin Vuagnoux (EPFL, Ilion)*
12600 * Make the no-err option work as intended. The intention with no-err
12601 is not to have the whole error stack handling routines removed from
12602 libcrypto, it's only intended to remove all the function name and
12603 reason texts, thereby removing some of the footprint that may not
12604 be interesting if those errors aren't displayed anyway.
12606 NOTE: it's still possible for any application or module to have its
12607 own set of error texts inserted. The routines are there, just not
12608 used by default when no-err is given.
12612 * Add support for FreeBSD on IA64.
12614 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12616 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12617 Kerberos function mit_des_cbc_cksum(). Before this change,
12618 the value returned by DES_cbc_cksum() was like the one from
12619 mit_des_cbc_cksum(), except the bytes were swapped.
12621 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12623 * Allow an application to disable the automatic SSL chain building.
12624 Before this a rather primitive chain build was always performed in
12625 ssl3_output_cert_chain(): an application had no way to send the
12626 correct chain if the automatic operation produced an incorrect result.
12628 Now the chain builder is disabled if either:
12630 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12632 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12634 The reasoning behind this is that an application would not want the
12635 auto chain building to take place if extra chain certificates are
12636 present and it might also want a means of sending no additional
12637 certificates (for example the chain has two certificates and the
12642 * Add the possibility to build without the ENGINE framework.
12644 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12646 * Under Win32 gmtime() can return NULL: check return value in
12647 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12651 * DSA routines: under certain error conditions uninitialized BN objects
12652 could be freed. Solution: make sure initialization is performed early
12653 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12654 Nils Larsch <nla@trustcenter.de> via PR#459)
12658 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12659 checked on reconnect on the client side, therefore session resumption
12660 could still fail with a "ssl session id is different" error. This
12661 behaviour is masked when SSL_OP_ALL is used due to
12662 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12663 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12664 followup to PR #377.
12668 * IA-32 assembler support enhancements: unified ELF targets, support
12669 for SCO/Caldera platforms, fix for Cygwin shared build.
12673 * Add support for FreeBSD on sparc64. As a consequence, support for
12674 FreeBSD on non-x86 processors is separate from x86 processors on
12675 the config script, much like the NetBSD support.
12677 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12679 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12681 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12684 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12685 code (06) was taken as the first octet of the session ID and the last
12686 octet was ignored consequently. As a result SSLv2 client side session
12687 caching could not have worked due to the session ID mismatch between
12689 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12694 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12695 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12700 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12701 seems that in spite of existing for more than a year, many application
12702 author have done nothing to provide the necessary callbacks, which
12703 means that this particular engine will not work properly anywhere.
12704 This is a very unfortunate situation which forces us, in the name
12705 of usability, to give the hw_ncipher.c a static lock, which is part
12707 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12708 appear in 0.9.8 or later. We EXPECT application authors to have
12709 dealt properly with this when 0.9.8 is released (unless we actually
12710 make such changes in the libcrypto locking code that changes will
12711 have to be made anyway).
12715 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12716 octets have been read, EOF or an error occurs. Without this change
12717 some truncated ASN1 structures will not produce an error.
12721 * Disable Heimdal support, since it hasn't been fully implemented.
12722 Still give the possibility to force the use of Heimdal, but with
12723 warnings and a request that patches get sent to openssl-dev.
12727 * Add the VC-CE target, introduce the WINCE sysname, and add
12728 INSTALL.WCE and appropriate conditionals to make it build.
12730 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12732 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12733 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12734 edit numbers of the version.
12736 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12738 * Introduce safe string copy and catenation functions
12739 (BUF_strlcpy() and BUF_strlcat()).
12741 *Ben Laurie (CHATS) and Richard Levitte*
12743 * Avoid using fixed-size buffers for one-line DNs.
12745 *Ben Laurie (CHATS)*
12747 * Add BUF_MEM_grow_clean() to avoid information leakage when
12748 resizing buffers containing secrets, and use where appropriate.
12750 *Ben Laurie (CHATS)*
12752 * Avoid using fixed size buffers for configuration file location.
12754 *Ben Laurie (CHATS)*
12756 * Avoid filename truncation for various CA files.
12758 *Ben Laurie (CHATS)*
12760 * Use sizeof in preference to magic numbers.
12762 *Ben Laurie (CHATS)*
12764 * Avoid filename truncation in cert requests.
12766 *Ben Laurie (CHATS)*
12768 * Add assertions to check for (supposedly impossible) buffer
12771 *Ben Laurie (CHATS)*
12773 * Don't cache truncated DNS entries in the local cache (this could
12774 potentially lead to a spoofing attack).
12776 *Ben Laurie (CHATS)*
12778 * Fix various buffers to be large enough for hex/decimal
12779 representations in a platform independent manner.
12781 *Ben Laurie (CHATS)*
12783 * Add CRYPTO_realloc_clean() to avoid information leakage when
12784 resizing buffers containing secrets, and use where appropriate.
12786 *Ben Laurie (CHATS)*
12788 * Add BIO_indent() to avoid much slightly worrying code to do
12791 *Ben Laurie (CHATS)*
12793 * Convert sprintf()/BIO_puts() to BIO_printf().
12795 *Ben Laurie (CHATS)*
12797 * buffer_gets() could terminate with the buffer only half
12800 *Ben Laurie (CHATS)*
12802 * Add assertions to prevent user-supplied crypto functions from
12803 overflowing internal buffers by having large block sizes, etc.
12805 *Ben Laurie (CHATS)*
12807 * New OPENSSL_assert() macro (similar to assert(), but enabled
12810 *Ben Laurie (CHATS)*
12812 * Eliminate unused copy of key in RC4.
12814 *Ben Laurie (CHATS)*
12816 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12818 *Ben Laurie (CHATS)*
12820 * Fix off-by-one error in EGD path.
12822 *Ben Laurie (CHATS)*
12824 * If RANDFILE path is too long, ignore instead of truncating.
12826 *Ben Laurie (CHATS)*
12828 * Eliminate unused and incorrectly sized X.509 structure
12831 *Ben Laurie (CHATS)*
12833 * Eliminate unused and dangerous function knumber().
12835 *Ben Laurie (CHATS)*
12837 * Eliminate unused and dangerous structure, KSSL_ERR.
12839 *Ben Laurie (CHATS)*
12841 * Protect against overlong session ID context length in an encoded
12842 session object. Since these are local, this does not appear to be
12845 *Ben Laurie (CHATS)*
12847 * Change from security patch (see 0.9.6e below) that did not affect
12848 the 0.9.6 release series:
12850 Remote buffer overflow in SSL3 protocol - an attacker could
12851 supply an oversized master key in Kerberos-enabled versions.
12854 *Ben Laurie (CHATS)*
12856 * Change the SSL kerb5 codes to match RFC 2712.
12860 * Make -nameopt work fully for req and add -reqopt switch.
12862 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12864 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12866 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12868 * Make sure tests can be performed even if the corresponding algorithms
12869 have been removed entirely. This was also the last step to make
12870 OpenSSL compilable with DJGPP under all reasonable conditions.
12872 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12874 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12875 to allow version independent disabling of normally unselected ciphers,
12876 which may be activated as a side-effect of selecting a single cipher.
12878 (E.g., cipher list string "RSA" enables ciphersuites that are left
12879 out of "ALL" because they do not provide symmetric encryption.
12880 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12882 *Lutz Jaenicke, Bodo Moeller*
12884 * Add appropriate support for separate platform-dependent build
12885 directories. The recommended way to make a platform-dependent
12886 build directory is the following (tested on Linux), maybe with
12889 # Place yourself outside of the OpenSSL source tree. In
12890 # this example, the environment variable OPENSSL_SOURCE
12891 # is assumed to contain the absolute OpenSSL source directory.
12892 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12893 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12894 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12895 mkdir -p `dirname $F`
12896 ln -s $OPENSSL_SOURCE/$F $F
12899 To be absolutely sure not to disturb the source tree, a "make clean"
12900 is a good thing. If it isn't successful, don't worry about it,
12901 it probably means the source directory is very clean.
12905 * Make sure any ENGINE control commands make local copies of string
12906 pointers passed to them whenever necessary. Otherwise it is possible
12907 the caller may have overwritten (or deallocated) the original string
12908 data when a later ENGINE operation tries to use the stored values.
12910 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12912 * Improve diagnostics in file reading and command-line digests.
12914 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12916 * Add AES modes CFB and OFB to the object database. Correct an
12917 error in AES-CFB decryption.
12921 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12922 allows existing EVP_CIPHER_CTX structures to be reused after
12923 calling `EVP_*Final()`. This behaviour is used by encryption
12924 BIOs and some applications. This has the side effect that
12925 applications must explicitly clean up cipher contexts with
12926 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12930 * Check the values of dna and dnb in bn_mul_recursive before calling
12931 bn_mul_comba (a non zero value means the a or b arrays do not contain
12932 n2 elements) and fallback to bn_mul_normal if either is not zero.
12936 * Fix escaping of non-ASCII characters when using the -subj option
12937 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12941 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12942 form for "surname", serialNumber has no short form.
12943 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12944 therefore remove "mail" short name for "internet 7".
12945 The OID for unique identifiers in X509 certificates is
12946 x500UniqueIdentifier, not uniqueIdentifier.
12947 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12951 * Add an "init" command to the ENGINE config module and auto initialize
12952 ENGINEs. Without any "init" command the ENGINE will be initialized
12953 after all ctrl commands have been executed on it. If init=1 the
12954 ENGINE is initialized at that point (ctrls before that point are run
12955 on the uninitialized ENGINE and after on the initialized one). If
12956 init=0 then the ENGINE will not be initialized at all.
12960 * Fix the 'app_verify_callback' interface so that the user-defined
12961 argument is actually passed to the callback: In the
12962 SSL_CTX_set_cert_verify_callback() prototype, the callback
12963 declaration has been changed from
12966 int (*cb)(X509_STORE_CTX *,void *);
12967 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12968 i=s->ctx->app_verify_callback(&ctx)
12969 has been changed into
12970 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12972 To update applications using SSL_CTX_set_cert_verify_callback(),
12973 a dummy argument can be added to their callback functions.
12975 *D. K. Smetters <smetters@parc.xerox.com>*
12977 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12979 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12981 * Add and OPENSSL_LOAD_CONF define which will cause
12982 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12983 This allows older applications to transparently support certain
12984 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12985 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12986 load the config file and OPENSSL_add_all_algorithms_conf() which will
12987 always load it have also been added.
12991 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12992 Adjust NIDs and EVP layer.
12994 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12996 * Config modules support in openssl utility.
12998 Most commands now load modules from the config file,
12999 though in a few (such as version) this isn't done
13000 because it couldn't be used for anything.
13002 In the case of ca and req the config file used is
13003 the same as the utility itself: that is the -config
13004 command line option can be used to specify an
13009 * Move default behaviour from OPENSSL_config(). If appname is NULL
13010 use "openssl_conf" if filename is NULL use default openssl config file.
13014 * Add an argument to OPENSSL_config() to allow the use of an alternative
13015 config section name. Add a new flag to tolerate a missing config file
13016 and move code to CONF_modules_load_file().
13020 * Support for crypto accelerator cards from Accelerated Encryption
13021 Processing, www.aep.ie. (Use engine 'aep')
13022 The support was copied from 0.9.6c [engine] and adapted/corrected
13023 to work with the new engine framework.
13025 *AEP Inc. and Richard Levitte*
13027 * Support for SureWare crypto accelerator cards from Baltimore
13028 Technologies. (Use engine 'sureware')
13029 The support was copied from 0.9.6c [engine] and adapted
13030 to work with the new engine framework.
13034 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13035 make the newer ENGINE framework commands for the CHIL engine work.
13037 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13039 * Make it possible to produce shared libraries on ReliantUNIX.
13041 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13043 * Add the configuration target debug-linux-ppro.
13044 Make 'openssl rsa' use the general key loading routines
13045 implemented in `apps.c`, and make those routines able to
13046 handle the key format FORMAT_NETSCAPE and the variant
13049 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13051 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13053 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13055 * Add -keyform to rsautl, and document -engine.
13057 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13059 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13060 BIO_R_NO_SUCH_FILE error code rather than the generic
13061 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13065 * Add new functions
13066 ERR_peek_last_error
13067 ERR_peek_last_error_line
13068 ERR_peek_last_error_line_data.
13069 These are similar to
13071 ERR_peek_error_line
13072 ERR_peek_error_line_data,
13073 but report on the latest error recorded rather than the first one
13074 still in the error queue.
13076 *Ben Laurie, Bodo Moeller*
13078 * default_algorithms option in ENGINE config module. This allows things
13080 default_algorithms = ALL
13081 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13085 * Preliminary ENGINE config module.
13089 * New experimental application configuration code.
13093 * Change the AES code to follow the same name structure as all other
13094 symmetric ciphers, and behave the same way. Move everything to
13095 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13097 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13099 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13101 *Ben Laurie and Theo de Raadt*
13103 * Add option to output public keys in req command.
13105 *Massimiliano Pala madwolf@openca.org*
13107 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13108 (up to about 10% better than before for P-192 and P-224).
13112 * New functions/macros
13114 SSL_CTX_set_msg_callback(ctx, cb)
13115 SSL_CTX_set_msg_callback_arg(ctx, arg)
13116 SSL_set_msg_callback(ssl, cb)
13117 SSL_set_msg_callback_arg(ssl, arg)
13119 to request calling a callback function
13121 void cb(int write_p, int version, int content_type,
13122 const void *buf, size_t len, SSL *ssl, void *arg)
13124 whenever a protocol message has been completely received
13125 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13126 protocol version according to which the SSL library interprets
13127 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13128 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13129 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13130 specification (change_cipher_spec(20), alert(21), handshake(22)).
13131 'buf' and 'len' point to the actual message, 'ssl' to the
13132 SSL object, and 'arg' is the application-defined value set by
13133 SSL[_CTX]_set_msg_callback_arg().
13135 'openssl s_client' and 'openssl s_server' have new '-msg' options
13136 to enable a callback that displays all protocol messages.
13140 * Change the shared library support so shared libraries are built as
13141 soon as the corresponding static library is finished, and thereby get
13142 openssl and the test programs linked against the shared library.
13143 This still only happens when the keyword "shard" has been given to
13144 the configuration scripts.
13146 NOTE: shared library support is still an experimental thing, and
13147 backward binary compatibility is still not guaranteed.
13149 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13151 * Add support for Subject Information Access extension.
13153 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13155 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13156 additional bytes when new memory had to be allocated, not just
13157 when reusing an existing buffer.
13161 * New command line and configuration option 'utf8' for the req command.
13162 This allows field values to be specified as UTF8 strings.
13166 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13167 runs for the former and machine-readable output for the latter.
13171 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13172 of the e-mail address in the DN (i.e., it will go into a certificate
13173 extension only). The new configuration file option 'email_in_dn = no'
13174 has the same effect.
13176 *Massimiliano Pala madwolf@openca.org*
13178 * Change all functions with names starting with `des_` to be starting
13179 with `DES_` instead. Add wrappers that are compatible with libdes,
13180 but are named `_ossl_old_des_*`. Finally, add macros that map the
13181 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13182 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13183 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13186 Since we provide two compatibility mappings, the user needs to
13187 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13188 compatibility is desired. The default (i.e., when that macro
13189 isn't defined) is OpenSSL 0.9.6c compatibility.
13191 There are also macros that enable and disable the support of old
13192 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13193 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13194 are defined, the default will apply: to support the old des routines.
13196 In either case, one must include openssl/des.h to get the correct
13197 definitions. Do not try to just include openssl/des_old.h, that
13200 NOTE: This is a major break of an old API into a new one. Software
13201 authors are encouraged to switch to the `DES_` style functions. Some
13202 time in the future, des_old.h and the libdes compatibility functions
13203 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13204 default), and then completely removed.
13208 * Test for certificates which contain unsupported critical extensions.
13209 If such a certificate is found during a verify operation it is
13210 rejected by default: this behaviour can be overridden by either
13211 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13212 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13213 X509_supported_extension() has also been added which returns 1 if a
13214 particular extension is supported.
13218 * Modify the behaviour of EVP cipher functions in similar way to digests
13219 to retain compatibility with existing code.
13223 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13224 compatibility with existing code. In particular the 'ctx' parameter does
13225 not have to be to be initialized before the call to EVP_DigestInit() and
13226 it is tidied up after a call to EVP_DigestFinal(). New function
13227 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13228 EVP_MD_CTX_copy() changed to not require the destination to be
13229 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13230 requires the destination to be valid.
13232 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13233 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13237 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13238 so that complete 'Handshake' protocol structures are kept in memory
13239 instead of overwriting 'msg_type' and 'length' with 'body' data.
13243 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13245 *Massimo Santin via Richard Levitte*
13247 * Major restructuring to the underlying ENGINE code. This includes
13248 reduction of linker bloat, separation of pure "ENGINE" manipulation
13249 (initialisation, etc) from functionality dealing with implementations
13250 of specific crypto interfaces. This change also introduces integrated
13251 support for symmetric ciphers and digest implementations - so ENGINEs
13252 can now accelerate these by providing EVP_CIPHER and EVP_MD
13253 implementations of their own. This is detailed in
13254 [crypto/engine/README.md](crypto/engine/README.md)
13255 as it couldn't be adequately described here. However, there are a few
13256 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13257 were changed in the original introduction of ENGINE code have now
13258 reverted back - the hooking from this code to ENGINE is now a good
13259 deal more passive and at run-time, operations deal directly with
13260 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13261 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13262 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13263 they were not being used by the framework as there is no concept of a
13264 BIGNUM_METHOD and they could not be generalised to the new
13265 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13266 ENGINE_cpy() has been removed as it cannot be consistently defined in
13271 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13275 * Change mkdef.pl to sort symbols that get the same entry number,
13276 and make sure the automatically generated functions `ERR_load_*`
13277 become part of libeay.num as well.
13281 * New function SSL_renegotiate_pending(). This returns true once
13282 renegotiation has been requested (either SSL_renegotiate() call
13283 or HelloRequest/ClientHello received from the peer) and becomes
13284 false once a handshake has been completed.
13285 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13286 sends a HelloRequest, but does not ensure that a handshake takes
13287 place. SSL_renegotiate_pending() is useful for checking if the
13288 client has followed the request.)
13292 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13293 By default, clients may request session resumption even during
13294 renegotiation (if session ID contexts permit); with this option,
13295 session resumption is possible only in the first handshake.
13297 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13298 more bits available for options that should not be part of
13299 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13303 * Add some demos for certificate and certificate request creation.
13307 * Make maximum certificate chain size accepted from the peer application
13308 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13309 "Douglas E. Engert" <deengert@anl.gov>.
13313 * Add support for shared libraries for Unixware-7
13314 (Boyd Lynn Gerber <gerberb@zenez.com>).
13318 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13319 be done prior to destruction. Use this to unload error strings from
13320 ENGINEs that load their own error strings. NB: This adds two new API
13321 functions to "get" and "set" this destroy handler in an ENGINE.
13325 * Alter all existing ENGINE implementations (except "openssl" and
13326 "openbsd") to dynamically instantiate their own error strings. This
13327 makes them more flexible to be built both as statically-linked ENGINEs
13328 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13329 Also, add stub code to each that makes building them as self-contained
13330 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13334 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13335 implementations into applications that are completely implemented in
13336 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13337 commands that can be used to configure what shared-library to load and
13338 to control aspects of the way it is handled. Also, made an update to
13339 the [README-Engine.md](README-Engine.md) file
13340 that brings its information up-to-date and
13341 provides some information and instructions on the "dynamic" ENGINE
13342 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13346 * Make it possible to unload ranges of ERR strings with a new
13347 "ERR_unload_strings" function.
13351 * Add a copy() function to EVP_MD.
13355 * Make EVP_MD routines take a context pointer instead of just the
13356 md_data void pointer.
13360 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13361 that the digest can only process a single chunk of data
13362 (typically because it is provided by a piece of
13363 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13364 is only going to provide a single chunk of data, and hence the
13365 framework needn't accumulate the data for oneshot drivers.
13369 * As with "ERR", make it possible to replace the underlying "ex_data"
13370 functions. This change also alters the storage and management of global
13371 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13372 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13373 index counters. The API functions that use this state have been changed
13374 to take a "class_index" rather than pointers to the class's local STACK
13375 and counter, and there is now an API function to dynamically create new
13376 classes. This centralisation allows us to (a) plug a lot of the
13377 thread-safety problems that existed, and (b) makes it possible to clean
13378 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13379 such data would previously have always leaked in application code and
13380 workarounds were in place to make the memory debugging turn a blind eye
13381 to it. Application code that doesn't use this new function will still
13382 leak as before, but their memory debugging output will announce it now
13383 rather than letting it slide.
13385 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13386 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13387 has a return value to indicate success or failure.
13391 * Make it possible to replace the underlying "ERR" functions such that the
13392 global state (2 LHASH tables and 2 locks) is only used by the "default"
13393 implementation. This change also adds two functions to "get" and "set"
13394 the implementation prior to it being automatically set the first time
13395 any other ERR function takes place. Ie. an application can call "get",
13396 pass the return value to a module it has just loaded, and that module
13397 can call its own "set" function using that value. This means the
13398 module's "ERR" operations will use (and modify) the error state in the
13399 application and not in its own statically linked copy of OpenSSL code.
13403 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13404 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13405 the operation, and provides a more encapsulated way for external code
13406 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13407 to use these functions rather than manually incrementing the counts.
13409 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13413 * Add EVP test program.
13417 * Add symmetric cipher support to ENGINE. Expect the API to change!
13421 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13422 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13423 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13424 These allow a CRL to be built without having to access X509_CRL fields
13425 directly. Modify 'ca' application to use new functions.
13429 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13430 bug workarounds. Rollback attack detection is a security feature.
13431 The problem will only arise on OpenSSL servers when TLSv1 is not
13432 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13433 Software authors not wanting to support TLSv1 will have special reasons
13434 for their choice and can explicitly enable this option.
13436 *Bodo Moeller, Lutz Jaenicke*
13438 * Rationalise EVP so it can be extended: don't include a union of
13439 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13440 (similar to those existing for EVP_CIPHER_CTX).
13445 EVP_MD_CTX_init(&md); /* new function call */
13446 EVP_DigestInit(&md, EVP_sha1());
13447 EVP_DigestUpdate(&md, in, len);
13448 EVP_DigestFinal(&md, out, NULL);
13449 EVP_MD_CTX_cleanup(&md); /* new function call */
13453 * Make DES key schedule conform to the usual scheme, as well as
13454 correcting its structure. This means that calls to DES functions
13455 now have to pass a pointer to a des_key_schedule instead of a
13456 plain des_key_schedule (which was actually always a pointer
13459 des_key_schedule ks;
13461 des_set_key_checked(..., &ks);
13462 des_ncbc_encrypt(..., &ks, ...);
13464 (Note that a later change renames 'des_...' into 'DES_...'.)
13468 * Initial reduction of linker bloat: the use of some functions, such as
13469 PEM causes large amounts of unused functions to be linked in due to
13470 poor organisation. For example pem_all.c contains every PEM function
13471 which has a knock on effect of linking in large amounts of (unused)
13472 ASN1 code. Grouping together similar functions and splitting unrelated
13473 functions prevents this.
13477 * Cleanup of EVP macros.
13481 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13482 correct `_ecb suffix`.
13486 * Add initial OCSP responder support to ocsp application. The
13487 revocation information is handled using the text based index
13488 use by the ca application. The responder can either handle
13489 requests generated internally, supplied in files (for example
13490 via a CGI script) or using an internal minimal server.
13494 * Add configuration choices to get zlib compression for TLS.
13498 * Changes to Kerberos SSL for RFC 2712 compliance:
13499 1. Implemented real KerberosWrapper, instead of just using
13500 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13501 2. Implemented optional authenticator field of KerberosWrapper.
13503 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13504 and authenticator structs; see crypto/krb5/.
13506 Generalized Kerberos calls to support multiple Kerberos libraries.
13507 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13508 via Richard Levitte*
13510 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13511 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13512 values for each of the key sizes rather than having just
13513 parameters (and 'speed' generating keys each time).
13517 * Speed up EVP routines.
13520 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13521 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13522 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13523 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13525 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13526 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13527 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13530 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13532 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13536 * Added the OS2-EMX target.
13538 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13540 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13541 New functions to support `NCONF` routines in extension code.
13542 New function `CONF_set_nconf()`
13543 to allow functions which take an `NCONF` to also handle the old `LHASH`
13544 structure: this means that the old `CONF` compatible routines can be
13545 retained (in particular w.rt. extensions) without having to duplicate the
13546 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13550 * Enhance the general user interface with mechanisms for inner control
13551 and with possibilities to have yes/no kind of prompts.
13555 * Change all calls to low-level digest routines in the library and
13556 applications to use EVP. Add missing calls to HMAC_cleanup() and
13557 don't assume HMAC_CTX can be copied using memcpy().
13559 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13561 * Add the possibility to control engines through control names but with
13562 arbitrary arguments instead of just a string.
13563 Change the key loaders to take a UI_METHOD instead of a callback
13564 function pointer. NOTE: this breaks binary compatibility with earlier
13565 versions of OpenSSL [engine].
13566 Adapt the nCipher code for these new conditions and add a card insertion
13571 * Enhance the general user interface with mechanisms to better support
13572 dialog box interfaces, application-defined prompts, the possibility
13573 to use defaults (for example default passwords from somewhere else)
13574 and interrupts/cancellations.
13578 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13579 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13583 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13584 tidy up some unnecessarily weird code in 'sk_new()').
13586 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13588 * Change the key loading routines for ENGINEs to use the same kind
13589 callback (pem_password_cb) as all other routines that need this
13594 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13595 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13596 than this minimum value is recommended.
13600 * New random seeder for OpenVMS, using the system process statistics
13601 that are easily reachable.
13605 * Windows apparently can't transparently handle global
13606 variables defined in DLLs. Initialisations such as:
13608 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13610 won't compile. This is used by the any applications that need to
13611 declare their own ASN1 modules. This was fixed by adding the option
13612 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13613 needed for static libraries under Win32.
13617 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13618 setting of purpose and trust fields. New X509_STORE trust and
13619 purpose functions and tidy up setting in other SSL functions.
13623 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13624 structure. These are inherited by X509_STORE_CTX when it is
13625 initialised. This allows various defaults to be set in the
13626 X509_STORE structure (such as flags for CRL checking and custom
13627 purpose or trust settings) for functions which only use X509_STORE_CTX
13628 internally such as S/MIME.
13630 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13631 trust settings if they are not set in X509_STORE. This allows X509_STORE
13632 purposes and trust (in S/MIME for example) to override any set by default.
13634 Add command line options for CRL checking to smime, s_client and s_server
13639 * Initial CRL based revocation checking. If the CRL checking flag(s)
13640 are set then the CRL is looked up in the X509_STORE structure and
13641 its validity and signature checked, then if the certificate is found
13642 in the CRL the verify fails with a revoked error.
13644 Various new CRL related callbacks added to X509_STORE_CTX structure.
13646 Command line options added to 'verify' application to support this.
13648 This needs some additional work, such as being able to handle multiple
13649 CRLs with different times, extension based lookup (rather than just
13650 by subject name) and ultimately more complete V2 CRL extension
13655 * Add a general user interface API (crypto/ui/). This is designed
13656 to replace things like des_read_password and friends (backward
13657 compatibility functions using this new API are provided).
13658 The purpose is to remove prompting functions from the DES code
13659 section as well as provide for prompting through dialog boxes in
13660 a window system and the like.
13664 * Add "ex_data" support to ENGINE so implementations can add state at a
13665 per-structure level rather than having to store it globally.
13669 * Make it possible for ENGINE structures to be copied when retrieved by
13670 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13671 This causes the "original" ENGINE structure to act like a template,
13672 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13673 operational state can be localised to each ENGINE structure, despite the
13674 fact they all share the same "methods". New ENGINE structures returned in
13675 this case have no functional references and the return value is the single
13676 structural reference. This matches the single structural reference returned
13677 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13682 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13683 needs to match any other type at all we need to manually clear the
13688 * Changes to the "openssl engine" utility to include;
13689 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13690 about an ENGINE's available control commands.
13691 - executing control commands from command line arguments using the
13692 '-pre' and '-post' switches. '-post' is only used if '-t' is
13693 specified and the ENGINE is successfully initialised. The syntax for
13694 the individual commands are colon-separated, for example;
13695 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13699 * New dynamic control command support for ENGINEs. ENGINEs can now
13700 declare their own commands (numbers), names (strings), descriptions,
13701 and input types for run-time discovery by calling applications. A
13702 subset of these commands are implicitly classed as "executable"
13703 depending on their input type, and only these can be invoked through
13704 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13705 can be based on user input, config files, etc). The distinction is
13706 that "executable" commands cannot return anything other than a boolean
13707 result and can only support numeric or string input, whereas some
13708 discoverable commands may only be for direct use through
13709 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13710 pointers, or other custom uses. The "executable" commands are to
13711 support parameterisations of ENGINE behaviour that can be
13712 unambiguously defined by ENGINEs and used consistently across any
13713 OpenSSL-based application. Commands have been added to all the
13714 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13715 control over shared-library paths without source code alterations.
13719 * Changed all ENGINE implementations to dynamically allocate their
13720 ENGINEs rather than declaring them statically. Apart from this being
13721 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13722 this also allows the implementations to compile without using the
13723 internal engine_int.h header.
13727 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13728 'const' value. Any code that should be able to modify a RAND_METHOD
13729 should already have non-const pointers to it (ie. they should only
13730 modify their own ones).
13734 * Made a variety of little tweaks to the ENGINE code.
13735 - "atalla" and "ubsec" string definitions were moved from header files
13736 to C code. "nuron" string definitions were placed in variables
13737 rather than hard-coded - allowing parameterisation of these values
13738 later on via ctrl() commands.
13739 - Removed unused "#if 0"'d code.
13740 - Fixed engine list iteration code so it uses ENGINE_free() to release
13741 structural references.
13742 - Constified the RAND_METHOD element of ENGINE structures.
13743 - Constified various get/set functions as appropriate and added
13744 missing functions (including a catch-all ENGINE_cpy that duplicates
13745 all ENGINE values onto a new ENGINE except reference counts/state).
13746 - Removed NULL parameter checks in get/set functions. Setting a method
13747 or function to NULL is a way of cancelling out a previously set
13748 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13749 and doesn't justify the extra error symbols and code.
13750 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13751 flags from engine_int.h to engine.h.
13752 - Changed prototypes for ENGINE handler functions (init(), finish(),
13753 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13757 * Implement binary inversion algorithm for BN_mod_inverse in addition
13758 to the algorithm using long division. The binary algorithm can be
13759 used only if the modulus is odd. On 32-bit systems, it is faster
13760 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13761 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13762 up to 450 bits. In 64-bit environments, the binary algorithm
13763 appears to be advantageous for much longer moduli; here we use it
13764 for moduli up to 2048 bits.
13768 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13769 could not support the combine flag in choice fields.
13773 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13774 extensions from a certificate request to the certificate.
13778 * Allow multiple 'certopt' and 'nameopt' options to be separated
13779 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13780 file: this allows the display of the certificate about to be
13781 signed to be customised, to allow certain fields to be included
13782 or excluded and extension details. The old system didn't display
13783 multicharacter strings properly, omitted fields not in the policy
13784 and couldn't display additional details such as extensions.
13788 * Function EC_POINTs_mul for multiple scalar multiplication
13789 of an arbitrary number of elliptic curve points
13790 \sum scalars[i]*points[i],
13791 optionally including the generator defined for the EC_GROUP:
13792 scalar*generator + \sum scalars[i]*points[i].
13794 EC_POINT_mul is a simple wrapper function for the typical case
13795 that the point list has just one item (besides the optional
13800 * First EC_METHODs for curves over GF(p):
13802 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13803 operations and provides various method functions that can also
13804 operate with faster implementations of modular arithmetic.
13806 EC_GFp_mont_method() reuses most functions that are part of
13807 EC_GFp_simple_method, but uses Montgomery arithmetic.
13809 *Bodo Moeller; point addition and point doubling
13810 implementation directly derived from source code provided by
13811 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13813 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13814 crypto/ec/ec_lib.c):
13816 Curves are EC_GROUP objects (with an optional group generator)
13817 based on EC_METHODs that are built into the library.
13819 Points are EC_POINT objects based on EC_GROUP objects.
13821 Most of the framework would be able to handle curves over arbitrary
13822 finite fields, but as there are no obvious types for fields other
13823 than GF(p), some functions are limited to that for now.
13827 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13828 that the file contains a complete HTTP response.
13832 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13833 change the def and num file printf format specifier from "%-40sXXX"
13834 to "%-39s XXX". The latter will always guarantee a space after the
13835 field while the former will cause them to run together if the field
13836 is 40 of more characters long.
13840 * Constify the cipher and digest 'method' functions and structures
13841 and modify related functions to take constant EVP_MD and EVP_CIPHER
13846 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13847 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13851 * Modify `EVP_Digest*()` routines so they now return values. Although the
13852 internal software routines can never fail additional hardware versions
13857 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13859 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13860 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13863 ERR_R_NESTED_ASN1_ERROR
13865 ERR_R_MISSING_ASN1_EOS
13866 were 4 .. 9, conflicting with
13867 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13869 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13870 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13872 Add new error code 'ERR_R_INTERNAL_ERROR'.
13876 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13881 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13882 sets the subject name for a new request or supersedes the
13883 subject name in a given request. Formats that can be parsed are
13884 'CN=Some Name, OU=myOU, C=IT'
13886 'CN=Some Name/OU=myOU/C=IT'.
13888 Add options '-batch' and '-verbose' to 'openssl req'.
13890 *Massimiliano Pala <madwolf@hackmasters.net>*
13892 * Introduce the possibility to access global variables through
13893 functions on platform were that's the best way to handle exporting
13894 global variables in shared libraries. To enable this functionality,
13895 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13896 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13897 is normally done by Configure or something similar).
13899 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13900 in the source file (foo.c) like this:
13902 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13903 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13905 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13906 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13908 OPENSSL_DECLARE_GLOBAL(int,foo);
13909 #define foo OPENSSL_GLOBAL_REF(foo)
13910 OPENSSL_DECLARE_GLOBAL(double,bar);
13911 #define bar OPENSSL_GLOBAL_REF(bar)
13913 The #defines are very important, and therefore so is including the
13914 header file everywhere where the defined globals are used.
13916 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13917 of ASN.1 items, but that structure is a bit different.
13919 The largest change is in util/mkdef.pl which has been enhanced with
13920 better and easier to understand logic to choose which symbols should
13921 go into the Windows .def files as well as a number of fixes and code
13922 cleanup (among others, algorithm keywords are now sorted
13923 lexicographically to avoid constant rewrites).
13927 * In BN_div() keep a copy of the sign of 'num' before writing the
13928 result to 'rm' because if rm==num the value will be overwritten
13929 and produce the wrong result if 'num' is negative: this caused
13930 problems with BN_mod() and BN_nnmod().
13934 * Function OCSP_request_verify(). This checks the signature on an
13935 OCSP request and verifies the signer certificate. The signer
13936 certificate is just checked for a generic purpose and OCSP request
13941 * Add OCSP_check_validity() function to check the validity of OCSP
13942 responses. OCSP responses are prepared in real time and may only
13943 be a few seconds old. Simply checking that the current time lies
13944 between thisUpdate and nextUpdate max reject otherwise valid responses
13945 caused by either OCSP responder or client clock inaccuracy. Instead
13946 we allow thisUpdate and nextUpdate to fall within a certain period of
13947 the current time. The age of the response can also optionally be
13948 checked. Two new options -validity_period and -status_age added to
13953 * If signature or public key algorithm is unrecognized print out its
13954 OID rather that just UNKNOWN.
13958 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13959 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13960 ID to be generated from the issuer certificate alone which can then be
13961 passed to OCSP_id_issuer_cmp().
13965 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13966 ASN1 modules to export functions returning ASN1_ITEM pointers
13967 instead of the ASN1_ITEM structures themselves. This adds several
13968 new macros which allow the underlying ASN1 function/structure to
13969 be accessed transparently. As a result code should not use ASN1_ITEM
13970 references directly (such as &X509_it) but instead use the relevant
13971 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13972 use of the new ASN1 code on platforms where exporting structures
13973 is problematical (for example in shared libraries) but exporting
13974 functions returning pointers to structures is not.
13978 * Add support for overriding the generation of SSL/TLS session IDs.
13979 These callbacks can be registered either in an SSL_CTX or per SSL.
13980 The purpose of this is to allow applications to control, if they wish,
13981 the arbitrary values chosen for use as session IDs, particularly as it
13982 can be useful for session caching in multiple-server environments. A
13983 command-line switch for testing this (and any client code that wishes
13984 to use such a feature) has been added to "s_server".
13986 *Geoff Thorpe, Lutz Jaenicke*
13988 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13989 of the form `#if defined(...) || defined(...) || ...` and
13990 `#if !defined(...) && !defined(...) && ...`. This also avoids
13991 the growing number of special cases it was previously handling.
13995 * Make all configuration macros available for application by making
13996 sure they are available in opensslconf.h, by giving them names starting
13997 with `OPENSSL_` to avoid conflicts with other packages and by making
13998 sure e_os2.h will cover all platform-specific cases together with
14000 Additionally, it is now possible to define configuration/platform-
14001 specific names (called "system identities"). In the C code, these
14002 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14003 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14004 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14009 * New option -set_serial to 'req' and 'x509' this allows the serial
14010 number to use to be specified on the command line. Previously self
14011 signed certificates were hard coded with serial number 0 and the
14012 CA options of 'x509' had to use a serial number in a file which was
14017 * New options to 'ca' utility to support V2 CRL entry extensions.
14018 Currently CRL reason, invalidity date and hold instruction are
14019 supported. Add new CRL extensions to V3 code and some new objects.
14023 * New function EVP_CIPHER_CTX_set_padding() this is used to
14024 disable standard block padding (aka PKCS#5 padding) in the EVP
14025 API, which was previously mandatory. This means that the data is
14026 not padded in any way and so the total length much be a multiple
14027 of the block size, otherwise an error occurs.
14031 * Initial (incomplete) OCSP SSL support.
14035 * New function OCSP_parse_url(). This splits up a URL into its host,
14036 port and path components: primarily to parse OCSP URLs. New -url
14037 option to ocsp utility.
14041 * New nonce behavior. The return value of OCSP_check_nonce() now
14042 reflects the various checks performed. Applications can decide
14043 whether to tolerate certain situations such as an absent nonce
14044 in a response when one was present in a request: the ocsp application
14045 just prints out a warning. New function OCSP_add1_basic_nonce()
14046 this is to allow responders to include a nonce in a response even if
14047 the request is nonce-less.
14051 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14052 skipped when using openssl x509 multiple times on a single input file,
14053 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14057 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14058 set string type: to handle setting ASN1_TIME structures. Fix ca
14059 utility to correctly initialize revocation date of CRLs.
14063 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14064 the clients preferred ciphersuites and rather use its own preferences.
14065 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14066 Internet Explorer by ensuring unchanged hash method during stepup.
14067 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14071 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14072 to aes and add a new 'exist' option to print out symbols that don't
14077 * Additional options to ocsp utility to allow flags to be set and
14078 additional certificates supplied.
14082 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14083 OCSP client a number of certificate to only verify the response
14088 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14089 handle the new API. Currently only ECB, CBC modes supported. Add new
14092 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14093 Encryption Standard (AES) Ciphersuites for Transport Layer
14094 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14095 not enabled by default and were not part of the "ALL" ciphersuite
14096 alias because they were not yet official; they could be
14097 explicitly requested by specifying the "AESdraft" ciphersuite
14098 group alias. In the final release of OpenSSL 0.9.7, the group
14099 alias is called "AES" and is part of "ALL".)
14101 *Ben Laurie, Steve Henson, Bodo Moeller*
14103 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14104 request to response.
14108 * Functions for OCSP responders. OCSP_request_onereq_count(),
14109 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14110 extract information from a certificate request. OCSP_response_create()
14111 creates a response and optionally adds a basic response structure.
14112 OCSP_basic_add1_status() adds a complete single response to a basic
14113 response and returns the OCSP_SINGLERESP structure just added (to allow
14114 extensions to be included for example). OCSP_basic_add1_cert() adds a
14115 certificate to a basic response and OCSP_basic_sign() signs a basic
14116 response with various flags. New helper functions ASN1_TIME_check()
14117 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14118 (converts ASN1_TIME to GeneralizedTime).
14122 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14123 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14124 structure from a certificate. X509_pubkey_digest() digests the public_key
14125 contents: this is used in various key identifiers.
14129 * Make sk_sort() tolerate a NULL argument.
14131 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14133 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14134 passed by the function are trusted implicitly. If any of them signed the
14135 response then it is assumed to be valid and is not verified.
14139 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14140 to data. This was previously part of the PKCS7 ASN1 code. This
14141 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14142 *Steve Henson, reported by Kenneth R. Robinette
14143 <support@securenetterm.com>*
14145 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14146 routines: without these tracing memory leaks is very painful.
14147 Fix leaks in PKCS12 and PKCS7 routines.
14151 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14152 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14153 effectively meant GeneralizedTime would never be used. Now it
14154 is initialised to -1 but X509_time_adj() now has to check the value
14155 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14156 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14157 *Steve Henson, reported by Kenneth R. Robinette
14158 <support@securenetterm.com>*
14160 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14161 result in a zero length in the ASN1_INTEGER structure which was
14162 not consistent with the structure when d2i_ASN1_INTEGER() was used
14163 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14164 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14165 where it did not print out a minus for negative ASN1_INTEGER.
14169 * Add summary printout to ocsp utility. The various functions which
14170 convert status values to strings have been renamed to:
14171 OCSP_response_status_str(), OCSP_cert_status_str() and
14172 OCSP_crl_reason_str() and are no longer static. New options
14173 to verify nonce values and to disable verification. OCSP response
14174 printout format cleaned up.
14178 * Add additional OCSP certificate checks. These are those specified
14179 in RFC2560. This consists of two separate checks: the CA of the
14180 certificate being checked must either be the OCSP signer certificate
14181 or the issuer of the OCSP signer certificate. In the latter case the
14182 OCSP signer certificate must contain the OCSP signing extended key
14183 usage. This check is performed by attempting to match the OCSP
14184 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14185 in the OCSP_CERTID structures of the response.
14189 * Initial OCSP certificate verification added to OCSP_basic_verify()
14190 and related routines. This uses the standard OpenSSL certificate
14191 verify routines to perform initial checks (just CA validity) and
14192 to obtain the certificate chain. Then additional checks will be
14193 performed on the chain. Currently the root CA is checked to see
14194 if it is explicitly trusted for OCSP signing. This is used to set
14195 a root CA as a global signing root: that is any certificate that
14196 chains to that CA is an acceptable OCSP signing certificate.
14200 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14201 extensions from a separate configuration file.
14202 As when reading extensions from the main configuration file,
14203 the '-extensions ...' option may be used for specifying the
14206 *Massimiliano Pala <madwolf@comune.modena.it>*
14208 * New OCSP utility. Allows OCSP requests to be generated or
14209 read. The request can be sent to a responder and the output
14210 parsed, outputted or printed in text form. Not complete yet:
14211 still needs to check the OCSP response validity.
14215 * New subcommands for 'openssl ca':
14216 `openssl ca -status <serial>` prints the status of the cert with
14217 the given serial number (according to the index file).
14218 `openssl ca -updatedb` updates the expiry status of certificates
14221 *Massimiliano Pala <madwolf@comune.modena.it>*
14223 * New '-newreq-nodes' command option to CA.pl. This is like
14224 '-newreq', but calls 'openssl req' with the '-nodes' option
14225 so that the resulting key is not encrypted.
14227 *Damien Miller <djm@mindrot.org>*
14229 * New configuration for the GNU Hurd.
14231 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14233 * Initial code to implement OCSP basic response verify. This
14234 is currently incomplete. Currently just finds the signer's
14235 certificate and verifies the signature on the response.
14239 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14240 value of OPENSSLDIR. This is available via the new '-d' option
14241 to 'openssl version', and is also included in 'openssl version -a'.
14245 * Allowing defining memory allocation callbacks that will be given
14246 file name and line number information in additional arguments
14247 (a `const char*` and an int). The basic functionality remains, as
14248 well as the original possibility to just replace malloc(),
14249 realloc() and free() by functions that do not know about these
14250 additional arguments. To register and find out the current
14251 settings for extended allocation functions, the following
14252 functions are provided:
14254 CRYPTO_set_mem_ex_functions
14255 CRYPTO_set_locked_mem_ex_functions
14256 CRYPTO_get_mem_ex_functions
14257 CRYPTO_get_locked_mem_ex_functions
14259 These work the same way as CRYPTO_set_mem_functions and friends.
14260 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14261 extended allocation function is enabled.
14262 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14263 a conventional allocation function is enabled.
14265 *Richard Levitte, Bodo Moeller*
14267 * Finish off removing the remaining LHASH function pointer casts.
14268 There should no longer be any prototype-casting required when using
14269 the LHASH abstraction, and any casts that remain are "bugs". See
14270 the callback types and macros at the head of lhash.h for details
14271 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14275 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14276 If /dev/[u]random devices are not available or do not return enough
14277 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14279 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14280 /etc/entropy will be queried once each in this sequence, querying stops
14281 when enough entropy was collected without querying more sockets.
14285 * Change the Unix RAND_poll() variant to be able to poll several
14286 random devices, as specified by DEVRANDOM, until a sufficient amount
14287 of data has been collected. We spend at most 10 ms on each file
14288 (select timeout) and read in non-blocking mode. DEVRANDOM now
14289 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14290 (previously it was just the string "/dev/urandom"), so on typical
14291 platforms the 10 ms delay will never occur.
14292 Also separate out the Unix variant to its own file, rand_unix.c.
14293 For VMS, there's a currently-empty rand_vms.c.
14297 * Move OCSP client related routines to ocsp_cl.c. These
14298 provide utility functions which an application needing
14299 to issue a request to an OCSP responder and analyse the
14300 response will typically need: as opposed to those which an
14301 OCSP responder itself would need which will be added later.
14303 OCSP_request_sign() signs an OCSP request with an API similar
14304 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14305 response. OCSP_response_get1_basic() extracts basic response
14306 from response. OCSP_resp_find_status(): finds and extracts status
14307 information from an OCSP_CERTID structure (which will be created
14308 when the request structure is built). These are built from lower
14309 level functions which work on OCSP_SINGLERESP structures but
14310 won't normally be used unless the application wishes to examine
14311 extensions in the OCSP response for example.
14313 Replace nonce routines with a pair of functions.
14314 OCSP_request_add1_nonce() adds a nonce value and optionally
14315 generates a random value. OCSP_check_nonce() checks the
14316 validity of the nonce in an OCSP response.
14320 * Change function OCSP_request_add() to OCSP_request_add0_id().
14321 This doesn't copy the supplied OCSP_CERTID and avoids the
14322 need to free up the newly created id. Change return type
14323 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14324 This can then be used to add extensions to the request.
14325 Deleted OCSP_request_new(), since most of its functionality
14326 is now in OCSP_REQUEST_new() (and the case insensitive name
14327 clash) apart from the ability to set the request name which
14328 will be added elsewhere.
14332 * Update OCSP API. Remove obsolete extensions argument from
14333 various functions. Extensions are now handled using the new
14334 OCSP extension code. New simple OCSP HTTP function which
14335 can be used to send requests and parse the response.
14339 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14340 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14341 uses the special reorder version of SET OF to sort the attributes
14342 and reorder them to match the encoded order. This resolves a long
14343 standing problem: a verify on a PKCS7 structure just after signing
14344 it used to fail because the attribute order did not match the
14345 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14346 it uses the received order. This is necessary to tolerate some broken
14347 software that does not order SET OF. This is handled by encoding
14348 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14349 to produce the required SET OF.
14353 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14354 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14355 files to get correct declarations of the ASN.1 item variables.
14359 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14360 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14361 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14362 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14363 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14364 ASN1_ITEM and no wrapper functions.
14368 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14369 replace the old function pointer based I/O routines. Change most of
14370 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14374 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14375 lines, recognize more "algorithms" that can be deselected, and make
14376 it complain about algorithm deselection that isn't recognised.
14380 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14381 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14382 to use new functions. Add NO_ASN1_OLD which can be set to remove
14383 some old style ASN1 functions: this can be used to determine if old
14384 code will still work when these eventually go away.
14388 * New extension functions for OCSP structures, these follow the
14389 same conventions as certificates and CRLs.
14393 * New function X509V3_add1_i2d(). This automatically encodes and
14394 adds an extension. Its behaviour can be customised with various
14395 flags to append, replace or delete. Various wrappers added for
14396 certificates and CRLs.
14400 * Fix to avoid calling the underlying ASN1 print routine when
14401 an extension cannot be parsed. Correct a typo in the
14402 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14406 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14407 entries for variables.
14411 * Add functionality to `apps/openssl.c` for detecting locking
14412 problems: As the program is single-threaded, all we have
14413 to do is register a locking callback using an array for
14414 storing which locks are currently held by the program.
14418 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14419 SSL_get_ex_data_X509_STORE_idx(), which is used in
14420 ssl_verify_cert_chain() and thus can be called at any time
14421 during TLS/SSL handshakes so that thread-safety is essential.
14422 Unfortunately, the ex_data design is not at all suited
14423 for multi-threaded use, so it probably should be abolished.
14427 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14429 *Broadcom, tweaked and integrated by Geoff Thorpe*
14431 * Move common extension printing code to new function
14432 X509V3_print_extensions(). Reorganise OCSP print routines and
14433 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14437 * New function X509_signature_print() to remove duplication in some
14442 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14443 set (this was treated exactly the same as SET OF previously). This
14444 is used to reorder the STACK representing the structure to match the
14445 encoding. This will be used to get round a problem where a PKCS7
14446 structure which was signed could not be verified because the STACK
14447 order did not reflect the encoded order.
14451 * Reimplement the OCSP ASN1 module using the new code.
14455 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14456 for its ASN1 operations. The old style function pointers still exist
14457 for now but they will eventually go away.
14461 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14462 completely replaces the old ASN1 functionality with a table driven
14463 encoder and decoder which interprets an ASN1_ITEM structure describing
14464 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14465 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14466 has also been converted to the new form.
14470 * Change BN_mod_exp_recp so that negative moduli are tolerated
14471 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14472 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14473 for negative moduli.
14477 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14478 of not touching the result's sign bit.
14482 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14487 * Changed the LHASH code to use prototypes for callbacks, and created
14488 macros to declare and implement thin (optionally static) functions
14489 that provide type-safety and avoid function pointer casting for the
14490 type-specific callbacks.
14494 * Added Kerberos Cipher Suites to be used with TLS, as written in
14496 *Veers Staats <staatsvr@asc.hpc.mil>,
14497 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14499 * Reformat the FAQ so the different questions and answers can be divided
14500 in sections depending on the subject.
14504 * Have the zlib compression code load ZLIB.DLL dynamically under
14509 * New function BN_mod_sqrt for computing square roots modulo a prime
14510 (using the probabilistic Tonelli-Shanks algorithm unless
14511 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14512 be handled deterministically).
14514 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14516 * Make BN_mod_inverse faster by explicitly handling small quotients
14517 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14518 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14522 * New function BN_kronecker.
14526 * Fix BN_gcd so that it works on negative inputs; the result is
14527 positive unless both parameters are zero.
14528 Previously something reasonably close to an infinite loop was
14529 possible because numbers could be growing instead of shrinking
14530 in the implementation of Euclid's algorithm.
14534 * Fix BN_is_word() and BN_is_one() macros to take into account the
14535 sign of the number in question.
14537 Fix BN_is_word(a,w) to work correctly for w == 0.
14539 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14540 because its test if the absolute value of 'a' equals 'w'.
14541 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14542 it exists mostly for use in the implementations of BN_is_zero(),
14543 BN_is_one(), and BN_is_word().
14547 * New function BN_swap.
14551 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14552 the exponentiation functions are more likely to produce reasonable
14553 results on negative inputs.
14557 * Change BN_mod_mul so that the result is always non-negative.
14558 Previously, it could be negative if one of the factors was negative;
14559 I don't think anyone really wanted that behaviour.
14563 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14564 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14565 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14566 and add new functions:
14575 BN_mod_lshift1_quick
14577 BN_mod_lshift_quick
14579 These functions always generate non-negative results.
14581 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14582 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14584 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14585 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14586 be reduced modulo `m`.
14588 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14591 The following entry accidentally appeared in the CHANGES file
14592 distributed with OpenSSL 0.9.7. The modifications described in
14593 it do *not* apply to OpenSSL 0.9.7.
14595 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14596 was actually never needed) and in BN_mul(). The removal in BN_mul()
14597 required a small change in bn_mul_part_recursive() and the addition
14598 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14599 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14600 bn_sub_words() and bn_add_words() except they take arrays with
14606 * In 'openssl passwd', verify passwords read from the terminal
14607 unless the '-salt' option is used (which usually means that
14608 verification would just waste user's time since the resulting
14609 hash is going to be compared with some given password hash)
14610 or the new '-noverify' option is used.
14612 This is an incompatible change, but it does not affect
14613 non-interactive use of 'openssl passwd' (passwords on the command
14614 line, '-stdin' option, '-in ...' option) and thus should not
14615 cause any problems.
14619 * Remove all references to RSAref, since there's no more need for it.
14623 * Make DSO load along a path given through an environment variable
14624 (SHLIB_PATH) with shl_load().
14628 * Constify the ENGINE code as a result of BIGNUM constification.
14629 Also constify the RSA code and most things related to it. In a
14630 few places, most notable in the depth of the ASN.1 code, ugly
14631 casts back to non-const were required (to be solved at a later
14636 * Make it so the openssl application has all engines loaded by default.
14640 * Constify the BIGNUM routines a little more.
14644 * Add the following functions:
14646 ENGINE_load_cswift()
14648 ENGINE_load_atalla()
14649 ENGINE_load_nuron()
14650 ENGINE_load_builtin_engines()
14652 That way, an application can itself choose if external engines that
14653 are built-in in OpenSSL shall ever be used or not. The benefit is
14654 that applications won't have to be linked with libdl or other dso
14655 libraries unless it's really needed.
14657 Changed 'openssl engine' to load all engines on demand.
14658 Changed the engine header files to avoid the duplication of some
14659 declarations (they differed!).
14663 * 'openssl engine' can now list capabilities.
14667 * Better error reporting in 'openssl engine'.
14671 * Never call load_dh_param(NULL) in s_server.
14675 * Add engine application. It can currently list engines by name and
14676 identity, and test if they are actually available.
14680 * Improve RPM specification file by forcing symbolic linking and making
14681 sure the installed documentation is also owned by root.root.
14683 *Damien Miller <djm@mindrot.org>*
14685 * Give the OpenSSL applications more possibilities to make use of
14686 keys (public as well as private) handled by engines.
14690 * Add OCSP code that comes from CertCo.
14694 * Add VMS support for the Rijndael code.
14698 * Added untested support for Nuron crypto accelerator.
14702 * Add support for external cryptographic devices. This code was
14703 previously distributed separately as the "engine" branch.
14705 *Geoff Thorpe, Richard Levitte*
14707 * Rework the filename-translation in the DSO code. It is now possible to
14708 have far greater control over how a "name" is turned into a filename
14709 depending on the operating environment and any oddities about the
14710 different shared library filenames on each system.
14714 * Support threads on FreeBSD-elf in Configure.
14718 * Fix for SHA1 assembly problem with MASM: it produces
14719 warnings about corrupt line number information when assembling
14720 with debugging information. This is caused by the overlapping
14723 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14726 NCONF_get_number() has no error checking at all. As a replacement,
14727 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14728 promoted strongly. The old NCONF_get_number is kept around for
14729 binary backward compatibility.
14730 Make it possible for methods to load from something other than a BIO,
14731 by providing a function pointer that is given a name instead of a BIO.
14732 For example, this could be used to load configuration data from an
14737 * Fix for non blocking accept BIOs. Added new I/O special reason
14738 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14739 with non blocking I/O was not possible because no retry code was
14740 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14745 * Added the beginnings of Rijndael support.
14749 * Fix for bug in DirectoryString mask setting. Add support for
14750 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14751 to allow certificate printing to more controllable, additional
14752 'certopt' option to 'x509' to allow new printing options to be
14757 * Clean old EAY MD5 hack from e_os.h.
14761 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14763 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14764 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14766 *Joe Orton, Steve Henson*
14768 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14770 * Fix additional bug revealed by the NISCC test suite:
14772 Stop bug triggering large recursion when presented with
14773 certain ASN.1 tags ([CVE-2003-0851])
14777 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14779 * Fix various bugs revealed by running the NISCC test suite:
14781 Stop out of bounds reads in the ASN1 code when presented with
14782 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14784 If verify callback ignores invalid public key errors don't try to check
14785 certificate signature with the NULL public key.
14789 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14790 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14795 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14796 extra data after the compression methods not only for TLS 1.0
14797 but also for SSL 3.0 (as required by the specification).
14799 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14801 * Change X509_certificate_type() to mark the key as exported/exportable
14802 when it's 512 *bits* long, not 512 bytes.
14806 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14808 * Countermeasure against the Klima-Pokorny-Rosa extension of
14809 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14810 a protocol version number mismatch like a decryption error
14811 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14815 * Turn on RSA blinding by default in the default implementation
14816 to avoid a timing attack. Applications that don't want it can call
14817 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14818 They would be ill-advised to do so in most cases.
14820 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14822 * Change RSA blinding code so that it works when the PRNG is not
14823 seeded (in this case, the secret RSA exponent is abused as
14824 an unpredictable seed -- if it is not unpredictable, there
14825 is no point in blinding anyway). Make RSA blinding thread-safe
14826 by remembering the creator's thread ID in rsa->blinding and
14827 having all other threads use local one-time blinding factors
14828 (this requires more computation than sharing rsa->blinding, but
14829 avoids excessive locking; and if an RSA object is not shared
14830 between threads, blinding will still be very fast).
14834 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14836 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14837 via timing by performing a MAC computation even if incorrect
14838 block cipher padding has been found. This is a countermeasure
14839 against active attacks where the attacker has to distinguish
14840 between bad padding and a MAC verification error. ([CVE-2003-0078])
14842 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14843 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14844 Martin Vuagnoux (EPFL, Ilion)*
14846 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14848 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14849 memory from its contents. This is done with a counter that will
14850 place alternating values in each byte. This can be used to solve
14851 two issues: 1) the removal of calls to memset() by highly optimizing
14852 compilers, and 2) cleansing with other values than 0, since those can
14853 be read through on certain media, for example a swap space on disk.
14857 * Bugfix: client side session caching did not work with external caching,
14858 because the session->cipher setting was not restored when reloading
14859 from the external cache. This problem was masked, when
14860 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14861 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14865 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14866 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14868 *Zeev Lieber <zeev-l@yahoo.com>*
14870 * Undo an undocumented change introduced in 0.9.6e which caused
14871 repeated calls to OpenSSL_add_all_ciphers() and
14872 OpenSSL_add_all_digests() to be ignored, even after calling
14877 * Change the default configuration reader to deal with last line not
14878 being properly terminated.
14882 * Change X509_NAME_cmp() so it applies the special rules on handling
14883 DN values that are of type PrintableString, as well as RDNs of type
14884 emailAddress where the value has the type ia5String.
14886 *stefank@valicert.com via Richard Levitte*
14888 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14889 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14890 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14891 the bitwise-OR of the two for use by the majority of applications
14892 wanting this behaviour, and update the docs. The documented
14893 behaviour and actual behaviour were inconsistent and had been
14894 changing anyway, so this is more a bug-fix than a behavioural
14897 *Geoff Thorpe, diagnosed by Nadav Har'El*
14899 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14900 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14904 * Fix initialization code race conditions in
14905 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14906 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14907 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14908 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14909 ssl2_get_cipher_by_char(),
14910 ssl3_get_cipher_by_char().
14912 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14914 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14915 the cached sessions are flushed, as the remove_cb() might use ex_data
14916 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14917 (see [openssl.org #212]).
14919 *Geoff Thorpe, Lutz Jaenicke*
14921 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14922 length, instead of the encoding length to d2i_ASN1_OBJECT.
14926 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14928 * [In 0.9.6g-engine release:]
14929 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14931 *Lynn Gazis <lgazis@rainbow.com>*
14933 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14935 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14936 and get fix the header length calculation.
14937 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14938 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14940 * Use proper error handling instead of 'assertions' in buffer
14941 overflow checks added in 0.9.6e. This prevents DoS (the
14942 assertions could call abort()).
14944 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14946 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14948 * Add various sanity checks to asn1_get_length() to reject
14949 the ASN1 length bytes if they exceed sizeof(long), will appear
14950 negative or the content length exceeds the length of the
14953 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14955 * Fix cipher selection routines: ciphers without encryption had no flags
14956 for the cipher strength set and where therefore not handled correctly
14957 by the selection routines (PR #130).
14961 * Fix EVP_dsa_sha macro.
14966 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14967 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14968 that was added in OpenSSL 0.9.6d.
14970 As the countermeasure turned out to be incompatible with some
14971 broken SSL implementations, the new option is part of SSL_OP_ALL.
14972 SSL_OP_ALL is usually employed when compatibility with weird SSL
14973 implementations is desired (e.g. '-bugs' option to 's_client' and
14974 's_server'), so the new option is automatically set in many
14979 * Changes in security patch:
14981 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14982 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14983 Air Force Materiel Command, USAF, under agreement number
14986 * Add various sanity checks to asn1_get_length() to reject
14987 the ASN1 length bytes if they exceed sizeof(long), will appear
14988 negative or the content length exceeds the length of the
14989 supplied buffer. ([CVE-2002-0659])
14991 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14993 * Assertions for various potential buffer overflows, not known to
14994 happen in practice.
14996 *Ben Laurie (CHATS)*
14998 * Various temporary buffers to hold ASCII versions of integers were
14999 too small for 64 bit platforms. ([CVE-2002-0655])
15000 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15002 * Remote buffer overflow in SSL3 protocol - an attacker could
15003 supply an oversized session ID to a client. ([CVE-2002-0656])
15005 *Ben Laurie (CHATS)*
15007 * Remote buffer overflow in SSL2 protocol - an attacker could
15008 supply an oversized client master key. ([CVE-2002-0656])
15010 *Ben Laurie (CHATS)*
15012 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15014 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15015 encoded as NULL) with id-dsa-with-sha1.
15017 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15019 * Check various `X509_...()` return values in `apps/req.c`.
15021 *Nils Larsch <nla@trustcenter.de>*
15023 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15024 an end-of-file condition would erroneously be flagged, when the CRLF
15025 was just at the end of a processed block. The bug was discovered when
15026 processing data through a buffering memory BIO handing the data to a
15027 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15028 <ptsekov@syntrex.com> and Nedelcho Stanev.
15032 * Implement a countermeasure against a vulnerability recently found
15033 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15034 before application data chunks to avoid the use of known IVs
15035 with data potentially chosen by the attacker.
15039 * Fix length checks in ssl3_get_client_hello().
15043 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15044 to prevent ssl3_read_internal() from incorrectly assuming that
15045 ssl3_read_bytes() found application data while handshake
15046 processing was enabled when in fact s->s3->in_read_app_data was
15047 merely automatically cleared during the initial handshake.
15049 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15051 * Fix object definitions for Private and Enterprise: they were not
15052 recognized in their shortname (=lowercase) representation. Extend
15053 obj_dat.pl to issue an error when using undefined keywords instead
15054 of silently ignoring the problem (Svenning Sorensen
15055 <sss@sss.dnsalias.net>).
15059 * Fix DH_generate_parameters() so that it works for 'non-standard'
15060 generators, i.e. generators other than 2 and 5. (Previously, the
15061 code did not properly initialise the 'add' and 'rem' values to
15062 BN_generate_prime().)
15064 In the new general case, we do not insist that 'generator' is
15065 actually a primitive root: This requirement is rather pointless;
15066 a generator of the order-q subgroup is just as good, if not
15071 * Map new X509 verification errors to alerts. Discovered and submitted by
15072 Tom Wu <tom@arcot.com>.
15076 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15077 returning non-zero before the data has been completely received
15078 when using non-blocking I/O.
15080 *Bodo Moeller; problem pointed out by John Hughes*
15082 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15084 *Ben Laurie, Lutz Jaenicke*
15086 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15087 Yoram Zahavi <YoramZ@gilian.com>).
15091 * Add information about CygWin 1.3 and on, and preserve proper
15092 configuration for the versions before that.
15094 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15096 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15097 check whether we deal with a copy of a session and do not delete from
15098 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15099 <izhar@checkpoint.com>.
15103 * Do not store session data into the internal session cache, if it
15104 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15105 flag is set). Proposed by Aslam <aslam@funk.com>.
15109 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15114 * [In 0.9.6d-engine release:]
15115 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15117 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15119 * Add the configuration target linux-s390x.
15121 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15123 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15124 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15125 variable as an indication that a ClientHello message has been
15126 received. As the flag value will be lost between multiple
15127 invocations of ssl3_accept when using non-blocking I/O, the
15128 function may not be aware that a handshake has actually taken
15129 place, thus preventing a new session from being added to the
15132 To avoid this problem, we now set s->new_session to 2 instead of
15133 using a local variable.
15135 *Lutz Jaenicke, Bodo Moeller*
15137 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15138 if the SSL_R_LENGTH_MISMATCH error is detected.
15140 *Geoff Thorpe, Bodo Moeller*
15142 * New 'shared_ldflag' column in Configure platform table.
15146 * Fix EVP_CIPHER_mode macro.
15148 *"Dan S. Camper" <dan@bti.net>*
15150 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15151 type, we must throw them away by setting rr->length to 0.
15153 *D P Chang <dpc@qualys.com>*
15155 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15157 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15158 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15159 worked incorrectly for those cases where range = `10..._2` and
15160 `3*range` is two bits longer than range.)
15164 * Only add signing time to PKCS7 structures if it is not already
15169 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15170 OBJ_ld_ce should be OBJ_id_ce.
15171 Also some ip-pda OIDs in crypto/objects/objects.txt were
15172 incorrect (cf. RFC 3039).
15174 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15176 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15177 returns early because it has nothing to do.
15179 *Andy Schneider <andy.schneider@bjss.co.uk>*
15181 * [In 0.9.6c-engine release:]
15182 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15184 *Andy Schneider <andy.schneider@bjss.co.uk>*
15186 * [In 0.9.6c-engine release:]
15187 Add support for Cryptographic Appliance's keyserver technology.
15188 (Use engine 'keyclient')
15190 *Cryptographic Appliances and Geoff Thorpe*
15192 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15193 is called via tools/c89.sh because arguments have to be
15194 rearranged (all '-L' options must appear before the first object
15197 *Richard Shapiro <rshapiro@abinitio.com>*
15199 * [In 0.9.6c-engine release:]
15200 Add support for Broadcom crypto accelerator cards, backported
15203 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15205 * [In 0.9.6c-engine release:]
15206 Add support for SureWare crypto accelerator cards from
15207 Baltimore Technologies. (Use engine 'sureware')
15209 *Baltimore Technologies and Mark Cox*
15211 * [In 0.9.6c-engine release:]
15212 Add support for crypto accelerator cards from Accelerated
15213 Encryption Processing, www.aep.ie. (Use engine 'aep')
15215 *AEP Inc. and Mark Cox*
15217 * Add a configuration entry for gcc on UnixWare.
15219 *Gary Benson <gbenson@redhat.com>*
15221 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15222 messages are stored in a single piece (fixed-length part and
15223 variable-length part combined) and fix various bugs found on the way.
15227 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15228 instead. BIO_gethostbyname() does not know what timeouts are
15229 appropriate, so entries would stay in cache even when they have
15231 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15233 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15234 faced with a pathologically small ClientHello fragment that does
15235 not contain client_version: Instead of aborting with an error,
15236 simply choose the highest available protocol version (i.e.,
15237 TLS 1.0 unless it is disabled). In practice, ClientHello
15238 messages are never sent like this, but this change gives us
15239 strictly correct behaviour at least for TLS.
15243 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15244 never resets s->method to s->ctx->method when called from within
15245 one of the SSL handshake functions.
15247 *Bodo Moeller; problem pointed out by Niko Baric*
15249 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15250 (sent using the client's version number) if client_version is
15251 smaller than the protocol version in use. Also change
15252 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15253 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15254 the client will at least see that alert.
15258 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15263 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15264 client receives HelloRequest while in a handshake.
15266 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15268 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15269 should end in 'break', not 'goto end' which circumvents various
15270 cleanups done in state SSL_ST_OK. But session related stuff
15271 must be disabled for SSL_ST_OK in the case that we just sent a
15274 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15275 before just sending a HelloRequest.
15277 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15279 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15280 reveal whether illegal block cipher padding was found or a MAC
15281 verification error occurred. (Neither SSLerr() codes nor alerts
15282 are directly visible to potential attackers, but the information
15283 may leak via logfiles.)
15285 Similar changes are not required for the SSL 2.0 implementation
15286 because the number of padding bytes is sent in clear for SSL 2.0,
15287 and the extra bytes are just ignored. However ssl/s2_pkt.c
15288 failed to verify that the purported number of padding bytes is in
15293 * Add OpenUNIX-8 support including shared libraries
15294 (Boyd Lynn Gerber <gerberb@zenez.com>).
15298 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15299 'wristwatch attack' using huge encoding parameters (cf.
15300 James H. Manger's CRYPTO 2001 paper). Note that the
15301 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15302 encoding parameters and hence was not vulnerable.
15306 * BN_sqr() bug fix.
15308 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15310 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15311 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15312 followed by modular reduction.
15314 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15316 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15317 equivalent based on BN_pseudo_rand() instead of BN_rand().
15321 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15322 This function was broken, as the check for a new client hello message
15323 to handle SGC did not allow these large messages.
15324 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15328 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15332 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15333 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15337 * Rework the configuration and shared library support for Tru64 Unix.
15338 The configuration part makes use of modern compiler features and
15339 still retains old compiler behavior for those that run older versions
15340 of the OS. The shared library support part includes a variant that
15341 uses the RPATH feature, and is available through the special
15342 configuration target "alpha-cc-rpath", which will never be selected
15345 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15347 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15348 with the same message size as in ssl3_get_certificate_request().
15349 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15350 messages might inadvertently be reject as too long.
15352 *Petr Lampa <lampa@fee.vutbr.cz>*
15354 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15358 * Modified SSL library such that the verify_callback that has been set
15359 specifically for an SSL object with SSL_set_verify() is actually being
15360 used. Before the change, a verify_callback set with this function was
15361 ignored and the verify_callback() set in the SSL_CTX at the time of
15362 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15363 to allow the necessary settings.
15367 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15368 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15369 done automatically (in contradiction to the requirements of the C
15370 standard). This made problems when used from OpenSSH.
15374 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15375 dh->length and always used
15377 BN_rand_range(priv_key, dh->p).
15379 BN_rand_range() is not necessary for Diffie-Hellman, and this
15380 specific range makes Diffie-Hellman unnecessarily inefficient if
15381 dh->length (recommended exponent length) is much smaller than the
15382 length of dh->p. We could use BN_rand_range() if the order of
15383 the subgroup was stored in the DH structure, but we only have
15388 BN_rand(priv_key, l, ...)
15390 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15397 RSA_eay_public_encrypt
15398 RSA_eay_private_decrypt
15399 RSA_eay_private_encrypt (signing)
15400 RSA_eay_public_decrypt (signature verification)
15402 (default implementations for RSA_public_encrypt,
15403 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15404 always reject numbers >= n.
15408 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15409 to synchronize access to 'locking_thread'. This is necessary on
15410 systems where access to 'locking_thread' (an 'unsigned long'
15411 variable) is not atomic.
15415 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15416 *before* setting the 'crypto_lock_rand' flag. The previous code had
15417 a race condition if 0 is a valid thread ID.
15419 *Travis Vitek <vitek@roguewave.com>*
15421 * Add support for shared libraries under Irix.
15423 *Albert Chin-A-Young <china@thewrittenword.com>*
15425 * Add configuration option to build on Linux on both big-endian and
15426 little-endian MIPS.
15428 *Ralf Baechle <ralf@uni-koblenz.de>*
15430 * Add the possibility to create shared libraries on HP-UX.
15434 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15436 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15437 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15438 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15439 PRNG state recovery was possible based on the output of
15440 one PRNG request appropriately sized to gain knowledge on
15441 'md' followed by enough consecutive 1-byte PRNG requests
15442 to traverse all of 'state'.
15444 1. When updating 'md_local' (the current thread's copy of 'md')
15445 during PRNG output generation, hash all of the previous
15446 'md_local' value, not just the half used for PRNG output.
15448 2. Make the number of bytes from 'state' included into the hash
15449 independent from the number of PRNG bytes requested.
15451 The first measure alone would be sufficient to avoid
15452 Markku-Juhani's attack. (Actually it had never occurred
15453 to me that the half of 'md_local' used for chaining was the
15454 half from which PRNG output bytes were taken -- I had always
15455 assumed that the secret half would be used.) The second
15456 measure makes sure that additional data from 'state' is never
15457 mixed into 'md_local' in small portions; this heuristically
15458 further strengthens the PRNG.
15462 * Fix crypto/bn/asm/mips3.s.
15466 * When only the key is given to "enc", the IV is undefined. Print out
15467 an error message in this case.
15471 * Handle special case when X509_NAME is empty in X509 printing routines.
15475 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15476 positive and less than q.
15480 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15481 used: it isn't thread safe and the add_lock_callback should handle
15484 *Paul Rose <Paul.Rose@bridge.com>*
15486 * Verify that incoming data obeys the block size in
15487 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15493 *Ulf Möller, Bodo Möller*
15495 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15496 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15497 when fixing the server behaviour for backwards-compatible 'client
15498 hello' messages. (Note that the attack is impractical against
15499 SSL 3.0 and TLS 1.0 anyway because length and version checking
15500 means that the probability of guessing a valid ciphertext is
15501 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15504 Before 0.9.5, the countermeasure (hide the error by generating a
15505 random 'decryption result') did not work properly because
15506 ERR_clear_error() was missing, meaning that SSL_get_error() would
15507 detect the supposedly ignored error.
15509 Both problems are now fixed.
15513 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15514 (previously it was 1024).
15518 * Fix for compatibility mode trust settings: ignore trust settings
15519 unless some valid trust or reject settings are present.
15523 * Fix for blowfish EVP: its a variable length cipher.
15527 * Fix various bugs related to DSA S/MIME verification. Handle missing
15528 parameters in DSA public key structures and return an error in the
15529 DSA routines if parameters are absent.
15533 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15534 in the current directory if neither $RANDFILE nor $HOME was set.
15535 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15536 caused some confusion to Windows users who haven't defined $HOME.
15537 Thus RAND_file_name() is changed again: e_os.h can define a
15538 DEFAULT_HOME, which will be used if $HOME is not set.
15539 For Windows, we use "C:"; on other platforms, we still require
15540 environment variables.
15542 * Move 'if (!initialized) RAND_poll()' into regions protected by
15543 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15544 having multiple threads call RAND_poll() concurrently.
15548 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15549 combination of a flag and a thread ID variable.
15550 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15551 flag), *other* threads can enter ssleay_add_bytes without obeying
15552 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15553 that they do not hold after the first thread unsets add_do_not_lock).
15557 * Change bctest again: '-x' expressions are not available in all
15558 versions of 'test'.
15562 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15564 * Fix a couple of memory leaks in PKCS7_dataDecode()
15566 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15568 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15569 the default extension for executables, if any. Also, make the perl
15570 scripts that use symlink() to test if it really exists and use "cp"
15571 if it doesn't. All this made OpenSSL compilable and installable in
15576 * Fix for asn1_GetSequence() for indefinite length constructed data.
15577 If SEQUENCE is length is indefinite just set c->slen to the total
15578 amount of data available.
15580 *Steve Henson, reported by shige@FreeBSD.org*
15582 *This change does not apply to 0.9.7.*
15584 * Change bctest to avoid here-documents inside command substitution
15585 (workaround for FreeBSD /bin/sh bug).
15586 For compatibility with Ultrix, avoid shell functions (introduced
15587 in the bctest version that searches along $PATH).
15591 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15592 with des_encrypt() defined on some operating systems, like Solaris
15597 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15598 On the Importance of Eliminating Errors in Cryptographic
15599 Computations, J. Cryptology 14 (2001) 2, 101-119,
15600 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15604 * MIPS assembler BIGNUM division bug fix.
15608 * Disabled incorrect Alpha assembler code.
15612 * Fix PKCS#7 decode routines so they correctly update the length
15613 after reading an EOC for the EXPLICIT tag.
15617 *This change does not apply to 0.9.7.*
15619 * Fix bug in PKCS#12 key generation routines. This was triggered
15620 if a 3DES key was generated with a 0 initial byte. Include
15621 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15622 (but broken) behaviour.
15626 * Enhance bctest to search for a working bc along $PATH and print
15629 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15631 * Fix memory leaks in err.c: free err_data string if necessary;
15632 don't write to the wrong index in ERR_set_error_data.
15636 * Implement ssl23_peek (analogous to ssl23_read), which previously
15641 * Replace rdtsc with `_emit` statements for VC++ version 5.
15643 *Jeremy Cooper <jeremy@baymoo.org>*
15645 * Make it possible to reuse SSLv2 sessions.
15649 * In copy_email() check for >= 0 as a return value for
15650 X509_NAME_get_index_by_NID() since 0 is a valid index.
15652 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15654 * Avoid coredump with unsupported or invalid public keys by checking if
15655 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15656 PKCS7_verify() fails with non detached data.
15660 * Don't use getenv in library functions when run as setuid/setgid.
15661 New function OPENSSL_issetugid().
15665 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15666 due to incorrect handling of multi-threading:
15668 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15670 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15672 3. Count how many times MemCheck_off() has been called so that
15673 nested use can be treated correctly. This also avoids
15674 inband-signalling in the previous code (which relied on the
15675 assumption that thread ID 0 is impossible).
15679 * Add "-rand" option also to s_client and s_server.
15683 * Fix CPU detection on Irix 6.x.
15684 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15685 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15687 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15692 *This change does not apply to 0.9.7.*
15694 * Use the cached encoding of an X509_NAME structure rather than
15695 copying it. This is apparently the reason for the libsafe "errors"
15696 but the code is actually correct.
15700 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15701 Bleichenbacher's DSA attack.
15702 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15703 to be set and top=0 forces the highest bit to be set; top=-1 is new
15704 and leaves the highest bit random.
15706 *Ulf Moeller, Bodo Moeller*
15708 * In the `NCONF_...`-based implementations for `CONF_...` queries
15709 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15710 a temporary CONF structure with the data component set to NULL
15711 (which gives segmentation faults in lh_retrieve).
15712 Instead, use NULL for the CONF pointer in CONF_get_string and
15713 CONF_get_number (which may use environment variables) and directly
15714 return NULL from CONF_get_section.
15718 * Fix potential buffer overrun for EBCDIC.
15722 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15723 keyUsage if basicConstraints absent for a CA.
15727 * Make SMIME_write_PKCS7() write mail header values with a format that
15728 is more generally accepted (no spaces before the semicolon), since
15729 some programs can't parse those values properly otherwise. Also make
15730 sure BIO's that break lines after each write do not create invalid
15735 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15736 macros previously used would not encode an empty SEQUENCE OF
15737 and break the signature.
15741 *This change does not apply to 0.9.7.*
15743 * Zero the premaster secret after deriving the master secret in
15748 * Add some EVP_add_digest_alias registrations (as found in
15749 OpenSSL_add_all_digests()) to SSL_library_init()
15750 aka OpenSSL_add_ssl_algorithms(). This provides improved
15751 compatibility with peers using X.509 certificates
15752 with unconventional AlgorithmIdentifier OIDs.
15756 * Fix for Irix with NO_ASM.
15758 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15760 * ./config script fixes.
15762 *Ulf Moeller, Richard Levitte*
15764 * Fix 'openssl passwd -1'.
15768 * Change PKCS12_key_gen_asc() so it can cope with non null
15769 terminated strings whose length is passed in the passlen
15770 parameter, for example from PEM callbacks. This was done
15771 by adding an extra length parameter to asc2uni().
15773 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15775 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15776 call failed, free the DSA structure.
15780 * Fix to uni2asc() to cope with zero length Unicode strings.
15781 These are present in some PKCS#12 files.
15785 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15786 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15787 when writing a 32767 byte record.
15789 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15791 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15792 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15794 (RSA objects have a reference count access to which is protected
15795 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15796 so they are meant to be shared between threads.)
15797 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15798 "Reddie, Steven" <Steven.Reddie@ca.com>*
15800 * Fix a deadlock in CRYPTO_mem_leaks().
15804 * Use better test patterns in bntest.
15808 * rand_win.c fix for Borland C.
15812 * BN_rshift bugfix for n == 0.
15816 * Add a 'bctest' script that checks for some known 'bc' bugs
15817 so that 'make test' does not abort just because 'bc' is broken.
15821 * Store verify_result within SSL_SESSION also for client side to
15822 avoid potential security hole. (Re-used sessions on the client side
15823 always resulted in verify_result==X509_V_OK, not using the original
15824 result of the server certificate verification.)
15828 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15829 SSL3_RT_APPLICATION_DATA, return 0.
15830 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15835 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15836 releases, have been re-implemented by renaming the previous
15837 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15838 and ssl3_read_internal, respectively, and adding 'peek' parameters
15839 to them. The new ssl[23]_{read,peek} functions are calls to
15840 ssl[23]_read_internal with the 'peek' flag set appropriately.
15841 A 'peek' parameter has also been added to ssl3_read_bytes, which
15842 does the actual work for ssl3_read_internal.
15846 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15847 the method-specific "init()" handler. Also clean up ex_data after
15848 calling the method-specific "finish()" handler. Previously, this was
15849 happening the other way round.
15853 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15854 The previous value, 12, was not always sufficient for BN_mod_exp().
15858 * Make sure that shared libraries get the internal name engine with
15859 the full version number and not just 0. This should mark the
15860 shared libraries as not backward compatible. Of course, this should
15861 be changed again when we can guarantee backward binary compatibility.
15865 * Fix typo in get_cert_by_subject() in by_dir.c
15867 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15869 * Rework the system to generate shared libraries:
15871 - Make note of the expected extension for the shared libraries and
15872 if there is a need for symbolic links from for example libcrypto.so.0
15873 to libcrypto.so.0.9.7. There is extended info in Configure for
15876 - Make as few rebuilds of the shared libraries as possible.
15878 - Still avoid linking the OpenSSL programs with the shared libraries.
15880 - When installing, install the shared libraries separately from the
15885 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15887 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15888 and not in SSL_clear because the latter is also used by the
15889 accept/connect functions; previously, the settings made by
15890 SSL_set_read_ahead would be lost during the handshake.
15892 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15894 * Correct util/mkdef.pl to be selective about disabled algorithms.
15895 Previously, it would create entries for disabled algorithms no
15900 * Added several new manual pages for SSL_* function.
15904 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15906 * In ssl23_get_client_hello, generate an error message when faced
15907 with an initial SSL 3.0/TLS record that is too small to contain the
15908 first two bytes of the ClientHello message, i.e. client_version.
15909 (Note that this is a pathologic case that probably has never happened
15910 in real life.) The previous approach was to use the version number
15911 from the record header as a substitute; but our protocol choice
15912 should not depend on that one because it is not authenticated
15913 by the Finished messages.
15917 * More robust randomness gathering functions for Windows.
15919 *Jeffrey Altman <jaltman@columbia.edu>*
15921 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15922 not set then we don't setup the error code for issuer check errors
15923 to avoid possibly overwriting other errors which the callback does
15924 handle. If an application does set the flag then we assume it knows
15925 what it is doing and can handle the new informational codes
15930 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15931 a general "ANY" type, as such it should be able to decode anything
15932 including tagged types. However it didn't check the class so it would
15933 wrongly interpret tagged types in the same way as their universal
15934 counterpart and unknown types were just rejected. Changed so that the
15935 tagged and unknown types are handled in the same way as a SEQUENCE:
15936 that is the encoding is stored intact. There is also a new type
15937 "V_ASN1_OTHER" which is used when the class is not universal, in this
15938 case we have no idea what the actual type is so we just lump them all
15943 * On VMS, stdout may very well lead to a file that is written to
15944 in a record-oriented fashion. That means that every write() will
15945 write a separate record, which will be read separately by the
15946 programs trying to read from it. This can be very confusing.
15948 The solution is to put a BIO filter in the way that will buffer
15949 text until a linefeed is reached, and then write everything a
15950 line at a time, so every record written will be an actual line,
15951 not chunks of lines and not (usually doesn't happen, but I've
15952 seen it once) several lines in one record. BIO_f_linebuffer() is
15955 Currently, it's a VMS-only method, because that's where it has
15956 been tested well enough.
15960 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15961 it can return incorrect results.
15962 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15963 but it was in 0.9.6-beta[12].)
15967 * Disable the check for content being present when verifying detached
15968 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15969 include zero length content when signing messages.
15973 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15974 BIO_ctrl (for BIO pairs).
15978 * Add DSO method for VMS.
15982 * Bug fix: Montgomery multiplication could produce results with the
15987 * Add RPM specification openssl.spec and modify it to build three
15988 packages. The default package contains applications, application
15989 documentation and run-time libraries. The devel package contains
15990 include files, static libraries and function documentation. The
15991 doc package contains the contents of the doc directory. The original
15992 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15996 * Add a large number of documentation files for many SSL routines.
15998 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16000 * Add a configuration entry for Sony News 4.
16002 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16004 * Don't set the two most significant bits to one when generating a
16005 random number < q in the DSA library.
16009 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16010 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16011 the underlying transport is blocking) if a handshake took place.
16012 (The default behaviour is needed by applications such as s_client
16013 and s_server that use select() to determine when to use SSL_read;
16014 but for applications that know in advance when to expect data, it
16015 just makes things more complicated.)
16019 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16024 * Add a few more EBCDIC conditionals that make `req` and `x509`
16025 work better on such systems.
16027 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16029 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16030 Update PKCS12_parse() so it copies the friendlyName and the
16031 keyid to the certificates aux info.
16035 * Fix bug in PKCS7_verify() which caused an infinite loop
16036 if there was more than one signature.
16038 *Sven Uszpelkat <su@celocom.de>*
16040 * Major change in util/mkdef.pl to include extra information
16041 about each symbol, as well as presenting variables as well
16042 as functions. This change means that there's n more need
16043 to rebuild the .num files when some algorithms are excluded.
16047 * Allow the verify time to be set by an application,
16048 rather than always using the current time.
16052 * Phase 2 verify code reorganisation. The certificate
16053 verify code now looks up an issuer certificate by a
16054 number of criteria: subject name, authority key id
16055 and key usage. It also verifies self signed certificates
16056 by the same criteria. The main comparison function is
16057 X509_check_issued() which performs these checks.
16059 Lot of changes were necessary in order to support this
16060 without completely rewriting the lookup code.
16062 Authority and subject key identifier are now cached.
16064 The LHASH 'certs' is X509_STORE has now been replaced
16065 by a STACK_OF(X509_OBJECT). This is mainly because an
16066 LHASH can't store or retrieve multiple objects with
16067 the same hash value.
16069 As a result various functions (which were all internal
16070 use only) have changed to handle the new X509_STORE
16071 structure. This will break anything that messed round
16072 with X509_STORE internally.
16074 The functions X509_STORE_add_cert() now checks for an
16075 exact match, rather than just subject name.
16077 The X509_STORE API doesn't directly support the retrieval
16078 of multiple certificates matching a given criteria, however
16079 this can be worked round by performing a lookup first
16080 (which will fill the cache with candidate certificates)
16081 and then examining the cache for matches. This is probably
16082 the best we can do without throwing out X509_LOOKUP
16083 entirely (maybe later...).
16085 The X509_VERIFY_CTX structure has been enhanced considerably.
16087 All certificate lookup operations now go via a get_issuer()
16088 callback. Although this currently uses an X509_STORE it
16089 can be replaced by custom lookups. This is a simple way
16090 to bypass the X509_STORE hackery necessary to make this
16091 work and makes it possible to use more efficient techniques
16092 in future. A very simple version which uses a simple
16093 STACK for its trusted certificate store is also provided
16094 using X509_STORE_CTX_trusted_stack().
16096 The verify_cb() and verify() callbacks now have equivalents
16097 in the X509_STORE_CTX structure.
16099 X509_STORE_CTX also has a 'flags' field which can be used
16100 to customise the verify behaviour.
16104 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16105 excludes S/MIME capabilities.
16109 * When a certificate request is read in keep a copy of the
16110 original encoding of the signed data and use it when outputting
16111 again. Signatures then use the original encoding rather than
16112 a decoded, encoded version which may cause problems if the
16113 request is improperly encoded.
16117 * For consistency with other BIO_puts implementations, call
16118 buffer_write(b, ...) directly in buffer_puts instead of calling
16121 In BIO_puts, increment b->num_write as in BIO_write.
16123 *Peter.Sylvester@EdelWeb.fr*
16125 * Fix BN_mul_word for the case where the word is 0. (We have to use
16126 BN_zero, we may not return a BIGNUM with an array consisting of
16127 words set to zero.)
16131 * Avoid calling abort() from within the library when problems are
16132 detected, except if preprocessor symbols have been defined
16133 (such as REF_CHECK, BN_DEBUG etc.).
16137 * New openssl application 'rsautl'. This utility can be
16138 used for low-level RSA operations. DER public key
16139 BIO/fp routines also added.
16143 * New Configure entry and patches for compiling on QNX 4.
16145 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16147 * A demo state-machine implementation was sponsored by
16148 Nuron (<http://www.nuron.com/>) and is now available in
16149 demos/state_machine.
16153 * New options added to the 'dgst' utility for signature
16154 generation and verification.
16158 * Unrecognized PKCS#7 content types are now handled via a
16159 catch all ASN1_TYPE structure. This allows unsupported
16160 types to be stored as a "blob" and an application can
16161 encode and decode it manually.
16165 * Fix various signed/unsigned issues to make a_strex.c
16166 compile under VC++.
16168 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16170 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16171 length if passed a buffer. ASN1_INTEGER_to_BN failed
16172 if passed a NULL BN and its argument was negative.
16174 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16176 * Modification to PKCS#7 encoding routines to output definite
16177 length encoding. Since currently the whole structures are in
16178 memory there's not real point in using indefinite length
16179 constructed encoding. However if OpenSSL is compiled with
16180 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16184 * Added BIO_vprintf() and BIO_vsnprintf().
16188 * Added more prefixes to parse for in the strings written
16189 through a logging bio, to cover all the levels that are available
16190 through syslog. The prefixes are now:
16192 PANIC, EMERG, EMR => LOG_EMERG
16193 ALERT, ALR => LOG_ALERT
16194 CRIT, CRI => LOG_CRIT
16195 ERROR, ERR => LOG_ERR
16196 WARNING, WARN, WAR => LOG_WARNING
16197 NOTICE, NOTE, NOT => LOG_NOTICE
16198 INFO, INF => LOG_INFO
16199 DEBUG, DBG => LOG_DEBUG
16201 and as before, if none of those prefixes are present at the
16202 beginning of the string, LOG_ERR is chosen.
16204 On Win32, the `LOG_*` levels are mapped according to this:
16206 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16207 LOG_WARNING => EVENTLOG_WARNING_TYPE
16208 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16212 * Made it possible to reconfigure with just the configuration
16213 argument "reconf" or "reconfigure". The command line arguments
16214 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16215 and are retrieved from there when reconfiguring.
16221 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16223 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16227 * The obj_dat.pl script was messing up the sorting of object
16228 names. The reason was that it compared the quoted version
16229 of strings as a result "OCSP" > "OCSP Signing" because
16230 " > SPACE. Changed script to store unquoted versions of
16231 names and add quotes on output. It was also omitting some
16232 names from the lookup table if they were given a default
16233 value (that is if SN is missing it is given the same
16234 value as LN and vice versa), these are now added on the
16235 grounds that if an object has a name we should be able to
16236 look it up. Finally added warning output when duplicate
16237 short or long names are found.
16241 * Changes needed for Tandem NSK.
16243 *Scott Uroff <scott@xypro.com>*
16245 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16246 RSA_padding_check_SSLv23(), special padding was never detected
16247 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16248 version rollback attacks was not effective.
16250 In s23_clnt.c, don't use special rollback-attack detection padding
16251 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16252 client; similarly, in s23_srvr.c, don't do the rollback check if
16253 SSL 2.0 is the only protocol enabled in the server.
16257 * Make it possible to get hexdumps of unprintable data with 'openssl
16258 asn1parse'. By implication, the functions ASN1_parse_dump() and
16259 BIO_dump_indent() are added.
16263 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16264 these print out strings and name structures based on various
16265 flags including RFC2253 support and proper handling of
16266 multibyte characters. Added options to the 'x509' utility
16267 to allow the various flags to be set.
16271 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16272 Also change the functions X509_cmp_current_time() and
16273 X509_gmtime_adj() work with an ASN1_TIME structure,
16274 this will enable certificates using GeneralizedTime in validity
16275 dates to be checked.
16279 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16280 negative public key encodings) on by default,
16281 NO_NEG_PUBKEY_BUG can be set to disable it.
16285 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16286 content octets. An i2c_ASN1_OBJECT is unnecessary because
16287 the encoding can be trivially obtained from the structure.
16291 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16292 not read locks (`CRYPTO_r_[un]lock`).
16296 * A first attempt at creating official support for shared
16297 libraries through configuration. I've kept it so the
16298 default is static libraries only, and the OpenSSL programs
16299 are always statically linked for now, but there are
16300 preparations for dynamic linking in place.
16301 This has been tested on Linux and Tru64.
16305 * Randomness polling function for Win9x, as described in:
16306 Peter Gutmann, Software Generation of Practically Strong
16311 * Fix so PRNG is seeded in req if using an already existing
16316 * New options to smime application. -inform and -outform
16317 allow alternative formats for the S/MIME message including
16318 PEM and DER. The -content option allows the content to be
16319 specified separately. This should allow things like Netscape
16320 form signing output easier to verify.
16324 * Fix the ASN1 encoding of tags using the 'long form'.
16328 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16329 STRING types. These convert content octets to and from the
16330 underlying type. The actual tag and length octets are
16331 already assumed to have been read in and checked. These
16332 are needed because all other string types have virtually
16333 identical handling apart from the tag. By having versions
16334 of the ASN1 functions that just operate on content octets
16335 IMPLICIT tagging can be handled properly. It also allows
16336 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16337 and ASN1_INTEGER are identical apart from the tag.
16341 * Change the handling of OID objects as follows:
16343 - New object identifiers are inserted in objects.txt, following
16344 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16345 - objects.pl is used to process obj_mac.num and create a new
16347 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16350 This is currently kind of a hack, and the perl code in objects.pl
16351 isn't very elegant, but it works as I intended. The simplest way
16352 to check that it worked correctly is to look in obj_dat.h and
16353 check the array nid_objs and make sure the objects haven't moved
16354 around (this is important!). Additions are OK, as well as
16355 consistent name changes.
16359 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16363 * Addition of the command line parameter '-rand file' to 'openssl req'.
16364 The given file adds to whatever has already been seeded into the
16365 random pool through the RANDFILE configuration file option or
16366 environment variable, or the default random state file.
16370 * mkstack.pl now sorts each macro group into lexical order.
16371 Previously the output order depended on the order the files
16372 appeared in the directory, resulting in needless rewriting
16377 * Patches to make OpenSSL compile under Win32 again. Mostly
16378 work arounds for the VC++ problem that it treats func() as
16379 func(void). Also stripped out the parts of mkdef.pl that
16380 added extra typesafe functions: these no longer exist.
16384 * Reorganisation of the stack code. The macros are now all
16385 collected in safestack.h . Each macro is defined in terms of
16386 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16387 DEBUG_SAFESTACK is now handled in terms of function casts,
16388 this has the advantage of retaining type safety without the
16389 use of additional functions. If DEBUG_SAFESTACK is not defined
16390 then the non typesafe macros are used instead. Also modified the
16391 mkstack.pl script to handle the new form. Needs testing to see
16392 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16393 the default if no major problems. Similar behaviour for ASN1_SET_OF
16394 and PKCS12_STACK_OF.
16398 * When some versions of IIS use the 'NET' form of private key the
16399 key derivation algorithm is different. Normally MD5(password) is
16400 used as a 128 bit RC4 key. In the modified case
16401 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16402 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16403 as the old Netscape_RSA functions except they have an additional
16404 'sgckey' parameter which uses the modified algorithm. Also added
16405 an -sgckey command line option to the rsa utility. Thanks to
16406 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16407 algorithm to openssl-dev.
16411 * The evp_local.h macros were using 'c.##kname' which resulted in
16412 invalid expansion on some systems (SCO 5.0.5 for example).
16413 Corrected to 'c.kname'.
16415 *Phillip Porch <root@theporch.com>*
16417 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16418 a STACK of email addresses from a certificate or request, these look
16419 in the subject name and the subject alternative name extensions and
16420 omit any duplicate addresses.
16424 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16425 This makes DSA verification about 2 % faster.
16429 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16430 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16431 plus overhead for 1024 bit moduli).
16432 This makes exponentiations about 0.5 % faster for 1024 bit
16433 exponents (as measured by "openssl speed rsa2048").
16437 * Rename memory handling macros to avoid conflicts with other
16439 Malloc => OPENSSL_malloc
16440 Malloc_locked => OPENSSL_malloc_locked
16441 Realloc => OPENSSL_realloc
16442 Free => OPENSSL_free
16446 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16447 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16451 * CygWin32 support.
16453 *John Jarvie <jjarvie@newsguy.com>*
16455 * The type-safe stack code has been rejigged. It is now only compiled
16456 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16457 by default all type-specific stack functions are "#define"d back to
16458 standard stack functions. This results in more streamlined output
16459 but retains the type-safety checking possibilities of the original
16464 * The STACK code has been cleaned up, and certain type declarations
16465 that didn't make a lot of sense have been brought in line. This has
16466 also involved a cleanup of sorts in safestack.h to more correctly
16467 map type-safe stack functions onto their plain stack counterparts.
16468 This work has also resulted in a variety of "const"ifications of
16469 lots of the code, especially `_cmp` operations which should normally
16470 be prototyped with "const" parameters anyway.
16474 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16475 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16476 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16477 where all of 'md' is used each time the PRNG is used, but 'state'
16478 is used only indexed by a cyclic counter. As entropy may not be
16479 well distributed from the beginning, 'md' is important as a
16480 chaining variable. However, the output function chains only half
16481 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16482 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16483 in all of 'state' being rewritten, with the new values depending
16484 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16488 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16489 the handshake is continued after ssl_verify_cert_chain();
16490 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16491 can lead to 'unexplainable' connection aborts later.
16493 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16495 * Major EVP API cipher revision.
16496 Add hooks for extra EVP features. This allows various cipher
16497 parameters to be set in the EVP interface. Support added for variable
16498 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16499 setting of RC2 and RC5 parameters.
16501 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16504 Remove lots of duplicated code from the EVP library. For example *every*
16505 cipher init() function handles the 'iv' in the same way according to the
16506 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16507 for CFB and OFB modes they zero ctx->num.
16509 New functionality allows removal of S/MIME code RC2 hack.
16511 Most of the routines have the same form and so can be declared in terms
16514 By shifting this to the top level EVP_CipherInit() it can be removed from
16515 all individual ciphers. If the cipher wants to handle IVs or keys
16516 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16519 Change lots of functions like EVP_EncryptUpdate() to now return a
16520 value: although software versions of the algorithms cannot fail
16521 any installed hardware versions can.
16525 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16526 this option is set, tolerate broken clients that send the negotiated
16527 protocol version number instead of the requested protocol version
16532 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16533 i.e. non-zero for export ciphersuites, zero otherwise.
16534 Previous versions had this flag inverted, inconsistent with
16535 rsa_tmp_cb (..._TMP_RSA_CB).
16537 *Bodo Moeller; problem reported by Amit Chopra*
16539 * Add missing DSA library text string. Work around for some IIS
16540 key files with invalid SEQUENCE encoding.
16544 * Add a document (doc/standards.txt) that list all kinds of standards
16545 and so on that are implemented in OpenSSL.
16549 * Enhance c_rehash script. Old version would mishandle certificates
16550 with the same subject name hash and wouldn't handle CRLs at all.
16551 Added -fingerprint option to crl utility, to support new c_rehash
16556 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16560 * Fix for SSL server purpose checking. Server checking was
16561 rejecting certificates which had extended key usage present
16562 but no ssl client purpose.
16564 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16566 * Make PKCS#12 code work with no password. The PKCS#12 spec
16567 is a little unclear about how a blank password is handled.
16568 Since the password in encoded as a BMPString with terminating
16569 double NULL a zero length password would end up as just the
16570 double NULL. However no password at all is different and is
16571 handled differently in the PKCS#12 key generation code. NS
16572 treats a blank password as zero length. MSIE treats it as no
16573 password on export: but it will try both on import. We now do
16574 the same: PKCS12_parse() tries zero length and no password if
16575 the password is set to "" or NULL (NULL is now a valid password:
16576 it wasn't before) as does the pkcs12 application.
16580 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16581 perror when PEM_read_bio_X509_REQ fails, the error message must
16582 be obtained from the error queue.
16586 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16587 it in ERR_remove_state if appropriate, and change ERR_get_state
16588 accordingly to avoid race conditions (this is necessary because
16589 thread_hash is no longer constant once set).
16593 * Bugfix for linux-elf makefile.one.
16597 * RSA_get_default_method() will now cause a default
16598 RSA_METHOD to be chosen if one doesn't exist already.
16599 Previously this was only set during a call to RSA_new()
16600 or RSA_new_method(NULL) meaning it was possible for
16601 RSA_get_default_method() to return NULL.
16605 * Added native name translation to the existing DSO code
16606 that will convert (if the flag to do so is set) filenames
16607 that are sufficiently small and have no path information
16608 into a canonical native form. Eg. "blah" converted to
16609 "libblah.so" or "blah.dll" etc.
16613 * New function ERR_error_string_n(e, buf, len) which is like
16614 ERR_error_string(e, buf), but writes at most 'len' bytes
16615 including the 0 terminator. For ERR_error_string_n, 'buf'
16618 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16620 * CONF library reworked to become more general. A new CONF
16621 configuration file reader "class" is implemented as well as a
16622 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16623 old `CONF_*` functions are still there, but are reimplemented to
16624 work in terms of the new functions. Also, a set of functions
16625 to handle the internal storage of the configuration data is
16626 provided to make it easier to write new configuration file
16627 reader "classes" (I can definitely see something reading a
16628 configuration file in XML format, for example), called `_CONF_*`,
16629 or "the configuration storage API"...
16631 The new configuration file reading functions are:
16633 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16634 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16636 NCONF_default, NCONF_WIN32
16638 NCONF_dump_fp, NCONF_dump_bio
16640 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16641 NCONF_new creates a new CONF object. This works in the same way
16642 as other interfaces in OpenSSL, like the BIO interface.
16643 `NCONF_dump_*` dump the internal storage of the configuration file,
16644 which is useful for debugging. All other functions take the same
16645 arguments as the old `CONF_*` functions with the exception of the
16646 first that must be a `CONF *` instead of a `LHASH *`.
16648 To make it easier to use the new classes with the old `CONF_*` functions,
16649 the function CONF_set_default_method is provided.
16653 * Add '-tls1' option to 'openssl ciphers', which was already
16654 mentioned in the documentation but had not been implemented.
16655 (This option is not yet really useful because even the additional
16656 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16660 * Initial DSO code added into libcrypto for letting OpenSSL (and
16661 OpenSSL-based applications) load shared libraries and bind to
16662 them in a portable way.
16664 *Geoff Thorpe, with contributions from Richard Levitte*
16666 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16668 * Make sure _lrotl and _lrotr are only used with MSVC.
16670 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16671 (the default implementation of RAND_status).
16673 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16674 to '-clrext' (= clear extensions), as intended and documented.
16675 *Bodo Moeller; inconsistency pointed out by Michael Attili
16676 <attili@amaxo.com>*
16678 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16679 was larger than the MD block size.
16681 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16683 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16684 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16685 using the passed key: if the passed key was a private key the result
16686 of X509_print(), for example, would be to print out all the private key
16691 * des_quad_cksum() byte order bug fix.
16692 *Ulf Möller, using the problem description in krb4-0.9.7, where
16693 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16695 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16698 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16700 * For easily testing in shell scripts whether some command
16701 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16702 returns with exit code 0 iff no command of the given name is available.
16703 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16704 the output goes to stdout and nothing is printed to stderr.
16705 Additional arguments are always ignored.
16707 Since for each cipher there is a command of the same name,
16708 the 'no-cipher' compilation switches can be tested this way.
16710 ('openssl no-XXX' is not able to detect pseudo-commands such
16711 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16715 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16719 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16720 is set; it will be thrown away anyway because each handshake creates
16722 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16723 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16724 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16725 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16729 * New s_client option -ign_eof: EOF at stdin is ignored, and
16730 'Q' and 'R' lose their special meanings (quit/renegotiate).
16731 This is part of what -quiet does; unlike -quiet, -ign_eof
16732 does not suppress any output.
16736 * Add compatibility options to the purpose and trust code. The
16737 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16738 accepts a certificate or CA, this was the previous behaviour,
16739 with all the associated security issues.
16741 X509_TRUST_COMPAT is the old trust behaviour: only and
16742 automatically trust self signed roots in certificate store. A
16743 new trust setting X509_TRUST_DEFAULT is used to specify that
16744 a purpose has no associated trust setting and it should instead
16745 use the value in the default purpose.
16749 * Fix the PKCS#8 DSA private key code so it decodes keys again
16750 and fix a memory leak.
16754 * In util/mkerr.pl (which implements 'make errors'), preserve
16755 reason strings from the previous version of the .c file, as
16756 the default to have only downcase letters (and digits) in
16757 automatically generated reasons codes is not always appropriate.
16761 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16762 using strerror. Previously, ERR_reason_error_string() returned
16763 library names as reason strings for SYSerr; but SYSerr is a special
16764 case where small numbers are errno values, not library numbers.
16768 * Add '-dsaparam' option to 'openssl dhparam' application. This
16769 converts DSA parameters into DH parameters. (When creating parameters,
16770 DSA_generate_parameters is used.)
16774 * Include 'length' (recommended exponent length) in C code generated
16775 by 'openssl dhparam -C'.
16779 * The second argument to set_label in perlasm was already being used
16780 so couldn't be used as a "file scope" flag. Moved to third argument
16785 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16786 instead of RAND_bytes for encryption IVs and salts.
16790 * Include RAND_status() into RAND_METHOD instead of implementing
16791 it only for md_rand.c Otherwise replacing the PRNG by calling
16792 RAND_set_rand_method would be impossible.
16796 * Don't let DSA_generate_key() enter an infinite loop if the random
16797 number generation fails.
16801 * New 'rand' application for creating pseudo-random output.
16805 * Added configuration support for Linux/IA64
16807 *Rolf Haberrecker <rolf@suse.de>*
16809 * Assembler module support for Mingw32.
16813 * Shared library support for HPUX (in shlib/).
16815 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16817 * Shared library support for Solaris gcc.
16819 *Lutz Behnke <behnke@trustcenter.de>*
16821 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16823 * PKCS7_encrypt() was adding text MIME headers twice because they
16824 were added manually and by SMIME_crlf_copy().
16828 * In bntest.c don't call BN_rand with zero bits argument.
16830 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16832 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16833 case was implemented. This caused BN_div_recp() to fail occasionally.
16837 * Add an optional second argument to the set_label() in the perl
16838 assembly language builder. If this argument exists and is set
16839 to 1 it signals that the assembler should use a symbol whose
16840 scope is the entire file, not just the current function. This
16841 is needed with MASM which uses the format label:: for this scope.
16843 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16845 * Change the ASN1 types so they are typedefs by default. Before
16846 almost all types were #define'd to ASN1_STRING which was causing
16847 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16852 * Change names of new functions to the new get1/get0 naming
16853 convention: After 'get1', the caller owns a reference count
16854 and has to call `..._free`; 'get0' returns a pointer to some
16855 data structure without incrementing reference counters.
16856 (Some of the existing 'get' functions increment a reference
16857 counter, some don't.)
16858 Similarly, 'set1' and 'add1' functions increase reference
16859 counters or duplicate objects.
16863 * Allow for the possibility of temp RSA key generation failure:
16864 the code used to assume it always worked and crashed on failure.
16868 * Fix potential buffer overrun problem in BIO_printf().
16869 *Ulf Möller, using public domain code by Patrick Powell; problem
16870 pointed out by David Sacerdote <das33@cornell.edu>*
16872 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16873 RAND_egd() and RAND_status(). In the command line application,
16874 the EGD socket can be specified like a seed file using RANDFILE
16879 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16880 Some CAs (e.g. Verisign) distribute certificates in this form.
16884 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16885 list to exclude them. This means that no special compilation option
16886 is needed to use anonymous DH: it just needs to be included in the
16891 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16892 EVP_MD_type. The old functionality is available in a new macro called
16893 EVP_MD_md(). Change code that uses it and update docs.
16897 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16898 where the `void *` argument is replaced by a function pointer argument.
16899 Previously `void *` was abused to point to functions, which works on
16900 many platforms, but is not correct. As these functions are usually
16901 called by macros defined in OpenSSL header files, most source code
16902 should work without changes.
16906 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16907 sections with information on -D... compiler switches used for
16908 compiling the library so that applications can see them. To enable
16909 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16910 must be defined. E.g.,
16911 #define OPENSSL_ALGORITHM_DEFINES
16912 #include <openssl/opensslconf.h>
16913 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16915 *Richard Levitte, Ulf and Bodo Möller*
16917 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16922 * Change the 'other' type in certificate aux info to a STACK_OF
16923 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16924 the required ASN1 format: arbitrary types determined by an OID.
16928 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16929 argument to 'req'. This is not because the function is newer or
16930 better than others it just uses the work 'NEW' in the certificate
16931 request header lines. Some software needs this.
16935 * Reorganise password command line arguments: now passwords can be
16936 obtained from various sources. Delete the PEM_cb function and make
16937 it the default behaviour: i.e. if the callback is NULL and the
16938 usrdata argument is not NULL interpret it as a null terminated pass
16939 phrase. If usrdata and the callback are NULL then the pass phrase
16940 is prompted for as usual.
16944 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16945 the support is automatically enabled. The resulting binaries will
16946 autodetect the card and use it if present.
16948 *Ben Laurie and Compaq Inc.*
16950 * Work around for Netscape hang bug. This sends certificate request
16951 and server done in one record. Since this is perfectly legal in the
16952 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16953 the bugs/SSLv3 entry for more info.
16957 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16961 * Add -rand argument to smime and pkcs12 applications and read/write
16966 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16970 * Add command line password options to the remaining applications.
16974 * Bug fix for BN_div_recp() for numerators with an even number of
16979 * More tests in bntest.c, and changed test_bn output.
16983 * ./config recognizes MacOS X now.
16987 * Bug fix for BN_div() when the first words of num and divisor are
16988 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16992 * Add support for various broken PKCS#8 formats, and command line
16993 options to produce them.
16997 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16998 get temporary BIGNUMs from a BN_CTX.
17002 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17007 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17008 include a #define from the old name to the new. The original intent
17009 was that statically linked binaries could for example just call
17010 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17011 link with digests. This never worked because SSLeay_add_all_digests()
17012 and SSLeay_add_all_ciphers() were in the same source file so calling
17013 one would link with the other. They are now in separate source files.
17017 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17021 * Use a less unusual form of the Miller-Rabin primality test (it used
17022 a binary algorithm for exponentiation integrated into the Miller-Rabin
17023 loop, our standard modexp algorithms are faster).
17027 * Support for the EBCDIC character set completed.
17029 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17031 * Source code cleanups: use const where appropriate, eliminate casts,
17032 use `void *` instead of `char *` in lhash.
17036 * Bugfix: ssl3_send_server_key_exchange was not restartable
17037 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17038 this the server could overwrite ephemeral keys that the client
17043 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17044 using 50 iterations of the Rabin-Miller test.
17046 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17047 iterations of the Rabin-Miller test as required by the appendix
17048 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17049 As BN_is_prime_fasttest includes trial division, DSA parameter
17050 generation becomes much faster.
17052 This implies a change for the callback functions in DSA_is_prime
17053 and DSA_generate_parameters: The callback function is called once
17054 for each positive witness in the Rabin-Miller test, not just
17055 occasionally in the inner loop; and the parameters to the
17056 callback function now provide an iteration count for the outer
17057 loop rather than for the current invocation of the inner loop.
17058 DSA_generate_parameters additionally can call the callback
17059 function with an 'iteration count' of -1, meaning that a
17060 candidate has passed the trial division test (when q is generated
17061 from an application-provided seed, trial division is skipped).
17065 * New function BN_is_prime_fasttest that optionally does trial
17066 division before starting the Rabin-Miller test and has
17067 an additional BN_CTX * argument (whereas BN_is_prime always
17068 has to allocate at least one BN_CTX).
17069 'callback(1, -1, cb_arg)' is called when a number has passed the
17070 trial division stage.
17074 * Fix for bug in CRL encoding. The validity dates weren't being handled
17079 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17083 * New function BN_pseudo_rand().
17087 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17088 bignum version of BN_from_montgomery() with the working code from
17089 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17094 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17095 made it impossible to use the same SSL_SESSION data structure in
17096 SSL2 clients in multiple threads.
17100 * The return value of RAND_load_file() no longer counts bytes obtained
17101 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17102 to seed the PRNG (previously an explicit byte count was required).
17104 *Ulf Möller, Bodo Möller*
17106 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17107 used `char *` instead of `void *` and had casts all over the place.
17111 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17115 * Retain source code compatibility for BN_prime_checks macro:
17116 BN_is_prime(..., BN_prime_checks, ...) now uses
17117 BN_prime_checks_for_size to determine the appropriate number of
17118 Rabin-Miller iterations.
17122 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17123 DH_CHECK_P_NOT_SAFE_PRIME.
17124 (Check if this is true? OpenPGP calls them "strong".)
17128 * Merge the functionality of "dh" and "gendh" programs into a new program
17129 "dhparam". The old programs are retained for now but will handle DH keys
17130 (instead of parameters) in future.
17134 * Make the ciphers, s_server and s_client programs check the return values
17135 when a new cipher list is set.
17139 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17140 ciphers. Before when the 56bit ciphers were enabled the sorting was
17143 The syntax for the cipher sorting has been extended to support sorting by
17144 cipher-strength (using the strength_bits hard coded in the tables).
17145 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17147 Fix a bug in the cipher-command parser: when supplying a cipher command
17148 string with an "undefined" symbol (neither command nor alphanumeric
17149 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17150 an error is flagged.
17152 Due to the strength-sorting extension, the code of the
17153 ssl_create_cipher_list() function was completely rearranged. I hope that
17154 the readability was also increased :-)
17156 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17158 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17159 for the first serial number and places 2 in the serial number file. This
17160 avoids problems when the root CA is created with serial number zero and
17161 the first user certificate has the same issuer name and serial number
17166 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17167 the new code. Add documentation for this stuff.
17171 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17172 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17173 structures and behave in an analogous way to the X509v3 functions:
17174 they shouldn't be called directly but wrapper functions should be used
17177 So we also now have some wrapper functions that call the X509at functions
17178 when passed certificate requests. (TO DO: similar things can be done with
17179 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17180 things. Some of these need some d2i or i2d and print functionality
17181 because they handle more complex structures.)
17185 * Add missing #ifndefs that caused missing symbols when building libssl
17186 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17187 NO_RSA in `ssl/s2*.c`.
17189 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17191 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17192 has a return value which indicates the quality of the random data
17193 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17194 error queue. New function RAND_pseudo_bytes() generates output that is
17195 guaranteed to be unique but not unpredictable. RAND_add is like
17196 RAND_seed, but takes an extra argument for an entropy estimate
17197 (RAND_seed always assumes full entropy).
17201 * Do more iterations of Rabin-Miller probable prime test (specifically,
17202 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17203 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17204 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17205 false-positive rate of at most 2^-80 for random input.
17209 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17213 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17214 in the 0.9.5 release), this returns the chain
17215 from an X509_CTX structure with a dup of the stack and all
17216 the X509 reference counts upped: so the stack will exist
17217 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17220 Also make SSL_SESSION_print() print out the verify return
17225 * Add manpage for the pkcs12 command. Also change the default
17226 behaviour so MAC iteration counts are used unless the new
17227 -nomaciter option is used. This improves file security and
17228 only older versions of MSIE (4.0 for example) need it.
17232 * Honor the no-xxx Configure options when creating .DEF files.
17236 * Add PKCS#10 attributes to field table: challengePassword,
17237 unstructuredName and unstructuredAddress. These are taken from
17238 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17239 international characters are used.
17241 More changes to X509_ATTRIBUTE code: allow the setting of types
17242 based on strings. Remove the 'loc' parameter when adding
17243 attributes because these will be a SET OF encoding which is sorted
17248 * Initial changes to the 'req' utility to allow request generation
17249 automation. This will allow an application to just generate a template
17250 file containing all the field values and have req construct the
17253 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17254 used all over the place including certificate requests and PKCS#7
17255 structures. They are currently handled manually where necessary with
17256 some primitive wrappers for PKCS#7. The new functions behave in a
17257 manner analogous to the X509 extension functions: they allow
17258 attributes to be looked up by NID and added.
17260 Later something similar to the X509V3 code would be desirable to
17261 automatically handle the encoding, decoding and printing of the
17262 more complex types. The string types like challengePassword can
17263 be handled by the string table functions.
17265 Also modified the multi byte string table handling. Now there is
17266 a 'global mask' which masks out certain types. The table itself
17267 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17268 is useful when for example there is only one permissible type
17269 (as in countryName) and using the mask might result in no valid
17274 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17275 SSL_get_peer_finished to allow applications to obtain the latest
17276 Finished messages sent to the peer or expected from the peer,
17277 respectively. (SSL_get_peer_finished is usually the Finished message
17278 actually received from the peer, otherwise the protocol will be aborted.)
17280 As the Finished message are message digests of the complete handshake
17281 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17282 be used for external authentication procedures when the authentication
17283 provided by SSL/TLS is not desired or is not enough.
17287 * Enhanced support for Alpha Linux is added. Now ./config checks if
17288 the host supports BWX extension and if Compaq C is present on the
17289 $PATH. Just exploiting of the BWX extension results in 20-30%
17290 performance kick for some algorithms, e.g. DES and RC4 to mention
17291 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17296 * Add support for MS "fast SGC". This is arguably a violation of the
17297 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17298 weak crypto and after checking the certificate is SGC a second one
17299 with strong crypto. MS SGC stops the first handshake after receiving
17300 the server certificate message and sends a second client hello. Since
17301 a server will typically do all the time consuming operations before
17302 expecting any further messages from the client (server key exchange
17303 is the most expensive) there is little difference between the two.
17305 To get OpenSSL to support MS SGC we have to permit a second client
17306 hello message after we have sent server done. In addition we have to
17307 reset the MAC if we do get this second client hello.
17311 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17312 if a DER encoded private key is RSA or DSA traditional format. Changed
17313 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17314 format DER encoded private key. Newer code should use PKCS#8 format which
17315 has the key type encoded in the ASN1 structure. Added DER private key
17316 support to pkcs8 application.
17320 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17321 ciphersuites has been selected (as required by the SSL 3/TLS 1
17322 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17323 is set, we interpret this as a request to violate the specification
17324 (the worst that can happen is a handshake failure, and 'correct'
17325 behaviour would result in a handshake failure anyway).
17329 * In SSL_CTX_add_session, take into account that there might be multiple
17330 SSL_SESSION structures with the same session ID (e.g. when two threads
17331 concurrently obtain them from an external cache).
17332 The internal cache can handle only one SSL_SESSION with a given ID,
17333 so if there's a conflict, we now throw out the old one to achieve
17338 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17339 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17340 some routines that use cipher OIDs: some ciphers do not have OIDs
17341 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17346 * Simplify the trust setting structure and code. Now we just have
17347 two sequences of OIDs for trusted and rejected settings. These will
17348 typically have values the same as the extended key usage extension
17349 and any application specific purposes.
17351 The trust checking code now has a default behaviour: it will just
17352 check for an object with the same NID as the passed id. Functions can
17353 be provided to override either the default behaviour or the behaviour
17354 for a given id. SSL client, server and email already have functions
17355 in place for compatibility: they check the NID and also return "trusted"
17356 if the certificate is self signed.
17360 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17361 traditional format into an EVP_PKEY structure.
17365 * Add a password callback function PEM_cb() which either prompts for
17366 a password if usr_data is NULL or otherwise assumes it is a null
17367 terminated password. Allow passwords to be passed on command line
17368 environment or config files in a few more utilities.
17372 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17373 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17374 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17375 Update documentation.
17379 * Support for ASN1 "NULL" type. This could be handled before by using
17380 ASN1_TYPE but there wasn't any function that would try to read a NULL
17381 and produce an error if it couldn't. For compatibility we also have
17382 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17383 don't allocate anything because they don't need to.
17387 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17390 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17392 * Rebuild of the memory allocation routines used by OpenSSL code and
17393 possibly others as well. The purpose is to make an interface that
17394 provide hooks so anyone can build a separate set of allocation and
17395 deallocation routines to be used by OpenSSL, for example memory
17396 pool implementations, or something else, which was previously hard
17397 since Malloc(), Realloc() and Free() were defined as macros having
17398 the values malloc, realloc and free, respectively (except for Win32
17399 compilations). The same is provided for memory debugging code.
17400 OpenSSL already comes with functionality to find memory leaks, but
17401 this gives people a chance to debug other memory problems.
17403 With these changes, a new set of functions and macros have appeared:
17405 CRYPTO_set_mem_debug_functions() [F]
17406 CRYPTO_get_mem_debug_functions() [F]
17407 CRYPTO_dbg_set_options() [F]
17408 CRYPTO_dbg_get_options() [F]
17409 CRYPTO_malloc_debug_init() [M]
17411 The memory debug functions are NULL by default, unless the library
17412 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17413 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17414 gives the standard debugging functions that come with OpenSSL) or
17415 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17416 provided by the library user) must be used. When the standard
17417 debugging functions are used, CRYPTO_dbg_set_options can be used to
17418 request additional information:
17419 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17420 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17422 Also, things like CRYPTO_set_mem_functions will always give the
17423 expected result (the new set of functions is used for allocation
17424 and deallocation) at all times, regardless of platform and compiler
17427 To finish it up, some functions that were never use in any other
17428 way than through macros have a new API and new semantic:
17430 CRYPTO_dbg_malloc()
17431 CRYPTO_dbg_realloc()
17434 All macros of value have retained their old syntax.
17436 *Richard Levitte and Bodo Moeller*
17438 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17439 ordering of SMIMECapabilities wasn't in "strength order" and there
17440 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17445 * Some ASN1 types with illegal zero length encoding (INTEGER,
17446 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17448 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17450 * Merge in my S/MIME library for OpenSSL. This provides a simple
17451 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17452 functionality to handle multipart/signed properly) and a utility
17453 called 'smime' to call all this stuff. This is based on code I
17454 originally wrote for Celo who have kindly allowed it to be
17455 included in OpenSSL.
17459 * Add variants des_set_key_checked and des_set_key_unchecked of
17460 des_set_key (aka des_key_sched). Global variable des_check_key
17461 decides which of these is called by des_set_key; this way
17462 des_check_key behaves as it always did, but applications and
17463 the library itself, which was buggy for des_check_key == 1,
17464 have a cleaner way to pick the version they need.
17468 * New function PKCS12_newpass() which changes the password of a
17473 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17474 dynamic mix. In both cases the ids can be used as an index into the
17475 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17476 functions so they accept a list of the field values and the
17477 application doesn't need to directly manipulate the X509_TRUST
17482 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17487 * Modify the way the V3 extension code looks up extensions. This now
17488 works in a similar way to the object code: we have some "standard"
17489 extensions in a static table which is searched with OBJ_bsearch()
17490 and the application can add dynamic ones if needed. The file
17491 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17492 updated whenever a new extension is added to the core code and kept
17493 in ext_nid order. There is a simple program 'tabtest.c' which checks
17494 this. New extensions are not added too often so this file can readily
17495 be maintained manually.
17497 There are two big advantages in doing things this way. The extensions
17498 can be looked up immediately and no longer need to be "added" using
17499 X509V3_add_standard_extensions(): this function now does nothing.
17500 Side note: I get *lots* of email saying the extension code doesn't
17501 work because people forget to call this function.
17502 Also no dynamic allocation is done unless new extensions are added:
17503 so if we don't add custom extensions there is no need to call
17504 X509V3_EXT_cleanup().
17508 * Modify enc utility's salting as follows: make salting the default. Add a
17509 magic header, so unsalted files fail gracefully instead of just decrypting
17510 to garbage. This is because not salting is a big security hole, so people
17511 should be discouraged from doing it.
17515 * Fixes and enhancements to the 'x509' utility. It allowed a message
17516 digest to be passed on the command line but it only used this
17517 parameter when signing a certificate. Modified so all relevant
17518 operations are affected by the digest parameter including the
17519 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17520 DSA key was used because it didn't fix the digest.
17524 * Initial certificate chain verify code. Currently tests the untrusted
17525 certificates for consistency with the verify purpose (which is set
17526 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17528 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17529 this is because it will reject chains with invalid extensions whereas
17530 every previous version of OpenSSL and SSLeay made no checks at all.
17532 Trust code: checks the root CA for the relevant trust settings. Trust
17533 settings have an initial value consistent with the verify purpose: e.g.
17534 if the verify purpose is for SSL client use it expects the CA to be
17535 trusted for SSL client use. However the default value can be changed to
17536 permit custom trust settings: one example of this would be to only trust
17537 certificates from a specific "secure" set of CAs.
17539 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17540 which should be used for version portability: especially since the
17541 verify structure is likely to change more often now.
17543 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17544 to set them. If not set then assume SSL clients will verify SSL servers
17547 Two new options to the verify program: -untrusted allows a set of
17548 untrusted certificates to be passed in and -purpose which sets the
17549 intended purpose of the certificate. If a purpose is set then the
17550 new chain verify code is used to check extension consistency.
17554 * Support for the authority information access extension.
17558 * Modify RSA and DSA PEM read routines to transparently handle
17559 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17560 public keys in a format compatible with certificate
17561 SubjectPublicKeyInfo structures. Unfortunately there were already
17562 functions called *_PublicKey_* which used various odd formats so
17563 these are retained for compatibility: however the DSA variants were
17564 never in a public release so they have been deleted. Changed dsa/rsa
17565 utilities to handle the new format: note no releases ever handled public
17566 keys so we should be OK.
17568 The primary motivation for this change is to avoid the same fiasco
17569 that dogs private keys: there are several incompatible private key
17570 formats some of which are standard and some OpenSSL specific and
17571 require various evil hacks to allow partial transparent handling and
17572 even then it doesn't work with DER formats. Given the option anything
17573 other than PKCS#8 should be dumped: but the other formats have to
17574 stay in the name of compatibility.
17576 With public keys and the benefit of hindsight one standard format
17577 is used which works with EVP_PKEY, RSA or DSA structures: though
17578 it clearly returns an error if you try to read the wrong kind of key.
17580 Added a -pubkey option to the 'x509' utility to output the public key.
17581 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17582 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17583 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17584 that do the same as the `EVP_PKEY_assign_*()` except they up the
17585 reference count of the added key (they don't "swallow" the
17590 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17591 CRLs would fail if the file contained no certificates or no CRLs:
17592 added a new function to read in both types and return the number
17593 read: this means that if none are read it will be an error. The
17594 DER versions of the certificate and CRL reader would always fail
17595 because it isn't possible to mix certificates and CRLs in DER format
17596 without choking one or the other routine. Changed this to just read
17597 a certificate: this is the best we can do. Also modified the code
17598 in `apps/verify.c` to take notice of return codes: it was previously
17599 attempting to read in certificates from NULL pointers and ignoring
17600 any errors: this is one reason why the cert and CRL reader seemed
17601 to work. It doesn't check return codes from the default certificate
17602 routines: these may well fail if the certificates aren't installed.
17606 * Code to support otherName option in GeneralName.
17610 * First update to verify code. Change the verify utility
17611 so it warns if it is passed a self signed certificate:
17612 for consistency with the normal behaviour. X509_verify
17613 has been modified to it will now verify a self signed
17614 certificate if *exactly* the same certificate appears
17615 in the store: it was previously impossible to trust a
17616 single self signed certificate. This means that:
17617 openssl verify ss.pem
17618 now gives a warning about a self signed certificate but
17619 openssl verify -CAfile ss.pem ss.pem
17624 * For servers, store verify_result in SSL_SESSION data structure
17625 (and add it to external session representation).
17626 This is needed when client certificate verifications fails,
17627 but an application-provided verification callback (set by
17628 SSL_CTX_set_cert_verify_callback) allows accepting the session
17629 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17630 but returns 1): When the session is reused, we have to set
17631 ssl->verify_result to the appropriate error code to avoid
17634 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17636 * Fix a bug in the new PKCS#7 code: it didn't consider the
17637 case in PKCS7_dataInit() where the signed PKCS7 structure
17638 didn't contain any existing data because it was being created.
17640 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17642 * Add a salt to the key derivation routines in enc.c. This
17643 forms the first 8 bytes of the encrypted file. Also add a
17644 -S option to allow a salt to be input on the command line.
17648 * New function X509_cmp(). Oddly enough there wasn't a function
17649 to compare two certificates. We do this by working out the SHA1
17650 hash and comparing that. X509_cmp() will be needed by the trust
17655 * SSL_get1_session() is like SSL_get_session(), but increments
17656 the reference count in the SSL_SESSION returned.
17658 *Geoff Thorpe <geoff@eu.c2.net>*
17660 * Fix for 'req': it was adding a null to request attributes.
17661 Also change the X509_LOOKUP and X509_INFO code to handle
17662 certificate auxiliary information.
17666 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17671 * Add the possibility to add extra information to the memory leak
17672 detecting output, to form tracebacks, showing from where each
17673 allocation was originated: CRYPTO_push_info("constant string") adds
17674 the string plus current file name and line number to a per-thread
17675 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17676 is like calling CYRPTO_pop_info() until the stack is empty.
17677 Also updated memory leak detection code to be multi-thread-safe.
17681 * Add options -text and -noout to pkcs7 utility and delete the
17682 encryption options which never did anything. Update docs.
17686 * Add options to some of the utilities to allow the pass phrase
17687 to be included on either the command line (not recommended on
17688 OSes like Unix) or read from the environment. Update the
17689 manpages and fix a few bugs.
17693 * Add a few manpages for some of the openssl commands.
17697 * Fix the -revoke option in ca. It was freeing up memory twice,
17698 leaking and not finding already revoked certificates.
17702 * Extensive changes to support certificate auxiliary information.
17703 This involves the use of X509_CERT_AUX structure and X509_AUX
17704 functions. An X509_AUX function such as PEM_read_X509_AUX()
17705 can still read in a certificate file in the usual way but it
17706 will also read in any additional "auxiliary information". By
17707 doing things this way a fair degree of compatibility can be
17708 retained: existing certificates can have this information added
17709 using the new 'x509' options.
17711 Current auxiliary information includes an "alias" and some trust
17712 settings. The trust settings will ultimately be used in enhanced
17713 certificate chain verification routines: currently a certificate
17714 can only be trusted if it is self signed and then it is trusted
17719 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17720 The problem was that one of the replacement routines had not been working
17721 since SSLeay releases. For now the offending routine has been replaced
17722 with non-optimised assembler. Even so, this now gives around 95%
17723 performance improvement for 1024 bit RSA signs.
17727 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17728 handling. Most clients have the effective key size in bits equal to
17729 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17730 A few however don't do this and instead use the size of the decrypted key
17731 to determine the RC2 key length and the AlgorithmIdentifier to determine
17732 the effective key length. In this case the effective key length can still
17733 be 40 bits but the key length can be 168 bits for example. This is fixed
17734 by manually forcing an RC2 key into the EVP_PKEY structure because the
17735 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17736 the key length and effective key length are equal.
17740 * Add a bunch of functions that should simplify the creation of
17741 X509_NAME structures. Now you should be able to do:
17742 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17743 and have it automatically work out the correct field type and fill in
17744 the structures. The more adventurous can try:
17745 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17746 and it will (hopefully) work out the correct multibyte encoding.
17750 * Change the 'req' utility to use the new field handling and multibyte
17751 copy routines. Before the DN field creation was handled in an ad hoc
17752 way in req, ca, and x509 which was rather broken and didn't support
17753 BMPStrings or UTF8Strings. Since some software doesn't implement
17754 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17755 using the dirstring_type option. See the new comment in the default
17756 openssl.cnf for more info.
17760 * Make crypto/rand/md_rand.c more robust:
17761 - Assure unique random numbers after fork().
17762 - Make sure that concurrent threads access the global counter and
17763 md serializably so that we never lose entropy in them
17764 or use exactly the same state in multiple threads.
17765 Access to the large state is not always serializable because
17766 the additional locking could be a performance killer, and
17767 md should be large enough anyway.
17771 * New file `apps/app_rand.c` with commonly needed functionality
17772 for handling the random seed file.
17774 Use the random seed file in some applications that previously did not:
17776 dsaparam -genkey (which also ignored its '-rand' option),
17779 x509 (when signing).
17780 Except on systems with /dev/urandom, it is crucial to have a random
17781 seed file at least for key creation, DSA signing, and for DH exchanges;
17782 for RSA signatures we could do without one.
17784 gendh and gendsa (unlike genrsa) used to read only the first byte
17785 of each file listed in the '-rand' option. The function as previously
17786 found in genrsa is now in app_rand.c and is used by all programs
17787 that support '-rand'.
17791 * In RAND_write_file, use mode 0600 for creating files;
17792 don't just chmod when it may be too late.
17796 * Report an error from X509_STORE_load_locations
17797 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17801 * New function ASN1_mbstring_copy() this copies a string in either
17802 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17803 into an ASN1_STRING type. A mask of permissible types is passed
17804 and it chooses the "minimal" type to use or an error if not type
17809 * Add function equivalents to the various macros in asn1.h. The old
17810 macros are retained with an `M_` prefix. Code inside the library can
17811 use the `M_` macros. External code (including the openssl utility)
17812 should *NOT* in order to be "shared library friendly".
17816 * Add various functions that can check a certificate's extensions
17817 to see if it usable for various purposes such as SSL client,
17818 server or S/MIME and CAs of these types. This is currently
17819 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17820 verification. Also added a -purpose flag to x509 utility to
17821 print out all the purposes.
17825 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17830 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17831 for, obtain and decode and extension and obtain its critical flag.
17832 This allows all the necessary extension code to be handled in a
17833 single function call.
17837 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17838 platforms. See crypto/rc4/rc4_enc.c for further details.
17842 * New -noout option to asn1parse. This causes no output to be produced
17843 its main use is when combined with -strparse and -out to extract data
17844 from a file (which may not be in ASN.1 format).
17848 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17849 when producing the local key id.
17851 *Richard Levitte <levitte@stacken.kth.se>*
17853 * New option -dhparam in s_server. This allows a DH parameter file to be
17854 stated explicitly. If it is not stated then it tries the first server
17855 certificate file. The previous behaviour hard coded the filename
17860 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17861 a public key to be input or output. For example:
17862 openssl rsa -in key.pem -pubout -out pubkey.pem
17863 Also added necessary DSA public key functions to handle this.
17867 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17868 in the message. This was handled by allowing
17869 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17871 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17873 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17874 to the end of the strings whereas this didn't. This would cause problems
17875 if strings read with d2i_ASN1_bytes() were later modified.
17877 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17879 * Fix for base64 decode bug. When a base64 bio reads only one line of
17880 data and it contains EOF it will end up returning an error. This is
17881 caused by input 46 bytes long. The cause is due to the way base64
17882 BIOs find the start of base64 encoded data. They do this by trying a
17883 trial decode on each line until they find one that works. When they
17884 do a flag is set and it starts again knowing it can pass all the
17885 data directly through the decoder. Unfortunately it doesn't reset
17886 the context it uses. This means that if EOF is reached an attempt
17887 is made to pass two EOFs through the context and this causes the
17888 resulting error. This can also cause other problems as well. As is
17889 usual with these problems it takes *ages* to find and the fix is
17890 trivial: move one line.
17892 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17894 * Ugly workaround to get s_client and s_server working under Windows. The
17895 old code wouldn't work because it needed to select() on sockets and the
17896 tty (for keypresses and to see if data could be written). Win32 only
17897 supports select() on sockets so we select() with a 1s timeout on the
17898 sockets and then see if any characters are waiting to be read, if none
17899 are present then we retry, we also assume we can always write data to
17900 the tty. This isn't nice because the code then blocks until we've
17901 received a complete line of data and it is effectively polling the
17902 keyboard at 1s intervals: however it's quite a bit better than not
17903 working at all :-) A dedicated Windows application might handle this
17904 with an event loop for example.
17908 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17909 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17910 will be called when RSA_sign() and RSA_verify() are used. This is useful
17911 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17912 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17913 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17914 This necessitated the support of an extra signature type NID_md5_sha1
17915 for SSL signatures and modifications to the SSL library to use it instead
17916 of calling RSA_public_decrypt() and RSA_private_encrypt().
17920 * Add new -verify -CAfile and -CApath options to the crl program, these
17921 will lookup a CRL issuers certificate and verify the signature in a
17922 similar way to the verify program. Tidy up the crl program so it
17923 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17924 less strict. It will now permit CRL extensions even if it is not
17925 a V2 CRL: this will allow it to tolerate some broken CRLs.
17929 * Initialize all non-automatic variables each time one of the openssl
17930 sub-programs is started (this is necessary as they may be started
17931 multiple times from the "OpenSSL>" prompt).
17933 *Lennart Bang, Bodo Moeller*
17935 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17936 removing all other RSA functionality (this is what NO_RSA does). This
17937 is so (for example) those in the US can disable those operations covered
17938 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17943 * Non-copying interface to BIO pairs.
17944 (still largely untested)
17948 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17949 ASCII string. This was handled independently in various places before.
17953 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17954 UTF8 strings a character at a time.
17958 * Use client_version from client hello to select the protocol
17959 (s23_srvr.c) and for RSA client key exchange verification
17960 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17964 * Add various utility functions to handle SPKACs, these were previously
17965 handled by poking round in the structure internals. Added new function
17966 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17967 print, verify and generate SPKACs. Based on an original idea from
17968 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17972 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17976 * Allow the config file extension section to be overwritten on the
17977 command line. Based on an original idea from Massimiliano Pala
17978 <madwolf@comune.modena.it>. The new option is called -extensions
17979 and can be applied to ca, req and x509. Also -reqexts to override
17980 the request extensions in req and -crlexts to override the crl extensions
17985 * Add new feature to the SPKAC handling in ca. Now you can include
17986 the same field multiple times by preceding it by "XXXX." for example:
17989 this is the same syntax as used in the req config file.
17993 * Allow certificate extensions to be added to certificate requests. These
17994 are specified in a 'req_extensions' option of the req section of the
17995 config file. They can be printed out with the -text option to req but
17996 are otherwise ignored at present.
18000 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18001 data read consists of only the final block it would not decrypted because
18002 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18003 A misplaced 'break' also meant the decrypted final block might not be
18004 copied until the next read.
18008 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18009 a few extra parameters to the DH structure: these will be useful if
18010 for example we want the value of 'q' or implement X9.42 DH.
18014 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18015 provides hooks that allow the default DSA functions or functions on a
18016 "per key" basis to be replaced. This allows hardware acceleration and
18017 hardware key storage to be handled without major modification to the
18018 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18019 associated functions.
18023 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18024 as "read only": it can't be written to and the buffer it points to will
18025 not be freed. Reading from a read only BIO is much more efficient than
18026 a normal memory BIO. This was added because there are several times when
18027 an area of memory needs to be read from a BIO. The previous method was
18028 to create a memory BIO and write the data to it, this results in two
18029 copies of the data and an O(n^2) reading algorithm. There is a new
18030 function BIO_new_mem_buf() which creates a read only memory BIO from
18031 an area of memory. Also modified the PKCS#7 routines to use read only
18036 * Bugfix: ssl23_get_client_hello did not work properly when called in
18037 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18038 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18039 but a retry condition occurred while trying to read the rest.
18043 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18044 NID_pkcs7_encrypted by default: this was wrong since this should almost
18045 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18046 the encrypted data type: this is a more sensible place to put it and it
18047 allows the PKCS#12 code to be tidied up that duplicated this
18052 * Changed obj_dat.pl script so it takes its input and output files on
18053 the command line. This should avoid shell escape redirection problems
18058 * Initial support for certificate extension requests, these are included
18059 in things like Xenroll certificate requests. Included functions to allow
18060 extensions to be obtained and added.
18064 * -crlf option to s_client and s_server for sending newlines as
18065 CRLF (as required by many protocols).
18069 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18071 * Install libRSAglue.a when OpenSSL is built with RSAref.
18073 *Ralf S. Engelschall*
18075 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18077 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18079 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18084 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18085 DH parameters/keys (q is lost during that conversion, but the resulting
18086 DH parameters contain its length).
18088 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18089 much faster than DH_generate_parameters (which creates parameters
18090 where `p = 2*q + 1`), and also the smaller q makes DH computations
18091 much more efficient (160-bit exponentiation instead of 1024-bit
18092 exponentiation); so this provides a convenient way to support DHE
18093 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18094 utter importance to use
18095 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18097 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18098 when such DH parameters are used, because otherwise small subgroup
18099 attacks may become possible!
18103 * Avoid memory leak in i2d_DHparams.
18107 * Allow the -k option to be used more than once in the enc program:
18108 this allows the same encrypted message to be read by multiple recipients.
18112 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18113 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18114 it will always use the numerical form of the OID, even if it has a short
18119 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18120 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18121 otherwise bn_mod_exp was called. In the case of hardware keys for example
18122 no private key components need be present and it might store extra data
18123 in the RSA structure, which cannot be accessed from bn_mod_exp.
18124 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18125 private key operations.
18129 * Added support for SPARC Linux.
18133 * pem_password_cb function type incompatibly changed from
18134 typedef int pem_password_cb(char *buf, int size, int rwflag);
18136 ....(char *buf, int size, int rwflag, void *userdata);
18137 so that applications can pass data to their callbacks:
18138 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18139 additional void * argument, which is just handed through whenever
18140 the password callback is called.
18142 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18144 New function SSL_CTX_set_default_passwd_cb_userdata.
18146 Compatibility note: As many C implementations push function arguments
18147 onto the stack in reverse order, the new library version is likely to
18148 interoperate with programs that have been compiled with the old
18149 pem_password_cb definition (PEM_whatever takes some data that
18150 happens to be on the stack as its last argument, and the callback
18151 just ignores this garbage); but there is no guarantee whatsoever that
18154 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18155 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18156 problems not only on Windows, but also on some Unix platforms.
18157 To avoid problematic command lines, these definitions are now in an
18158 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18159 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18163 * MIPS III/IV assembler module is reimplemented.
18167 * More DES library cleanups: remove references to srand/rand and
18168 delete an unused file.
18172 * Add support for the free Netwide assembler (NASM) under Win32,
18173 since not many people have MASM (ml) and it can be hard to obtain.
18174 This is currently experimental but it seems to work OK and pass all
18175 the tests. Check out INSTALL.W32 for info.
18179 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18180 without temporary keys kept an extra copy of the server key,
18181 and connections with temporary keys did not free everything in case
18186 * New function RSA_check_key and new openssl rsa option -check
18187 for verifying the consistency of RSA keys.
18189 *Ulf Moeller, Bodo Moeller*
18191 * Various changes to make Win32 compile work:
18192 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18193 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18194 comparison" warnings.
18195 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18199 * Add a debugging option to PKCS#5 v2 key generation function: when
18200 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18201 derived keys are printed to stderr.
18205 * Copy the flags in ASN1_STRING_dup().
18207 *Roman E. Pavlov <pre@mo.msk.ru>*
18209 * The x509 application mishandled signing requests containing DSA
18210 keys when the signing key was also DSA and the parameters didn't match.
18212 It was supposed to omit the parameters when they matched the signing key:
18213 the verifying software was then supposed to automatically use the CA's
18214 parameters if they were absent from the end user certificate.
18216 Omitting parameters is no longer recommended. The test was also
18217 the wrong way round! This was probably due to unusual behaviour in
18218 EVP_cmp_parameters() which returns 1 if the parameters match.
18219 This meant that parameters were omitted when they *didn't* match and
18220 the certificate was useless. Certificates signed with 'ca' didn't have
18223 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18225 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18226 The interface is as follows:
18227 Applications can use
18228 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18229 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18230 "off" is now the default.
18231 The library internally uses
18232 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18233 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18234 to disable memory-checking temporarily.
18236 Some inconsistent states that previously were possible (and were
18237 even the default) are now avoided.
18239 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18240 with each memory chunk allocated; this is occasionally more helpful
18241 than just having a counter.
18243 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18245 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18250 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18251 which largely parallels "options", but is for changing API behaviour,
18252 whereas "options" are about protocol behaviour.
18253 Initial "mode" flags are:
18255 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18256 a single record has been written.
18257 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18258 retries use the same buffer location.
18259 (But all of the contents must be
18264 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18267 * Fix problems with no-hmac etc.
18269 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18271 * New functions RSA_get_default_method(), RSA_set_method() and
18272 RSA_get_method(). These allows replacement of RSA_METHODs without having
18273 to mess around with the internals of an RSA structure.
18277 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18278 Also really enable memory leak checks in openssl.c and in some
18281 *Chad C. Mulligan, Bodo Moeller*
18283 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18284 up the length of negative integers. This has now been simplified to just
18285 store the length when it is first determined and use it later, rather
18286 than trying to keep track of where data is copied and updating it to
18288 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18290 * Add a new function PKCS7_signatureVerify. This allows the verification
18291 of a PKCS#7 signature but with the signing certificate passed to the
18292 function itself. This contrasts with PKCS7_dataVerify which assumes the
18293 certificate is present in the PKCS#7 structure. This isn't always the
18294 case: certificates can be omitted from a PKCS#7 structure and be
18295 distributed by "out of band" means (such as a certificate database).
18299 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18300 function prototypes in pem.h, also change util/mkdef.pl to add the
18301 necessary function names.
18305 * mk1mf.pl (used by Windows builds) did not properly read the
18306 options set by Configure in the top level Makefile, and Configure
18307 was not even able to write more than one option correctly.
18308 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18312 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18313 file to be loaded from a BIO or FILE pointer. The BIO version will
18314 for example allow memory BIOs to contain config info.
18318 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18319 Whoever hopes to achieve shared-library compatibility across versions
18320 must use this, not the compile-time macro.
18321 (Exercise 0.9.4: Which is the minimum library version required by
18323 Note: All this applies only to multi-threaded programs, others don't
18328 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18329 through a BIO pair triggered the default case, i.e.
18330 SSLerr(...,SSL_R_UNKNOWN_STATE).
18334 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18335 can use the SSL library even if none of the specific BIOs is
18340 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18341 for the encoded length.
18343 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18345 * Add initial documentation of the X509V3 functions.
18349 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18350 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18351 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18352 secure PKCS#8 private key format with a high iteration count.
18356 * Fix determination of Perl interpreter: A perl or perl5
18357 *directory* in $PATH was also accepted as the interpreter.
18359 *Ralf S. Engelschall*
18361 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18362 wrong with it but it was very old and did things like calling
18363 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18364 unusual formatting.
18368 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18369 to use the new extension code.
18373 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18374 with macros. This should make it easier to change their form, add extra
18375 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18380 * Add to configuration table a new entry that can specify an alternative
18381 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18382 according to Mark Crispin <MRC@Panda.COM>.
18386 * DES CBC did not update the IV. Weird.
18390 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18391 Changing the behaviour of the former might break existing programs --
18392 where IV updating is needed, des_ncbc_encrypt can be used.
18395 * When bntest is run from "make test" it drives bc to check its
18396 calculations, as well as internally checking them. If an internal check
18397 fails, it needs to cause bc to give a non-zero result or make test carries
18398 on without noticing the failure. Fixed.
18402 * DES library cleanups.
18406 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18407 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18408 ciphers. NOTE: although the key derivation function has been verified
18409 against some published test vectors it has not been extensively tested
18410 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18415 * Instead of "mkdir -p", which is not fully portable, use new
18416 Perl script "util/mkdir-p.pl".
18420 * Rewrite the way password based encryption (PBE) is handled. It used to
18421 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18422 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18423 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18424 the 'parameter' field of the AlgorithmIdentifier is passed to the
18425 underlying key generation function so it must do its own ASN1 parsing.
18426 This has also changed the EVP_PBE_CipherInit() function which now has a
18427 'parameter' argument instead of literal salt and iteration count values
18428 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18432 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18433 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18434 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18435 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18436 value was just used as a "magic string" and not used directly its
18437 value doesn't matter.
18441 * Introduce some semblance of const correctness to BN. Shame C doesn't
18446 * "linux-sparc64" configuration (ultrapenguin).
18448 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18449 "linux-sparc" configuration.
18451 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18453 * config now generates no-xxx options for missing ciphers.
18457 * Support the EBCDIC character set (work in progress).
18458 File ebcdic.c not yet included because it has a different license.
18460 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18462 * Support BS2000/OSD-POSIX.
18464 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18466 * Make callbacks for key generation use `void *` instead of `char *`.
18470 * Make S/MIME samples compile (not yet tested).
18474 * Additional typesafe stacks.
18478 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18482 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18484 * New configuration variant "sco5-gcc".
18486 * Updated some demos.
18488 *Sean O Riordain, Wade Scholine*
18490 * Add missing BIO_free at exit of pkcs12 application.
18494 * Fix memory leak in conf.c.
18498 * Updates for Win32 to assembler version of MD5.
18502 * Set #! path to perl in `apps/der_chop` to where we found it
18503 instead of using a fixed path.
18507 * SHA library changes for irix64-mips4-cc.
18511 * Improvements for VMS support.
18515 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18517 * Bignum library bug fix. IRIX 6 passes "make test" now!
18518 This also avoids the problems with SC4.2 and unpatched SC5.
18520 *Andy Polyakov <appro@fy.chalmers.se>*
18522 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18523 These are required because of the typesafe stack would otherwise break
18524 existing code. If old code used a structure member which used to be STACK
18525 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18526 sk_num or sk_value it would produce an error because the num, data members
18527 are not present in STACK_OF. Now it just produces a warning. sk_set
18528 replaces the old method of assigning a value to sk_value
18529 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18530 that does this will no longer work (and should use sk_set instead) but
18531 this could be regarded as a "questionable" behaviour anyway.
18535 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18536 correctly handle encrypted S/MIME data.
18540 * Change type of various DES function arguments from des_cblock
18541 (which means, in function argument declarations, pointer to char)
18542 to des_cblock * (meaning pointer to array with 8 char elements),
18543 which allows the compiler to do more typechecking; it was like
18544 that back in SSLeay, but with lots of ugly casts.
18546 Introduce new type const_des_cblock.
18550 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18551 problems: find RecipientInfo structure that matches recipient certificate
18552 and initialise the ASN1 structures properly based on passed cipher.
18556 * Belatedly make the BN tests actually check the results.
18560 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18561 to and from BNs: it was completely broken. New compilation option
18562 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18563 key elements as negative integers.
18567 * Reorganize and speed up MD5.
18569 *Andy Polyakov <appro@fy.chalmers.se>*
18573 *Richard Levitte <richard@levitte.org>*
18575 * New option -out to asn1parse to allow the parsed structure to be
18576 output to a file. This is most useful when combined with the -strparse
18577 option to examine the output of things like OCTET STRINGS.
18581 * Make SSL library a little more fool-proof by not requiring any longer
18582 that `SSL_set_{accept,connect}_state` be called before
18583 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18584 in many applications because usually everything *appeared* to work as
18585 intended anyway -- now it really works as intended).
18589 * Move openssl.cnf out of lib/.
18593 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18594 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18595 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18597 *Ralf S. Engelschall*
18599 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18600 handle PKCS#7 enveloped data properly.
18602 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18604 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18605 copying pointers. The cert_st handling is changed by this in
18606 various ways (and thus what used to be known as ctx->default_cert
18607 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18608 any longer when s->cert does not give us what we need).
18609 ssl_cert_instantiate becomes obsolete by this change.
18610 As soon as we've got the new code right (possibly it already is?),
18611 we have solved a couple of bugs of the earlier code where s->cert
18612 was used as if it could not have been shared with other SSL structures.
18614 Note that using the SSL API in certain dirty ways now will result
18615 in different behaviour than observed with earlier library versions:
18616 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18617 does not influence s as it used to.
18619 In order to clean up things more thoroughly, inside SSL_SESSION
18620 we don't use CERT any longer, but a new structure SESS_CERT
18621 that holds per-session data (if available); currently, this is
18622 the peer's certificate chain and, for clients, the server's certificate
18623 and temporary key. CERT holds only those values that can have
18624 meaningful defaults in an SSL_CTX.
18628 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18629 from the internal representation. Various PKCS#7 fixes: remove some
18630 evil casts and set the enc_dig_alg field properly based on the signing
18635 * Allow PKCS#12 password to be set from the command line or the
18636 environment. Let 'ca' get its config file name from the environment
18637 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18642 * Allow certificate policies extension to use an IA5STRING for the
18643 organization field. This is contrary to the PKIX definition but
18644 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18649 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18650 without disallowing inline assembler and the like for non-pedantic builds.
18654 * Support Borland C++ builder.
18656 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18662 * SHA-1 cleanups and performance enhancements.
18664 *Andy Polyakov <appro@fy.chalmers.se>*
18666 * Sparc v8plus assembler for the bignum library.
18668 *Andy Polyakov <appro@fy.chalmers.se>*
18670 * Accept any -xxx and +xxx compiler options in Configure.
18674 * Update HPUX configuration.
18678 * Add missing `sk_<type>_unshift()` function to safestack.h
18680 *Ralf S. Engelschall*
18682 * New function SSL_CTX_use_certificate_chain_file that sets the
18683 "extra_cert"s in addition to the certificate. (This makes sense
18684 only for "PEM" format files, as chains as a whole are not
18689 * Support verify_depth from the SSL API.
18690 x509_vfy.c had what can be considered an off-by-one-error:
18691 Its depth (which was not part of the external interface)
18692 was actually counting the number of certificates in a chain;
18693 now it really counts the depth.
18697 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18698 instead of X509err, which often resulted in confusing error
18699 messages since the error codes are not globally unique
18700 (e.g. an alleged error in ssl3_accept when a certificate
18701 didn't match the private key).
18703 * New function SSL_CTX_set_session_id_context that allows to set a default
18704 value (so that you don't need SSL_set_session_id_context for each
18705 connection using the SSL_CTX).
18709 * OAEP decoding bug fix.
18713 * Support INSTALL_PREFIX for package builders, as proposed by
18718 * New Configure options "threads" and "no-threads". For systems
18719 where the proper compiler options are known (currently Solaris
18720 and Linux), "threads" is the default.
18724 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18728 * Install various scripts to $(OPENSSLDIR)/misc, not to
18729 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18730 such as /usr/local/bin.
18734 * "make linux-shared" to build shared libraries.
18736 *Niels Poppe <niels@netbox.org>*
18738 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18742 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18743 extension adding in x509 utility.
18747 * Remove NOPROTO sections and error code comments.
18751 * Partial rewrite of the DEF file generator to now parse the ANSI
18756 * New Configure options --prefix=DIR and --openssldir=DIR.
18760 * Complete rewrite of the error code script(s). It is all now handled
18761 by one script at the top level which handles error code gathering,
18762 header rewriting and C source file generation. It should be much better
18763 than the old method: it now uses a modified version of Ulf's parser to
18764 read the ANSI prototypes in all header files (thus the old K&R definitions
18765 aren't needed for error creation any more) and do a better job of
18766 translating function codes into names. The old 'ASN1 error code embedded
18767 in a comment' is no longer necessary and it doesn't use .err files which
18768 have now been deleted. Also the error code call doesn't have to appear all
18769 on one line (which resulted in some large lines...).
18773 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18777 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18778 0 (which usually indicates a closed connection), but continue reading.
18782 * Fix some race conditions.
18786 * Add support for CRL distribution points extension. Add Certificate
18787 Policies and CRL distribution points documentation.
18791 * Move the autogenerated header file parts to crypto/opensslconf.h.
18795 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18796 8 of keying material. Merlin has also confirmed interop with this fix
18797 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18799 *Merlin Hughes <merlin@baltimore.ie>*
18801 * Fix lots of warnings.
18803 *Richard Levitte <levitte@stacken.kth.se>*
18805 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18806 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18808 *Richard Levitte <levitte@stacken.kth.se>*
18810 * Fix problems with sizeof(long) == 8.
18812 *Andy Polyakov <appro@fy.chalmers.se>*
18814 * Change functions to ANSI C.
18818 * Fix typos in error codes.
18820 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18822 * Remove defunct assembler files from Configure.
18826 * SPARC v8 assembler BIGNUM implementation.
18828 *Andy Polyakov <appro@fy.chalmers.se>*
18830 * Support for Certificate Policies extension: both print and set.
18831 Various additions to support the r2i method this uses.
18835 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18836 return a const string when you are expecting an allocated buffer.
18840 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18841 types DirectoryString and DisplayText.
18845 * Add code to allow r2i extensions to access the configuration database,
18846 add an LHASH database driver and add several ctx helper functions.
18850 * Fix an evil bug in bn_expand2() which caused various BN functions to
18851 fail when they extended the size of a BIGNUM.
18855 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18856 support typesafe stack.
18860 * Fix typo in SSL_[gs]et_options().
18862 *Nils Frostberg <nils@medcom.se>*
18864 * Delete various functions and files that belonged to the (now obsolete)
18865 old X509V3 handling code.
18869 * New Configure option "rsaref".
18873 * Don't auto-generate pem.h.
18877 * Introduce type-safe ASN.1 SETs.
18881 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18883 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18885 * Introduce type-safe STACKs. This will almost certainly break lots of code
18886 that links with OpenSSL (well at least cause lots of warnings), but fear
18887 not: the conversion is trivial, and it eliminates loads of evil casts. A
18888 few STACKed things have been converted already. Feel free to convert more.
18889 In the fullness of time, I'll do away with the STACK type altogether.
18893 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18894 specified in `<certfile>` by updating the entry in the index.txt file.
18895 This way one no longer has to edit the index.txt file manually for
18896 revoking a certificate. The -revoke option does the gory details now.
18898 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18900 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18901 `-text` option at all and this way the `-noout -text` combination was
18902 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18904 *Ralf S. Engelschall*
18906 * Make sure a corresponding plain text error message exists for the
18907 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18908 verify callback function determined that a certificate was revoked.
18910 *Ralf S. Engelschall*
18912 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18913 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18914 all available ciphers including rc5, which was forgotten until now.
18915 In order to let the testing shell script know which algorithms
18916 are available, a new (up to now undocumented) command
18917 `openssl list-cipher-commands` is used.
18921 * Bugfix: s_client occasionally would sleep in select() when
18922 it should have checked SSL_pending() first.
18926 * New functions DSA_do_sign and DSA_do_verify to provide access to
18927 the raw DSA values prior to ASN.1 encoding.
18931 * Tweaks to Configure
18933 *Niels Poppe <niels@netbox.org>*
18935 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18940 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18944 * New config option to avoid instructions that are illegal on the 80386.
18945 The default code is faster, but requires at least a 486.
18949 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18950 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18951 same as SSL2_VERSION anyway.
18955 * New "-showcerts" option for s_client.
18959 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18960 application. Various cleanups and fixes.
18964 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18965 modify error routines to work internally. Add error codes and PBE init
18966 to library startup routines.
18970 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18971 packing functions to asn1 and evp. Changed function names and error
18972 codes along the way.
18976 * PKCS12 integration: and so it begins... First of several patches to
18977 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18978 objects to objects.h
18982 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18983 and display support for Thawte strong extranet extension.
18987 * Add LinuxPPC support.
18989 *Jeff Dubrule <igor@pobox.org>*
18991 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18992 bn_div_words in alpha.s.
18994 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18996 * Make sure the RSA OAEP test is skipped under -DRSAref because
18997 OAEP isn't supported when OpenSSL is built with RSAref.
18999 *Ulf Moeller <ulf@fitug.de>*
19001 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19002 so they no longer are missing under -DNOPROTO.
19004 *Soren S. Jorvang <soren@t.dk>*
19006 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19008 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19009 doesn't work when the session is reused. Coming soon!
19013 * Fix a security hole, that allows sessions to be reused in the wrong
19014 context thus bypassing client cert protection! All software that uses
19015 client certs and session caches in multiple contexts NEEDS PATCHING to
19016 allow session reuse! A fuller solution is in the works.
19018 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19020 * Some more source tree cleanups (removed obsolete files
19021 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19022 permission on "config" script to be executable) and a fix for the INSTALL
19025 *Ulf Moeller <ulf@fitug.de>*
19027 * Remove some legacy and erroneous uses of malloc, free instead of
19030 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19032 * Make rsa_oaep_test return non-zero on error.
19034 *Ulf Moeller <ulf@fitug.de>*
19036 * Add support for native Solaris shared libraries. Configure
19037 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19038 if someone would make that last step automatic.
19040 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19042 * ctx_size was not built with the right compiler during "make links". Fixed.
19046 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19047 except NULL ciphers". This means the default cipher list will no longer
19048 enable NULL ciphers. They need to be specifically enabled e.g. with
19049 the string "DEFAULT:eNULL".
19053 * Fix to RSA private encryption routines: if p < q then it would
19054 occasionally produce an invalid result. This will only happen with
19055 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19059 * Be less restrictive and allow also `perl util/perlpath.pl
19060 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19061 because this way one can also use an interpreter named `perl5` (which is
19062 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19063 installed as `perl`).
19065 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19067 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19069 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19071 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19072 advapi32.lib to Win32 build and change the pem test comparison
19073 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19074 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19075 and crypto/des/ede_cbcm_enc.c.
19079 * DES quad checksum was broken on big-endian architectures. Fixed.
19083 * Comment out two functions in bio.h that aren't implemented. Fix up the
19084 Win32 test batch file so it (might) work again. The Win32 test batch file
19085 is horrible: I feel ill....
19089 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19090 in e_os.h. Audit of header files to check ANSI and non ANSI
19091 sections: 10 functions were absent from non ANSI section and not exported
19092 from Windows DLLs. Fixed up libeay.num for new functions.
19096 * Make `openssl version` output lines consistent.
19098 *Ralf S. Engelschall*
19100 * Fix Win32 symbol export lists for BIO functions: Added
19101 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19102 to ms/libeay{16,32}.def.
19104 *Ralf S. Engelschall*
19106 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19107 fine under Unix and passes some trivial tests I've now added. But the
19108 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19109 added to make sure no one expects that this stuff really works in the
19110 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19111 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19114 *Ralf S. Engelschall*
19116 * Fix the generation of two part addresses in perl.
19118 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19120 * Add config entry for Linux on MIPS.
19122 *John Tobey <jtobey@channel1.com>*
19124 * Make links whenever Configure is run, unless we are on Windoze.
19128 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19129 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19134 * Add a useful kludge to allow package maintainers to specify compiler and
19135 other platforms details on the command line without having to patch the
19136 Configure script every time: One now can use
19137 `perl Configure <id>:<details>`,
19138 i.e. platform ids are allowed to have details appended
19139 to them (separated by colons). This is treated as there would be a static
19140 pre-configured entry in Configure's %table under key `<id>` with value
19141 `<details>` and `perl Configure <id>` is called. So, when you want to
19142 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19143 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19144 now, which overrides the FreeBSD-elf entry on-the-fly.
19146 *Ralf S. Engelschall*
19148 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19152 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19153 on the `perl Configure ...` command line. This way one can compile
19154 OpenSSL libraries with Position Independent Code (PIC) which is needed
19155 for linking it into DSOs.
19157 *Ralf S. Engelschall*
19159 * Remarkably, export ciphers were totally broken and no-one had noticed!
19164 * Cleaned up the LICENSE document: The official contact for any license
19165 questions now is the OpenSSL core team under openssl-core@openssl.org.
19166 And add a paragraph about the dual-license situation to make sure people
19167 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19168 to the OpenSSL toolkit.
19170 *Ralf S. Engelschall*
19172 * General source tree makefile cleanups: Made `making xxx in yyy...`
19173 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19174 Additionally cleaned up the `make links` target: Remove unnecessary
19175 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19176 to speed processing and no longer clutter the display with confusing
19177 stuff. Instead only the actually done links are displayed.
19179 *Ralf S. Engelschall*
19181 * Permit null encryption ciphersuites, used for authentication only. It used
19182 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19183 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19188 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19189 signed attributes when verifying signatures (this would break them),
19190 the detached data encoding was wrong and public keys obtained using
19191 X509_get_pubkey() weren't freed.
19195 * Add text documentation for the BUFFER functions. Also added a work around
19196 to a Win95 console bug. This was triggered by the password read stuff: the
19197 last character typed gets carried over to the next fread(). If you were
19198 generating a new cert request using 'req' for example then the last
19199 character of the passphrase would be CR which would then enter the first
19204 * Added the new 'Includes OpenSSL Cryptography Software' button as
19205 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19206 button and can be used by applications based on OpenSSL to show the
19207 relationship to the OpenSSL project.
19209 *Ralf S. Engelschall*
19211 * Remove confusing variables in function signatures in files
19212 ssl/ssl_lib.c and ssl/ssl.h.
19214 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19216 * Don't install bss_file.c under PREFIX/include/
19218 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19220 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19221 functions that return function pointers and has support for NT specific
19222 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19223 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19224 unsigned to signed types: this was killing the Win32 compile.
19228 * Add new certificate file to stack functions,
19229 SSL_add_dir_cert_subjects_to_stack() and
19230 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19231 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19232 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19233 This means that Apache-SSL and similar packages don't have to mess around
19234 to add as many CAs as they want to the preferred list.
19238 * Experiment with doxygen documentation. Currently only partially applied to
19240 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19241 openssl.doxy as the configuration file.
19245 * Get rid of remaining C++-style comments which strict C compilers hate.
19247 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19249 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19250 compiled in by default: it has problems with large keys.
19254 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19255 DH private keys and/or callback functions which directly correspond to
19256 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19257 is needed for applications which have to configure certificates on a
19258 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19260 For the RSA certificate situation is makes no difference, but
19261 for the DSA certificate situation this fixes the "no shared cipher"
19262 problem where the OpenSSL cipher selection procedure failed because the
19263 temporary keys were not overtaken from the context and the API provided
19264 no way to reconfigure them.
19265 The new functions now let applications reconfigure the stuff and they
19266 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19267 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19268 non-public-API function ssl_cert_instantiate() is used as a helper
19269 function and also to reduce code redundancy inside ssl_rsa.c.
19271 *Ralf S. Engelschall*
19273 * Move s_server -dcert and -dkey options out of the undocumented feature
19274 area because they are useful for the DSA situation and should be
19275 recognized by the users.
19277 *Ralf S. Engelschall*
19279 * Fix the cipher decision scheme for export ciphers: the export bits are
19280 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19281 SSL_EXP_MASK. So, the original variable has to be used instead of the
19282 already masked variable.
19284 *Richard Levitte <levitte@stacken.kth.se>*
19286 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19288 *Richard Levitte <levitte@stacken.kth.se>*
19290 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19291 from `int` to `unsigned int` because it is a length and initialized by
19292 EVP_DigestFinal() which expects an `unsigned int *`.
19294 *Richard Levitte <levitte@stacken.kth.se>*
19296 * Don't hard-code path to Perl interpreter on shebang line of Configure
19297 script. Instead use the usual Shell->Perl transition trick.
19299 *Ralf S. Engelschall*
19301 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19302 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19303 -noout -modulus` as it's already the case for `openssl rsa -noout
19304 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19305 currently the public key is printed (a decision which was already done by
19306 `openssl dsa -modulus` in the past) which serves a similar purpose.
19307 Additionally the NO_RSA no longer completely removes the whole -modulus
19308 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19311 *Ralf S. Engelschall*
19313 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19314 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19316 *Arne Ansper <arne@ats.cyber.ee>*
19318 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19319 to be added. Now both 'req' and 'ca' can use new objects defined in the
19324 * Add cool BIO that does syslog (or event log on NT).
19326 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19328 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19329 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19330 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19331 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19335 * Add preliminary config info for new extension code.
19339 * Make RSA_NO_PADDING really use no padding.
19341 *Ulf Moeller <ulf@fitug.de>*
19343 * Generate errors when private/public key check is done.
19347 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19348 for some CRL extensions and new objects added.
19352 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19353 key usage extension and fuller support for authority key id.
19357 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19358 padding method for RSA, which is recommended for new applications in PKCS
19359 #1 v2.0 (RFC 2437, October 1998).
19360 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19361 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19362 against Bleichbacher's attack on RSA.
19363 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19366 * Updates to the new SSL compression code
19368 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19370 * Fix so that the version number in the master secret, when passed
19371 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19372 (because the server will not accept higher), that the version number
19373 is 0x03,0x01, not 0x03,0x00
19375 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19377 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19378 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19379 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19383 * Support for RAW extensions where an arbitrary extension can be
19384 created by including its DER encoding. See `apps/openssl.cnf` for
19389 * Make sure latest Perl versions don't interpret some generated C array
19390 code as Perl array code in the crypto/err/err_genc.pl script.
19392 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19394 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19395 not many people have the assembler. Various Win32 compilation fixes and
19396 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19397 build instructions.
19401 * Modify configure script 'Configure' to automatically create crypto/date.h
19402 file under Win32 and also build pem.h from pem.org. New script
19403 util/mkfiles.pl to create the MINFO file on environments that can't do a
19404 'make files': perl util/mkfiles.pl >MINFO should work.
19408 * Major rework of DES function declarations, in the pursuit of correctness
19409 and purity. As a result, many evil casts evaporated, and some weirdness,
19410 too. You may find this causes warnings in your code. Zapping your evil
19411 casts will probably fix them. Mostly.
19415 * Fix for a typo in asn1.h. Bug fix to object creation script
19416 obj_dat.pl. It considered a zero in an object definition to mean
19417 "end of object": none of the objects in objects.h have any zeros
19418 so it wasn't spotted.
19420 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19422 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19423 Masking (CBCM). In the absence of test vectors, the best I have been able
19424 to do is check that the decrypt undoes the encrypt, so far. Send me test
19425 vectors if you have them.
19429 * Correct calculation of key length for export ciphers (too much space was
19430 allocated for null ciphers). This has not been tested!
19434 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19435 message is now correct (it understands "crypto" and "ssl" on its
19436 command line). There is also now an "update" option. This will update
19437 the util/ssleay.num and util/libeay.num files with any new functions.
19439 perl util/mkdef.pl crypto ssl update
19440 it will update them.
19444 * Overhauled the Perl interface:
19445 - ported BN stuff to OpenSSL's different BN library
19446 - made the perl/ source tree CVS-aware
19447 - renamed the package from SSLeay to OpenSSL (the files still contain
19448 their history because I've copied them in the repository)
19449 - removed obsolete files (the test scripts will be replaced
19450 by better Test::Harness variants in the future)
19452 *Ralf S. Engelschall*
19454 * First cut for a very conservative source tree cleanup:
19455 1. merge various obsolete readme texts into doc/ssleay.txt
19456 where we collect the old documents and readme texts.
19457 2. remove the first part of files where I'm already sure that we no
19458 longer need them because of three reasons: either they are just temporary
19459 files which were left by Eric or they are preserved original files where
19460 I've verified that the diff is also available in the CVS via "cvs diff
19461 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19462 the crypto/md/ stuff).
19464 *Ralf S. Engelschall*
19466 * More extension code. Incomplete support for subject and issuer alt
19467 name, issuer and authority key id. Change the i2v function parameters
19468 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19469 what that's for :-) Fix to ASN1 macro which messed up
19470 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19474 * Preliminary support for ENUMERATED type. This is largely copied from the
19479 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19481 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19483 * Make sure `make rehash` target really finds the `openssl` program.
19485 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19487 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19488 like to hear about it if this slows down other processors.
19492 * Add CygWin32 platform information to Configure script.
19494 *Alan Batie <batie@aahz.jf.intel.com>*
19496 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19498 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19500 * New program nseq to manipulate netscape certificate sequences
19504 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19509 * Fixes to BN code. Previously the default was to define BN_RECURSION
19510 but the BN code had some problems that would cause failures when
19511 doing certificate verification and some other functions.
19513 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19515 * Add ASN1 and PEM code to support netscape certificate sequences.
19519 * Add ASN1 and PEM code to support netscape certificate sequences.
19523 * Add several PKIX and private extended key usage OIDs.
19527 * Modify the 'ca' program to handle the new extension code. Modify
19528 openssl.cnf for new extension format, add comments.
19532 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19533 and add a sample to openssl.cnf so req -x509 now adds appropriate
19538 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19539 error code, add initial support to X509_print() and x509 application.
19543 * Takes a deep breath and start adding X509 V3 extension support code. Add
19544 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19545 stuff is currently isolated and isn't even compiled yet.
19549 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19550 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19551 Removed the versions check from X509 routines when loading extensions:
19552 this allows certain broken certificates that don't set the version
19553 properly to be processed.
19557 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19558 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19559 can still be regenerated with "make depend".
19563 * Spelling mistake in C version of CAST-128.
19565 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19567 * Changes to the error generation code. The perl script err-code.pl
19568 now reads in the old error codes and retains the old numbers, only
19569 adding new ones if necessary. It also only changes the .err files if new
19570 codes are added. The makefiles have been modified to only insert errors
19571 when needed (to avoid needlessly modifying header files). This is done
19572 by only inserting errors if the .err file is newer than the auto generated
19573 C file. To rebuild all the error codes from scratch (the old behaviour)
19574 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19575 or delete all the .err files.
19579 * CAST-128 was incorrectly implemented for short keys. The C version has
19580 been fixed, but is untested. The assembler versions are also fixed, but
19581 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19582 to regenerate it if needed.
19583 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19584 Hagino <itojun@kame.net>*
19586 * File was opened incorrectly in randfile.c.
19588 *Ulf Möller <ulf@fitug.de>*
19590 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19591 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19592 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19593 al: it's just almost always a UTCTime. Note this patch adds new error
19594 codes so do a "make errors" if there are problems.
19598 * Correct Linux 1 recognition in config.
19600 *Ulf Möller <ulf@fitug.de>*
19602 * Remove pointless MD5 hash when using DSA keys in ca.
19604 *Anonymous <nobody@replay.com>*
19606 * Generate an error if given an empty string as a cert directory. Also
19607 generate an error if handed NULL (previously returned 0 to indicate an
19608 error, but didn't set one).
19610 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19612 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19616 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19617 parameters. This was causing a warning which killed off the Win32 compile.
19621 * Remove C++ style comments from crypto/bn/bn_local.h.
19623 *Neil Costigan <neil.costigan@celocom.com>*
19625 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19626 based on a text string, looking up short and long names and finally
19627 "dot" format. The "dot" format stuff didn't work. Added new function
19628 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19629 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19630 OID is not part of the table.
19634 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19635 X509_LOOKUP_by_alias().
19639 * Sort openssl functions by name.
19643 * Get the `gendsa` command working and add it to the `list` command. Remove
19644 encryption from sample DSA keys (in case anyone is interested the password
19649 * Make *all* `*_free` functions accept a NULL pointer.
19651 *Frans Heymans <fheymans@isaserver.be>*
19653 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19656 *Anonymous <nobody@replay.com>*
19658 * s_server should send the CAfile as acceptable CAs, not its own cert.
19660 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19662 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19664 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19666 * Temp key "for export" tests were wrong in s3_srvr.c.
19668 *Anonymous <nobody@replay.com>*
19670 * Add prototype for temp key callback functions
19671 SSL_CTX_set_tmp_{rsa,dh}_callback().
19675 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19676 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19680 * X509_name_add_entry() freed the wrong thing after an error.
19682 *Arne Ansper <arne@ats.cyber.ee>*
19684 * rsa_eay.c would attempt to free a NULL context.
19686 *Arne Ansper <arne@ats.cyber.ee>*
19688 * BIO_s_socket() had a broken should_retry() on Windoze.
19690 *Arne Ansper <arne@ats.cyber.ee>*
19692 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19694 *Arne Ansper <arne@ats.cyber.ee>*
19696 * Make sure the already existing X509_STORE->depth variable is initialized
19697 in X509_STORE_new(), but document the fact that this variable is still
19698 unused in the certificate verification process.
19700 *Ralf S. Engelschall*
19702 * Fix the various library and `apps/` files to free up pkeys obtained from
19703 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19707 * Fix reference counting in X509_PUBKEY_get(). This makes
19708 demos/maurice/example2.c work, amongst others, probably.
19710 *Steve Henson and Ben Laurie*
19712 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19713 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19714 are no longer created. This way we have a single and consistent command
19715 line interface `openssl <command>`, similar to `cvs <command>`.
19717 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19719 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19720 BIT STRING wrapper always have zero unused bits.
19724 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19728 * Make the top-level INSTALL documentation easier to understand.
19732 * Makefiles updated to exit if an error occurs in a sub-directory
19733 make (including if user presses ^C) [Paul Sutton]
19735 * Make Montgomery context stuff explicit in RSA data structure.
19739 * Fix build order of pem and err to allow for generated pem.h.
19743 * Fix renumbering bug in X509_NAME_delete_entry().
19747 * Enhanced the err-ins.pl script so it makes the error library number
19748 global and can add a library name. This is needed for external ASN1 and
19749 other error libraries.
19753 * Fixed sk_insert which never worked properly.
19757 * Fix ASN1 macros so they can handle indefinite length constructed
19758 EXPLICIT tags. Some non standard certificates use these: they can now
19763 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19764 into a single doc/ssleay.txt bundle. This way the information is still
19765 preserved but no longer messes up this directory. Now it's new room for
19766 the new set of documentation files.
19768 *Ralf S. Engelschall*
19770 * SETs were incorrectly DER encoded. This was a major pain, because they
19771 shared code with SEQUENCEs, which aren't coded the same. This means that
19772 almost everything to do with SETs or SEQUENCEs has either changed name or
19773 number of arguments.
19775 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19777 * Fix test data to work with the above.
19781 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19782 was already fixed by Eric for 0.9.1 it seems.
19784 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19786 * Autodetect FreeBSD3.
19790 * Fix various bugs in Configure. This affects the following platforms:
19794 unixware-2.0-pentium
19799 * Eliminate generated files from CVS. Reorder tests to regenerate files
19800 before they are needed.
19804 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19808 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19810 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19811 changed SSLeay to OpenSSL in version strings.
19813 *Ralf S. Engelschall*
19815 * Some fixups to the top-level documents.
19819 * Fixed the nasty bug where rsaref.h was not found under compile-time
19820 because the symlink to include/ was missing.
19822 *Ralf S. Engelschall*
19824 * Incorporated the popular no-RSA/DSA-only patches
19825 which allow to compile an RSA-free SSLeay.
19827 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19829 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19830 when "ssleay" is still not found.
19832 *Ralf S. Engelschall*
19834 * Added more platforms to Configure: Cray T3E, HPUX 11,
19836 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19838 * Updated the README file.
19840 *Ralf S. Engelschall*
19842 * Added various .cvsignore files in the CVS repository subdirs
19843 to make a "cvs update" really silent.
19845 *Ralf S. Engelschall*
19847 * Recompiled the error-definition header files and added
19848 missing symbols to the Win32 linker tables.
19850 *Ralf S. Engelschall*
19852 * Cleaned up the top-level documents;
19853 o new files: CHANGES and LICENSE
19854 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19855 o merged COPYRIGHT into LICENSE
19856 o removed obsolete TODO file
19857 o renamed MICROSOFT to INSTALL.W32
19859 *Ralf S. Engelschall*
19861 * Removed dummy files from the 0.9.1b source tree:
19862 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19863 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19864 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19865 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19866 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19868 *Ralf S. Engelschall*
19870 * Added various platform portability fixes.
19874 * The Genesis of the OpenSSL rpject:
19875 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19876 Young and Tim J. Hudson created while they were working for C2Net until
19879 *The OpenSSL Project*
19881 ### Changes between 0.9.0b and 0.9.1b [not released]
19883 * Updated a few CA certificates under certs/
19887 * Changed some BIGNUM api stuff.
19891 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19892 DGUX x86, Linux Alpha, etc.
19896 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19897 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19902 * Add -strparse option to asn1pars program which parses nested
19905 *Dr Stephen Henson <shenson@bigfoot.com>*
19907 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19911 * DSA fix for "ca" program.
19915 * Added "-genkey" option to "dsaparam" program.
19919 * Added RIPE MD160 (rmd160) message digest.
19923 * Added -a (all) option to "ssleay version" command.
19927 * Added PLATFORM define which is the id given to Configure.
19931 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19935 * Extended the ASN.1 parser routines.
19939 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19943 * Added a BN_CTX to the BN library.
19947 * Fixed the weak key values in DES library
19951 * Changed API in EVP library for cipher aliases.
19955 * Added support for RC2/64bit cipher.
19959 * Converted the lhash library to the crypto/mem.c functions.
19963 * Added more recognized ASN.1 object ids.
19967 * Added more RSA padding checks for SSL/TLS.
19971 * Added BIO proxy/filter functionality.
19975 * Added extra_certs to SSL_CTX which can be used
19976 send extra CA certificates to the client in the CA cert chain sending
19977 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19981 * Now Fortezza is denied in the authentication phase because
19982 this is key exchange mechanism is not supported by SSLeay at all.
19986 * Additional PKCS1 checks.
19990 * Support the string "TLSv1" for all TLS v1 ciphers.
19994 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19995 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19999 * Fixed a few memory leaks.
20003 * Fixed various code and comment typos.
20007 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20008 bytes sent in the client random.
20010 *Edward Bishop <ebishop@spyglass.com>*
20014 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20015 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20016 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20017 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20018 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20019 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20020 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20021 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20022 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20023 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20024 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20025 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20026 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20027 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20028 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20029 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20030 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20031 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20032 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20033 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20034 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20035 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20036 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20037 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20038 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20039 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20040 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20041 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20042 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20043 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20044 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20045 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20046 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20047 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20048 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20049 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20050 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20051 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20052 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20053 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20054 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20055 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20056 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20057 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20058 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20059 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20060 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20061 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20062 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20063 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20064 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20065 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20066 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20067 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20068 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20069 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20070 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20071 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20072 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20073 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20074 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20075 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20076 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20077 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20078 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20079 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20080 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20081 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20082 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20083 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20084 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20085 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20086 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20087 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20088 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20089 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20090 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20091 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20092 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20093 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20094 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20095 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20096 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20097 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20098 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20099 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20100 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20101 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20102 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20103 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20104 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20105 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20106 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20107 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20108 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20109 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20110 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20111 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20112 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20113 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20114 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20115 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20116 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20117 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20118 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20119 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20120 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20121 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20122 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20123 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20124 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20125 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20126 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20127 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20128 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20129 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20130 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20131 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20132 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20133 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20134 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20135 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20136 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20137 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20138 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20139 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20140 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20141 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20142 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20143 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20144 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20145 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20146 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20147 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20148 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20149 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20150 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20151 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20152 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20153 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20154 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20155 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20156 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20157 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20158 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20159 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20160 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20161 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20162 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20163 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20164 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20165 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20166 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20167 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20168 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20169 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20170 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20171 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20172 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20173 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20174 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20175 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20176 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20177 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20178 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20179 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20180 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20181 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20182 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20183 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20184 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20185 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20186 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20187 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20188 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20189 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20190 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20191 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20192 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655