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 support for Hybrid Public Key Encryption (HPKE) as defined
29 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
30 Message Layer Security (MLS) and other IETF specifications.
31 HPKE can also be used by other applications that require
32 encrypting "to" an ECDH public key. External APIs are defined in
33 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
37 * Add support for certificate compression (RFC8879), including
38 library support for Brotli and Zstandard compression.
42 * Add the ability to add custom attributes to PKCS12 files. Add a new API
43 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
44 for a user specified callback and optional argument.
45 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
46 added to the existing STACK_OF attrs.
50 * Major refactor of the libssl record layer
54 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
58 * Add a mac salt length option for the pkcs12 command.
62 * Add more SRTP protection profiles from RFC8723 and RFC8269.
66 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
68 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
70 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
71 supported and enabled.
75 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
76 to the list of ciphersuites providing Perfect Forward Secrecy as
77 required by SECLEVEL >= 3.
79 *Dmitry Belyavskiy, Nicola Tuveri*
81 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
82 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
83 SSL_get0_iana_groups() function-like macro, retrieves the list of
84 supported groups sent by the peer.
85 The function SSL_client_hello_get_extension_order() populates
86 a caller-supplied array with the list of extension types present in the
87 ClientHello, in order of appearance.
91 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
92 to make it possible to use empty passphrase strings.
96 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
97 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
98 of 160 bits and above and less than 224 bits were previously accepted by
99 default but are now no longer allowed. By default TLS compression was
100 already disabled in previous OpenSSL versions. At security level 2 it cannot
105 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
110 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
111 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
112 will need to load the legacy crypto provider.
115 * CCM8 cipher suites in TLS have been downgraded to security level zero
116 because they use a short authentication tag which lowers their strength.
120 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
125 * Add X.509 certificate codeSigning purpose and related checks on key usage and
126 extended key usage of the leaf certificate according to the CA/Browser Forum.
130 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
131 such as adding a trace facility for debugging certificate chain building.
135 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
136 in particular supporting requests for central key generation, generalized
137 polling, and various types of genm/genp exchanges defined in CMP Updates.
141 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
142 like correcting the TLS and proxy support and adding tracing for debugging.
146 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
150 * Fixed and extended `util/check-format.pl` for checking adherence to the
151 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
152 The checks are meanwhile more complete and yield fewer false positives.
156 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
157 sending and receiving multiple messages in a single call. An implementation
158 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
162 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
163 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
164 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
165 paths which are searched for root certificates.
167 The existing `SSL_CERT_DIR` environment variable is deprecated.
168 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
169 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
170 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
171 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
172 for the purposes of determining root certificate stores.
176 * Support for loading root certificates from the Windows certificate store
177 has been added. The support is in the form of a store which recognises the
178 URI string of `org.openssl.winstore://`. This store is enabled by default and
179 can be disabled using the new compile-time option `no-winstore`.
183 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
184 kernel versions that support KTLS have a known bug in CCM processing. That
185 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
186 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
191 * Zerocopy KTLS sendfile() support on Linux.
195 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
196 decryption as a protection against Bleichenbacher-like attacks.
197 The RSA decryption API will now return a randomly generated deterministic
198 message instead of an error in case it detects an error when checking
199 padding during PKCS#1 v1.5 decryption. This is a general protection against
200 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
202 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
203 on the RSA decryption context.
210 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
212 * Added support for KMAC in KBKDF.
216 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
217 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
218 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
219 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
220 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
221 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
222 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
223 for legacy EC and SM2 keys is also changed similarly to honor the
224 equivalent conversion format flag as specified in the underlying
225 `EC_KEY` object being exported to a provider, when this function is
226 called through `EVP_PKEY_export()`.
230 * RNDR and RNDRRS support in provider functions to provide
231 random number generation for Arm CPUs (aarch64).
235 * s_client and s_server apps now explicitly say when the TLS version
236 does not include the renegotiation mechanism. This avoids confusion
237 between that scenario versus when the TLS version includes secure
238 renegotiation but the peer lacks support for it.
242 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
244 *Tomasz Kantecki, Andrey Matyukov*
246 * The various OBJ_* functions have been made thread safe.
250 * Parallel dual-prime 1536/2048-bit modular exponentiation for
251 AVX512_IFMA capable processors.
253 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
255 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
256 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
257 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
258 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
259 `OPENSSL_NO_DEPRECATED_3_1`.
261 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
262 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
263 definitions for these functions regardless of whether
264 `OPENSSL_NO_DEPRECATED_3_1` is defined.
266 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
267 functions regardless of whether they are using them. It is recommended that
268 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
272 * When generating safe-prime DH parameters set the recommended private key
273 length equivalent to minimum key lengths as in RFC 7919.
277 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
278 maximum size that is smaller or equal to the digest length to comply with
279 FIPS 186-4 section 5. This is implemented by a new option
280 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
281 `rsa_pss_saltlen` parameter, which is now the default. Signature
282 verification is not affected by this change and continues to work as before.
289 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
290 listed here are only a brief description.
291 The migration guide contains more detailed information related to new features,
292 breaking changes, and mappings for the large list of deprecated functions.
294 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
296 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
298 * Fixed two buffer overflows in punycode decoding functions.
300 A buffer overrun can be triggered in X.509 certificate verification,
301 specifically in name constraint checking. Note that this occurs after
302 certificate chain signature verification and requires either a CA to
303 have signed the malicious certificate or for the application to continue
304 certificate verification despite failure to construct a path to a trusted
307 In a TLS client, this can be triggered by connecting to a malicious
308 server. In a TLS server, this can be triggered if the server requests
309 client authentication and a malicious client connects.
311 An attacker can craft a malicious email address to overflow
312 an arbitrary number of bytes containing the `.` character (decimal 46)
313 on the stack. This buffer overflow could result in a crash (causing a
317 An attacker can craft a malicious email address to overflow four
318 attacker-controlled bytes on the stack. This buffer overflow could
319 result in a crash (causing a denial of service) or potentially remote code
320 execution depending on stack layout for any given platform/compiler.
325 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
326 parameters in OpenSSL code.
327 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
328 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
329 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
330 Using these invalid names may cause algorithms to use slower methods
331 that ignore the CRT parameters.
335 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
340 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
341 data to be signed before signing the certificate.
345 * Added RIPEMD160 to the default provider.
349 * Ensured that the key share group sent or accepted for the key exchange
350 is allowed for the protocol version.
354 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
356 * OpenSSL supports creating a custom cipher via the legacy
357 EVP_CIPHER_meth_new() function and associated function calls. This function
358 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
359 to use the new provider mechanism in order to implement custom ciphers.
361 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
362 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
363 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
364 and decryption initialisation functions). Instead of using the custom cipher
365 directly it incorrectly tries to fetch an equivalent cipher from the
366 available providers. An equivalent cipher is found based on the NID passed to
367 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
368 given cipher. However it is possible for an application to incorrectly pass
369 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
370 is used in this way the OpenSSL encryption/decryption initialisation function
371 will match the NULL cipher as being equivalent and will fetch this from the
372 available providers. This will succeed if the default provider has been
373 loaded (or if a third party provider has been loaded that offers this
374 cipher). Using the NULL cipher means that the plaintext is emitted as the
377 Applications are only affected by this issue if they call
378 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
379 encryption/decryption initialisation function. Applications that only use
380 SSL/TLS are not impacted by this issue.
385 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
390 * Fixed the linux-mips64 Configure target which was missing the
391 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
396 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
401 * Correctly handle a retransmitted ClientHello in DTLS
405 * Fixed detection of ktls support in cross-compile environment on Linux
409 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
414 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
415 report correct results in some cases
419 * Fix UWP builds by defining VirtualLock
423 * For known safe primes use the minimum key length according to RFC 7919.
424 Longer private key sizes unnecessarily raise the cycles needed to compute the
425 shared secret without any increase of the real security. This fixes a
426 regression from 1.1.1 where these shorter keys were generated for the known
431 * Added the loongarch64 target
435 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
436 only passed to the FIPS provider and not to the default or legacy provider.
440 * Fixed reported performance degradation on aarch64. Restored the
441 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
442 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
443 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
444 The new algorithm is still used for 32 bit targets.
448 * Added a missing header for memcmp that caused compilation failure on some
453 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
455 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
456 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
457 This issue makes the RSA implementation with 2048 bit private keys
458 incorrect on such machines and memory corruption will happen during
459 the computation. As a consequence of the memory corruption an attacker
460 may be able to trigger a remote code execution on the machine performing
463 SSL/TLS servers or other servers using 2048 bit RSA private keys running
464 on machines supporting AVX512IFMA instructions of the X86_64 architecture
465 are affected by this issue.
470 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
471 implementation would not encrypt the entirety of the data under some
472 circumstances. This could reveal sixteen bytes of data that was
473 preexisting in the memory that wasn't written. In the special case of
474 "in place" encryption, sixteen bytes of the plaintext would be revealed.
476 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
477 they are both unaffected.
480 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
482 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
484 * In addition to the c_rehash shell command injection identified in
485 CVE-2022-1292, further bugs where the c_rehash script does not
486 properly sanitise shell metacharacters to prevent command injection have been
489 When the CVE-2022-1292 was fixed it was not discovered that there
490 are other places in the script where the file names of certificates
491 being hashed were possibly passed to a command executed through the shell.
493 This script is distributed by some operating systems in a manner where
494 it is automatically executed. On such operating systems, an attacker
495 could execute arbitrary commands with the privileges of the script.
497 Use of the c_rehash script is considered obsolete and should be replaced
498 by the OpenSSL rehash command line tool.
501 *Daniel Fiala, Tomáš Mráz*
503 * Case insensitive string comparison no longer uses locales. It has instead
504 been directly implemented.
508 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
510 * Case insensitive string comparison is reimplemented via new locale-agnostic
511 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
512 comparison. The previous implementation had problems when the Turkish locale
517 * Fixed a bug in the c_rehash script which was not properly sanitising shell
518 metacharacters to prevent command injection. This script is distributed by
519 some operating systems in a manner where it is automatically executed. On
520 such operating systems, an attacker could execute arbitrary commands with the
521 privileges of the script.
523 Use of the c_rehash script is considered obsolete and should be replaced
524 by the OpenSSL rehash command line tool.
529 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
530 certificate on an OCSP response. The bug caused the function in the case
531 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
532 response (meaning a successful verification) even in the case where the
533 response signing certificate fails to verify.
535 It is anticipated that most users of `OCSP_basic_verify` will not use the
536 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
537 a negative value (indicating a fatal error) in the case of a certificate
538 verification failure. The normal expected return value in this case would be
541 This issue also impacts the command line OpenSSL "ocsp" application. When
542 verifying an ocsp response with the "-no_cert_checks" option the command line
543 application will report that the verification is successful even though it
544 has in fact failed. In this case the incorrect successful response will also
545 be accompanied by error messages showing the failure and contradicting the
546 apparently successful result.
551 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
552 AAD data as the MAC key. This made the MAC key trivially predictable.
554 An attacker could exploit this issue by performing a man-in-the-middle attack
555 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
556 that the modified data would still pass the MAC integrity check.
558 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
559 endpoint will always be rejected by the recipient and the connection will
560 fail at that point. Many application protocols require data to be sent from
561 the client to the server first. Therefore, in such a case, only an OpenSSL
562 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
564 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
565 sent in both directions. In this case both clients and servers could be
566 affected, regardless of the application protocol.
568 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
569 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
570 the handshake when using this ciphersuite.
572 The confidentiality of data is not impacted by this issue, i.e. an attacker
573 cannot decrypt data that has been encrypted using this ciphersuite - they can
576 In order for this attack to work both endpoints must legitimately negotiate
577 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
578 OpenSSL 3.0, and is not available within the default provider or the default
579 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
580 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
581 following must have occurred:
583 1) OpenSSL must have been compiled with the (non-default) compile time option
584 enable-weak-ssl-ciphers
586 2) OpenSSL must have had the legacy provider explicitly loaded (either
587 through application code or via configuration)
589 3) The ciphersuite must have been explicitly added to the ciphersuite list
591 4) The libssl security level must have been set to 0 (default is 1)
593 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
595 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
596 others that both endpoints have in common
601 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
602 occupied by the removed hash table entries.
604 This function is used when decoding certificates or keys. If a long lived
605 process periodically decodes certificates or keys its memory usage will
606 expand without bounds and the process might be terminated by the operating
607 system causing a denial of service. Also traversing the empty hash table
608 entries will take increasingly more time.
610 Typically such long lived processes might be TLS clients or TLS servers
611 configured to accept client certificate authentication.
614 *Hugo Landau, Aliaksei Levin*
616 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
617 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
618 statistics are no longer supported. For compatibility, these statistics are
619 still listed in the output but are now always reported as zero.
623 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
625 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
626 for non-prime moduli.
628 Internally this function is used when parsing certificates that contain
629 elliptic curve public keys in compressed form or explicit elliptic curve
630 parameters with a base point encoded in compressed form.
632 It is possible to trigger the infinite loop by crafting a certificate that
633 has invalid explicit curve parameters.
635 Since certificate parsing happens prior to verification of the certificate
636 signature, any process that parses an externally supplied certificate may thus
637 be subject to a denial of service attack. The infinite loop can also be
638 reached when parsing crafted private keys as they can contain explicit
639 elliptic curve parameters.
641 Thus vulnerable situations include:
643 - TLS clients consuming server certificates
644 - TLS servers consuming client certificates
645 - Hosting providers taking certificates or private keys from customers
646 - Certificate authorities parsing certification requests from subscribers
647 - Anything else which parses ASN.1 elliptic curve parameters
649 Also any other applications that use the BN_mod_sqrt() where the attacker
650 can control the parameter values are vulnerable to this DoS issue.
655 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
656 to the list of ciphersuites providing Perfect Forward Secrecy as
657 required by SECLEVEL >= 3.
659 *Dmitry Belyavskiy, Nicola Tuveri*
661 * Made the AES constant time code for no-asm configurations
662 optional due to the resulting 95% performance degradation.
663 The AES constant time code can be enabled, for no assembly
664 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
668 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
673 * The negative return value handling of the certificate verification callback
674 was reverted. The replacement is to set the verification retry state with
675 the SSL_set_retry_verify() function.
679 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
681 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
682 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
683 verify a certificate supplied by a server. That function may return a
684 negative return value to indicate an internal error (for example out of
685 memory). Such a negative return value is mishandled by OpenSSL and will cause
686 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
687 success and a subsequent call to SSL_get_error() to return the value
688 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
689 returned by OpenSSL if the application has previously called
690 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
691 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
692 totally unexpected and applications may not behave correctly as a result. The
693 exact behaviour will depend on the application but it could result in
694 crashes, infinite loops or other similar incorrect responses.
696 This issue is made more serious in combination with a separate bug in OpenSSL
697 3.0 that will cause X509_verify_cert() to indicate an internal error when
698 processing a certificate chain. This will occur where a certificate does not
699 include the Subject Alternative Name extension but where a Certificate
700 Authority has enforced name constraints. This issue can occur even with valid
706 * Corrected a few file name and file reference bugs in the build,
707 installation and setup scripts, which lead to installation verification
708 failures. Slightly enhanced the installation verification script.
712 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
717 * Fixed PVK encoder to properly query for the passphrase.
721 * Multiple fixes in the OSSL_HTTP API functions.
725 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
726 OSSL_PARAM_INTEGER data type and return error on negative numbers
727 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
728 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
732 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
736 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
740 * Multiple threading fixes.
744 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
748 * Allow fetching an operation from the provider that owns an unexportable key
749 as a fallback if that is still allowed by the property query.
753 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
755 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
760 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
761 S390X capability vector to zero. This simplifies testing of different code
762 paths on S390X architecture.
766 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
767 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
768 SP 800-38D". The communication will fail at this point.
772 * The EC_GROUP_clear_free() function is deprecated as there is nothing
773 confidential in EC_GROUP data.
777 * The byte order mark (BOM) character is ignored if encountered at the
778 beginning of a PEM-formatted file.
782 * Added CMS support for the Russian GOST algorithms.
786 * Due to move of the implementation of cryptographic operations
787 to the providers, validation of various operation parameters can
788 be postponed until the actual operation is executed where previously
789 it happened immediately when an operation parameter was set.
791 For example when setting an unsupported curve with
792 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
793 fail but later keygen operations with the EVP_PKEY_CTX will fail.
795 *OpenSSL team members and many third party contributors*
797 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
798 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
799 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
800 instead to retrieve these algorithms from a provider.
804 * On build targets where the multilib postfix is set in the build
805 configuration the libdir directory was changing based on whether
806 the lib directory with the multilib postfix exists on the system
807 or not. This unpredictable behavior was removed and eventual
808 multilib postfix is now always added to the default libdir. Use
809 `--libdir=lib` to override the libdir if adding the postfix is
814 * The triple DES key wrap functionality now conforms to RFC 3217 but is
815 no longer interoperable with OpenSSL 1.1.1.
819 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
820 function codes, this function can only cause problems for calling
825 * Add a configurable flag to output date formats as ISO 8601. Does not
826 change the default date format.
830 * Version of MSVC earlier than 1300 could get link warnings, which could
831 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
832 Support for this flag has been removed.
836 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
837 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
838 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
839 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
840 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
844 * The signatures of the functions to get and set options on SSL and
845 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
846 Some source code changes may be required.
850 * The public definitions of conf_method_st and conf_st have been
851 deprecated. They will be made opaque in a future release.
853 *Rich Salz and Tomáš Mráz*
855 * Client-initiated renegotiation is disabled by default. To allow it, use
856 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
857 flag, or the "ClientRenegotiation" config parameter as appropriate.
861 * Add "abspath" and "includedir" pragma's to config files, to prevent,
862 or modify relative pathname inclusion.
866 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
867 validated. Please consult the README-FIPS and
868 README-PROVIDERS files, as well as the migration guide.
870 *OpenSSL team members and many third party contributors*
872 * For the key types DH and DHX the allowed settable parameters are now different.
876 * The openssl commands that read keys, certificates, and CRLs now
877 automatically detect the PEM or DER format of the input files.
879 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
881 * Added enhanced PKCS#12 APIs which accept a library context.
885 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
889 * Added support for Kernel TLS (KTLS).
891 *Boris Pismenny, John Baldwin and Andrew Gallatin*
893 * Support for RFC 5746 secure renegotiation is now required by default for
894 SSL or TLS connections to succeed.
898 * The signature of the `copy` functional parameter of the
899 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
900 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
901 the signature of the `pub_decode` functional parameter of the
902 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
903 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
907 * The error return values from some control calls (ctrl) have changed.
911 * A public key check is now performed during EVP_PKEY_derive_set_peer().
915 * Many functions in the EVP_ namespace that are getters of values from
916 implementations or contexts were renamed to include get or get0 in their
917 names. Old names are provided as macro aliases for compatibility and
922 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
923 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
924 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
929 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
932 * The output from the command line applications may have minor
937 * The output from numerous "printing" may have minor changes.
941 * Windows thread synchronization uses read/write primitives (SRWLock) when
942 supported by the OS, otherwise CriticalSection continues to be used.
946 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
947 work on read only BIO source/sinks that do not support these functions.
948 This allows piping or redirection of a file BIO using stdin to be buffered
949 into memory. This is used internally in OSSL_DECODER_from_bio().
953 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
954 this function would return one of the values OSSL_STORE_INFO_NAME,
955 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
956 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
957 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
958 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
959 using this function should be amended to handle the changed return value.
963 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
964 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
965 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
966 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
967 contains more than one certificate identifier: This means that all
968 certificates referenced there MUST be part of the validation chain.
972 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
973 RC5, DESX and DES have been moved to the legacy provider.
977 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
978 RIPEMD-160 have been moved to the legacy provider.
982 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
987 * The deprecated functions EVP_PKEY_get0_RSA(),
988 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
989 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
990 well as the similarly named "get1" functions behave differently in
995 * A number of functions handling low-level keys or engines were deprecated
996 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
997 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
998 EVP_PKEY_get0_siphash().
1002 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1003 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1004 will need to load the legacy crypto provider. This includes these PBE
1005 algorithms which use this KDF:
1006 - NID_pbeWithMD2AndDES_CBC
1007 - NID_pbeWithMD5AndDES_CBC
1008 - NID_pbeWithSHA1AndRC2_CBC
1009 - NID_pbeWithMD2AndRC2_CBC
1010 - NID_pbeWithMD5AndRC2_CBC
1011 - NID_pbeWithSHA1AndDES_CBC
1015 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1016 BIO_debug_callback() functions.
1020 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1021 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1025 * The RAND_METHOD APIs have been deprecated.
1029 * The SRP APIs have been deprecated.
1033 * Add a compile time option to prevent the caching of provider fetched
1034 algorithms. This is enabled by including the no-cached-fetch option
1035 at configuration time.
1039 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1040 count of PKCS12_DEFAULT_ITER.
1042 *Tomáš Mráz and Sahana Prasad*
1044 * The openssl speed command does not use low-level API calls anymore.
1048 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1051 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1053 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1057 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1058 providers may supply their own group implementations (using either the "key
1059 exchange" or the "key encapsulation" methods) which will automatically be
1060 detected and used by libssl.
1062 *Matt Caswell, Nicola Tuveri*
1064 * The undocumented function X509_certificate_type() has been deprecated;
1068 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1072 * Removed RSA padding mode for SSLv23 (which was only used for
1073 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1074 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1079 * Deprecated the obsolete X9.31 RSA key generation related functions.
1081 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1082 is not allowed to return a value > 1, this is no more taken as failure.
1084 *Viktor Dukhovni and David von Oheimb*
1086 * Deprecated the obsolete X9.31 RSA key generation related functions
1087 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1088 BN_X931_generate_prime_ex().
1092 * The default key generation method for the regular 2-prime RSA keys was
1093 changed to the FIPS 186-4 B.3.6 method.
1097 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1101 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1105 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1106 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1108 *Rich Salz, Richard Levitte, and David von Oheimb*
1110 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1114 * Deprecated `OCSP_parse_url()`.
1118 * Validation of SM2 keys has been separated from the validation of regular EC
1123 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1124 switches: a validation failure triggers an early exit, returning a failure
1125 exit status to the parent process.
1129 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1130 to ignore unknown ciphers.
1134 * The `-cipher-commands` and `-digest-commands` options
1135 of the command line utility `list` have been deprecated.
1136 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1140 * Added convenience functions for generating asymmetric key pairs:
1141 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1142 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1146 * All of the low-level EC_KEY functions have been deprecated.
1148 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1150 * Deprecated all the libcrypto and libssl error string loading
1155 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1156 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1161 * The `-crypt` option to the `passwd` command line tool has been removed.
1165 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1170 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1174 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1175 EVP_PKEY_get1_tls_encodedpoint().
1179 * The security callback, which can be customised by application code, supports
1180 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1181 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1185 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1186 interface. Their functionality remains unchanged.
1190 * Added new option for 'openssl list', '-providers', which will display the
1191 list of loaded providers, their names, version and status. It optionally
1192 displays their gettable parameters.
1196 * Removed EVP_PKEY_set_alias_type().
1200 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1201 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1205 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1206 parameters are still checked for type safety at compile time via helper
1211 * Remove the RAND_DRBG API
1213 *Paul Dale and Matthias St. Pierre*
1215 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1216 as well as actual hostnames.
1220 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1221 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1222 conversely, silently ignore DTLS protocol version bounds when configuring
1223 TLS-based contexts. The commands can be repeated to set bounds of both
1224 types. The same applies with the corresponding "min_protocol" and
1225 "max_protocol" command-line switches, in case some application uses both TLS
1228 SSL_CTX instances that are created for a fixed protocol version (e.g.
1229 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1230 attempts to apply bounds to these protocol versions would result in an
1231 error. Now only the "version-flexible" SSL_CTX instances are subject to
1232 limits in configuration files in command-line options.
1236 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1241 * Reworked the recorded ERR codes to make better space for system errors.
1242 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1243 given code is a system error (true) or an OpenSSL error (false).
1247 * Reworked the test perl framework to better allow parallel testing.
1249 *Nicola Tuveri and David von Oheimb*
1251 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1252 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1256 * 'Configure' has been changed to figure out the configuration target if
1257 none is given on the command line. Consequently, the 'config' script is
1258 now only a mere wrapper. All documentation is changed to only mention
1261 *Rich Salz and Richard Levitte*
1263 * Added a library context `OSSL_LIB_CTX` that applications as well as
1264 other libraries can use to form a separate context within which
1265 libcrypto operations are performed.
1269 * Added various `_ex` functions to the OpenSSL API that support using
1270 a non-default `OSSL_LIB_CTX`.
1274 * Handshake now fails if Extended Master Secret extension is dropped
1279 * Dropped interactive mode from the `openssl` program.
1283 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1285 *David von Oheimb and Shane Lontis*
1287 * Deprecated `EC_METHOD_get_field_type()`.
1291 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1292 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1293 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1297 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1301 * Add CAdES-BES signature verification support, mostly derived
1302 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1304 *Filipe Raimundo da Silva*
1306 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1310 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1311 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1315 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1319 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1320 EC_KEY_precompute_mult().
1324 * Deprecated EC_POINTs_mul().
1328 * Removed FIPS_mode() and FIPS_mode_set().
1332 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1336 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1337 EC_POINT_get_Jprojective_coordinates_GFp().
1341 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1342 arrays to be more easily constructed via a series of utility functions.
1343 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1344 the various push functions and finally convert to a passable OSSL_PARAM
1345 array using OSSL_PARAM_BLD_to_param().
1349 * The security strength of SHA1 and MD5 based signatures in TLS has been
1354 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1355 contain a provider side internal key.
1359 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1363 * Project text documents not yet having a proper file name extension
1364 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1365 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1369 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1370 have been converted to Markdown with the goal to produce documents
1371 which not only look pretty when viewed online in the browser, but
1372 remain well readable inside a plain text editor.
1374 To achieve this goal, a 'minimalistic' Markdown style has been applied
1375 which avoids formatting elements that interfere too much with the
1376 reading flow in the text file. For example, it
1378 * avoids [ATX headings][] and uses [setext headings][] instead
1379 (which works for `<h1>` and `<h2>` headings only).
1380 * avoids [inline links][] and uses [reference links][] instead.
1381 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1383 [ATX headings]: https://github.github.com/gfm/#atx-headings
1384 [setext headings]: https://github.github.com/gfm/#setext-headings
1385 [inline links]: https://github.github.com/gfm/#inline-link
1386 [reference links]: https://github.github.com/gfm/#reference-link
1387 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1388 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1390 *Matthias St. Pierre*
1392 * The test suite is changed to preserve results of each test recipe.
1393 A new directory test-runs/ with subdirectories named like the
1394 test recipes are created in the build tree for this purpose.
1398 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1399 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1400 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1402 *David von Oheimb, Martin Peylo*
1404 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1405 It supports arbitrary request and response content types, GET redirection,
1406 TLS, connections via HTTP(S) proxies, connections and exchange via
1407 user-defined BIOs (allowing implicit connections), persistent connections,
1408 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1409 The legacy OCSP-focused (and only partly documented) API
1410 is retained for backward compatibility, while most of it is deprecated.
1414 * Added `util/check-format.pl`, a tool for checking adherence to the
1415 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1416 The checks performed are incomplete and yield some false positives.
1417 Still the tool should be useful for detecting most typical glitches.
1421 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1422 If domain name resolution yields multiple IP addresses all of them are tried
1423 after `connect()` failures.
1427 * All of the low-level RSA functions have been deprecated.
1431 * X509 certificates signed using SHA1 are no longer allowed at security
1436 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1437 modified to use PKEY APIs. These commands are now in maintenance mode
1438 and no new features will be added to them.
1442 * The command line utility rsautl has been deprecated.
1446 * The command line utilities genrsa and rsa have been modified to use PKEY
1447 APIs. They now write PKCS#8 keys by default. These commands are now in
1448 maintenance mode and no new features will be added to them.
1452 * All of the low-level DH functions have been deprecated.
1454 *Paul Dale and Matt Caswell*
1456 * All of the low-level DSA functions have been deprecated.
1460 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1461 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1465 * Deprecated low-level ECDH and ECDSA functions.
1469 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1473 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1474 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1475 a new formulation to include all the things it can be used for,
1476 as well as words of caution.
1480 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1484 * All of the low-level HMAC functions have been deprecated.
1486 *Paul Dale and David von Oheimb*
1488 * Over two thousand fixes were made to the documentation, including:
1489 - Common options (such as -rand/-writerand, TLS version control, etc)
1490 were refactored and point to newly-enhanced descriptions in openssl.pod.
1491 - Added style conformance for all options (with help from Richard Levitte),
1492 documented all reported missing options, added a CI build to check
1493 that all options are documented and that no unimplemented options
1495 - Documented some internals, such as all use of environment variables.
1496 - Addressed all internal broken L<> references.
1500 * All of the low-level CMAC functions have been deprecated.
1504 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1505 functions have been deprecated.
1507 *Paul Dale and David von Oheimb*
1509 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1510 set of functions. The documentation mentioned negative values for some
1511 errors, but this was never the case, so the mention of negative values
1514 Code that followed the documentation and thereby check with something
1515 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1519 * All of the low-level cipher functions have been deprecated.
1521 *Matt Caswell and Paul Dale*
1523 * Removed include/openssl/opensslconf.h.in and replaced it with
1524 include/openssl/configuration.h.in, which differs in not including
1525 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1526 was added to include both.
1528 This allows internal hacks where one might need to modify the set
1529 of configured macros, for example this if deprecated symbols are
1530 still supposed to be available internally:
1532 #include <openssl/configuration.h>
1534 #undef OPENSSL_NO_DEPRECATED
1535 #define OPENSSL_SUPPRESS_DEPRECATED
1537 #include <openssl/macros.h>
1539 This should not be used by applications that use the exported
1540 symbols, as that will lead to linking errors.
1544 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1545 used in exponentiation with 512-bit moduli. No EC algorithms are
1546 affected. Analysis suggests that attacks against 2-prime RSA1024,
1547 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1548 difficult to perform and are not believed likely. Attacks against DH512
1549 are considered just feasible. However, for an attack the target would
1550 have to re-use the DH512 private key, which is not recommended anyway.
1551 Also applications directly using the low-level API BN_mod_exp may be
1552 affected if they use BN_FLG_CONSTTIME.
1557 * Most memory-debug features have been deprecated, and the functionality
1558 replaced with no-ops.
1562 * Added documentation for the STACK API.
1566 * Introduced a new method type and API, OSSL_ENCODER, to represent
1567 generic encoders. These do the same sort of job that PEM writers
1568 and d2i functions do, but with support for methods supplied by
1569 providers, and the possibility for providers to support other
1574 * Introduced a new method type and API, OSSL_DECODER, to represent
1575 generic decoders. These do the same sort of job that PEM readers
1576 and i2d functions do, but with support for methods supplied by
1577 providers, and the possibility for providers to support other
1582 * Added a .pragma directive to the syntax of configuration files, to
1583 allow varying behavior in a supported and predictable manner.
1584 Currently added pragma:
1588 This allows dollar signs to be a keyword character unless it's
1589 followed by a opening brace or parenthesis. This is useful for
1590 platforms where dollar signs are commonly used in names, such as
1591 volume names and system directory names on VMS.
1595 * Added functionality to create an EVP_PKEY from user data.
1599 * Change the interpretation of the '--api' configuration option to
1600 mean that this is a desired API compatibility level with no
1601 further meaning. The previous interpretation, that this would
1602 also mean to remove all deprecated symbols up to and including
1603 the given version, no requires that 'no-deprecated' is also used
1604 in the configuration.
1606 When building applications, the desired API compatibility level
1607 can be set with the OPENSSL_API_COMPAT macro like before. For
1608 API compatibility version below 3.0, the old style numerical
1609 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1610 For version 3.0 and on, the value is expected to be the decimal
1611 value calculated from the major and minor version like this:
1613 MAJOR * 10000 + MINOR * 100
1617 -DOPENSSL_API_COMPAT=30000 For 3.0
1618 -DOPENSSL_API_COMPAT=30200 For 3.2
1620 To hide declarations that are deprecated up to and including the
1621 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1622 given when building the application as well.
1626 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1627 access to certificate and CRL stores via URIs and OSSL_STORE
1630 This adds the following functions:
1632 - X509_LOOKUP_store()
1633 - X509_STORE_load_file()
1634 - X509_STORE_load_path()
1635 - X509_STORE_load_store()
1636 - SSL_add_store_cert_subjects_to_stack()
1637 - SSL_CTX_set_default_verify_store()
1638 - SSL_CTX_load_verify_file()
1639 - SSL_CTX_load_verify_dir()
1640 - SSL_CTX_load_verify_store()
1644 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1645 The presence of this system service is determined at run-time.
1649 * Added functionality to create an EVP_PKEY context based on data
1650 for methods from providers. This takes an algorithm name and a
1651 property query string and simply stores them, with the intent
1652 that any operation that uses this context will use those strings
1653 to fetch the needed methods implicitly, thereby making the port
1654 of application written for pre-3.0 OpenSSL easier.
1658 * The undocumented function NCONF_WIN32() has been deprecated; for
1659 conversion details see the HISTORY section of doc/man5/config.pod
1663 * Introduced the new functions EVP_DigestSignInit_ex() and
1664 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1665 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1666 pages for further details.
1670 * Over two thousand fixes were made to the documentation, including:
1671 adding missing command flags, better style conformance, documentation
1674 *Rich Salz, Richard Levitte*
1676 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1677 X25519, X448, Ed25519 and Ed448.
1681 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1686 * Deprecated the public definition of `ERR_STATE` as well as the function
1687 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1692 * Added ERR functionality to give callers access to the stored function
1693 names that have replaced the older function code based functions.
1695 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1696 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1697 ERR_peek_error_all() and ERR_peek_last_error_all().
1699 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1700 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1701 ERR_func_error_string().
1705 * Extended testing to be verbose for failing tests only. The make variables
1706 VERBOSE_FAILURE or VF can be used to enable this:
1708 $ make VF=1 test # Unix
1709 $ mms /macro=(VF=1) test ! OpenVMS
1710 $ nmake VF=1 test # Windows
1714 * Added the `-copy_extensions` option to the `x509` command for use with
1715 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1716 all extensions in the request are copied to the certificate or vice versa.
1718 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1720 * Added the `-copy_extensions` option to the `req` command for use with
1721 `-x509`. When given with the `copy` or `copyall` argument,
1722 all extensions in the certification request are copied to the certificate.
1726 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1727 they generate are by default RFC 5280 compliant in the following sense:
1728 There is a subjectKeyIdentifier extension with a hash value of the public key
1729 and for not self-signed certs there is an authorityKeyIdentifier extension
1730 with a keyIdentifier field or issuer information identifying the signing key.
1731 This is done unless some configuration overrides the new default behavior,
1732 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1736 * Added several checks to `X509_verify_cert()` according to requirements in
1737 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1738 (which may be done by using the CLI option `-x509_strict`):
1739 * The basicConstraints of CA certificates must be marked critical.
1740 * CA certificates must explicitly include the keyUsage extension.
1741 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1742 * The issuer name of any certificate must not be empty.
1743 * The subject name of CA certs, certs with keyUsage crlSign,
1744 and certs without subjectAlternativeName must not be empty.
1745 * If a subjectAlternativeName extension is given it must not be empty.
1746 * The signatureAlgorithm field and the cert signature must be consistent.
1747 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1748 must not be marked critical.
1749 * The authorityKeyIdentifier must be given for X.509v3 certs
1750 unless they are self-signed.
1751 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1755 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1756 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1760 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1761 used even when parsing explicit parameters, when loading a encoded key
1762 or calling `EC_GROUP_new_from_ecpkparameters()`/
1763 `EC_GROUP_new_from_ecparameters()`.
1764 This prevents bypass of security hardening and performance gains,
1765 especially for curves with specialized EC_METHODs.
1766 By default, if a key encoded with explicit parameters is loaded and later
1767 encoded, the output is still encoded with explicit parameters, even if
1768 internally a "named" EC_GROUP is used for computation.
1772 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1773 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1774 NULL. After this change, only the cofactor parameter can be NULL. It also
1775 does some minimal sanity checks on the passed order.
1780 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1781 An attack is simple, if the first CMS_recipientInfo is valid but the
1782 second CMS_recipientInfo is chosen ciphertext. If the second
1783 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1784 encryption key will be replaced by garbage, and the message cannot be
1785 decoded, but if the RSA decryption fails, the correct encryption key is
1786 used and the recipient will not notice the attack.
1787 As a work around for this potential attack the length of the decrypted
1788 key must be equal to the cipher default key length, in case the
1789 certificate is not given and all recipientInfo are tried out.
1790 The old behaviour can be re-enabled in the CMS code by setting the
1791 CMS_DEBUG_DECRYPT flag.
1795 * Early start up entropy quality from the DEVRANDOM seed source has been
1796 improved for older Linux systems. The RAND subsystem will wait for
1797 /dev/random to be producing output before seeding from /dev/urandom.
1798 The seeded state is stored for future library initialisations using
1799 a system global shared memory segment. The shared memory identifier
1800 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1801 the desired value. The default identifier is 114.
1805 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1806 when primes for RSA keys are computed.
1807 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1808 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1809 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1810 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1811 This avoids possible fingerprinting of newly generated RSA modules.
1815 * Correct the extended master secret constant on EBCDIC systems. Without this
1816 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1817 negotiate EMS will fail. Unfortunately this also means that TLS connections
1818 between EBCDIC systems with this fix, and EBCDIC systems without this
1819 fix will fail if they negotiate EMS.
1823 * Changed the library initialisation so that the config file is now loaded
1824 by default. This was already the case for libssl. It now occurs for both
1825 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1826 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1830 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1831 where the former acts as a replacement for `ERR_put_error()`, and the
1832 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1833 `ERR_raise_data()` adds more flexibility by taking a format string and
1834 an arbitrary number of arguments following it, to be processed with
1839 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1840 to check if a named provider is loaded and available. When called, it
1841 will also activate all fallback providers if such are still present.
1845 * Enforce a minimum DH modulus size of 512 bits.
1849 * Changed DH parameters to generate the order q subgroup instead of 2q.
1850 Previously generated DH parameters are still accepted by DH_check
1851 but DH_generate_key works around that by clearing bit 0 of the
1852 private key for those. This avoids leaking bit 0 of the private key.
1856 * Significantly reduce secure memory usage by the randomness pools.
1860 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1865 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1866 algorithms. An implementation of a key exchange algorithm can be obtained
1867 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1868 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1869 the older EVP_PKEY_derive_init() function. See the man pages for the new
1870 functions for further details.
1874 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1878 * Removed the function names from error messages and deprecated the
1883 * Removed NextStep support and the macro OPENSSL_UNISTD
1887 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1888 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1889 Also removed "export var as function" capability; we do not export
1890 variables, only functions.
1894 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1895 an error and 1 indicating success. In previous versions of OpenSSL this
1896 was a void type. If a key was set longer than the maximum possible this
1901 * Support SM2 signing and verification schemes with X509 certificate.
1905 * Use SHA256 as the default digest for TS query in the `ts` app.
1909 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1913 * Default cipher lists/suites are now available via a function, the
1914 #defines are deprecated.
1918 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1919 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1920 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1924 * Join the directories crypto/x509 and crypto/x509v3
1928 * Added command 'openssl kdf' that uses the EVP_KDF API.
1932 * Added command 'openssl mac' that uses the EVP_MAC API.
1936 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1937 as default directories. Also added the command 'openssl info'
1938 for scripting purposes.
1942 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1947 * Add prediction resistance to the DRBG reseeding process.
1951 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1952 mandated by IEEE Std 1619-2018.
1956 * Added newline escaping functionality to a filename when using openssl dgst.
1957 This output format is to replicate the output format found in the `*sum`
1958 checksum programs. This aims to preserve backward compatibility.
1960 *Matt Eaton, Richard Levitte, and Paul Dale*
1962 * Removed the heartbeat message in DTLS feature, as it has very
1963 little usage and doesn't seem to fulfill a valuable purpose.
1964 The configuration option is now deprecated.
1968 * Changed the output of 'openssl {digestname} < file' to display the
1969 digest name in its output.
1973 * Added a new generic trace API which provides support for enabling
1974 instrumentation through trace output.
1976 *Richard Levitte & Matthias St. Pierre*
1978 * Added build tests for C++. These are generated files that only do one
1979 thing, to include one public OpenSSL head file each. This tests that
1980 the public header files can be usefully included in a C++ application.
1982 This test isn't enabled by default. It can be enabled with the option
1983 'enable-buildtest-c++'.
1987 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1991 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1995 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1999 * Added KMAC to EVP_MAC.
2003 * Added property based algorithm implementation selection framework to
2008 * Added SCA hardening for modular field inversion in EC_GROUP through
2009 a new dedicated field_inv() pointer in EC_METHOD.
2010 This also addresses a leakage affecting conversions from projective
2011 to affine coordinates.
2013 *Billy Bob Brumley, Nicola Tuveri*
2015 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2016 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2017 those algorithms that were already supported through the EVP_PKEY API
2018 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2019 and scrypt are now wrappers that call EVP_KDF.
2023 * Build devcrypto engine as a dynamic engine.
2025 *Eneas U de Queiroz*
2027 * Add keyed BLAKE2 to EVP_MAC.
2031 * Fix a bug in the computation of the endpoint-pair shared secret used
2032 by DTLS over SCTP. This breaks interoperability with older versions
2033 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2034 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2035 interoperability with such broken implementations. However, enabling
2036 this switch breaks interoperability with correct implementations.
2038 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2039 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2043 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2047 * Changed the license to the Apache License v2.0.
2051 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2053 - Major releases (indicated by incrementing the MAJOR release number)
2054 may introduce incompatible API/ABI changes.
2055 - Minor releases (indicated by incrementing the MINOR release number)
2056 may introduce new features but retain API/ABI compatibility.
2057 - Patch releases (indicated by incrementing the PATCH number)
2058 are intended for bug fixes and other improvements of existing
2059 features only (like improving performance or adding documentation)
2060 and retain API/ABI compatibility.
2064 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2068 * Remove the 'dist' target and add a tarball building script. The
2069 'dist' target has fallen out of use, and it shouldn't be
2070 necessary to configure just to create a source distribution.
2074 * Recreate the OS390-Unix config target. It no longer relies on a
2075 special script like it did for OpenSSL pre-1.1.0.
2079 * Instead of having the source directories listed in Configure, add
2080 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2085 * Add GMAC to EVP_MAC.
2089 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2093 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2094 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2095 to facilitate the continued use of MACs through raw private keys in
2096 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2100 * Deprecate ECDH_KDF_X9_62().
2104 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2105 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2106 are retained for backwards compatibility.
2110 * AES-XTS mode now enforces that its two keys are different to mitigate
2111 the attacked described in "Efficient Instantiations of Tweakable
2112 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2113 Details of this attack can be obtained from:
2114 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2118 * Rename the object files, i.e. give them other names than in previous
2119 versions. Their names now include the name of the final product, as
2120 well as its type mnemonic (bin, lib, shlib).
2124 * Added new option for 'openssl list', '-objects', which will display the
2125 list of built in objects, i.e. OIDs with names.
2129 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2130 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2135 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2136 improves application performance by removing data copies and providing
2137 applications with zero-copy system calls such as sendfile and splice.
2141 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2145 * `PKCS12_parse` now maintains the order of the parsed certificates
2146 when outputting them via `*ca` (rather than reversing it).
2150 * Deprecated pthread fork support methods.
2154 * Added support for FFDHE key exchange in TLS 1.3.
2158 * Added a new concept for OpenSSL plugability: providers. This
2159 functionality is designed to replace the ENGINE API and ENGINE
2160 implementations, and to be much more dynamic, allowing provider
2161 authors to introduce new algorithms among other things, as long as
2162 there's an API that supports the algorithm type.
2164 With this concept comes a new core API for interaction between
2165 libcrypto and provider implementations. Public libcrypto functions
2166 that want to use providers do so through this core API.
2168 The main documentation for this core API is found in
2169 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2170 refer to other manuals describing the API specific for supported
2171 algorithm types (also called operations).
2178 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2180 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2182 * Avoid loading of a dynamic engine twice.
2186 * Prioritise DANE TLSA issuer certs over peer certs
2190 * Fixed random API for MacOS prior to 10.12
2192 These MacOS versions don't support the CommonCrypto APIs
2196 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2198 * Fixed an SM2 Decryption Buffer Overflow.
2200 In order to decrypt SM2 encrypted data an application is expected to
2201 call the API function EVP_PKEY_decrypt(). Typically an application will
2202 call this function twice. The first time, on entry, the "out" parameter
2203 can be NULL and, on exit, the "outlen" parameter is populated with the
2204 buffer size required to hold the decrypted plaintext. The application
2205 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2206 again, but this time passing a non-NULL value for the "out" parameter.
2208 A bug in the implementation of the SM2 decryption code means that the
2209 calculation of the buffer size required to hold the plaintext returned
2210 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2211 size required by the second call. This can lead to a buffer overflow
2212 when EVP_PKEY_decrypt() is called by the application a second time with
2213 a buffer that is too small.
2215 A malicious attacker who is able present SM2 content for decryption to
2216 an application could cause attacker chosen data to overflow the buffer
2217 by up to a maximum of 62 bytes altering the contents of other data held
2218 after the buffer, possibly changing application behaviour or causing
2219 the application to crash. The location of the buffer is application
2220 dependent but is typically heap allocated.
2225 * Fixed various read buffer overruns processing ASN.1 strings
2227 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2228 structure which contains a buffer holding the string data and a field
2229 holding the buffer length. This contrasts with normal C strings which
2230 are represented as a buffer for the string data which is terminated
2231 with a NUL (0) byte.
2233 Although not a strict requirement, ASN.1 strings that are parsed using
2234 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2235 well as any string whose value has been set with the ASN1_STRING_set()
2236 function will additionally NUL terminate the byte array in the
2237 ASN1_STRING structure.
2239 However, it is possible for applications to directly construct valid
2240 ASN1_STRING structures which do not NUL terminate the byte array by
2241 directly setting the "data" and "length" fields in the ASN1_STRING
2242 array. This can also happen by using the ASN1_STRING_set0() function.
2244 Numerous OpenSSL functions that print ASN.1 data have been found to
2245 assume that the ASN1_STRING byte array will be NUL terminated, even
2246 though this is not guaranteed for strings that have been directly
2247 constructed. Where an application requests an ASN.1 structure to be
2248 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2249 been directly constructed by the application without NUL terminating
2250 the "data" field, then a read buffer overrun can occur.
2252 The same thing can also occur during name constraints processing
2253 of certificates (for example if a certificate has been directly
2254 constructed by the application instead of loading it via the OpenSSL
2255 parsing functions, and the certificate contains non NUL terminated
2256 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2257 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2259 If a malicious actor can cause an application to directly construct an
2260 ASN1_STRING and then process it through one of the affected OpenSSL
2261 functions then this issue could be hit. This might result in a crash
2262 (causing a Denial of Service attack). It could also result in the
2263 disclosure of private memory contents (such as private keys, or
2264 sensitive plaintext).
2269 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2271 * Fixed a problem with verifying a certificate chain when using the
2272 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2273 the certificates present in a certificate chain. It is not set by default.
2275 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2276 the chain that have explicitly encoded elliptic curve parameters was added
2277 as an additional strict check.
2279 An error in the implementation of this check meant that the result of a
2280 previous check to confirm that certificates in the chain are valid CA
2281 certificates was overwritten. This effectively bypasses the check
2282 that non-CA certificates must not be able to issue other certificates.
2284 If a "purpose" has been configured then there is a subsequent opportunity
2285 for checks that the certificate is a valid CA. All of the named "purpose"
2286 values implemented in libcrypto perform this check. Therefore, where
2287 a purpose is set the certificate chain will still be rejected even when the
2288 strict flag has been used. A purpose is set by default in libssl client and
2289 server certificate verification routines, but it can be overridden or
2290 removed by an application.
2292 In order to be affected, an application must explicitly set the
2293 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2294 for the certificate verification or, in the case of TLS client or server
2295 applications, override the default purpose.
2300 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2301 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2302 renegotiation ClientHello omits the signature_algorithms extension (where it
2303 was present in the initial ClientHello), but includes a
2304 signature_algorithms_cert extension then a NULL pointer dereference will
2305 result, leading to a crash and a denial of service attack.
2307 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2308 (which is the default configuration). OpenSSL TLS clients are not impacted by
2312 *Peter Kästle and Samuel Sapalski*
2314 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2316 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2317 create a unique hash value based on the issuer and serial number data
2318 contained within an X509 certificate. However, it was failing to correctly
2319 handle any errors that may occur while parsing the issuer field (which might
2320 occur if the issuer field is maliciously constructed). This may subsequently
2321 result in a NULL pointer deref and a crash leading to a potential denial of
2327 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2328 padding mode to correctly check for rollback attacks. This is considered a
2329 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2334 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2335 functions. Previously they could overflow the output length argument in some
2336 cases where the input length is close to the maximum permissible length for
2337 an integer on the platform. In such cases the return value from the function
2338 call would be 1 (indicating success), but the output length value would be
2339 negative. This could cause applications to behave incorrectly or crash.
2344 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2345 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2346 could be exploited in a side channel attack to recover the password. Since
2347 the attack is local host only this is outside of the current OpenSSL
2348 threat model and therefore no CVE is assigned.
2350 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2355 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2357 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2358 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2359 If an attacker can control both items being compared then this could lead
2360 to a possible denial of service attack. OpenSSL itself uses the
2361 GENERAL_NAME_cmp function for two purposes:
2362 1) Comparing CRL distribution point names between an available CRL and a
2363 CRL distribution point embedded in an X509 certificate
2364 2) When verifying that a timestamp response token signer matches the
2365 timestamp authority name (exposed via the API functions
2366 TS_RESP_verify_response and TS_RESP_verify_token)
2371 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2373 * Certificates with explicit curve parameters are now disallowed in
2374 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2378 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2379 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2380 conversely, silently ignore DTLS protocol version bounds when configuring
2381 TLS-based contexts. The commands can be repeated to set bounds of both
2382 types. The same applies with the corresponding "min_protocol" and
2383 "max_protocol" command-line switches, in case some application uses both TLS
2386 SSL_CTX instances that are created for a fixed protocol version (e.g.
2387 TLSv1_server_method()) also silently ignore version bounds. Previously
2388 attempts to apply bounds to these protocol versions would result in an
2389 error. Now only the "version-flexible" SSL_CTX instances are subject to
2390 limits in configuration files in command-line options.
2394 * Handshake now fails if Extended Master Secret extension is dropped
2399 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2401 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2403 * Fixed segmentation fault in SSL_check_chain()
2404 Server or client applications that call the SSL_check_chain() function
2405 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2406 dereference as a result of incorrect handling of the
2407 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2408 or unrecognised signature algorithm is received from the peer. This could
2409 be exploited by a malicious peer in a Denial of Service attack.
2414 * Added AES consttime code for no-asm configurations
2415 an optional constant time support for AES was added
2416 when building openssl for no-asm.
2417 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2418 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2419 At this time this feature is by default disabled.
2420 It will be enabled by default in 3.0.
2424 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2426 * Revert the change of EOF detection while reading in libssl to avoid
2427 regressions in applications depending on the current way of reporting
2428 the EOF. As the existing method is not fully accurate the change to
2429 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2430 branch and will be present in the 3.0 release.
2434 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2435 when primes for RSA keys are computed.
2436 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2437 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2438 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2439 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2440 This avoids possible fingerprinting of newly generated RSA modules.
2444 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2446 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2447 while reading in libssl then we would report an error back to the
2448 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2449 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2450 therefore give a hint as to what went wrong.
2454 * Check that ed25519 and ed448 are allowed by the security level. Previously
2455 signature algorithms not using an MD were not being checked that they were
2456 allowed by the security level.
2460 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2461 was not quite right. The behaviour was not consistent between resumption
2462 and normal handshakes, and also not quite consistent with historical
2463 behaviour. The behaviour in various scenarios has been clarified and
2464 it has been updated to make it match historical behaviour as closely as
2469 * *[VMS only]* The header files that the VMS compilers include automatically,
2470 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2471 that the C++ compiler doesn't understand. This is a shortcoming in the
2472 compiler, but can be worked around with `__cplusplus` guards.
2474 C++ applications that use OpenSSL libraries must be compiled using the
2475 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2476 functions. Otherwise, only functions with symbols of less than 31
2477 characters can be used, as the linker will not be able to successfully
2478 resolve symbols with longer names.
2482 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2483 The presence of this system service is determined at run-time.
2487 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2492 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2494 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2495 number generator (RNG). This was intended to include protection in the
2496 event of a fork() system call in order to ensure that the parent and child
2497 processes did not share the same RNG state. However, this protection was not
2498 being used in the default case.
2500 A partial mitigation for this issue is that the output from a high
2501 precision timer is mixed into the RNG state so the likelihood of a parent
2502 and child process sharing state is significantly reduced.
2504 If an application already calls OPENSSL_init_crypto() explicitly using
2505 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2508 *Matthias St. Pierre*
2510 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2511 used even when parsing explicit parameters, when loading a encoded key
2512 or calling `EC_GROUP_new_from_ecpkparameters()`/
2513 `EC_GROUP_new_from_ecparameters()`.
2514 This prevents bypass of security hardening and performance gains,
2515 especially for curves with specialized EC_METHODs.
2516 By default, if a key encoded with explicit parameters is loaded and later
2517 encoded, the output is still encoded with explicit parameters, even if
2518 internally a "named" EC_GROUP is used for computation.
2522 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2523 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2524 NULL. After this change, only the cofactor parameter can be NULL. It also
2525 does some minimal sanity checks on the passed order.
2530 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2531 An attack is simple, if the first CMS_recipientInfo is valid but the
2532 second CMS_recipientInfo is chosen ciphertext. If the second
2533 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2534 encryption key will be replaced by garbage, and the message cannot be
2535 decoded, but if the RSA decryption fails, the correct encryption key is
2536 used and the recipient will not notice the attack.
2537 As a work around for this potential attack the length of the decrypted
2538 key must be equal to the cipher default key length, in case the
2539 certificate is not given and all recipientInfo are tried out.
2540 The old behaviour can be re-enabled in the CMS code by setting the
2541 CMS_DEBUG_DECRYPT flag.
2546 * Early start up entropy quality from the DEVRANDOM seed source has been
2547 improved for older Linux systems. The RAND subsystem will wait for
2548 /dev/random to be producing output before seeding from /dev/urandom.
2549 The seeded state is stored for future library initialisations using
2550 a system global shared memory segment. The shared memory identifier
2551 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2552 the desired value. The default identifier is 114.
2556 * Correct the extended master secret constant on EBCDIC systems. Without this
2557 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2558 negotiate EMS will fail. Unfortunately this also means that TLS connections
2559 between EBCDIC systems with this fix, and EBCDIC systems without this
2560 fix will fail if they negotiate EMS.
2564 * Use Windows installation paths in the mingw builds
2566 Mingw isn't a POSIX environment per se, which means that Windows
2567 paths should be used for installation.
2572 * Changed DH_check to accept parameters with order q and 2q subgroups.
2573 With order 2q subgroups the bit 0 of the private key is not secret
2574 but DH_generate_key works around that by clearing bit 0 of the
2575 private key for those. This avoids leaking bit 0 of the private key.
2579 * Significantly reduce secure memory usage by the randomness pools.
2583 * Revert the DEVRANDOM_WAIT feature for Linux systems
2585 The DEVRANDOM_WAIT feature added a select() call to wait for the
2586 /dev/random device to become readable before reading from the
2587 /dev/urandom device.
2589 It turned out that this change had negative side effects on
2590 performance which were not acceptable. After some discussion it
2591 was decided to revert this feature and leave it up to the OS
2592 resp. the platform maintainer to ensure a proper initialization
2593 during early boot time.
2595 *Matthias St. Pierre*
2597 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2599 * Add build tests for C++. These are generated files that only do one
2600 thing, to include one public OpenSSL head file each. This tests that
2601 the public header files can be usefully included in a C++ application.
2603 This test isn't enabled by default. It can be enabled with the option
2604 'enable-buildtest-c++'.
2608 * Enable SHA3 pre-hashing for ECDSA and DSA.
2612 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2613 This changes the size when using the `genpkey` command when no size is given.
2614 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2615 generation commands to use 2048 bits by default.
2619 * Reorganize the manual pages to consistently have RETURN VALUES,
2620 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2621 util/fix-doc-nits accordingly.
2623 *Paul Yang, Joshua Lock*
2625 * Add the missing accessor EVP_PKEY_get0_engine()
2629 * Have commands like `s_client` and `s_server` output the signature scheme
2630 along with other cipher suite parameters when debugging.
2634 * Make OPENSSL_config() error agnostic again.
2638 * Do the error handling in RSA decryption constant time.
2642 * Prevent over long nonces in ChaCha20-Poly1305.
2644 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2645 for every encryption operation. RFC 7539 specifies that the nonce value
2646 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2647 and front pads the nonce with 0 bytes if it is less than 12
2648 bytes. However it also incorrectly allows a nonce to be set of up to 16
2649 bytes. In this case only the last 12 bytes are significant and any
2650 additional leading bytes are ignored.
2652 It is a requirement of using this cipher that nonce values are
2653 unique. Messages encrypted using a reused nonce value are susceptible to
2654 serious confidentiality and integrity attacks. If an application changes
2655 the default nonce length to be longer than 12 bytes and then makes a
2656 change to the leading bytes of the nonce expecting the new value to be a
2657 new unique nonce then such an application could inadvertently encrypt
2658 messages with a reused nonce.
2660 Additionally the ignored bytes in a long nonce are not covered by the
2661 integrity guarantee of this cipher. Any application that relies on the
2662 integrity of these ignored leading bytes of a long nonce may be further
2663 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2664 is safe because no such use sets such a long nonce value. However user
2665 applications that use this cipher directly and set a non-default nonce
2666 length to be longer than 12 bytes may be vulnerable.
2668 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2674 * Add DEVRANDOM_WAIT feature for Linux systems
2676 On older Linux systems where the getrandom() system call is not available,
2677 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2678 Contrary to getrandom(), the /dev/urandom device will not block during
2679 early boot when the kernel CSPRNG has not been seeded yet.
2681 To mitigate this known weakness, use select() to wait for /dev/random to
2682 become readable before reading from /dev/urandom.
2684 * Ensure that SM2 only uses SM3 as digest algorithm
2688 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2690 * Change the info callback signals for the start and end of a post-handshake
2691 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2692 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2693 confused by this and assume that a TLSv1.2 renegotiation has started. This
2694 can break KeyUpdate handling. Instead we no longer signal the start and end
2695 of a post handshake message exchange (although the messages themselves are
2696 still signalled). This could break some applications that were expecting
2697 the old signals. However without this KeyUpdate is not usable for many
2702 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2704 * Timing vulnerability in DSA signature generation
2706 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2707 timing side channel attack. An attacker could use variations in the signing
2708 algorithm to recover the private key.
2710 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2715 * Timing vulnerability in ECDSA signature generation
2717 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2718 timing side channel attack. An attacker could use variations in the signing
2719 algorithm to recover the private key.
2721 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2726 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2727 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2728 of two gigabytes and the error handling improved.
2730 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2731 categorized as a normal bug, not a security issue, because the DRBG reseeds
2732 automatically and is fully functional even without additional randomness
2733 provided by the application.
2735 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2737 * Add a new ClientHello callback. Provides a callback interface that gives
2738 the application the ability to adjust the nascent SSL object at the
2739 earliest stage of ClientHello processing, immediately after extensions have
2740 been collected but before they have been processed. In particular, this
2741 callback can adjust the supported TLS versions in response to the contents
2746 * Add SM2 base algorithm support.
2750 * s390x assembly pack: add (improved) hardware-support for the following
2751 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2752 aes-cfb/cfb8, aes-ecb.
2756 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2757 parameter is no longer accepted, as it leads to a corrupt table. NULL
2758 pem_str is reserved for alias entries only.
2762 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2763 step for prime curves. The new implementation is based on formulae from
2764 differential addition-and-doubling in homogeneous projective coordinates
2765 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2766 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2767 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2768 to work in projective coordinates.
2770 *Billy Bob Brumley, Nicola Tuveri*
2772 * Change generating and checking of primes so that the error rate of not
2773 being prime depends on the intended use based on the size of the input.
2774 For larger primes this will result in more rounds of Miller-Rabin.
2775 The maximal error rate for primes with more than 1080 bits is lowered
2778 *Kurt Roeckx, Annie Yousar*
2780 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2784 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2785 moving between systems, and to avoid confusion when a Windows build is
2786 done with mingw vs with MSVC. For POSIX installs, there's still a
2787 symlink or copy named 'tsget' to avoid that confusion as well.
2791 * Revert blinding in ECDSA sign and instead make problematic addition
2792 length-invariant. Switch even to fixed-length Montgomery multiplication.
2796 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2797 step for binary curves. The new implementation is based on formulae from
2798 differential addition-and-doubling in mixed Lopez-Dahab projective
2799 coordinates, modified to independently blind the operands.
2801 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2803 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2804 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2805 EC_METHODs to implement their own specialized "ladder step", to take
2806 advantage of more favorable coordinate systems or more efficient
2807 differential addition-and-doubling algorithms.
2809 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2811 * Modified the random device based seed sources to keep the relevant
2812 file descriptors open rather than reopening them on each access.
2813 This allows such sources to operate in a chroot() jail without
2814 the associated device nodes being available. This behaviour can be
2815 controlled using RAND_keep_random_devices_open().
2819 * Numerous side-channel attack mitigations have been applied. This may have
2820 performance impacts for some algorithms for the benefit of improved
2821 security. Specific changes are noted in this change log by their respective
2826 * AIX shared library support overhaul. Switch to AIX "natural" way of
2827 handling shared libraries, which means collecting shared objects of
2828 different versions and bitnesses in one common archive. This allows to
2829 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2830 doesn't affect the way 3rd party applications are linked, only how
2831 multi-version installation is managed.
2835 * Make ec_group_do_inverse_ord() more robust and available to other
2836 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2837 mitigations are applied to the fallback BN_mod_inverse().
2838 When using this function rather than BN_mod_inverse() directly, new
2839 EC cryptosystem implementations are then safer-by-default.
2843 * Add coordinate blinding for EC_POINT and implement projective
2844 coordinate blinding for generic prime curves as a countermeasure to
2845 chosen point SCA attacks.
2847 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2849 * Add blinding to ECDSA and DSA signatures to protect against side channel
2850 attacks discovered by Keegan Ryan (NCC Group).
2854 * Enforce checking in the `pkeyutl` command to ensure that the input
2855 length does not exceed the maximum supported digest length when performing
2856 a sign, verify or verifyrecover operation.
2860 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2861 I/O in combination with something like select() or poll() will hang. This
2862 can be turned off again using SSL_CTX_clear_mode().
2863 Many applications do not properly handle non-application data records, and
2864 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2865 around the problems in those applications, but can also break some.
2866 It's recommended to read the manpages about SSL_read(), SSL_write(),
2867 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2868 SSL_CTX_set_read_ahead() again.
2872 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2873 now allow empty (zero character) pass phrases.
2877 * Apply blinding to binary field modular inversion and remove patent
2878 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2882 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2883 binary and prime elliptic curves.
2887 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2888 constant time fixed point multiplication.
2892 * Revise elliptic curve scalar multiplication with timing attack
2893 defenses: ec_wNAF_mul redirects to a constant time implementation
2894 when computing fixed point and variable point multiplication (which
2895 in OpenSSL are mostly used with secret scalars in keygen, sign,
2896 ECDH derive operations).
2897 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2900 * Updated CONTRIBUTING
2904 * Updated DRBG / RAND to request nonce and additional low entropy
2905 randomness from the system.
2907 *Matthias St. Pierre*
2909 * Updated 'openssl rehash' to use OpenSSL consistent default.
2913 * Moved the load of the ssl_conf module to libcrypto, which helps
2914 loading engines that libssl uses before libssl is initialised.
2918 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2922 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2924 *Ingo Schwarze, Rich Salz*
2926 * Added output of accepting IP address and port for 'openssl s_server'
2930 * Added a new API for TLSv1.3 ciphersuites:
2931 SSL_CTX_set_ciphersuites()
2932 SSL_set_ciphersuites()
2936 * Memory allocation failures consistently add an error to the error
2941 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2942 in libcrypto when run as setuid/setgid.
2946 * Load any config file by default when libssl is used.
2950 * Added new public header file <openssl/rand_drbg.h> and documentation
2951 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2953 *Matthias St. Pierre*
2955 * QNX support removed (cannot find contributors to get their approval
2956 for the license change).
2960 * TLSv1.3 replay protection for early data has been implemented. See the
2961 SSL_read_early_data() man page for further details.
2965 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2966 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2967 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2968 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2969 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2970 configuration has been separated out. See the ciphers man page or the
2971 SSL_CTX_set_ciphersuites() man page for more information.
2975 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2976 in responder mode now supports the new "-multi" option, which
2977 spawns the specified number of child processes to handle OCSP
2978 requests. The "-timeout" option now also limits the OCSP
2979 responder's patience to wait to receive the full client request
2980 on a newly accepted connection. Child processes are respawned
2981 as needed, and the CA index file is automatically reloaded
2982 when changed. This makes it possible to run the "ocsp" responder
2983 as a long-running service, making the OpenSSL CA somewhat more
2984 feature-complete. In this mode, most diagnostic messages logged
2985 after entering the event loop are logged via syslog(3) rather than
2990 * Added support for X448 and Ed448. Heavily based on original work by
2995 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2996 objects loaded. This adds the functions OSSL_STORE_expect() and
2997 OSSL_STORE_find() as well as needed tools to construct searches and
2998 get the search data out of them.
3002 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3003 version of OpenSSL should review their configuration settings to ensure
3004 that they are still appropriate for TLSv1.3. For further information see:
3005 <https://wiki.openssl.org/index.php/TLS1.3>
3009 * Grand redesign of the OpenSSL random generator
3011 The default RAND method now utilizes an AES-CTR DRBG according to
3012 NIST standard SP 800-90Ar1. The new random generator is essentially
3013 a port of the default random generator from the OpenSSL FIPS 2.0
3014 object module. It is a hybrid deterministic random bit generator
3015 using an AES-CTR bit stream and which seeds and reseeds itself
3016 automatically using trusted system entropy sources.
3018 Some of its new features are:
3019 - Support for multiple DRBG instances with seed chaining.
3020 - The default RAND method makes use of a DRBG.
3021 - There is a public and private DRBG instance.
3022 - The DRBG instances are fork-safe.
3023 - Keep all global DRBG instances on the secure heap if it is enabled.
3024 - The public and private DRBG instance are per thread for lock free
3027 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3029 * Changed Configure so it only says what it does and doesn't dump
3030 so much data. Instead, ./configdata.pm should be used as a script
3031 to display all sorts of configuration data.
3035 * Added processing of "make variables" to Configure.
3039 * Added SHA512/224 and SHA512/256 algorithm support.
3043 * The last traces of Netware support, first removed in 1.1.0, have
3048 * Get rid of Makefile.shared, and in the process, make the processing
3049 of certain files (rc.obj, or the .def/.map/.opt files produced from
3050 the ordinal files) more visible and hopefully easier to trace and
3051 debug (or make silent).
3055 * Make it possible to have environment variable assignments as
3056 arguments to config / Configure.
3060 * Add multi-prime RSA (RFC 8017) support.
3064 * Add SM3 implemented according to GB/T 32905-2016
3065 *Jack Lloyd <jack.lloyd@ribose.com>,*
3066 *Ronald Tse <ronald.tse@ribose.com>,*
3067 *Erick Borsboom <erick.borsboom@ribose.com>*
3069 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3070 as documented in RFC6066.
3071 Based on a patch from Tomasz Moń
3073 *Filipe Raimundo da Silva*
3075 * Add SM4 implemented according to GB/T 32907-2016.
3076 *Jack Lloyd <jack.lloyd@ribose.com>,*
3077 *Ronald Tse <ronald.tse@ribose.com>,*
3078 *Erick Borsboom <erick.borsboom@ribose.com>*
3080 * Reimplement -newreq-nodes and ERR_error_string_n; the
3081 original author does not agree with the license change.
3085 * Add ARIA AEAD TLS support.
3089 * Some macro definitions to support VS6 have been removed. Visual
3090 Studio 6 has not worked since 1.1.0
3094 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3095 without clearing the errors.
3099 * Add "atfork" functions. If building on a system that without
3100 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3101 requirements. The RAND facility now uses/requires this.
3109 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3110 not possible to disable entirely. However, it's still possible to
3111 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3114 To disable, configure with 'no-ui-console'. 'no-ui' is still
3115 possible to use as an alias. Check at compile time with the
3116 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3117 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3121 * Add a STORE module, which implements a uniform and URI based reader of
3122 stores that can contain keys, certificates, CRLs and numerous other
3123 objects. The main API is loosely based on a few stdio functions,
3124 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3125 OSSL_STORE_error and OSSL_STORE_close.
3126 The implementation uses backends called "loaders" to implement arbitrary
3127 URI schemes. There is one built in "loader" for the 'file' scheme.
3131 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3132 then adjusted to work on FreeBSD 8.4 as well.
3133 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3134 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3138 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3139 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3140 error code calls like this:
3142 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3144 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3145 that can be encoded in C. For the foreseeable future, this will only
3148 *Richard Levitte and Tim Hudson*
3150 * Removed BSD cryptodev engine.
3154 * Add a build target 'build_all_generated', to build all generated files
3155 and only that. This can be used to prepare everything that requires
3156 things like perl for a system that lacks perl and then move everything
3157 to that system and do the rest of the build there.
3161 * In the UI interface, make it possible to duplicate the user data. This
3162 can be used by engines that need to retain the data for a longer time
3163 than just the call where this user data is passed.
3167 * Ignore the '-named_curve auto' value for compatibility of applications
3170 *Tomáš Mráz <tmraz@fedoraproject.org>*
3172 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3173 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3174 alerts across multiple records (some of which could be empty). In practice
3175 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3176 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3177 support this at all. Supporting it adds significant complexity to the
3178 record layer, and its removal is unlikely to cause interoperability
3183 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3184 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3185 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3190 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3191 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3193 *Richard Levitte, Andy Polyakov*
3195 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3200 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3201 platform rather than 'mingw'.
3205 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3206 success if they are asked to add an object which already exists
3207 in the store. This change cascades to other functions which load
3208 certificates and CRLs.
3212 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3213 facilitate stack unwinding even from assembly subroutines.
3217 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3218 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3222 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3223 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3224 which is the minimum version we support.
3228 * Certificate time validation (X509_cmp_time) enforces stricter
3229 compliance with RFC 5280. Fractional seconds and timezone offsets
3230 are no longer allowed.
3234 * Add support for ARIA
3238 * s_client will now send the Server Name Indication (SNI) extension by
3239 default unless the new "-noservername" option is used. The server name is
3240 based on the host provided to the "-connect" option unless overridden by
3241 using "-servername".
3245 * Add support for SipHash
3249 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3250 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3251 prevent issues where no progress is being made and the peer continually
3252 sends unrecognised record types, using up resources processing them.
3256 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3257 using the algorithm defined in
3258 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3262 * Heartbeat support has been removed; the ABI is changed for now.
3264 *Richard Levitte, Rich Salz*
3266 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3270 * The RSA "null" method, which was partially supported to avoid patent
3271 issues, has been replaced to always returns NULL.
3278 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3280 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3281 used even when parsing explicit parameters, when loading a encoded key
3282 or calling `EC_GROUP_new_from_ecpkparameters()`/
3283 `EC_GROUP_new_from_ecparameters()`.
3284 This prevents bypass of security hardening and performance gains,
3285 especially for curves with specialized EC_METHODs.
3286 By default, if a key encoded with explicit parameters is loaded and later
3287 encoded, the output is still encoded with explicit parameters, even if
3288 internally a "named" EC_GROUP is used for computation.
3292 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3293 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3294 NULL. After this change, only the cofactor parameter can be NULL. It also
3295 does some minimal sanity checks on the passed order.
3300 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3301 An attack is simple, if the first CMS_recipientInfo is valid but the
3302 second CMS_recipientInfo is chosen ciphertext. If the second
3303 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3304 encryption key will be replaced by garbage, and the message cannot be
3305 decoded, but if the RSA decryption fails, the correct encryption key is
3306 used and the recipient will not notice the attack.
3307 As a work around for this potential attack the length of the decrypted
3308 key must be equal to the cipher default key length, in case the
3309 certificate is not given and all recipientInfo are tried out.
3310 The old behaviour can be re-enabled in the CMS code by setting the
3311 CMS_DEBUG_DECRYPT flag.
3316 * Use Windows installation paths in the mingw builds
3318 Mingw isn't a POSIX environment per se, which means that Windows
3319 paths should be used for installation.
3324 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3326 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3327 This changes the size when using the `genpkey` command when no size is given.
3328 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3329 generation commands to use 2048 bits by default.
3333 * Prevent over long nonces in ChaCha20-Poly1305.
3335 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3336 for every encryption operation. RFC 7539 specifies that the nonce value
3337 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3338 and front pads the nonce with 0 bytes if it is less than 12
3339 bytes. However it also incorrectly allows a nonce to be set of up to 16
3340 bytes. In this case only the last 12 bytes are significant and any
3341 additional leading bytes are ignored.
3343 It is a requirement of using this cipher that nonce values are
3344 unique. Messages encrypted using a reused nonce value are susceptible to
3345 serious confidentiality and integrity attacks. If an application changes
3346 the default nonce length to be longer than 12 bytes and then makes a
3347 change to the leading bytes of the nonce expecting the new value to be a
3348 new unique nonce then such an application could inadvertently encrypt
3349 messages with a reused nonce.
3351 Additionally the ignored bytes in a long nonce are not covered by the
3352 integrity guarantee of this cipher. Any application that relies on the
3353 integrity of these ignored leading bytes of a long nonce may be further
3354 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3355 is safe because no such use sets such a long nonce value. However user
3356 applications that use this cipher directly and set a non-default nonce
3357 length to be longer than 12 bytes may be vulnerable.
3359 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3365 * Added SCA hardening for modular field inversion in EC_GROUP through
3366 a new dedicated field_inv() pointer in EC_METHOD.
3367 This also addresses a leakage affecting conversions from projective
3368 to affine coordinates.
3370 *Billy Bob Brumley, Nicola Tuveri*
3372 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3373 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3377 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3381 * Remove the 'dist' target and add a tarball building script. The
3382 'dist' target has fallen out of use, and it shouldn't be
3383 necessary to configure just to create a source distribution.
3387 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3389 * Timing vulnerability in DSA signature generation
3391 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3392 timing side channel attack. An attacker could use variations in the signing
3393 algorithm to recover the private key.
3395 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3400 * Timing vulnerability in ECDSA signature generation
3402 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3403 timing side channel attack. An attacker could use variations in the signing
3404 algorithm to recover the private key.
3406 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3411 * Add coordinate blinding for EC_POINT and implement projective
3412 coordinate blinding for generic prime curves as a countermeasure to
3413 chosen point SCA attacks.
3415 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3417 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3419 * Client DoS due to large DH parameter
3421 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3422 malicious server can send a very large prime value to the client. This will
3423 cause the client to spend an unreasonably long period of time generating a
3424 key for this prime resulting in a hang until the client has finished. This
3425 could be exploited in a Denial Of Service attack.
3427 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3432 * Cache timing vulnerability in RSA Key Generation
3434 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3435 a cache timing side channel attack. An attacker with sufficient access to
3436 mount cache timing attacks during the RSA key generation process could
3437 recover the private key.
3439 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3440 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3445 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3446 parameter is no longer accepted, as it leads to a corrupt table. NULL
3447 pem_str is reserved for alias entries only.
3451 * Revert blinding in ECDSA sign and instead make problematic addition
3452 length-invariant. Switch even to fixed-length Montgomery multiplication.
3456 * Change generating and checking of primes so that the error rate of not
3457 being prime depends on the intended use based on the size of the input.
3458 For larger primes this will result in more rounds of Miller-Rabin.
3459 The maximal error rate for primes with more than 1080 bits is lowered
3462 *Kurt Roeckx, Annie Yousar*
3464 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3468 * Add blinding to ECDSA and DSA signatures to protect against side channel
3469 attacks discovered by Keegan Ryan (NCC Group).
3473 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3474 now allow empty (zero character) pass phrases.
3478 * Certificate time validation (X509_cmp_time) enforces stricter
3479 compliance with RFC 5280. Fractional seconds and timezone offsets
3480 are no longer allowed.
3484 * Fixed a text canonicalisation bug in CMS
3486 Where a CMS detached signature is used with text content the text goes
3487 through a canonicalisation process first prior to signing or verifying a
3488 signature. This process strips trailing space at the end of lines, converts
3489 line terminators to CRLF and removes additional trailing line terminators
3490 at the end of a file. A bug in the canonicalisation process meant that
3491 some characters, such as form-feed, were incorrectly treated as whitespace
3492 and removed. This is contrary to the specification (RFC5485). This fix
3493 could mean that detached text data signed with an earlier version of
3494 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3495 signed with a fixed OpenSSL may fail to verify with an earlier version of
3496 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3497 and use the "-binary" flag (for the "cms" command line application) or set
3498 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3502 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3504 * Constructed ASN.1 types with a recursive definition could exceed the stack
3506 Constructed ASN.1 types with a recursive definition (such as can be found
3507 in PKCS7) could eventually exceed the stack given malicious input with
3508 excessive recursion. This could result in a Denial Of Service attack. There
3509 are no such structures used within SSL/TLS that come from untrusted sources
3510 so this is considered safe.
3512 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3518 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3520 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3521 effectively reduced to only comparing the least significant bit of each
3522 byte. This allows an attacker to forge messages that would be considered as
3523 authenticated in an amount of tries lower than that guaranteed by the
3524 security claims of the scheme. The module can only be compiled by the
3525 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3527 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3533 * Add a build target 'build_all_generated', to build all generated files
3534 and only that. This can be used to prepare everything that requires
3535 things like perl for a system that lacks perl and then move everything
3536 to that system and do the rest of the build there.
3540 * Backport SSL_OP_NO_RENGOTIATION
3542 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3543 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3544 changes this is no longer possible in 1.1.0. Therefore, the new
3545 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3546 1.1.0 to provide equivalent functionality.
3548 Note that if an application built against 1.1.0h headers (or above) is run
3549 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3550 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3554 * Removed the OS390-Unix config target. It relied on a script that doesn't
3559 * rsaz_1024_mul_avx2 overflow bug on x86_64
3561 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3562 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3563 Analysis suggests that attacks against RSA and DSA as a result of this
3564 defect would be very difficult to perform and are not believed likely.
3565 Attacks against DH1024 are considered just feasible, because most of the
3566 work necessary to deduce information about a private key may be performed
3567 offline. The amount of resources required for such an attack would be
3568 significant. However, for an attack on TLS to be meaningful, the server
3569 would have to share the DH1024 private key among multiple clients, which is
3570 no longer an option since CVE-2016-0701.
3572 This only affects processors that support the AVX2 but not ADX extensions
3573 like Intel Haswell (4th generation).
3575 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3576 was originally found via the OSS-Fuzz project.
3581 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3583 * bn_sqrx8x_internal carry bug on x86_64
3585 There is a carry propagating bug in the x86_64 Montgomery squaring
3586 procedure. No EC algorithms are affected. Analysis suggests that attacks
3587 against RSA and DSA as a result of this defect would be very difficult to
3588 perform and are not believed likely. Attacks against DH are considered just
3589 feasible (although very difficult) because most of the work necessary to
3590 deduce information about a private key may be performed offline. The amount
3591 of resources required for such an attack would be very significant and
3592 likely only accessible to a limited number of attackers. An attacker would
3593 additionally need online access to an unpatched system using the target
3594 private key in a scenario with persistent DH parameters and a private
3595 key that is shared between multiple clients.
3597 This only affects processors that support the BMI1, BMI2 and ADX extensions
3598 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3600 This issue was reported to OpenSSL by the OSS-Fuzz project.
3605 * Malformed X.509 IPAddressFamily could cause OOB read
3607 If an X.509 certificate has a malformed IPAddressFamily extension,
3608 OpenSSL could do a one-byte buffer overread. The most likely result
3609 would be an erroneous display of the certificate in text format.
3611 This issue was reported to OpenSSL by the OSS-Fuzz project.
3616 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3618 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3619 platform rather than 'mingw'.
3623 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3624 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3625 which is the minimum version we support.
3629 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3631 * Encrypt-Then-Mac renegotiation crash
3633 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3634 negotiated where it was not in the original handshake (or vice-versa) then
3635 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3636 and servers are affected.
3638 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3643 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3645 * Truncated packet could crash via OOB read
3647 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3648 cipher is being used, then a truncated packet can cause that host to
3649 perform an out-of-bounds read, usually resulting in a crash.
3651 This issue was reported to OpenSSL by Robert Święcki of Google.
3656 * Bad (EC)DHE parameters cause a client crash
3658 If a malicious server supplies bad parameters for a DHE or ECDHE key
3659 exchange then this can result in the client attempting to dereference a
3660 NULL pointer leading to a client crash. This could be exploited in a Denial
3663 This issue was reported to OpenSSL by Guido Vranken.
3668 * BN_mod_exp may produce incorrect results on x86_64
3670 There is a carry propagating bug in the x86_64 Montgomery squaring
3671 procedure. No EC algorithms are affected. Analysis suggests that attacks
3672 against RSA and DSA as a result of this defect would be very difficult to
3673 perform and are not believed likely. Attacks against DH are considered just
3674 feasible (although very difficult) because most of the work necessary to
3675 deduce information about a private key may be performed offline. The amount
3676 of resources required for such an attack would be very significant and
3677 likely only accessible to a limited number of attackers. An attacker would
3678 additionally need online access to an unpatched system using the target
3679 private key in a scenario with persistent DH parameters and a private
3680 key that is shared between multiple clients. For example this can occur by
3681 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3682 similar to CVE-2015-3193 but must be treated as a separate problem.
3684 This issue was reported to OpenSSL by the OSS-Fuzz project.
3689 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3691 * ChaCha20/Poly1305 heap-buffer-overflow
3693 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3694 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3695 crash. This issue is not considered to be exploitable beyond a DoS.
3697 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3702 * CMS Null dereference
3704 Applications parsing invalid CMS structures can crash with a NULL pointer
3705 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3706 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3707 structure callback if an attempt is made to free certain invalid encodings.
3708 Only CHOICE structures using a callback which do not handle NULL value are
3711 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3716 * Montgomery multiplication may produce incorrect results
3718 There is a carry propagating bug in the Broadwell-specific Montgomery
3719 multiplication procedure that handles input lengths divisible by, but
3720 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3721 and DH private keys are impossible. This is because the subroutine in
3722 question is not used in operations with the private key itself and an input
3723 of the attacker's direct choice. Otherwise the bug can manifest itself as
3724 transient authentication and key negotiation failures or reproducible
3725 erroneous outcome of public-key operations with specially crafted input.
3726 Among EC algorithms only Brainpool P-512 curves are affected and one
3727 presumably can attack ECDH key negotiation. Impact was not analyzed in
3728 detail, because pre-requisites for attack are considered unlikely. Namely
3729 multiple clients have to choose the curve in question and the server has to
3730 share the private key among them, neither of which is default behaviour.
3731 Even then only clients that chose the curve will be affected.
3733 This issue was publicly reported as transient failures and was not
3734 initially recognized as a security issue. Thanks to Richard Morgan for
3735 providing reproducible case.
3740 * Removed automatic addition of RPATH in shared libraries and executables,
3741 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3745 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3747 * Fix Use After Free for large message sizes
3749 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3750 message larger than approx 16k is received then the underlying buffer to
3751 store the incoming message is reallocated and moved. Unfortunately a
3752 dangling pointer to the old location is left which results in an attempt to
3753 write to the previously freed location. This is likely to result in a
3754 crash, however it could potentially lead to execution of arbitrary code.
3756 This issue only affects OpenSSL 1.1.0a.
3758 This issue was reported to OpenSSL by Robert Święcki.
3763 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3765 * OCSP Status Request extension unbounded memory growth
3767 A malicious client can send an excessively large OCSP Status Request
3768 extension. If that client continually requests renegotiation, sending a
3769 large OCSP Status Request extension each time, then there will be unbounded
3770 memory growth on the server. This will eventually lead to a Denial Of
3771 Service attack through memory exhaustion. Servers with a default
3772 configuration are vulnerable even if they do not support OCSP. Builds using
3773 the "no-ocsp" build time option are not affected.
3775 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3780 * SSL_peek() hang on empty record
3782 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3783 sends an empty record. This could be exploited by a malicious peer in a
3784 Denial Of Service attack.
3786 This issue was reported to OpenSSL by Alex Gaynor.
3791 * Excessive allocation of memory in tls_get_message_header() and
3792 dtls1_preprocess_fragment()
3794 A (D)TLS message includes 3 bytes for its length in the header for the
3795 message. This would allow for messages up to 16Mb in length. Messages of
3796 this length are excessive and OpenSSL includes a check to ensure that a
3797 peer is sending reasonably sized messages in order to avoid too much memory
3798 being consumed to service a connection. A flaw in the logic of version
3799 1.1.0 means that memory for the message is allocated too early, prior to
3800 the excessive message length check. Due to way memory is allocated in
3801 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3802 to service a connection. This could lead to a Denial of Service through
3803 memory exhaustion. However, the excessive message length check still takes
3804 place, and this would cause the connection to immediately fail. Assuming
3805 that the application calls SSL_free() on the failed connection in a timely
3806 manner then the 21Mb of allocated memory will then be immediately freed
3807 again. Therefore, the excessive memory allocation will be transitory in
3808 nature. This then means that there is only a security impact if:
3810 1) The application does not call SSL_free() in a timely manner in the event
3811 that the connection fails
3813 2) The application is working in a constrained environment where there is
3814 very little free memory
3816 3) The attacker initiates multiple connection attempts such that there are
3817 multiple connections in a state where memory has been allocated for the
3818 connection; SSL_free() has not yet been called; and there is insufficient
3819 memory to service the multiple requests.
3821 Except in the instance of (1) above any Denial Of Service is likely to be
3822 transitory because as soon as the connection fails the memory is
3823 subsequently freed again in the SSL_free() call. However there is an
3824 increased risk during this period of application crashes due to the lack of
3825 memory - which would then mean a more serious Denial of Service.
3827 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3828 (CVE-2016-6307 and CVE-2016-6308)
3832 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3833 had to be removed. Primary reason is that vendor assembler can't
3834 assemble our modules with -KPIC flag. As result it, assembly
3835 support, was not even available as option. But its lack means
3836 lack of side-channel resistant code, which is incompatible with
3837 security by todays standards. Fortunately gcc is readily available
3838 prepackaged option, which we firmly point at...
3842 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3844 * Windows command-line tool supports UTF-8 opt-in option for arguments
3845 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3846 (to any value) allows Windows user to access PKCS#12 file generated
3847 with Windows CryptoAPI and protected with non-ASCII password, as well
3848 as files generated under UTF-8 locale on Linux also protected with
3853 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3854 have been disabled by default and removed from DEFAULT, just like RC4.
3855 See the RC4 item below to re-enable both.
3859 * The method for finding the storage location for the Windows RAND seed file
3860 has changed. First we check %RANDFILE%. If that is not set then we check
3861 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3862 all else fails we fall back to C:\.
3866 * The EVP_EncryptUpdate() function has had its return type changed from void
3867 to int. A return of 0 indicates and error while a return of 1 indicates
3872 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3873 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3874 off the constant time implementation for RSA, DSA and DH have been made
3875 no-ops and deprecated.
3879 * Windows RAND implementation was simplified to only get entropy by
3880 calling CryptGenRandom(). Various other RAND-related tickets
3883 *Joseph Wylie Yandle, Rich Salz*
3885 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3886 and `OPENSSL_LH_`, respectively. The old names are available
3887 with API compatibility. They new names are now completely documented.
3891 * Unify TYPE_up_ref(obj) methods signature.
3892 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3893 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3894 int (instead of void) like all others TYPE_up_ref() methods.
3895 So now these methods also check the return value of CRYPTO_atomic_add(),
3896 and the validity of object reference counter.
3898 *fdasilvayy@gmail.com*
3900 * With Windows Visual Studio builds, the .pdb files are installed
3901 alongside the installed libraries and executables. For a static
3902 library installation, ossl_static.pdb is the associate compiler
3903 generated .pdb file to be used when linking programs.
3907 * Remove openssl.spec. Packaging files belong with the packagers.
3911 * Automatic Darwin/OSX configuration has had a refresh, it will now
3912 recognise x86_64 architectures automatically. You can still decide
3913 to build for a different bitness with the environment variable
3914 KERNEL_BITS (can be 32 or 64), for example:
3916 KERNEL_BITS=32 ./config
3920 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3921 256 bit AES and HMAC with SHA256.
3925 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3929 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3933 * To enable users to have their own config files and build file templates,
3934 Configure looks in the directory indicated by the environment variable
3935 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3936 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3937 name and is used as is.
3941 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3942 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3943 X509_CERT_FILE_CTX was removed.
3947 * "shared" builds are now the default. To create only static libraries use
3948 the "no-shared" Configure option.
3952 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3953 All of these option have not worked for some while and are fundamental
3958 * Make various cleanup routines no-ops and mark them as deprecated. Most
3959 global cleanup functions are no longer required because they are handled
3960 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3961 Explicitly de-initing can cause problems (e.g. where a library that uses
3962 OpenSSL de-inits, but an application is still using it). The affected
3963 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3964 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3965 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3966 COMP_zlib_cleanup().
3970 * --strict-warnings no longer enables runtime debugging options
3971 such as REF_DEBUG. Instead, debug options are automatically
3972 enabled with '--debug' builds.
3974 *Andy Polyakov, Emilia Käsper*
3976 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3977 have been moved out of the public header files. New functions for managing
3978 these have been added.
3982 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3983 objects have been moved out of the public header files. New
3984 functions for managing these have been added.
3988 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3989 have been moved out of the public header files. New functions for managing
3990 these have been added.
3994 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3995 moved out of the public header files. New functions for managing these
4000 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4004 * Removed the mk1mf build scripts.
4008 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4009 it is always safe to #include a header now.
4013 * Removed the aged BC-32 config and all its supporting scripts
4017 * Removed support for Ultrix, Netware, and OS/2.
4021 * Add support for HKDF.
4023 *Alessandro Ghedini*
4025 * Add support for blake2b and blake2s
4029 * Added support for "pipelining". Ciphers that have the
4030 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4031 encryptions/decryptions simultaneously. There are currently no built-in
4032 ciphers with this property but the expectation is that engines will be able
4033 to offer it to significantly improve throughput. Support has been extended
4034 into libssl so that multiple records for a single connection can be
4035 processed in one go (for >=TLS 1.1).
4039 * Added the AFALG engine. This is an async capable engine which is able to
4040 offload work to the Linux kernel. In this initial version it only supports
4041 AES128-CBC. The kernel must be version 4.1.0 or greater.
4045 * OpenSSL now uses a new threading API. It is no longer necessary to
4046 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4047 are two supported threading models: pthreads and windows threads. It is
4048 also possible to configure OpenSSL at compile time for "no-threads". The
4049 old threading API should no longer be used. The functions have been
4050 replaced with "no-op" compatibility macros.
4052 *Alessandro Ghedini, Matt Caswell*
4054 * Modify behavior of ALPN to invoke callback after SNI/servername
4055 callback, such that updates to the SSL_CTX affect ALPN.
4059 * Add SSL_CIPHER queries for authentication and key-exchange.
4063 * Changes to the DEFAULT cipherlist:
4064 - Prefer (EC)DHE handshakes over plain RSA.
4065 - Prefer AEAD ciphers over legacy ciphers.
4066 - Prefer ECDSA over RSA when both certificates are available.
4067 - Prefer TLSv1.2 ciphers/PRF.
4068 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4073 * Change the ECC default curve list to be this, in order: x25519,
4074 secp256r1, secp521r1, secp384r1.
4078 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4079 disabled by default. They can be re-enabled using the
4080 enable-weak-ssl-ciphers option to Configure.
4084 * If the server has ALPN configured, but supports no protocols that the
4085 client advertises, send a fatal "no_application_protocol" alert.
4086 This behaviour is SHALL in RFC 7301, though it isn't universally
4087 implemented by other servers.
4091 * Add X25519 support.
4092 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4093 for public and private key encoding using the format documented in
4094 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4095 key generation and key derivation.
4097 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4102 * Deprecate SRP_VBASE_get_by_user.
4103 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4104 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4105 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4106 seed, even if the seed is configured.
4108 Users should use SRP_VBASE_get1_by_user instead. Note that in
4109 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4110 also that even though configuring the SRP seed attempts to hide
4111 invalid usernames by continuing the handshake with fake
4112 credentials, this behaviour is not constant time and no strong
4113 guarantees are made that the handshake is indistinguishable from
4114 that of a valid user.
4118 * Configuration change; it's now possible to build dynamic engines
4119 without having to build shared libraries and vice versa. This
4120 only applies to the engines in `engines/`, those in `crypto/engine/`
4121 will always be built into libcrypto (i.e. "static").
4123 Building dynamic engines is enabled by default; to disable, use
4124 the configuration option "disable-dynamic-engine".
4126 The only requirements for building dynamic engines are the
4127 presence of the DSO module and building with position independent
4128 code, so they will also automatically be disabled if configuring
4129 with "disable-dso" or "disable-pic".
4131 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4132 are also taken away from openssl/opensslconf.h, as they are
4137 * Configuration change; if there is a known flag to compile
4138 position independent code, it will always be applied on the
4139 libcrypto and libssl object files, and never on the application
4140 object files. This means other libraries that use routines from
4141 libcrypto / libssl can be made into shared libraries regardless
4142 of how OpenSSL was configured.
4144 If this isn't desirable, the configuration options "disable-pic"
4145 or "no-pic" can be used to disable the use of PIC. This will
4146 also disable building shared libraries and dynamic engines.
4150 * Removed JPAKE code. It was experimental and has no wide use.
4154 * The INSTALL_PREFIX Makefile variable has been renamed to
4155 DESTDIR. That makes for less confusion on what this variable
4156 is for. Also, the configuration option --install_prefix is
4161 * Heartbeat for TLS has been removed and is disabled by default
4162 for DTLS; configure with enable-heartbeats. Code that uses the
4163 old #define's might need to be updated.
4165 *Emilia Käsper, Rich Salz*
4167 * Rename REF_CHECK to REF_DEBUG.
4171 * New "unified" build system
4173 The "unified" build system is aimed to be a common system for all
4174 platforms we support. With it comes new support for VMS.
4176 This system builds supports building in a different directory tree
4177 than the source tree. It produces one Makefile (for unix family
4178 or lookalikes), or one descrip.mms (for VMS).
4180 The source of information to make the Makefile / descrip.mms is
4181 small files called 'build.info', holding the necessary
4182 information for each directory with source to compile, and a
4183 template in Configurations, like unix-Makefile.tmpl or
4186 With this change, the library names were also renamed on Windows
4187 and on VMS. They now have names that are closer to the standard
4188 on Unix, and include the major version number, and in certain
4189 cases, the architecture they are built for. See "Notes on shared
4190 libraries" in INSTALL.
4192 We rely heavily on the perl module Text::Template.
4196 * Added support for auto-initialisation and de-initialisation of the library.
4197 OpenSSL no longer requires explicit init or deinit routines to be called,
4198 except in certain circumstances. See the OPENSSL_init_crypto() and
4199 OPENSSL_init_ssl() man pages for further information.
4203 * The arguments to the DTLSv1_listen function have changed. Specifically the
4204 "peer" argument is now expected to be a BIO_ADDR object.
4206 * Rewrite of BIO networking library. The BIO library lacked consistent
4207 support of IPv6, and adding it required some more extensive
4208 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4209 which hold all types of addresses and chains of address information.
4210 It also introduces a new API, with functions like BIO_socket,
4211 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4212 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4213 have been adapted accordingly.
4217 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4222 * CRIME protection: disable compression by default, even if OpenSSL is
4223 compiled with zlib enabled. Applications can still enable compression
4224 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4225 using the SSL_CONF library to configure compression.
4229 * The signature of the session callback configured with
4230 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4231 was explicitly marked as `const unsigned char*` instead of
4236 * Always DPURIFY. Remove the use of uninitialized memory in the
4237 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4241 * Removed many obsolete configuration items, including
4242 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4243 MD2_CHAR, MD2_INT, MD2_LONG
4245 IDEA_SHORT, IDEA_LONG
4246 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4248 *Rich Salz, with advice from Andy Polyakov*
4250 * Many BN internals have been moved to an internal header file.
4252 *Rich Salz with help from Andy Polyakov*
4254 * Configuration and writing out the results from it has changed.
4255 Files such as Makefile include/openssl/opensslconf.h and are now
4256 produced through general templates, such as Makefile.in and
4257 crypto/opensslconf.h.in and some help from the perl module
4260 Also, the center of configuration information is no longer
4261 Makefile. Instead, Configure produces a perl module in
4262 configdata.pm which holds most of the config data (in the hash
4263 table %config), the target data that comes from the target
4264 configuration in one of the `Configurations/*.conf` files (in
4269 * To clarify their intended purposes, the Configure options
4270 --prefix and --openssldir change their semantics, and become more
4271 straightforward and less interdependent.
4273 --prefix shall be used exclusively to give the location INSTALLTOP
4274 where programs, scripts, libraries, include files and manuals are
4275 going to be installed. The default is now /usr/local.
4277 --openssldir shall be used exclusively to give the default
4278 location OPENSSLDIR where certificates, private keys, CRLs are
4279 managed. This is also where the default openssl.cnf gets
4281 If the directory given with this option is a relative path, the
4282 values of both the --prefix value and the --openssldir value will
4283 be combined to become OPENSSLDIR.
4284 The default for --openssldir is INSTALLTOP/ssl.
4286 Anyone who uses --openssldir to specify where OpenSSL is to be
4287 installed MUST change to use --prefix instead.
4291 * The GOST engine was out of date and therefore it has been removed. An up
4292 to date GOST engine is now being maintained in an external repository.
4293 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4294 support for GOST ciphersuites (these are only activated if a GOST engine
4299 * EGD is no longer supported by default; use enable-egd when
4302 *Ben Kaduk and Rich Salz*
4304 * The distribution now has Makefile.in files, which are used to
4305 create Makefile's when Configure is run. *Configure must be run
4306 before trying to build now.*
4310 * The return value for SSL_CIPHER_description() for error conditions
4315 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4317 Obtaining and performing DNSSEC validation of TLSA records is
4318 the application's responsibility. The application provides
4319 the TLSA records of its choice to OpenSSL, and these are then
4320 used to authenticate the peer.
4322 The TLSA records need not even come from DNS. They can, for
4323 example, be used to implement local end-entity certificate or
4324 trust-anchor "pinning", where the "pin" data takes the form
4325 of TLSA records, which can augment or replace verification
4326 based on the usual WebPKI public certification authorities.
4330 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4331 continues to support deprecated interfaces in default builds.
4332 However, applications are strongly advised to compile their
4333 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4334 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4335 or the 1.1.0 releases.
4337 In environments in which all applications have been ported to
4338 not use any deprecated interfaces OpenSSL's Configure script
4339 should be used with the --api=1.1.0 option to entirely remove
4340 support for the deprecated features from the library and
4341 unconditionally disable them in the installed headers.
4342 Essentially the same effect can be achieved with the "no-deprecated"
4343 argument to Configure, except that this will always restrict
4344 the build to just the latest API, rather than a fixed API
4347 As applications are ported to future revisions of the API,
4348 they should update their compile-time OPENSSL_API_COMPAT define
4349 accordingly, but in most cases should be able to continue to
4350 compile with later releases.
4352 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4353 0x10000000L and 0x00908000L, respectively. However those
4354 versions did not support the OPENSSL_API_COMPAT feature, and
4355 so applications are not typically tested for explicit support
4356 of just the undeprecated features of either release.
4360 * Add support for setting the minimum and maximum supported protocol.
4361 It can bet set via the SSL_set_min_proto_version() and
4362 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4363 MaxProtocol. It's recommended to use the new APIs to disable
4364 protocols instead of disabling individual protocols using
4365 SSL_set_options() or SSL_CONF's Protocol. This change also
4366 removes support for disabling TLS 1.2 in the OpenSSL TLS
4367 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4371 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4375 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4376 and integrates ECDSA and ECDH functionality into EC. Implementations can
4377 now redirect key generation and no longer need to convert to or from
4380 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4381 include the ec.h header file instead.
4385 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4386 ciphers who are no longer supported and drops support the ephemeral RSA key
4387 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4391 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4392 opaque. For HMAC_CTX, the following constructors and destructors
4395 HMAC_CTX *HMAC_CTX_new(void);
4396 void HMAC_CTX_free(HMAC_CTX *ctx);
4398 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4399 destroy such methods has been added. See EVP_MD_meth_new(3) and
4400 EVP_CIPHER_meth_new(3) for documentation.
4403 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4404 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4405 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4406 an already created structure.
4407 2) For consistency with the majority of our object creators and
4408 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4409 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4410 for deprecated builds.
4414 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4415 cryptographic operations to be performed asynchronously as long as an
4416 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4417 further details. Libssl has also had this capability integrated with the
4418 introduction of the new mode SSL_MODE_ASYNC and associated error
4419 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4420 pages. This work was developed in partnership with Intel Corp.
4424 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4425 always enabled now. If you want to disable the support you should
4426 exclude it using the list of supported ciphers. This also means that the
4427 "-no_ecdhe" option has been removed from s_server.
4431 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4432 SSL_{CTX_}set1_curves() which can set a list.
4436 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4437 curve you want to support using SSL_{CTX_}set1_curves().
4441 * State machine rewrite. The state machine code has been significantly
4442 refactored in order to remove much duplication of code and solve issues
4443 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4444 further details). This change does have some associated API changes.
4445 Notably the SSL_state() function has been removed and replaced by
4446 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4447 SSL_set_state() has been removed altogether. The previous handshake states
4448 defined in ssl.h and ssl3.h have also been removed.
4452 * All instances of the string "ssleay" in the public API were replaced
4453 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4454 Some error codes related to internal RSA_eay API's were renamed.
4458 * The demo files in crypto/threads were moved to demo/threads.
4462 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4465 *Matt Caswell, Rich Salz*
4467 * New ASN.1 embed macro.
4469 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4470 structure is not allocated: it is part of the parent. That is instead of
4478 This reduces memory fragmentation and make it impossible to accidentally
4479 set a mandatory field to NULL.
4481 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4482 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4483 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4488 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4492 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4493 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4494 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4495 DES and RC4 ciphersuites.
4499 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4500 This changes the decoding behaviour for some invalid messages,
4501 though the change is mostly in the more lenient direction, and
4502 legacy behaviour is preserved as much as possible.
4506 * Fix no-stdio build.
4507 *David Woodhouse <David.Woodhouse@intel.com> and also*
4508 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4510 * New testing framework
4511 The testing framework has been largely rewritten and is now using
4512 perl and the perl modules Test::Harness and an extended variant of
4513 Test::More called OpenSSL::Test to do its work. All test scripts in
4514 test/ have been rewritten into test recipes, and all direct calls to
4515 executables in test/Makefile have become individual recipes using the
4516 simplified testing OpenSSL::Test::Simple.
4518 For documentation on our testing modules, do:
4520 perldoc test/testlib/OpenSSL/Test/Simple.pm
4521 perldoc test/testlib/OpenSSL/Test.pm
4525 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4526 are used; the latter aborts on memory leaks (usually checked on exit).
4527 Some undocumented "set malloc, etc., hooks" functions were removed
4528 and others were changed. All are now documented.
4532 * In DSA_generate_parameters_ex, if the provided seed is too short,
4535 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4537 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4538 from RFC4279, RFC4785, RFC5487, RFC5489.
4540 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4541 original RSA_PSK patch.
4545 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4546 era flag was never set throughout the codebase (only read). Also removed
4547 SSL3_FLAGS_POP_BUFFER which was only used if
4548 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4552 * Changed the default name options in the "ca", "crl", "req" and "x509"
4553 to be "oneline" instead of "compat".
4557 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4558 not aware of clients that still exhibit this bug, and the workaround
4559 hasn't been working properly for a while.
4563 * The return type of BIO_number_read() and BIO_number_written() as well as
4564 the corresponding num_read and num_write members in the BIO structure has
4565 changed from unsigned long to uint64_t. On platforms where an unsigned
4566 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4571 * Given the pervasive nature of TLS extensions it is inadvisable to run
4572 OpenSSL without support for them. It also means that maintaining
4573 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4574 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4578 * Removed support for the two export grade static DH ciphersuites
4579 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4580 were newly added (along with a number of other static DH ciphersuites) to
4581 1.0.2. However the two export ones have *never* worked since they were
4582 introduced. It seems strange in any case to be adding new export
4583 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4587 * Version negotiation has been rewritten. In particular SSLv23_method(),
4588 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4589 and turned into macros which simply call the new preferred function names
4590 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4591 should use the new names instead. Also as part of this change the ssl23.h
4592 header file has been removed.
4596 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4597 code and the associated standard is no longer considered fit-for-purpose.
4601 * RT2547 was closed. When generating a private key, try to make the
4602 output file readable only by the owner. This behavior change might
4603 be noticeable when interacting with other software.
4605 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4610 * Added HTTP GET support to the ocsp command.
4614 * Changed default digest for the dgst and enc commands from MD5 to
4619 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4623 * Added support for TLS extended master secret from
4624 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4625 initial patch which was a great help during development.
4629 * All libssl internal structures have been removed from the public header
4630 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4631 now redundant). Users should not attempt to access internal structures
4632 directly. Instead they should use the provided API functions.
4636 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4637 Access to deprecated functions can be re-enabled by running config with
4638 "enable-deprecated". In addition applications wishing to use deprecated
4639 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4640 will, by default, disable some transitive includes that previously existed
4641 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4645 * Added support for OCB mode. OpenSSL has been granted a patent license
4646 compatible with the OpenSSL license for use of OCB. Details are available
4647 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4648 for OCB can be removed by calling config with no-ocb.
4652 * SSLv2 support has been removed. It still supports receiving an SSLv2
4653 compatible client hello.
4657 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4658 done while fixing the error code for the key-too-small case.
4660 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4662 * CA.sh has been removed; use CA.pl instead.
4666 * Removed old DES API.
4670 * Remove various unsupported platforms:
4676 Sinix/ReliantUNIX RM400
4681 16-bit platforms such as WIN16
4685 * Clean up OPENSSL_NO_xxx #define's
4686 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4687 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4688 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4689 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4690 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4691 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4692 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4693 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4694 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4695 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4699 * Cleaned up dead code
4700 Remove all but one '#ifdef undef' which is to be looked at.
4704 * Clean up calling of xxx_free routines.
4705 Just like free(), fix most of the xxx_free routines to accept
4706 NULL. Remove the non-null checks from callers. Save much code.
4710 * Add secure heap for storage of private keys (when possible).
4711 Add BIO_s_secmem(), CBIGNUM, etc.
4712 Contributed by Akamai Technologies under our Corporate CLA.
4716 * Experimental support for a new, fast, unbiased prime candidate generator,
4717 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4719 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4721 * New output format NSS in the sess_id command line tool. This allows
4722 exporting the session id and the master key in NSS keylog format.
4724 *Martin Kaiser <martin@kaiser.cx>*
4726 * Harmonize version and its documentation. -f flag is used to display
4729 *mancha <mancha1@zoho.com>*
4731 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4732 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4734 *mancha <mancha1@zoho.com>*
4736 * Fix some double frees. These are not thought to be exploitable.
4738 *mancha <mancha1@zoho.com>*
4740 * A missing bounds check in the handling of the TLS heartbeat extension
4741 can be used to reveal up to 64k of memory to a connected client or
4744 Thanks for Neel Mehta of Google Security for discovering this bug and to
4745 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4746 preparing the fix ([CVE-2014-0160])
4748 *Adam Langley, Bodo Moeller*
4750 * Fix for the attack described in the paper "Recovering OpenSSL
4751 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4752 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4753 <http://eprint.iacr.org/2014/140>
4755 Thanks to Yuval Yarom and Naomi Benger for discovering this
4756 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4758 *Yuval Yarom and Naomi Benger*
4760 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4761 this fixes a limitation in previous versions of OpenSSL.
4765 * Experimental encrypt-then-mac support.
4767 Experimental support for encrypt then mac from
4768 draft-gutmann-tls-encrypt-then-mac-02.txt
4770 To enable it set the appropriate extension number (0x42 for the test
4771 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4773 For non-compliant peers (i.e. just about everything) this should have no
4776 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4780 * Add EVP support for key wrapping algorithms, to avoid problems with
4781 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4782 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4783 algorithms and include tests cases.
4787 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4792 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4793 MGF1 digest and OAEP label.
4797 * Make openssl verify return errors.
4799 *Chris Palmer <palmer@google.com> and Ben Laurie*
4801 * New function ASN1_TIME_diff to calculate the difference between two
4802 ASN1_TIME structures or one structure and the current time.
4806 * Update fips_test_suite to support multiple command line options. New
4807 test to induce all self test errors in sequence and check expected
4812 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4813 sign or verify all in one operation.
4817 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4818 test programs and fips_test_suite. Includes functionality to parse
4819 the minimal script output of fipsalgest.pl directly.
4823 * Add authorisation parameter to FIPS_module_mode_set().
4827 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4831 * Use separate DRBG fields for internal and external flags. New function
4832 FIPS_drbg_health_check() to perform on demand health checking. Add
4833 generation tests to fips_test_suite with reduced health check interval to
4834 demonstrate periodic health checking. Add "nodh" option to
4835 fips_test_suite to skip very slow DH test.
4839 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4844 * More extensive health check for DRBG checking many more failure modes.
4845 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4846 combination: call this in fips_test_suite.
4850 * Add support for canonical generation of DSA parameter 'g'. See
4853 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4854 POST to handle HMAC cases.
4858 * Add functions FIPS_module_version() and FIPS_module_version_text()
4859 to return numerical and string versions of the FIPS module number.
4863 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4864 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4865 outside the validated module in the FIPS capable OpenSSL.
4869 * Minor change to DRBG entropy callback semantics. In some cases
4870 there is no multiple of the block length between min_len and
4871 max_len. Allow the callback to return more than max_len bytes
4872 of entropy but discard any extra: it is the callback's responsibility
4873 to ensure that the extra data discarded does not impact the
4874 requested amount of entropy.
4878 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4879 information in FIPS186-3, SP800-57 and SP800-131A.
4883 * CCM support via EVP. Interface is very similar to GCM case except we
4884 must supply all data in one chunk (i.e. no update, final) and the
4885 message length must be supplied if AAD is used. Add algorithm test
4890 * Initial version of POST overhaul. Add POST callback to allow the status
4891 of POST to be monitored and/or failures induced. Modify fips_test_suite
4892 to use callback. Always run all selftests even if one fails.
4896 * XTS support including algorithm test driver in the fips_gcmtest program.
4897 Note: this does increase the maximum key length from 32 to 64 bytes but
4898 there should be no binary compatibility issues as existing applications
4899 will never use XTS mode.
4903 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4904 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4905 performs algorithm blocking for unapproved PRNG types. Also do not
4906 set PRNG type in FIPS_mode_set(): leave this to the application.
4907 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4908 the standard OpenSSL PRNG: set additional data to a date time vector.
4912 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4913 This shouldn't present any incompatibility problems because applications
4914 shouldn't be using these directly and any that are will need to rethink
4915 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4919 * Extensive self tests and health checking required by SP800-90 DRBG.
4920 Remove strength parameter from FIPS_drbg_instantiate and always
4921 instantiate at maximum supported strength.
4925 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4929 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4933 * New function DH_compute_key_padded() to compute a DH key and pad with
4934 leading zeroes if needed: this complies with SP800-56A et al.
4938 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4939 anything, incomplete, subject to change and largely untested at present.
4943 * Modify fipscanisteronly build option to only build the necessary object
4944 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4948 * Add experimental option FIPSSYMS to give all symbols in
4949 fipscanister.o and FIPS or fips prefix. This will avoid
4950 conflicts with future versions of OpenSSL. Add perl script
4951 util/fipsas.pl to preprocess assembly language source files
4952 and rename any affected symbols.
4956 * Add selftest checks and algorithm block of non-fips algorithms in
4957 FIPS mode. Remove DES2 from selftests.
4961 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4962 return internal method without any ENGINE dependencies. Add new
4963 tiny fips sign and verify functions.
4967 * New build option no-ec2m to disable characteristic 2 code.
4971 * New build option "fipscanisteronly". This only builds fipscanister.o
4972 and (currently) associated fips utilities. Uses the file Makefile.fips
4973 instead of Makefile.org as the prototype.
4977 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4978 Update fips_gcmtest to use IV generator.
4982 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4983 setting output buffer to NULL. The `*Final` function must be
4984 called although it will not retrieve any additional data. The tag
4985 can be set or retrieved with a ctrl. The IV length is by default 12
4986 bytes (96 bits) but can be set to an alternative value. If the IV
4987 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4992 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4993 underlying do_cipher function handles all cipher semantics itself
4994 including padding and finalisation. This is useful if (for example)
4995 an ENGINE cipher handles block padding itself. The behaviour of
4996 do_cipher is subtly changed if this flag is set: the return value
4997 is the number of characters written to the output buffer (zero is
4998 no longer an error code) or a negative error code. Also if the
4999 input buffer is NULL and length 0 finalisation should be performed.
5003 * If a candidate issuer certificate is already part of the constructed
5004 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5008 * Improve forward-security support: add functions
5010 void SSL_CTX_set_not_resumable_session_callback(
5011 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5012 void SSL_set_not_resumable_session_callback(
5013 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5015 for use by SSL/TLS servers; the callback function will be called whenever a
5016 new session is created, and gets to decide whether the session may be
5017 cached to make it resumable (return 0) or not (return 1). (As by the
5018 SSL/TLS protocol specifications, the session_id sent by the server will be
5019 empty to indicate that the session is not resumable; also, the server will
5020 not generate RFC 4507 (RFC 5077) session tickets.)
5022 A simple reasonable callback implementation is to return is_forward_secure.
5023 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5024 by the SSL/TLS server library, indicating whether it can provide forward
5027 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5029 * New -verify_name option in command line utilities to set verification
5034 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5035 Add CMAC pkey methods.
5039 * Experimental renegotiation in s_server -www mode. If the client
5040 browses /reneg connection is renegotiated. If /renegcert it is
5041 renegotiated requesting a certificate.
5045 * Add an "external" session cache for debugging purposes to s_server. This
5046 should help trace issues which normally are only apparent in deployed
5047 multi-process servers.
5051 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5052 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5053 BIO_set_cipher() and some obscure PEM functions were changed so they
5054 can now return an error. The RAND changes required a change to the
5055 RAND_METHOD structure.
5059 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5060 a gcc attribute to warn if the result of a function is ignored. This
5061 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5062 whose return value is often ignored.
5066 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5067 These allow SCTs (signed certificate timestamps) to be requested and
5068 validated when establishing a connection.
5070 *Rob Percival <robpercival@google.com>*
5075 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5077 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5078 used even when parsing explicit parameters, when loading a encoded key
5079 or calling `EC_GROUP_new_from_ecpkparameters()`/
5080 `EC_GROUP_new_from_ecparameters()`.
5081 This prevents bypass of security hardening and performance gains,
5082 especially for curves with specialized EC_METHODs.
5083 By default, if a key encoded with explicit parameters is loaded and later
5084 encoded, the output is still encoded with explicit parameters, even if
5085 internally a "named" EC_GROUP is used for computation.
5089 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5090 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5091 NULL. After this change, only the cofactor parameter can be NULL. It also
5092 does some minimal sanity checks on the passed order.
5097 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5098 An attack is simple, if the first CMS_recipientInfo is valid but the
5099 second CMS_recipientInfo is chosen ciphertext. If the second
5100 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5101 encryption key will be replaced by garbage, and the message cannot be
5102 decoded, but if the RSA decryption fails, the correct encryption key is
5103 used and the recipient will not notice the attack.
5104 As a work around for this potential attack the length of the decrypted
5105 key must be equal to the cipher default key length, in case the
5106 certificate is not given and all recipientInfo are tried out.
5107 The old behaviour can be re-enabled in the CMS code by setting the
5108 CMS_DEBUG_DECRYPT flag.
5113 * Document issue with installation paths in diverse Windows builds
5115 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5116 binaries and run-time config file.
5121 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5123 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5124 This changes the size when using the `genpkey` command when no size is given.
5125 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5126 generation commands to use 2048 bits by default.
5130 * Add FIPS support for Android Arm 64-bit
5132 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5133 Module in Version 2.0.10. For some reason, the corresponding target
5134 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5135 built with FIPS support on Android Arm 64-bit. This omission has been
5138 *Matthias St. Pierre*
5140 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5142 * 0-byte record padding oracle
5144 If an application encounters a fatal protocol error and then calls
5145 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5146 then OpenSSL can respond differently to the calling application if a 0 byte
5147 record is received with invalid padding compared to if a 0 byte record is
5148 received with an invalid MAC. If the application then behaves differently
5149 based on that in a way that is detectable to the remote peer, then this
5150 amounts to a padding oracle that could be used to decrypt data.
5152 In order for this to be exploitable "non-stitched" ciphersuites must be in
5153 use. Stitched ciphersuites are optimised implementations of certain
5154 commonly used ciphersuites. Also the application must call SSL_shutdown()
5155 twice even if a protocol error has occurred (applications should not do
5156 this but some do anyway).
5158 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5159 Aviram, with additional investigation by Steven Collison and Andrew
5160 Hourselt. It was reported to OpenSSL on 10th December 2018.
5165 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5169 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5171 * Microarchitecture timing vulnerability in ECC scalar multiplication
5173 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5174 shown to be vulnerable to a microarchitecture timing side channel attack.
5175 An attacker with sufficient access to mount local timing attacks during
5176 ECDSA signature generation could recover the private key.
5178 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5179 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5185 * Timing vulnerability in DSA signature generation
5187 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5188 timing side channel attack. An attacker could use variations in the signing
5189 algorithm to recover the private key.
5191 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5196 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5197 Module, accidentally introduced while backporting security fixes from the
5198 development branch and hindering the use of ECC in FIPS mode.
5202 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5204 * Client DoS due to large DH parameter
5206 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5207 malicious server can send a very large prime value to the client. This will
5208 cause the client to spend an unreasonably long period of time generating a
5209 key for this prime resulting in a hang until the client has finished. This
5210 could be exploited in a Denial Of Service attack.
5212 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5217 * Cache timing vulnerability in RSA Key Generation
5219 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5220 a cache timing side channel attack. An attacker with sufficient access to
5221 mount cache timing attacks during the RSA key generation process could
5222 recover the private key.
5224 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5225 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5230 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5231 parameter is no longer accepted, as it leads to a corrupt table. NULL
5232 pem_str is reserved for alias entries only.
5236 * Revert blinding in ECDSA sign and instead make problematic addition
5237 length-invariant. Switch even to fixed-length Montgomery multiplication.
5241 * Change generating and checking of primes so that the error rate of not
5242 being prime depends on the intended use based on the size of the input.
5243 For larger primes this will result in more rounds of Miller-Rabin.
5244 The maximal error rate for primes with more than 1080 bits is lowered
5247 *Kurt Roeckx, Annie Yousar*
5249 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5253 * Add blinding to ECDSA and DSA signatures to protect against side channel
5254 attacks discovered by Keegan Ryan (NCC Group).
5258 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5259 now allow empty (zero character) pass phrases.
5263 * Certificate time validation (X509_cmp_time) enforces stricter
5264 compliance with RFC 5280. Fractional seconds and timezone offsets
5265 are no longer allowed.
5269 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5271 * Constructed ASN.1 types with a recursive definition could exceed the stack
5273 Constructed ASN.1 types with a recursive definition (such as can be found
5274 in PKCS7) could eventually exceed the stack given malicious input with
5275 excessive recursion. This could result in a Denial Of Service attack. There
5276 are no such structures used within SSL/TLS that come from untrusted sources
5277 so this is considered safe.
5279 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5285 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5287 * Read/write after SSL object in error state
5289 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5290 mechanism. The intent was that if a fatal error occurred during a handshake
5291 then OpenSSL would move into the error state and would immediately fail if
5292 you attempted to continue the handshake. This works as designed for the
5293 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5294 SSL_connect()), however due to a bug it does not work correctly if
5295 SSL_read() or SSL_write() is called directly. In that scenario, if the
5296 handshake fails then a fatal error will be returned in the initial function
5297 call. If SSL_read()/SSL_write() is subsequently called by the application
5298 for the same SSL object then it will succeed and the data is passed without
5299 being decrypted/encrypted directly from the SSL/TLS record layer.
5301 In order to exploit this issue an application bug would have to be present
5302 that resulted in a call to SSL_read()/SSL_write() being issued after having
5303 already received a fatal error.
5305 This issue was reported to OpenSSL by David Benjamin (Google).
5310 * rsaz_1024_mul_avx2 overflow bug on x86_64
5312 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5313 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5314 Analysis suggests that attacks against RSA and DSA as a result of this
5315 defect would be very difficult to perform and are not believed likely.
5316 Attacks against DH1024 are considered just feasible, because most of the
5317 work necessary to deduce information about a private key may be performed
5318 offline. The amount of resources required for such an attack would be
5319 significant. However, for an attack on TLS to be meaningful, the server
5320 would have to share the DH1024 private key among multiple clients, which is
5321 no longer an option since CVE-2016-0701.
5323 This only affects processors that support the AVX2 but not ADX extensions
5324 like Intel Haswell (4th generation).
5326 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5327 was originally found via the OSS-Fuzz project.
5332 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5334 * bn_sqrx8x_internal carry bug on x86_64
5336 There is a carry propagating bug in the x86_64 Montgomery squaring
5337 procedure. No EC algorithms are affected. Analysis suggests that attacks
5338 against RSA and DSA as a result of this defect would be very difficult to
5339 perform and are not believed likely. Attacks against DH are considered just
5340 feasible (although very difficult) because most of the work necessary to
5341 deduce information about a private key may be performed offline. The amount
5342 of resources required for such an attack would be very significant and
5343 likely only accessible to a limited number of attackers. An attacker would
5344 additionally need online access to an unpatched system using the target
5345 private key in a scenario with persistent DH parameters and a private
5346 key that is shared between multiple clients.
5348 This only affects processors that support the BMI1, BMI2 and ADX extensions
5349 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5351 This issue was reported to OpenSSL by the OSS-Fuzz project.
5356 * Malformed X.509 IPAddressFamily could cause OOB read
5358 If an X.509 certificate has a malformed IPAddressFamily extension,
5359 OpenSSL could do a one-byte buffer overread. The most likely result
5360 would be an erroneous display of the certificate in text format.
5362 This issue was reported to OpenSSL by the OSS-Fuzz project.
5366 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5368 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5369 platform rather than 'mingw'.
5373 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5375 * Truncated packet could crash via OOB read
5377 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5378 cipher is being used, then a truncated packet can cause that host to
5379 perform an out-of-bounds read, usually resulting in a crash.
5381 This issue was reported to OpenSSL by Robert Święcki of Google.
5386 * BN_mod_exp may produce incorrect results on x86_64
5388 There is a carry propagating bug in the x86_64 Montgomery squaring
5389 procedure. No EC algorithms are affected. Analysis suggests that attacks
5390 against RSA and DSA as a result of this defect would be very difficult to
5391 perform and are not believed likely. Attacks against DH are considered just
5392 feasible (although very difficult) because most of the work necessary to
5393 deduce information about a private key may be performed offline. The amount
5394 of resources required for such an attack would be very significant and
5395 likely only accessible to a limited number of attackers. An attacker would
5396 additionally need online access to an unpatched system using the target
5397 private key in a scenario with persistent DH parameters and a private
5398 key that is shared between multiple clients. For example this can occur by
5399 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5400 similar to CVE-2015-3193 but must be treated as a separate problem.
5402 This issue was reported to OpenSSL by the OSS-Fuzz project.
5407 * Montgomery multiplication may produce incorrect results
5409 There is a carry propagating bug in the Broadwell-specific Montgomery
5410 multiplication procedure that handles input lengths divisible by, but
5411 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5412 and DH private keys are impossible. This is because the subroutine in
5413 question is not used in operations with the private key itself and an input
5414 of the attacker's direct choice. Otherwise the bug can manifest itself as
5415 transient authentication and key negotiation failures or reproducible
5416 erroneous outcome of public-key operations with specially crafted input.
5417 Among EC algorithms only Brainpool P-512 curves are affected and one
5418 presumably can attack ECDH key negotiation. Impact was not analyzed in
5419 detail, because pre-requisites for attack are considered unlikely. Namely
5420 multiple clients have to choose the curve in question and the server has to
5421 share the private key among them, neither of which is default behaviour.
5422 Even then only clients that chose the curve will be affected.
5424 This issue was publicly reported as transient failures and was not
5425 initially recognized as a security issue. Thanks to Richard Morgan for
5426 providing reproducible case.
5431 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5432 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5433 prevent issues where no progress is being made and the peer continually
5434 sends unrecognised record types, using up resources processing them.
5438 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5440 * Missing CRL sanity check
5442 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5443 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5444 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5446 This issue only affects the OpenSSL 1.0.2i
5451 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5453 * OCSP Status Request extension unbounded memory growth
5455 A malicious client can send an excessively large OCSP Status Request
5456 extension. If that client continually requests renegotiation, sending a
5457 large OCSP Status Request extension each time, then there will be unbounded
5458 memory growth on the server. This will eventually lead to a Denial Of
5459 Service attack through memory exhaustion. Servers with a default
5460 configuration are vulnerable even if they do not support OCSP. Builds using
5461 the "no-ocsp" build time option are not affected.
5463 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5468 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5471 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5477 * OOB write in MDC2_Update()
5479 An overflow can occur in MDC2_Update() either if called directly or
5480 through the EVP_DigestUpdate() function using MDC2. If an attacker
5481 is able to supply very large amounts of input data after a previous
5482 call to EVP_EncryptUpdate() with a partial block then a length check
5483 can overflow resulting in a heap corruption.
5485 The amount of data needed is comparable to SIZE_MAX which is impractical
5488 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5493 * Malformed SHA512 ticket DoS
5495 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5496 DoS attack where a malformed ticket will result in an OOB read which will
5499 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5500 a custom server callback and ticket lookup mechanism.
5502 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5507 * OOB write in BN_bn2dec()
5509 The function BN_bn2dec() does not check the return value of BN_div_word().
5510 This can cause an OOB write if an application uses this function with an
5511 overly large BIGNUM. This could be a problem if an overly large certificate
5512 or CRL is printed out from an untrusted source. TLS is not affected because
5513 record limits will reject an oversized certificate before it is parsed.
5515 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5520 * OOB read in TS_OBJ_print_bio()
5522 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5523 the total length the OID text representation would use and not the amount
5524 of data written. This will result in OOB reads when large OIDs are
5527 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5532 * Pointer arithmetic undefined behaviour
5534 Avoid some undefined pointer arithmetic
5536 A common idiom in the codebase is to check limits in the following manner:
5539 Where "p" points to some malloc'd data of SIZE bytes and
5542 "len" here could be from some externally supplied data (e.g. from a TLS
5545 The rules of C pointer arithmetic are such that "p + len" is only well
5546 defined where len <= SIZE. Therefore the above idiom is actually
5547 undefined behaviour.
5549 For example this could cause problems if some malloc implementation
5550 provides an address for "p" such that "p + len" actually overflows for
5551 values of len that are too big and therefore p + len < limit.
5553 This issue was reported to OpenSSL by Guido Vranken
5558 * Constant time flag not preserved in DSA signing
5560 Operations in the DSA signing algorithm should run in constant time in
5561 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5562 implementation means that a non-constant time codepath is followed for
5563 certain operations. This has been demonstrated through a cache-timing
5564 attack to be sufficient for an attacker to recover the private DSA key.
5566 This issue was reported by César Pereida (Aalto University), Billy Brumley
5567 (Tampere University of Technology), and Yuval Yarom (The University of
5568 Adelaide and NICTA).
5573 * DTLS buffered message DoS
5575 In a DTLS connection where handshake messages are delivered out-of-order
5576 those messages that OpenSSL is not yet ready to process will be buffered
5577 for later use. Under certain circumstances, a flaw in the logic means that
5578 those messages do not get removed from the buffer even though the handshake
5579 has been completed. An attacker could force up to approx. 15 messages to
5580 remain in the buffer when they are no longer required. These messages will
5581 be cleared when the DTLS connection is closed. The default maximum size for
5582 a message is 100k. Therefore, the attacker could force an additional 1500k
5583 to be consumed per connection. By opening many simultaneous connections an
5584 attacker could cause a DoS attack through memory exhaustion.
5586 This issue was reported to OpenSSL by Quan Luo.
5591 * DTLS replay protection DoS
5593 A flaw in the DTLS replay attack protection mechanism means that records
5594 that arrive for future epochs update the replay protection "window" before
5595 the MAC for the record has been validated. This could be exploited by an
5596 attacker by sending a record for the next epoch (which does not have to
5597 decrypt or have a valid MAC), with a very large sequence number. This means
5598 that all subsequent legitimate packets are dropped causing a denial of
5599 service for a specific DTLS connection.
5601 This issue was reported to OpenSSL by the OCAP audit team.
5606 * Certificate message OOB reads
5608 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5609 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5610 theoretical DoS risk but this has not been observed in practice on common
5613 The messages affected are client certificate, client certificate request
5614 and server certificate. As a result the attack can only be performed
5615 against a client or a server which enables client authentication.
5617 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5622 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5624 * Prevent padding oracle in AES-NI CBC MAC check
5626 A MITM attacker can use a padding oracle attack to decrypt traffic
5627 when the connection uses an AES CBC cipher and the server support
5630 This issue was introduced as part of the fix for Lucky 13 padding
5631 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5632 constant time by making sure that always the same bytes are read and
5633 compared against either the MAC or padding bytes. But it no longer
5634 checked that there was enough data to have both the MAC and padding
5637 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5641 * Fix EVP_EncodeUpdate overflow
5643 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5644 Base64 encoding of binary data. If an attacker is able to supply very large
5645 amounts of input data then a length check can overflow resulting in a heap
5648 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5649 the `PEM_write_bio*` family of functions. These are mainly used within the
5650 OpenSSL command line applications, so any application which processes data
5651 from an untrusted source and outputs it as a PEM file should be considered
5652 vulnerable to this issue. User applications that call these APIs directly
5653 with large amounts of untrusted data may also be vulnerable.
5655 This issue was reported by Guido Vranken.
5660 * Fix EVP_EncryptUpdate overflow
5662 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5663 is able to supply very large amounts of input data after a previous call to
5664 EVP_EncryptUpdate() with a partial block then a length check can overflow
5665 resulting in a heap corruption. Following an analysis of all OpenSSL
5666 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5667 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5668 the first called function after an EVP_EncryptInit(), and therefore that
5669 specific call must be safe. The second form is where the length passed to
5670 EVP_EncryptUpdate() can be seen from the code to be some small value and
5671 therefore there is no possibility of an overflow. Since all instances are
5672 one of these two forms, it is believed that there can be no overflows in
5673 internal code due to this problem. It should be noted that
5674 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5675 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5676 of these calls have also been analysed too and it is believed there are no
5677 instances in internal usage where an overflow could occur.
5679 This issue was reported by Guido Vranken.
5684 * Prevent ASN.1 BIO excessive memory allocation
5686 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5687 a short invalid encoding can cause allocation of large amounts of memory
5688 potentially consuming excessive resources or exhausting memory.
5690 Any application parsing untrusted data through d2i BIO functions is
5691 affected. The memory based functions such as d2i_X509() are *not* affected.
5692 Since the memory based functions are used by the TLS library, TLS
5693 applications are not affected.
5695 This issue was reported by Brian Carpenter.
5702 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5703 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5704 in arbitrary stack data being returned in the buffer.
5706 This issue was reported by Guido Vranken.
5711 * Modify behavior of ALPN to invoke callback after SNI/servername
5712 callback, such that updates to the SSL_CTX affect ALPN.
5716 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5721 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5722 methods are enabled and ssl2 is disabled the methods return NULL.
5726 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5728 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5729 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5730 provide any "EXPORT" or "LOW" strength ciphers.
5734 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5735 is by default disabled at build-time. Builds that are not configured with
5736 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5737 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5738 will need to explicitly call either of:
5740 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5742 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5744 as appropriate. Even if either of those is used, or the application
5745 explicitly uses the version-specific SSLv2_method() or its client and
5746 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5747 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5748 ciphers, and SSLv2 56-bit DES are no longer available.
5753 * Fix a double-free in DSA code
5755 A double free bug was discovered when OpenSSL parses malformed DSA private
5756 keys and could lead to a DoS attack or memory corruption for applications
5757 that receive DSA private keys from untrusted sources. This scenario is
5760 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5766 * Disable SRP fake user seed to address a server memory leak.
5768 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5770 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5771 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5772 was changed to ignore the "fake user" SRP seed, even if the seed
5775 Users should use SRP_VBASE_get1_by_user instead. Note that in
5776 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5777 also that even though configuring the SRP seed attempts to hide
5778 invalid usernames by continuing the handshake with fake
5779 credentials, this behaviour is not constant time and no strong
5780 guarantees are made that the handshake is indistinguishable from
5781 that of a valid user.
5786 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5788 In the BN_hex2bn function the number of hex digits is calculated using an
5789 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5790 large values of `i` this can result in `bn_expand` not allocating any
5791 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5792 field as NULL leading to a subsequent NULL ptr deref. For very large values
5793 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5794 In this case memory is allocated to the internal BIGNUM data field, but it
5795 is insufficiently sized leading to heap corruption. A similar issue exists
5796 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5797 is ever called by user applications with very large untrusted hex/dec data.
5798 This is anticipated to be a rare occurrence.
5800 All OpenSSL internal usage of these functions use data that is not expected
5801 to be untrusted, e.g. config file data or application command line
5802 arguments. If user developed applications generate config file data based
5803 on untrusted data then it is possible that this could also lead to security
5804 consequences. This is also anticipated to be rare.
5806 This issue was reported to OpenSSL by Guido Vranken.
5811 * Fix memory issues in `BIO_*printf` functions
5813 The internal `fmtstr` function used in processing a "%s" format string in
5814 the `BIO_*printf` functions could overflow while calculating the length of a
5815 string and cause an OOB read when printing very long strings.
5817 Additionally the internal `doapr_outch` function can attempt to write to an
5818 OOB memory location (at an offset from the NULL pointer) in the event of a
5819 memory allocation failure. In 1.0.2 and below this could be caused where
5820 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5821 could be in processing a very long "%s" format string. Memory leaks can
5824 The first issue may mask the second issue dependent on compiler behaviour.
5825 These problems could enable attacks where large amounts of untrusted data
5826 is passed to the `BIO_*printf` functions. If applications use these functions
5827 in this way then they could be vulnerable. OpenSSL itself uses these
5828 functions when printing out human-readable dumps of ASN.1 data. Therefore
5829 applications that print this data could be vulnerable if the data is from
5830 untrusted sources. OpenSSL command line applications could also be
5831 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5832 as command line arguments.
5834 Libssl is not considered directly vulnerable. Additionally certificates etc
5835 received via remote connections via libssl are also unlikely to be able to
5836 trigger these issues because of message size limits enforced within libssl.
5838 This issue was reported to OpenSSL Guido Vranken.
5843 * Side channel attack on modular exponentiation
5845 A side-channel attack was found which makes use of cache-bank conflicts on
5846 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5847 of RSA keys. The ability to exploit this issue is limited as it relies on
5848 an attacker who has control of code in a thread running on the same
5849 hyper-threaded core as the victim thread which is performing decryptions.
5851 This issue was reported to OpenSSL by Yuval Yarom, The University of
5852 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5853 Nadia Heninger, University of Pennsylvania with more information at
5854 <http://cachebleed.info>.
5859 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5860 if no keysize is specified with default_bits. This fixes an
5861 omission in an earlier change that changed all RSA/DSA key generation
5862 commands to use 2048 bits by default.
5866 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5868 * DH small subgroups
5870 Historically OpenSSL only ever generated DH parameters based on "safe"
5871 primes. More recently (in version 1.0.2) support was provided for
5872 generating X9.42 style parameter files such as those required for RFC 5114
5873 support. The primes used in such files may not be "safe". Where an
5874 application is using DH configured with parameters based on primes that are
5875 not "safe" then an attacker could use this fact to find a peer's private
5876 DH exponent. This attack requires that the attacker complete multiple
5877 handshakes in which the peer uses the same private DH exponent. For example
5878 this could be used to discover a TLS server's private DH exponent if it's
5879 reusing the private DH exponent or it's using a static DH ciphersuite.
5881 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5882 TLS. It is not on by default. If the option is not set then the server
5883 reuses the same private DH exponent for the life of the server process and
5884 would be vulnerable to this attack. It is believed that many popular
5885 applications do set this option and would therefore not be at risk.
5887 The fix for this issue adds an additional check where a "q" parameter is
5888 available (as is the case in X9.42 based parameters). This detects the
5889 only known attack, and is the only possible defense for static DH
5890 ciphersuites. This could have some performance impact.
5892 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5893 default and cannot be disabled. This could have some performance impact.
5895 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5900 * SSLv2 doesn't block disabled ciphers
5902 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5903 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5904 been disabled, provided that the SSLv2 protocol was not also disabled via
5907 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5908 and Sebastian Schinzel.
5913 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5915 * BN_mod_exp may produce incorrect results on x86_64
5917 There is a carry propagating bug in the x86_64 Montgomery squaring
5918 procedure. No EC algorithms are affected. Analysis suggests that attacks
5919 against RSA and DSA as a result of this defect would be very difficult to
5920 perform and are not believed likely. Attacks against DH are considered just
5921 feasible (although very difficult) because most of the work necessary to
5922 deduce information about a private key may be performed offline. The amount
5923 of resources required for such an attack would be very significant and
5924 likely only accessible to a limited number of attackers. An attacker would
5925 additionally need online access to an unpatched system using the target
5926 private key in a scenario with persistent DH parameters and a private
5927 key that is shared between multiple clients. For example this can occur by
5928 default in OpenSSL DHE based SSL/TLS ciphersuites.
5930 This issue was reported to OpenSSL by Hanno Böck.
5935 * Certificate verify crash with missing PSS parameter
5937 The signature verification routines will crash with a NULL pointer
5938 dereference if presented with an ASN.1 signature using the RSA PSS
5939 algorithm and absent mask generation function parameter. Since these
5940 routines are used to verify certificate signature algorithms this can be
5941 used to crash any certificate verification operation and exploited in a
5942 DoS attack. Any application which performs certificate verification is
5943 vulnerable including OpenSSL clients and servers which enable client
5946 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5951 * X509_ATTRIBUTE memory leak
5953 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5954 memory. This structure is used by the PKCS#7 and CMS routines so any
5955 application which reads PKCS#7 or CMS data from untrusted sources is
5956 affected. SSL/TLS is not affected.
5958 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5964 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5965 This changes the decoding behaviour for some invalid messages,
5966 though the change is mostly in the more lenient direction, and
5967 legacy behaviour is preserved as much as possible.
5971 * In DSA_generate_parameters_ex, if the provided seed is too short,
5974 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5976 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5978 * Alternate chains certificate forgery
5980 During certificate verification, OpenSSL will attempt to find an
5981 alternative certificate chain if the first attempt to build such a chain
5982 fails. An error in the implementation of this logic can mean that an
5983 attacker could cause certain checks on untrusted certificates to be
5984 bypassed, such as the CA flag, enabling them to use a valid leaf
5985 certificate to act as a CA and "issue" an invalid certificate.
5987 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5992 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5994 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5995 incompatibility in the handling of HMAC. The previous ABI has now been
6000 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6002 * Malformed ECParameters causes infinite loop
6004 When processing an ECParameters structure OpenSSL enters an infinite loop
6005 if the curve specified is over a specially malformed binary polynomial
6008 This can be used to perform denial of service against any
6009 system which processes public keys, certificate requests or
6010 certificates. This includes TLS clients and TLS servers with
6011 client authentication enabled.
6013 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6018 * Exploitable out-of-bounds read in X509_cmp_time
6020 X509_cmp_time does not properly check the length of the ASN1_TIME
6021 string and can read a few bytes out of bounds. In addition,
6022 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6025 An attacker can use this to craft malformed certificates and CRLs of
6026 various sizes and potentially cause a segmentation fault, resulting in
6027 a DoS on applications that verify certificates or CRLs. TLS clients
6028 that verify CRLs are affected. TLS clients and servers with client
6029 authentication enabled may be affected if they use custom verification
6032 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6033 independently by Hanno Böck.
6038 * PKCS7 crash with missing EnvelopedContent
6040 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6041 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6042 with missing content and trigger a NULL pointer dereference on parsing.
6044 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6045 structures from untrusted sources are affected. OpenSSL clients and
6046 servers are not affected.
6048 This issue was reported to OpenSSL by Michal Zalewski (Google).
6053 * CMS verify infinite loop with unknown hash function
6055 When verifying a signedData message the CMS code can enter an infinite loop
6056 if presented with an unknown hash function OID. This can be used to perform
6057 denial of service against any system which verifies signedData messages using
6059 This issue was reported to OpenSSL by Johannes Bauer.
6064 * Race condition handling NewSessionTicket
6066 If a NewSessionTicket is received by a multi-threaded client when attempting to
6067 reuse a previous ticket then a race condition can occur potentially leading to
6068 a double free of the ticket data.
6073 * Only support 256-bit or stronger elliptic curves with the
6074 'ecdh_auto' setting (server) or by default (client). Of supported
6075 curves, prefer P-256 (both).
6079 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6081 * ClientHello sigalgs DoS fix
6083 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6084 invalid signature algorithms extension a NULL pointer dereference will
6085 occur. This can be exploited in a DoS attack against the server.
6087 This issue was was reported to OpenSSL by David Ramos of Stanford
6091 *Stephen Henson and Matt Caswell*
6093 * Multiblock corrupted pointer fix
6095 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6096 feature only applies on 64 bit x86 architecture platforms that support AES
6097 NI instructions. A defect in the implementation of "multiblock" can cause
6098 OpenSSL's internal write buffer to become incorrectly set to NULL when
6099 using non-blocking IO. Typically, when the user application is using a
6100 socket BIO for writing, this will only result in a failed connection.
6101 However if some other BIO is used then it is likely that a segmentation
6102 fault will be triggered, thus enabling a potential DoS attack.
6104 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6109 * Segmentation fault in DTLSv1_listen fix
6111 The DTLSv1_listen function is intended to be stateless and processes the
6112 initial ClientHello from many peers. It is common for user code to loop
6113 over the call to DTLSv1_listen until a valid ClientHello is received with
6114 an associated cookie. A defect in the implementation of DTLSv1_listen means
6115 that state is preserved in the SSL object from one invocation to the next
6116 that can lead to a segmentation fault. Errors processing the initial
6117 ClientHello can trigger this scenario. An example of such an error could be
6118 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6121 This issue was reported to OpenSSL by Per Allansson.
6126 * Segmentation fault in ASN1_TYPE_cmp fix
6128 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6129 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6130 certificate signature algorithm consistency this can be used to crash any
6131 certificate verification operation and exploited in a DoS attack. Any
6132 application which performs certificate verification is vulnerable including
6133 OpenSSL clients and servers which enable client authentication.
6138 * Segmentation fault for invalid PSS parameters fix
6140 The signature verification routines will crash with a NULL pointer
6141 dereference if presented with an ASN.1 signature using the RSA PSS
6142 algorithm and invalid parameters. Since these routines are used to verify
6143 certificate signature algorithms this can be used to crash any
6144 certificate verification operation and exploited in a DoS attack. Any
6145 application which performs certificate verification is vulnerable including
6146 OpenSSL clients and servers which enable client authentication.
6148 This issue was was reported to OpenSSL by Brian Carpenter.
6153 * ASN.1 structure reuse memory corruption fix
6155 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6156 memory corruption via an invalid write. Such reuse is and has been
6157 strongly discouraged and is believed to be rare.
6159 Applications that parse structures containing CHOICE or ANY DEFINED BY
6160 components may be affected. Certificate parsing (d2i_X509 and related
6161 functions) are however not affected. OpenSSL clients and servers are
6167 * PKCS7 NULL pointer dereferences fix
6169 The PKCS#7 parsing code does not handle missing outer ContentInfo
6170 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6171 missing content and trigger a NULL pointer dereference on parsing.
6173 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6174 otherwise parse PKCS#7 structures from untrusted sources are
6175 affected. OpenSSL clients and servers are not affected.
6177 This issue was reported to OpenSSL by Michal Zalewski (Google).
6182 * DoS via reachable assert in SSLv2 servers fix
6184 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6185 servers that both support SSLv2 and enable export cipher suites by sending
6186 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6188 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6189 (OpenSSL development team).
6194 * Empty CKE with client auth and DHE fix
6196 If client auth is used then a server can seg fault in the event of a DHE
6197 ciphersuite being selected and a zero length ClientKeyExchange message
6198 being sent by the client. This could be exploited in a DoS attack.
6203 * Handshake with unseeded PRNG fix
6205 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6206 with an unseeded PRNG. The conditions are:
6207 - The client is on a platform where the PRNG has not been seeded
6208 automatically, and the user has not seeded manually
6209 - A protocol specific client method version has been used (i.e. not
6210 SSL_client_methodv23)
6211 - A ciphersuite is used that does not require additional random data from
6212 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6214 If the handshake succeeds then the client random that has been used will
6215 have been generated from a PRNG with insufficient entropy and therefore the
6216 output may be predictable.
6218 For example using the following command with an unseeded openssl will
6219 succeed on an unpatched platform:
6221 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6226 * Use After Free following d2i_ECPrivatekey error fix
6228 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6229 could cause a use after free condition. This, in turn, could cause a double
6230 free in several private key parsing functions (such as d2i_PrivateKey
6231 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6232 for applications that receive EC private keys from untrusted
6233 sources. This scenario is considered rare.
6235 This issue was discovered by the BoringSSL project and fixed in their
6241 * X509_to_X509_REQ NULL pointer deref fix
6243 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6244 the certificate key is invalid. This function is rarely used in practice.
6246 This issue was discovered by Brian Carpenter.
6251 * Removed the export ciphers from the DEFAULT ciphers
6255 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6257 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6258 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6259 So far those who have to target multiple platforms would compromise
6260 and argue that binary targeting say ARMv5 would still execute on
6261 ARMv8. "Universal" build resolves this compromise by providing
6262 near-optimal performance even on newer platforms.
6266 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6267 (other platforms pending).
6269 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6271 * Add support for the SignedCertificateTimestampList certificate and
6272 OCSP response extensions from RFC6962.
6276 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6277 for corner cases. (Certain input points at infinity could lead to
6278 bogus results, with non-infinity inputs mapped to infinity too.)
6282 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6283 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6284 common cases are optimized and there still is room for further
6285 improvements. Vector Permutation AES for Altivec is also added.
6289 * Add support for little-endian ppc64 Linux target.
6291 *Marcelo Cerri (IBM)*
6293 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6294 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6295 are optimized and there still is room for further improvements.
6296 Both 32- and 64-bit modes are supported.
6298 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6300 * Improved ARMv7 NEON support.
6304 * Support for SPARC Architecture 2011 crypto extensions, first
6305 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6306 SHA256/512, MD5, GHASH and modular exponentiation.
6308 *Andy Polyakov, David Miller*
6310 * Accelerated modular exponentiation for Intel processors, a.k.a.
6313 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6315 * Support for new and upcoming Intel processors, including AVX2,
6316 BMI and SHA ISA extensions. This includes additional "stitched"
6317 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6320 This work was sponsored by Intel Corp.
6324 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6325 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6326 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6330 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6331 this fixes a limitation in previous versions of OpenSSL.
6335 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6336 MGF1 digest and OAEP label.
6340 * Add EVP support for key wrapping algorithms, to avoid problems with
6341 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6342 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6343 algorithms and include tests cases.
6347 * Add functions to allocate and set the fields of an ECDSA_METHOD
6350 *Douglas E. Engert, Steve Henson*
6352 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6353 difference in days and seconds between two tm or ASN1_TIME structures.
6357 * Add -rev test option to s_server to just reverse order of characters
6358 received by client and send back to server. Also prints an abbreviated
6359 summary of the connection parameters.
6363 * New option -brief for s_client and s_server to print out a brief summary
6364 of connection parameters.
6368 * Add callbacks for arbitrary TLS extensions.
6370 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6372 * New option -crl_download in several openssl utilities to download CRLs
6373 from CRLDP extension in certificates.
6377 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6381 * New function X509_CRL_diff to generate a delta CRL from the difference
6382 of two full CRLs. Add support to "crl" utility.
6386 * New functions to set lookup_crls function and to retrieve
6387 X509_STORE from X509_STORE_CTX.
6391 * Print out deprecated issuer and subject unique ID fields in
6396 * Extend OCSP I/O functions so they can be used for simple general purpose
6397 HTTP as well as OCSP. New wrapper function which can be used to download
6398 CRLs using the OCSP API.
6402 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6406 * `SSL_CONF*` functions. These provide a common framework for application
6407 configuration using configuration files or command lines.
6411 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6412 message callback and prints the results. Needs compile time option
6413 "enable-ssl-trace". New options to s_client and s_server to enable
6418 * New ctrl and macro to retrieve supported points extensions.
6419 Print out extension in s_server and s_client.
6423 * New functions to retrieve certificate signature and signature
6428 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6433 * New Suite B modes for TLS code. These use and enforce the requirements
6434 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6435 only use Suite B curves. The Suite B modes can be set by using the
6436 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6440 * New chain verification flags for Suite B levels of security. Check
6441 algorithms are acceptable when flags are set in X509_verify_cert.
6445 * Make tls1_check_chain return a set of flags indicating checks passed
6446 by a certificate chain. Add additional tests to handle client
6447 certificates: checks for matching certificate type and issuer name
6452 * If an attempt is made to use a signature algorithm not in the peer
6453 preference list abort the handshake. If client has no suitable
6454 signature algorithms in response to a certificate request do not
6455 use the certificate.
6459 * If server EC tmp key is not in client preference list abort handshake.
6463 * Add support for certificate stores in CERT structure. This makes it
6464 possible to have different stores per SSL structure or one store in
6465 the parent SSL_CTX. Include distinct stores for certificate chain
6466 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6467 to build and store a certificate chain in CERT structure: returning
6468 an error if the chain cannot be built: this will allow applications
6469 to test if a chain is correctly configured.
6471 Note: if the CERT based stores are not set then the parent SSL_CTX
6472 store is used to retain compatibility with existing behaviour.
6476 * New function ssl_set_client_disabled to set a ciphersuite disabled
6477 mask based on the current session, check mask when sending client
6478 hello and checking the requested ciphersuite.
6482 * New ctrls to retrieve and set certificate types in a certificate
6483 request message. Print out received values in s_client. If certificate
6484 types is not set with custom values set sensible values based on
6485 supported signature algorithms.
6489 * Support for distinct client and server supported signature algorithms.
6493 * Add certificate callback. If set this is called whenever a certificate
6494 is required by client or server. An application can decide which
6495 certificate chain to present based on arbitrary criteria: for example
6496 supported signature algorithms. Add very simple example to s_server.
6497 This fixes many of the problems and restrictions of the existing client
6498 certificate callback: for example you can now clear an existing
6499 certificate and specify the whole chain.
6503 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6504 the certificate can be used for (if anything). Set valid_flags field
6505 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6506 to have similar checks in it.
6508 Add new "cert_flags" field to CERT structure and include a "strict mode".
6509 This enforces some TLS certificate requirements (such as only permitting
6510 certificate signature algorithms contained in the supported algorithms
6511 extension) which some implementations ignore: this option should be used
6512 with caution as it could cause interoperability issues.
6516 * Update and tidy signature algorithm extension processing. Work out
6517 shared signature algorithms based on preferences and peer algorithms
6518 and print them out in s_client and s_server. Abort handshake if no
6519 shared signature algorithms.
6523 * Add new functions to allow customised supported signature algorithms
6524 for SSL and SSL_CTX structures. Add options to s_client and s_server
6529 * New function SSL_certs_clear() to delete all references to certificates
6530 from an SSL structure. Before this once a certificate had been added
6531 it couldn't be removed.
6535 * Integrate hostname, email address and IP address checking with certificate
6536 verification. New verify options supporting checking in openssl utility.
6540 * Fixes and wildcard matching support to hostname and email checking
6541 functions. Add manual page.
6543 *Florian Weimer (Red Hat Product Security Team)*
6545 * New functions to check a hostname email or IP address against a
6546 certificate. Add options x509 utility to print results of checks against
6551 * Fix OCSP checking.
6553 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6555 * Initial experimental support for explicitly trusted non-root CAs.
6556 OpenSSL still tries to build a complete chain to a root but if an
6557 intermediate CA has a trust setting included that is used. The first
6558 setting is used: whether to trust (e.g., -addtrust option to the x509
6563 * Add -trusted_first option which attempts to find certificates in the
6564 trusted store even if an untrusted chain is also supplied.
6568 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6569 platform support for Linux and Android.
6573 * Support for linux-x32, ILP32 environment in x86_64 framework.
6577 * Experimental multi-implementation support for FIPS capable OpenSSL.
6578 When in FIPS mode the approved implementations are used as normal,
6579 when not in FIPS mode the internal unapproved versions are used instead.
6580 This means that the FIPS capable OpenSSL isn't forced to use the
6581 (often lower performance) FIPS implementations outside FIPS mode.
6585 * Transparently support X9.42 DH parameters when calling
6586 PEM_read_bio_DHparameters. This means existing applications can handle
6587 the new parameter format automatically.
6591 * Initial experimental support for X9.42 DH parameter format: mainly
6592 to support use of 'q' parameter for RFC5114 parameters.
6596 * Add DH parameters from RFC5114 including test data to dhtest.
6600 * Support for automatic EC temporary key parameter selection. If enabled
6601 the most preferred EC parameters are automatically used instead of
6602 hardcoded fixed parameters. Now a server just has to call:
6603 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6604 support ECDH and use the most appropriate parameters.
6608 * Enhance and tidy EC curve and point format TLS extension code. Use
6609 static structures instead of allocation if default values are used.
6610 New ctrls to set curves we wish to support and to retrieve shared curves.
6611 Print out shared curves in s_server. New options to s_server and s_client
6612 to set list of supported curves.
6616 * New ctrls to retrieve supported signature algorithms and
6617 supported curve values as an array of NIDs. Extend openssl utility
6618 to print out received values.
6622 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6623 between NIDs and the more common NIST names such as "P-256". Enhance
6624 ecparam utility and ECC method to recognise the NIST names for curves.
6628 * Enhance SSL/TLS certificate chain handling to support different
6629 chains for each certificate instead of one chain in the parent SSL_CTX.
6633 * Support for fixed DH ciphersuite client authentication: where both
6634 server and client use DH certificates with common parameters.
6638 * Support for fixed DH ciphersuites: those requiring DH server
6643 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6645 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6646 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6647 X509_CINF_get_signature were reverted post internal team review.
6652 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6654 * OCSP Status Request extension unbounded memory growth
6656 A malicious client can send an excessively large OCSP Status Request
6657 extension. If that client continually requests renegotiation, sending a
6658 large OCSP Status Request extension each time, then there will be unbounded
6659 memory growth on the server. This will eventually lead to a Denial Of
6660 Service attack through memory exhaustion. Servers with a default
6661 configuration are vulnerable even if they do not support OCSP. Builds using
6662 the "no-ocsp" build time option are not affected.
6664 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6669 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6672 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6678 * OOB write in MDC2_Update()
6680 An overflow can occur in MDC2_Update() either if called directly or
6681 through the EVP_DigestUpdate() function using MDC2. If an attacker
6682 is able to supply very large amounts of input data after a previous
6683 call to EVP_EncryptUpdate() with a partial block then a length check
6684 can overflow resulting in a heap corruption.
6686 The amount of data needed is comparable to SIZE_MAX which is impractical
6689 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6694 * Malformed SHA512 ticket DoS
6696 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6697 DoS attack where a malformed ticket will result in an OOB read which will
6700 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6701 a custom server callback and ticket lookup mechanism.
6703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6708 * OOB write in BN_bn2dec()
6710 The function BN_bn2dec() does not check the return value of BN_div_word().
6711 This can cause an OOB write if an application uses this function with an
6712 overly large BIGNUM. This could be a problem if an overly large certificate
6713 or CRL is printed out from an untrusted source. TLS is not affected because
6714 record limits will reject an oversized certificate before it is parsed.
6716 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6721 * OOB read in TS_OBJ_print_bio()
6723 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6724 the total length the OID text representation would use and not the amount
6725 of data written. This will result in OOB reads when large OIDs are
6728 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6733 * Pointer arithmetic undefined behaviour
6735 Avoid some undefined pointer arithmetic
6737 A common idiom in the codebase is to check limits in the following manner:
6740 Where "p" points to some malloc'd data of SIZE bytes and
6743 "len" here could be from some externally supplied data (e.g. from a TLS
6746 The rules of C pointer arithmetic are such that "p + len" is only well
6747 defined where len <= SIZE. Therefore, the above idiom is actually
6748 undefined behaviour.
6750 For example this could cause problems if some malloc implementation
6751 provides an address for "p" such that "p + len" actually overflows for
6752 values of len that are too big and therefore p + len < limit.
6754 This issue was reported to OpenSSL by Guido Vranken
6759 * Constant time flag not preserved in DSA signing
6761 Operations in the DSA signing algorithm should run in constant time in
6762 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6763 implementation means that a non-constant time codepath is followed for
6764 certain operations. This has been demonstrated through a cache-timing
6765 attack to be sufficient for an attacker to recover the private DSA key.
6767 This issue was reported by César Pereida (Aalto University), Billy Brumley
6768 (Tampere University of Technology), and Yuval Yarom (The University of
6769 Adelaide and NICTA).
6774 * DTLS buffered message DoS
6776 In a DTLS connection where handshake messages are delivered out-of-order
6777 those messages that OpenSSL is not yet ready to process will be buffered
6778 for later use. Under certain circumstances, a flaw in the logic means that
6779 those messages do not get removed from the buffer even though the handshake
6780 has been completed. An attacker could force up to approx. 15 messages to
6781 remain in the buffer when they are no longer required. These messages will
6782 be cleared when the DTLS connection is closed. The default maximum size for
6783 a message is 100k. Therefore, the attacker could force an additional 1500k
6784 to be consumed per connection. By opening many simultaneous connections an
6785 attacker could cause a DoS attack through memory exhaustion.
6787 This issue was reported to OpenSSL by Quan Luo.
6792 * DTLS replay protection DoS
6794 A flaw in the DTLS replay attack protection mechanism means that records
6795 that arrive for future epochs update the replay protection "window" before
6796 the MAC for the record has been validated. This could be exploited by an
6797 attacker by sending a record for the next epoch (which does not have to
6798 decrypt or have a valid MAC), with a very large sequence number. This means
6799 that all subsequent legitimate packets are dropped causing a denial of
6800 service for a specific DTLS connection.
6802 This issue was reported to OpenSSL by the OCAP audit team.
6807 * Certificate message OOB reads
6809 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6810 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6811 theoretical DoS risk but this has not been observed in practice on common
6814 The messages affected are client certificate, client certificate request
6815 and server certificate. As a result the attack can only be performed
6816 against a client or a server which enables client authentication.
6818 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6823 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6825 * Prevent padding oracle in AES-NI CBC MAC check
6827 A MITM attacker can use a padding oracle attack to decrypt traffic
6828 when the connection uses an AES CBC cipher and the server support
6831 This issue was introduced as part of the fix for Lucky 13 padding
6832 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6833 constant time by making sure that always the same bytes are read and
6834 compared against either the MAC or padding bytes. But it no longer
6835 checked that there was enough data to have both the MAC and padding
6838 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6843 * Fix EVP_EncodeUpdate overflow
6845 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6846 Base64 encoding of binary data. If an attacker is able to supply very large
6847 amounts of input data then a length check can overflow resulting in a heap
6850 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6851 the `PEM_write_bio*` family of functions. These are mainly used within the
6852 OpenSSL command line applications, so any application which processes data
6853 from an untrusted source and outputs it as a PEM file should be considered
6854 vulnerable to this issue. User applications that call these APIs directly
6855 with large amounts of untrusted data may also be vulnerable.
6857 This issue was reported by Guido Vranken.
6862 * Fix EVP_EncryptUpdate overflow
6864 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6865 is able to supply very large amounts of input data after a previous call to
6866 EVP_EncryptUpdate() with a partial block then a length check can overflow
6867 resulting in a heap corruption. Following an analysis of all OpenSSL
6868 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6869 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6870 the first called function after an EVP_EncryptInit(), and therefore that
6871 specific call must be safe. The second form is where the length passed to
6872 EVP_EncryptUpdate() can be seen from the code to be some small value and
6873 therefore there is no possibility of an overflow. Since all instances are
6874 one of these two forms, it is believed that there can be no overflows in
6875 internal code due to this problem. It should be noted that
6876 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6877 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6878 of these calls have also been analysed too and it is believed there are no
6879 instances in internal usage where an overflow could occur.
6881 This issue was reported by Guido Vranken.
6886 * Prevent ASN.1 BIO excessive memory allocation
6888 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6889 a short invalid encoding can casuse allocation of large amounts of memory
6890 potentially consuming excessive resources or exhausting memory.
6892 Any application parsing untrusted data through d2i BIO functions is
6893 affected. The memory based functions such as d2i_X509() are *not* affected.
6894 Since the memory based functions are used by the TLS library, TLS
6895 applications are not affected.
6897 This issue was reported by Brian Carpenter.
6904 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6905 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6906 in arbitrary stack data being returned in the buffer.
6908 This issue was reported by Guido Vranken.
6913 * Modify behavior of ALPN to invoke callback after SNI/servername
6914 callback, such that updates to the SSL_CTX affect ALPN.
6918 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6923 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6924 methods are enabled and ssl2 is disabled the methods return NULL.
6928 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6930 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6931 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6932 provide any "EXPORT" or "LOW" strength ciphers.
6936 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6937 is by default disabled at build-time. Builds that are not configured with
6938 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6939 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6940 will need to explicitly call either of:
6942 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6944 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6946 as appropriate. Even if either of those is used, or the application
6947 explicitly uses the version-specific SSLv2_method() or its client and
6948 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6949 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6950 ciphers, and SSLv2 56-bit DES are no longer available.
6955 * Fix a double-free in DSA code
6957 A double free bug was discovered when OpenSSL parses malformed DSA private
6958 keys and could lead to a DoS attack or memory corruption for applications
6959 that receive DSA private keys from untrusted sources. This scenario is
6962 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6968 * Disable SRP fake user seed to address a server memory leak.
6970 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6972 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6973 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6974 was changed to ignore the "fake user" SRP seed, even if the seed
6977 Users should use SRP_VBASE_get1_by_user instead. Note that in
6978 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6979 also that even though configuring the SRP seed attempts to hide
6980 invalid usernames by continuing the handshake with fake
6981 credentials, this behaviour is not constant time and no strong
6982 guarantees are made that the handshake is indistinguishable from
6983 that of a valid user.
6988 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6990 In the BN_hex2bn function the number of hex digits is calculated using an
6991 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6992 large values of `i` this can result in `bn_expand` not allocating any
6993 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6994 field as NULL leading to a subsequent NULL ptr deref. For very large values
6995 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6996 In this case memory is allocated to the internal BIGNUM data field, but it
6997 is insufficiently sized leading to heap corruption. A similar issue exists
6998 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6999 is ever called by user applications with very large untrusted hex/dec data.
7000 This is anticipated to be a rare occurrence.
7002 All OpenSSL internal usage of these functions use data that is not expected
7003 to be untrusted, e.g. config file data or application command line
7004 arguments. If user developed applications generate config file data based
7005 on untrusted data then it is possible that this could also lead to security
7006 consequences. This is also anticipated to be rare.
7008 This issue was reported to OpenSSL by Guido Vranken.
7013 * Fix memory issues in `BIO_*printf` functions
7015 The internal `fmtstr` function used in processing a "%s" format string in
7016 the `BIO_*printf` functions could overflow while calculating the length of a
7017 string and cause an OOB read when printing very long strings.
7019 Additionally the internal `doapr_outch` function can attempt to write to an
7020 OOB memory location (at an offset from the NULL pointer) in the event of a
7021 memory allocation failure. In 1.0.2 and below this could be caused where
7022 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7023 could be in processing a very long "%s" format string. Memory leaks can
7026 The first issue may mask the second issue dependent on compiler behaviour.
7027 These problems could enable attacks where large amounts of untrusted data
7028 is passed to the `BIO_*printf` functions. If applications use these functions
7029 in this way then they could be vulnerable. OpenSSL itself uses these
7030 functions when printing out human-readable dumps of ASN.1 data. Therefore
7031 applications that print this data could be vulnerable if the data is from
7032 untrusted sources. OpenSSL command line applications could also be
7033 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7034 as command line arguments.
7036 Libssl is not considered directly vulnerable. Additionally certificates etc
7037 received via remote connections via libssl are also unlikely to be able to
7038 trigger these issues because of message size limits enforced within libssl.
7040 This issue was reported to OpenSSL Guido Vranken.
7045 * Side channel attack on modular exponentiation
7047 A side-channel attack was found which makes use of cache-bank conflicts on
7048 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7049 of RSA keys. The ability to exploit this issue is limited as it relies on
7050 an attacker who has control of code in a thread running on the same
7051 hyper-threaded core as the victim thread which is performing decryptions.
7053 This issue was reported to OpenSSL by Yuval Yarom, The University of
7054 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7055 Nadia Heninger, University of Pennsylvania with more information at
7056 <http://cachebleed.info>.
7061 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7062 if no keysize is specified with default_bits. This fixes an
7063 omission in an earlier change that changed all RSA/DSA key generation
7064 commands to use 2048 bits by default.
7068 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7070 * Protection for DH small subgroup attacks
7072 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7073 switched on by default and cannot be disabled. This could have some
7078 * SSLv2 doesn't block disabled ciphers
7080 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7081 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7082 been disabled, provided that the SSLv2 protocol was not also disabled via
7085 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7086 and Sebastian Schinzel.
7091 * Reject DH handshakes with parameters shorter than 1024 bits.
7095 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7097 * Certificate verify crash with missing PSS parameter
7099 The signature verification routines will crash with a NULL pointer
7100 dereference if presented with an ASN.1 signature using the RSA PSS
7101 algorithm and absent mask generation function parameter. Since these
7102 routines are used to verify certificate signature algorithms this can be
7103 used to crash any certificate verification operation and exploited in a
7104 DoS attack. Any application which performs certificate verification is
7105 vulnerable including OpenSSL clients and servers which enable client
7108 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7113 * X509_ATTRIBUTE memory leak
7115 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7116 memory. This structure is used by the PKCS#7 and CMS routines so any
7117 application which reads PKCS#7 or CMS data from untrusted sources is
7118 affected. SSL/TLS is not affected.
7120 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7126 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7127 This changes the decoding behaviour for some invalid messages,
7128 though the change is mostly in the more lenient direction, and
7129 legacy behaviour is preserved as much as possible.
7133 * In DSA_generate_parameters_ex, if the provided seed is too short,
7134 use a random seed, as already documented.
7136 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7138 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7140 * Alternate chains certificate forgery
7142 During certificate verfification, OpenSSL will attempt to find an
7143 alternative certificate chain if the first attempt to build such a chain
7144 fails. An error in the implementation of this logic can mean that an
7145 attacker could cause certain checks on untrusted certificates to be
7146 bypassed, such as the CA flag, enabling them to use a valid leaf
7147 certificate to act as a CA and "issue" an invalid certificate.
7149 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7155 * Race condition handling PSK identify hint
7157 If PSK identity hints are received by a multi-threaded client then
7158 the values are wrongly updated in the parent SSL_CTX structure. This can
7159 result in a race condition potentially leading to a double free of the
7165 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7167 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7168 incompatibility in the handling of HMAC. The previous ABI has now been
7171 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7173 * Malformed ECParameters causes infinite loop
7175 When processing an ECParameters structure OpenSSL enters an infinite loop
7176 if the curve specified is over a specially malformed binary polynomial
7179 This can be used to perform denial of service against any
7180 system which processes public keys, certificate requests or
7181 certificates. This includes TLS clients and TLS servers with
7182 client authentication enabled.
7184 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7189 * Exploitable out-of-bounds read in X509_cmp_time
7191 X509_cmp_time does not properly check the length of the ASN1_TIME
7192 string and can read a few bytes out of bounds. In addition,
7193 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7196 An attacker can use this to craft malformed certificates and CRLs of
7197 various sizes and potentially cause a segmentation fault, resulting in
7198 a DoS on applications that verify certificates or CRLs. TLS clients
7199 that verify CRLs are affected. TLS clients and servers with client
7200 authentication enabled may be affected if they use custom verification
7203 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7204 independently by Hanno Böck.
7209 * PKCS7 crash with missing EnvelopedContent
7211 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7212 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7213 with missing content and trigger a NULL pointer dereference on parsing.
7215 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7216 structures from untrusted sources are affected. OpenSSL clients and
7217 servers are not affected.
7219 This issue was reported to OpenSSL by Michal Zalewski (Google).
7224 * CMS verify infinite loop with unknown hash function
7226 When verifying a signedData message the CMS code can enter an infinite loop
7227 if presented with an unknown hash function OID. This can be used to perform
7228 denial of service against any system which verifies signedData messages using
7230 This issue was reported to OpenSSL by Johannes Bauer.
7235 * Race condition handling NewSessionTicket
7237 If a NewSessionTicket is received by a multi-threaded client when attempting to
7238 reuse a previous ticket then a race condition can occur potentially leading to
7239 a double free of the ticket data.
7244 * Reject DH handshakes with parameters shorter than 768 bits.
7246 *Kurt Roeckx and Emilia Kasper*
7248 * dhparam: generate 2048-bit parameters by default.
7250 *Kurt Roeckx and Emilia Kasper*
7252 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7254 * Segmentation fault in ASN1_TYPE_cmp fix
7256 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7257 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7258 certificate signature algorithm consistency this can be used to crash any
7259 certificate verification operation and exploited in a DoS attack. Any
7260 application which performs certificate verification is vulnerable including
7261 OpenSSL clients and servers which enable client authentication.
7266 * ASN.1 structure reuse memory corruption fix
7268 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7269 memory corruption via an invalid write. Such reuse is and has been
7270 strongly discouraged and is believed to be rare.
7272 Applications that parse structures containing CHOICE or ANY DEFINED BY
7273 components may be affected. Certificate parsing (d2i_X509 and related
7274 functions) are however not affected. OpenSSL clients and servers are
7280 * PKCS7 NULL pointer dereferences fix
7282 The PKCS#7 parsing code does not handle missing outer ContentInfo
7283 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7284 missing content and trigger a NULL pointer dereference on parsing.
7286 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7287 otherwise parse PKCS#7 structures from untrusted sources are
7288 affected. OpenSSL clients and servers are not affected.
7290 This issue was reported to OpenSSL by Michal Zalewski (Google).
7295 * DoS via reachable assert in SSLv2 servers fix
7297 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7298 servers that both support SSLv2 and enable export cipher suites by sending
7299 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7301 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7302 (OpenSSL development team).
7307 * Use After Free following d2i_ECPrivatekey error fix
7309 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7310 could cause a use after free condition. This, in turn, could cause a double
7311 free in several private key parsing functions (such as d2i_PrivateKey
7312 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7313 for applications that receive EC private keys from untrusted
7314 sources. This scenario is considered rare.
7316 This issue was discovered by the BoringSSL project and fixed in their
7322 * X509_to_X509_REQ NULL pointer deref fix
7324 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7325 the certificate key is invalid. This function is rarely used in practice.
7327 This issue was discovered by Brian Carpenter.
7332 * Removed the export ciphers from the DEFAULT ciphers
7336 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7338 * Build fixes for the Windows and OpenVMS platforms
7340 *Matt Caswell and Richard Levitte*
7342 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7344 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7345 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7346 dereference. This could lead to a Denial Of Service attack. Thanks to
7347 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7352 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7353 dtls1_buffer_record function under certain conditions. In particular this
7354 could occur if an attacker sent repeated DTLS records with the same
7355 sequence number but for the next epoch. The memory leak could be exploited
7356 by an attacker in a Denial of Service attack through memory exhaustion.
7357 Thanks to Chris Mueller for reporting this issue.
7362 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7363 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7364 method would be set to NULL which could later result in a NULL pointer
7365 dereference. Thanks to Frank Schmirler for reporting this issue.
7370 * Abort handshake if server key exchange message is omitted for ephemeral
7373 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7374 reporting this issue.
7379 * Remove non-export ephemeral RSA code on client and server. This code
7380 violated the TLS standard by allowing the use of temporary RSA keys in
7381 non-export ciphersuites and could be used by a server to effectively
7382 downgrade the RSA key length used to a value smaller than the server
7383 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7384 INRIA or reporting this issue.
7389 * Fixed issue where DH client certificates are accepted without verification.
7390 An OpenSSL server will accept a DH certificate for client authentication
7391 without the certificate verify message. This effectively allows a client to
7392 authenticate without the use of a private key. This only affects servers
7393 which trust a client certificate authority which issues certificates
7394 containing DH keys: these are extremely rare and hardly ever encountered.
7395 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7401 * Ensure that the session ID context of an SSL is updated when its
7402 SSL_CTX is updated via SSL_set_SSL_CTX.
7404 The session ID context is typically set from the parent SSL_CTX,
7405 and can vary with the CTX.
7409 * Fix various certificate fingerprint issues.
7411 By using non-DER or invalid encodings outside the signed portion of a
7412 certificate the fingerprint can be changed without breaking the signature.
7413 Although no details of the signed portion of the certificate can be changed
7414 this can cause problems with some applications: e.g. those using the
7415 certificate fingerprint for blacklists.
7417 1. Reject signatures with non zero unused bits.
7419 If the BIT STRING containing the signature has non zero unused bits reject
7420 the signature. All current signature algorithms require zero unused bits.
7422 2. Check certificate algorithm consistency.
7424 Check the AlgorithmIdentifier inside TBS matches the one in the
7425 certificate signature. NB: this will result in signature failure
7426 errors for some broken certificates.
7428 Thanks to Konrad Kraszewski from Google for reporting this issue.
7430 3. Check DSA/ECDSA signatures use DER.
7432 Re-encode DSA/ECDSA signatures and compare with the original received
7433 signature. Return an error if there is a mismatch.
7435 This will reject various cases including garbage after signature
7436 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7437 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7438 (negative or with leading zeroes).
7440 Further analysis was conducted and fixes were developed by Stephen Henson
7441 of the OpenSSL core team.
7447 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7448 results on some platforms, including x86_64. This bug occurs at random
7449 with a very low probability, and is not known to be exploitable in any
7450 way, though its exact impact is difficult to determine. Thanks to Pieter
7451 Wuille (Blockstream) who reported this issue and also suggested an initial
7452 fix. Further analysis was conducted by the OpenSSL development team and
7453 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7454 the OpenSSL core team.
7459 * Do not resume sessions on the server if the negotiated protocol
7460 version does not match the session's version. Resuming with a different
7461 version, while not strictly forbidden by the RFC, is of questionable
7462 sanity and breaks all known clients.
7464 *David Benjamin, Emilia Käsper*
7466 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7467 early CCS messages during renegotiation. (Note that because
7468 renegotiation is encrypted, this early CCS was not exploitable.)
7472 * Tighten client-side session ticket handling during renegotiation:
7473 ensure that the client only accepts a session ticket if the server sends
7474 the extension anew in the ServerHello. Previously, a TLS client would
7475 reuse the old extension state and thus accept a session ticket if one was
7476 announced in the initial ServerHello.
7478 Similarly, ensure that the client requires a session ticket if one
7479 was advertised in the ServerHello. Previously, a TLS client would
7480 ignore a missing NewSessionTicket message.
7484 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7488 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7489 sends a carefully crafted handshake message, to cause OpenSSL to fail
7490 to free up to 64k of memory causing a memory leak. This could be
7491 exploited in a Denial Of Service attack. This issue affects OpenSSL
7492 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7493 whether SRTP is used or configured. Implementations of OpenSSL that
7494 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7496 The fix was developed by the OpenSSL team.
7501 * Session Ticket Memory Leak.
7503 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7504 integrity of that ticket is first verified. In the event of a session
7505 ticket integrity check failing, OpenSSL will fail to free memory
7506 causing a memory leak. By sending a large number of invalid session
7507 tickets an attacker could exploit this issue in a Denial Of Service
7513 * Build option no-ssl3 is incomplete.
7515 When OpenSSL is configured with "no-ssl3" as a build option, servers
7516 could accept and complete an SSL 3.0 handshake, and clients could be
7517 configured to send them.
7520 *Akamai and the OpenSSL team*
7522 * Add support for TLS_FALLBACK_SCSV.
7523 Client applications doing fallback retries should call
7524 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7527 *Adam Langley, Bodo Moeller*
7529 * Add additional DigestInfo checks.
7531 Re-encode DigestInto in DER and check against the original when
7532 verifying RSA signature: this will reject any improperly encoded
7533 DigestInfo structures.
7535 Note: this is a precautionary measure and no attacks are currently known.
7539 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7541 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7542 SRP code can be overrun an internal buffer. Add sanity check that
7543 g, A, B < N to SRP code.
7545 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7546 Group for discovering this issue.
7551 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7552 TLS 1.0 instead of higher protocol versions when the ClientHello message
7553 is badly fragmented. This allows a man-in-the-middle attacker to force a
7554 downgrade to TLS 1.0 even if both the server and the client support a
7555 higher protocol version, by modifying the client's TLS records.
7557 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7558 researching this issue.
7563 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7564 to a denial of service attack. A malicious server can crash the client
7565 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7566 ciphersuite and sending carefully crafted handshake messages.
7568 Thanks to Felix Gröbert (Google) for discovering and researching this
7574 * By sending carefully crafted DTLS packets an attacker could cause openssl
7575 to leak memory. This can be exploited through a Denial of Service attack.
7576 Thanks to Adam Langley for discovering and researching this issue.
7581 * An attacker can force openssl to consume large amounts of memory whilst
7582 processing DTLS handshake messages. This can be exploited through a
7583 Denial of Service attack.
7584 Thanks to Adam Langley for discovering and researching this issue.
7589 * An attacker can force an error condition which causes openssl to crash
7590 whilst processing DTLS packets due to memory being freed twice. This
7591 can be exploited through a Denial of Service attack.
7592 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7598 * If a multithreaded client connects to a malicious server using a resumed
7599 session and the server sends an ec point format extension it could write
7600 up to 255 bytes to freed memory.
7602 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7608 * A malicious server can crash an OpenSSL client with a null pointer
7609 dereference (read) by specifying an SRP ciphersuite even though it was not
7610 properly negotiated with the client. This can be exploited through a
7611 Denial of Service attack.
7613 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7614 discovering and researching this issue.
7619 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7620 X509_name_oneline, X509_name_print_ex et al. to leak some information
7621 from the stack. Applications may be affected if they echo pretty printing
7622 output to the attacker.
7624 Thanks to Ivan Fratric (Google) for discovering this issue.
7627 *Emilia Käsper, and Steve Henson*
7629 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7630 for corner cases. (Certain input points at infinity could lead to
7631 bogus results, with non-infinity inputs mapped to infinity too.)
7635 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7637 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7638 handshake can force the use of weak keying material in OpenSSL
7639 SSL/TLS clients and servers.
7641 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7642 researching this issue. ([CVE-2014-0224])
7644 *KIKUCHI Masashi, Steve Henson*
7646 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7647 OpenSSL DTLS client the code can be made to recurse eventually crashing
7650 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7653 *Imre Rad, Steve Henson*
7655 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7656 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7657 client or server. This is potentially exploitable to run arbitrary
7658 code on a vulnerable client or server.
7660 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7662 *Jüri Aedla, Steve Henson*
7664 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7665 are subject to a denial of service attack.
7667 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7668 this issue. ([CVE-2014-3470])
7670 *Felix Gröbert, Ivan Fratric, Steve Henson*
7672 * Harmonize version and its documentation. -f flag is used to display
7675 *mancha <mancha1@zoho.com>*
7677 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7678 in i2d_ECPrivateKey.
7680 *mancha <mancha1@zoho.com>*
7682 * Fix some double frees. These are not thought to be exploitable.
7684 *mancha <mancha1@zoho.com>*
7686 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7688 * A missing bounds check in the handling of the TLS heartbeat extension
7689 can be used to reveal up to 64k of memory to a connected client or
7692 Thanks for Neel Mehta of Google Security for discovering this bug and to
7693 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7694 preparing the fix ([CVE-2014-0160])
7696 *Adam Langley, Bodo Moeller*
7698 * Fix for the attack described in the paper "Recovering OpenSSL
7699 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7700 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7701 <http://eprint.iacr.org/2014/140>
7703 Thanks to Yuval Yarom and Naomi Benger for discovering this
7704 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7706 *Yuval Yarom and Naomi Benger*
7708 * TLS pad extension: draft-agl-tls-padding-03
7710 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7711 TLS client Hello record length value would otherwise be > 255 and
7712 less that 512 pad with a dummy extension containing zeroes so it
7713 is at least 512 bytes long.
7715 *Adam Langley, Steve Henson*
7717 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7719 * Fix for TLS record tampering bug. A carefully crafted invalid
7720 handshake could crash OpenSSL with a NULL pointer exception.
7721 Thanks to Anton Johansson for reporting this issues.
7724 * Keep original DTLS digest and encryption contexts in retransmission
7725 structures so we can use the previous session parameters if they need
7726 to be resent. ([CVE-2013-6450])
7730 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7731 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7732 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7733 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7734 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7735 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7737 *Rob Stradling, Adam Langley*
7739 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7741 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7742 supporting platforms or when small records were transferred.
7744 *Andy Polyakov, Steve Henson*
7746 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7748 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7750 This addresses the flaw in CBC record processing discovered by
7751 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7752 at: <http://www.isg.rhul.ac.uk/tls/>
7754 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7755 Security Group at Royal Holloway, University of London
7756 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7757 Emilia Käsper for the initial patch.
7760 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7762 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7763 ciphersuites which can be exploited in a denial of service attack.
7764 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7765 and detecting this bug and to Wolfgang Ettlinger
7766 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7771 * Return an error when checking OCSP signatures when key is NULL.
7772 This fixes a DoS attack. ([CVE-2013-0166])
7776 * Make openssl verify return errors.
7778 *Chris Palmer <palmer@google.com> and Ben Laurie*
7780 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7781 the right response is stapled. Also change SSL_get_certificate()
7782 so it returns the certificate actually sent.
7783 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7785 *Rob Stradling <rob.stradling@comodo.com>*
7787 * Fix possible deadlock when decoding public keys.
7791 * Don't use TLS 1.0 record version number in initial client hello
7796 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7798 * Sanity check record length before skipping explicit IV in TLS
7799 1.2, 1.1 and DTLS to fix DoS attack.
7801 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7802 fuzzing as a service testing platform.
7807 * Initialise tkeylen properly when encrypting CMS messages.
7808 Thanks to Solar Designer of Openwall for reporting this issue.
7812 * In FIPS mode don't try to use composite ciphers as they are not
7817 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7819 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7820 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7821 mean any application compiled against OpenSSL 1.0.0 headers setting
7822 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7823 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7824 0x10000000L Any application which was previously compiled against
7825 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7826 will need to be recompiled as a result. Letting be results in
7827 inability to disable specifically TLS 1.1 and in client context,
7828 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7832 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7833 disable just protocol X, but all protocols above X *if* there are
7834 protocols *below* X still enabled. In more practical terms it means
7835 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7836 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7837 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7842 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7844 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7845 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7846 in CRYPTO_realloc_clean.
7848 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7849 issue and to Adam Langley <agl@chromium.org> for fixing it.
7852 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7854 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7858 * Workarounds for some broken servers that "hang" if a client hello
7859 record length exceeds 255 bytes.
7861 1. Do not use record version number > TLS 1.0 in initial client
7862 hello: some (but not all) hanging servers will now work.
7863 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7864 the number of ciphers sent in the client hello. This should be
7865 set to an even number, such as 50, for example by passing:
7866 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7867 Most broken servers should now work.
7868 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7869 TLS 1.2 client support entirely.
7873 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7877 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7879 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7880 STRING form instead of a DigestInfo.
7884 * The format used for MDC2 RSA signatures is inconsistent between EVP
7885 and the RSA_sign/RSA_verify functions. This was made more apparent when
7886 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7887 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7888 the correct format in RSA_verify so both forms transparently work.
7892 * Some servers which support TLS 1.0 can choke if we initially indicate
7893 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7894 encrypted premaster secret. As a workaround use the maximum permitted
7895 client version in client hello, this should keep such servers happy
7896 and still work with previous versions of OpenSSL.
7900 * Add support for TLS/DTLS heartbeats.
7902 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7904 * Add support for SCTP.
7906 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7908 * Improved PRNG seeding for VOS.
7910 *Paul Green <Paul.Green@stratus.com>*
7912 * Extensive assembler packs updates, most notably:
7914 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7915 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7916 - x86_64: bit-sliced AES implementation;
7917 - ARM: NEON support, contemporary platforms optimizations;
7918 - s390x: z196 support;
7919 - `*`: GHASH and GF(2^m) multiplication implementations;
7923 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7924 (removal of unnecessary code)
7926 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7928 * Add TLS key material exporter from RFC 5705.
7932 * Add DTLS-SRTP negotiation from RFC 5764.
7936 * Add Next Protocol Negotiation,
7937 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7938 disabled with a no-npn flag to config or Configure. Code donated
7941 *Adam Langley <agl@google.com> and Ben Laurie*
7943 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7944 NIST-P256, NIST-P521, with constant-time single point multiplication on
7945 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7946 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7947 Code made available under Apache License version 2.0.
7949 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7950 line to include this in your build of OpenSSL, and run "make depend" (or
7951 "make update"). This enables the following EC_METHODs:
7953 EC_GFp_nistp224_method()
7954 EC_GFp_nistp256_method()
7955 EC_GFp_nistp521_method()
7957 EC_GROUP_new_by_curve_name() will automatically use these (while
7958 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7961 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7963 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7964 all platforms. Move ssize_t definition from e_os.h to the public
7965 header file e_os2.h as it now appears in public header file cms.h
7969 * New -sigopt option to the ca, req and x509 utilities. Additional
7970 signature parameters can be passed using this option and in
7975 * Add RSA PSS signing function. This will generate and set the
7976 appropriate AlgorithmIdentifiers for PSS based on those in the
7977 corresponding EVP_MD_CTX structure. No application support yet.
7981 * Support for companion algorithm specific ASN1 signing routines.
7982 New function ASN1_item_sign_ctx() signs a pre-initialised
7983 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7984 the appropriate parameters.
7988 * Add new algorithm specific ASN1 verification initialisation function
7989 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7990 handling will be the same no matter what EVP_PKEY_METHOD is used.
7991 Add a PSS handler to support verification of PSS signatures: checked
7992 against a number of sample certificates.
7996 * Add signature printing for PSS. Add PSS OIDs.
7998 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8000 * Add algorithm specific signature printing. An individual ASN1 method
8001 can now print out signatures instead of the standard hex dump.
8003 More complex signatures (e.g. PSS) can print out more meaningful
8004 information. Include DSA version that prints out the signature
8009 * Password based recipient info support for CMS library: implementing
8014 * Split password based encryption into PBES2 and PBKDF2 functions. This
8015 neatly separates the code into cipher and PBE sections and is required
8016 for some algorithms that split PBES2 into separate pieces (such as
8017 password based CMS).
8021 * Session-handling fixes:
8022 - Fix handling of connections that are resuming with a session ID,
8023 but also support Session Tickets.
8024 - Fix a bug that suppressed issuing of a new ticket if the client
8025 presented a ticket with an expired session.
8026 - Try to set the ticket lifetime hint to something reasonable.
8027 - Make tickets shorter by excluding irrelevant information.
8028 - On the client side, don't ignore renewed tickets.
8030 *Adam Langley, Bodo Moeller (Google)*
8032 * Fix PSK session representation.
8036 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8038 This work was sponsored by Intel.
8042 * Add GCM support to TLS library. Some custom code is needed to split
8043 the IV between the fixed (from PRF) and explicit (from TLS record)
8044 portions. This adds all GCM ciphersuites supported by RFC5288 and
8045 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8046 add a special AESGCM string for GCM only.
8050 * Expand range of ctrls for AES GCM. Permit setting invocation
8051 field on decrypt and retrieval of invocation field only on encrypt.
8055 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8056 As required by RFC5289 these ciphersuites cannot be used if for
8057 versions of TLS earlier than 1.2.
8061 * For FIPS capable OpenSSL interpret a NULL default public key method
8062 as unset and return the appropriate default but do *not* set the default.
8063 This means we can return the appropriate method in applications that
8064 switch between FIPS and non-FIPS modes.
8068 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8069 ENGINE is used then we cannot handle that in the FIPS module so we
8070 keep original code iff non-FIPS operations are allowed.
8074 * Add -attime option to openssl utilities.
8076 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8078 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8082 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8083 FIPS EC methods unconditionally for now.
8087 * New build option no-ec2m to disable characteristic 2 code.
8091 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8092 all cases can be covered as some introduce binary incompatibilities.
8096 * Redirect RSA operations to FIPS module including keygen,
8097 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8101 * Add similar low-level API blocking to ciphers.
8105 * low-level digest APIs are not approved in FIPS mode: any attempt
8106 to use these will cause a fatal error. Applications that *really* want
8107 to use them can use the `private_*` version instead.
8111 * Redirect cipher operations to FIPS module for FIPS builds.
8115 * Redirect digest operations to FIPS module for FIPS builds.
8119 * Update build system to add "fips" flag which will link in fipscanister.o
8120 for static and shared library builds embedding a signature if needed.
8124 * Output TLS supported curves in preference order instead of numerical
8125 order. This is currently hardcoded for the highest order curves first.
8126 This should be configurable so applications can judge speed vs strength.
8130 * Add TLS v1.2 server support for client authentication.
8134 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8139 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8140 FIPS modules versions.
8144 * Add TLS v1.2 client side support for client authentication. Keep cache
8145 of handshake records longer as we don't know the hash algorithm to use
8146 until after the certificate request message is received.
8150 * Initial TLS v1.2 client support. Add a default signature algorithms
8151 extension including all the algorithms we support. Parse new signature
8152 format in client key exchange. Relax some ECC signing restrictions for
8153 TLS v1.2 as indicated in RFC5246.
8157 * Add server support for TLS v1.2 signature algorithms extension. Switch
8158 to new signature format when needed using client digest preference.
8159 All server ciphersuites should now work correctly in TLS v1.2. No client
8160 support yet and no support for client certificates.
8164 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8165 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8166 ciphersuites. At present only RSA key exchange ciphersuites work with
8167 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8168 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8169 and version checking.
8173 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8174 with this defined it will not be affected by any changes to ssl internal
8175 structures. Add several utility functions to allow openssl application
8176 to work with OPENSSL_NO_SSL_INTERN defined.
8180 * A long standing patch to add support for SRP from EdelWeb (Peter
8181 Sylvester and Christophe Renou) was integrated.
8182 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8183 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8186 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8190 * Permit abbreviated handshakes when renegotiating using the function
8191 SSL_renegotiate_abbreviated().
8193 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8195 * Add call to ENGINE_register_all_complete() to
8196 ENGINE_load_builtin_engines(), so some implementations get used
8197 automatically instead of needing explicit application support.
8201 * Add support for TLS key exporter as described in RFC5705.
8203 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8205 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8206 a few changes are required:
8208 Add SSL_OP_NO_TLSv1_1 flag.
8209 Add TLSv1_1 methods.
8210 Update version checking logic to handle version 1.1.
8211 Add explicit IV handling (ported from DTLS code).
8212 Add command line options to s_client/s_server.
8219 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8221 * X509_ATTRIBUTE memory leak
8223 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8224 memory. This structure is used by the PKCS#7 and CMS routines so any
8225 application which reads PKCS#7 or CMS data from untrusted sources is
8226 affected. SSL/TLS is not affected.
8228 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8234 * Race condition handling PSK identify hint
8236 If PSK identity hints are received by a multi-threaded client then
8237 the values are wrongly updated in the parent SSL_CTX structure. This can
8238 result in a race condition potentially leading to a double free of the
8244 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8246 * Malformed ECParameters causes infinite loop
8248 When processing an ECParameters structure OpenSSL enters an infinite loop
8249 if the curve specified is over a specially malformed binary polynomial
8252 This can be used to perform denial of service against any
8253 system which processes public keys, certificate requests or
8254 certificates. This includes TLS clients and TLS servers with
8255 client authentication enabled.
8257 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8262 * Exploitable out-of-bounds read in X509_cmp_time
8264 X509_cmp_time does not properly check the length of the ASN1_TIME
8265 string and can read a few bytes out of bounds. In addition,
8266 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8269 An attacker can use this to craft malformed certificates and CRLs of
8270 various sizes and potentially cause a segmentation fault, resulting in
8271 a DoS on applications that verify certificates or CRLs. TLS clients
8272 that verify CRLs are affected. TLS clients and servers with client
8273 authentication enabled may be affected if they use custom verification
8276 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8277 independently by Hanno Böck.
8282 * PKCS7 crash with missing EnvelopedContent
8284 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8285 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8286 with missing content and trigger a NULL pointer dereference on parsing.
8288 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8289 structures from untrusted sources are affected. OpenSSL clients and
8290 servers are not affected.
8292 This issue was reported to OpenSSL by Michal Zalewski (Google).
8297 * CMS verify infinite loop with unknown hash function
8299 When verifying a signedData message the CMS code can enter an infinite loop
8300 if presented with an unknown hash function OID. This can be used to perform
8301 denial of service against any system which verifies signedData messages using
8303 This issue was reported to OpenSSL by Johannes Bauer.
8308 * Race condition handling NewSessionTicket
8310 If a NewSessionTicket is received by a multi-threaded client when attempting to
8311 reuse a previous ticket then a race condition can occur potentially leading to
8312 a double free of the ticket data.
8317 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8319 * Segmentation fault in ASN1_TYPE_cmp fix
8321 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8322 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8323 certificate signature algorithm consistency this can be used to crash any
8324 certificate verification operation and exploited in a DoS attack. Any
8325 application which performs certificate verification is vulnerable including
8326 OpenSSL clients and servers which enable client authentication.
8331 * ASN.1 structure reuse memory corruption fix
8333 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8334 memory corruption via an invalid write. Such reuse is and has been
8335 strongly discouraged and is believed to be rare.
8337 Applications that parse structures containing CHOICE or ANY DEFINED BY
8338 components may be affected. Certificate parsing (d2i_X509 and related
8339 functions) are however not affected. OpenSSL clients and servers are
8345 * PKCS7 NULL pointer dereferences fix
8347 The PKCS#7 parsing code does not handle missing outer ContentInfo
8348 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8349 missing content and trigger a NULL pointer dereference on parsing.
8351 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8352 otherwise parse PKCS#7 structures from untrusted sources are
8353 affected. OpenSSL clients and servers are not affected.
8355 This issue was reported to OpenSSL by Michal Zalewski (Google).
8360 * DoS via reachable assert in SSLv2 servers fix
8362 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8363 servers that both support SSLv2 and enable export cipher suites by sending
8364 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8366 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8367 (OpenSSL development team).
8372 * Use After Free following d2i_ECPrivatekey error fix
8374 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8375 could cause a use after free condition. This, in turn, could cause a double
8376 free in several private key parsing functions (such as d2i_PrivateKey
8377 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8378 for applications that receive EC private keys from untrusted
8379 sources. This scenario is considered rare.
8381 This issue was discovered by the BoringSSL project and fixed in their
8387 * X509_to_X509_REQ NULL pointer deref fix
8389 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8390 the certificate key is invalid. This function is rarely used in practice.
8392 This issue was discovered by Brian Carpenter.
8397 * Removed the export ciphers from the DEFAULT ciphers
8401 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8403 * Build fixes for the Windows and OpenVMS platforms
8405 *Matt Caswell and Richard Levitte*
8407 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8409 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8410 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8411 dereference. This could lead to a Denial Of Service attack. Thanks to
8412 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8417 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8418 dtls1_buffer_record function under certain conditions. In particular this
8419 could occur if an attacker sent repeated DTLS records with the same
8420 sequence number but for the next epoch. The memory leak could be exploited
8421 by an attacker in a Denial of Service attack through memory exhaustion.
8422 Thanks to Chris Mueller for reporting this issue.
8427 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8428 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8429 method would be set to NULL which could later result in a NULL pointer
8430 dereference. Thanks to Frank Schmirler for reporting this issue.
8435 * Abort handshake if server key exchange message is omitted for ephemeral
8438 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8439 reporting this issue.
8444 * Remove non-export ephemeral RSA code on client and server. This code
8445 violated the TLS standard by allowing the use of temporary RSA keys in
8446 non-export ciphersuites and could be used by a server to effectively
8447 downgrade the RSA key length used to a value smaller than the server
8448 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8449 INRIA or reporting this issue.
8454 * Fixed issue where DH client certificates are accepted without verification.
8455 An OpenSSL server will accept a DH certificate for client authentication
8456 without the certificate verify message. This effectively allows a client to
8457 authenticate without the use of a private key. This only affects servers
8458 which trust a client certificate authority which issues certificates
8459 containing DH keys: these are extremely rare and hardly ever encountered.
8460 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8466 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8467 results on some platforms, including x86_64. This bug occurs at random
8468 with a very low probability, and is not known to be exploitable in any
8469 way, though its exact impact is difficult to determine. Thanks to Pieter
8470 Wuille (Blockstream) who reported this issue and also suggested an initial
8471 fix. Further analysis was conducted by the OpenSSL development team and
8472 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8473 the OpenSSL core team.
8478 * Fix various certificate fingerprint issues.
8480 By using non-DER or invalid encodings outside the signed portion of a
8481 certificate the fingerprint can be changed without breaking the signature.
8482 Although no details of the signed portion of the certificate can be changed
8483 this can cause problems with some applications: e.g. those using the
8484 certificate fingerprint for blacklists.
8486 1. Reject signatures with non zero unused bits.
8488 If the BIT STRING containing the signature has non zero unused bits reject
8489 the signature. All current signature algorithms require zero unused bits.
8491 2. Check certificate algorithm consistency.
8493 Check the AlgorithmIdentifier inside TBS matches the one in the
8494 certificate signature. NB: this will result in signature failure
8495 errors for some broken certificates.
8497 Thanks to Konrad Kraszewski from Google for reporting this issue.
8499 3. Check DSA/ECDSA signatures use DER.
8501 Reencode DSA/ECDSA signatures and compare with the original received
8502 signature. Return an error if there is a mismatch.
8504 This will reject various cases including garbage after signature
8505 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8506 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8507 (negative or with leading zeroes).
8509 Further analysis was conducted and fixes were developed by Stephen Henson
8510 of the OpenSSL core team.
8516 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8518 * Session Ticket Memory Leak.
8520 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8521 integrity of that ticket is first verified. In the event of a session
8522 ticket integrity check failing, OpenSSL will fail to free memory
8523 causing a memory leak. By sending a large number of invalid session
8524 tickets an attacker could exploit this issue in a Denial Of Service
8530 * Build option no-ssl3 is incomplete.
8532 When OpenSSL is configured with "no-ssl3" as a build option, servers
8533 could accept and complete an SSL 3.0 handshake, and clients could be
8534 configured to send them.
8537 *Akamai and the OpenSSL team*
8539 * Add support for TLS_FALLBACK_SCSV.
8540 Client applications doing fallback retries should call
8541 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8544 *Adam Langley, Bodo Moeller*
8546 * Add additional DigestInfo checks.
8548 Reencode DigestInto in DER and check against the original when
8549 verifying RSA signature: this will reject any improperly encoded
8550 DigestInfo structures.
8552 Note: this is a precautionary measure and no attacks are currently known.
8556 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8558 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8559 to a denial of service attack. A malicious server can crash the client
8560 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8561 ciphersuite and sending carefully crafted handshake messages.
8563 Thanks to Felix Gröbert (Google) for discovering and researching this
8569 * By sending carefully crafted DTLS packets an attacker could cause openssl
8570 to leak memory. This can be exploited through a Denial of Service attack.
8571 Thanks to Adam Langley for discovering and researching this issue.
8576 * An attacker can force openssl to consume large amounts of memory whilst
8577 processing DTLS handshake messages. This can be exploited through a
8578 Denial of Service attack.
8579 Thanks to Adam Langley for discovering and researching this issue.
8584 * An attacker can force an error condition which causes openssl to crash
8585 whilst processing DTLS packets due to memory being freed twice. This
8586 can be exploited through a Denial of Service attack.
8587 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8593 * If a multithreaded client connects to a malicious server using a resumed
8594 session and the server sends an ec point format extension it could write
8595 up to 255 bytes to freed memory.
8597 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8603 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8604 X509_name_oneline, X509_name_print_ex et al. to leak some information
8605 from the stack. Applications may be affected if they echo pretty printing
8606 output to the attacker.
8608 Thanks to Ivan Fratric (Google) for discovering this issue.
8611 *Emilia Käsper, and Steve Henson*
8613 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8614 for corner cases. (Certain input points at infinity could lead to
8615 bogus results, with non-infinity inputs mapped to infinity too.)
8619 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8621 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8622 handshake can force the use of weak keying material in OpenSSL
8623 SSL/TLS clients and servers.
8625 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8626 researching this issue. ([CVE-2014-0224])
8628 *KIKUCHI Masashi, Steve Henson*
8630 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8631 OpenSSL DTLS client the code can be made to recurse eventually crashing
8634 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8637 *Imre Rad, Steve Henson*
8639 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8640 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8641 client or server. This is potentially exploitable to run arbitrary
8642 code on a vulnerable client or server.
8644 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8646 *Jüri Aedla, Steve Henson*
8648 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8649 are subject to a denial of service attack.
8651 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8652 this issue. ([CVE-2014-3470])
8654 *Felix Gröbert, Ivan Fratric, Steve Henson*
8656 * Harmonize version and its documentation. -f flag is used to display
8659 *mancha <mancha1@zoho.com>*
8661 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8662 in i2d_ECPrivateKey.
8664 *mancha <mancha1@zoho.com>*
8666 * Fix some double frees. These are not thought to be exploitable.
8668 *mancha <mancha1@zoho.com>*
8670 * Fix for the attack described in the paper "Recovering OpenSSL
8671 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8672 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8673 <http://eprint.iacr.org/2014/140>
8675 Thanks to Yuval Yarom and Naomi Benger for discovering this
8676 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8678 *Yuval Yarom and Naomi Benger*
8680 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8682 * Keep original DTLS digest and encryption contexts in retransmission
8683 structures so we can use the previous session parameters if they need
8684 to be resent. ([CVE-2013-6450])
8688 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8689 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8690 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8691 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8692 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8693 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8695 *Rob Stradling, Adam Langley*
8697 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8699 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8701 This addresses the flaw in CBC record processing discovered by
8702 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8703 at: <http://www.isg.rhul.ac.uk/tls/>
8705 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8706 Security Group at Royal Holloway, University of London
8707 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8708 Emilia Käsper for the initial patch.
8711 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8713 * Return an error when checking OCSP signatures when key is NULL.
8714 This fixes a DoS attack. ([CVE-2013-0166])
8718 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8719 the right response is stapled. Also change SSL_get_certificate()
8720 so it returns the certificate actually sent.
8721 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8722 (This is a backport)
8724 *Rob Stradling <rob.stradling@comodo.com>*
8726 * Fix possible deadlock when decoding public keys.
8730 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8732 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8735 * Sanity check record length before skipping explicit IV in DTLS
8738 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8739 fuzzing as a service testing platform.
8744 * Initialise tkeylen properly when encrypting CMS messages.
8745 Thanks to Solar Designer of Openwall for reporting this issue.
8749 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8751 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8752 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8753 in CRYPTO_realloc_clean.
8755 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8756 issue and to Adam Langley <agl@chromium.org> for fixing it.
8759 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8761 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8763 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8764 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8765 content decryption and always return the same error. Note: this attack
8766 needs on average 2^20 messages so it only affects automated senders. The
8767 old behaviour can be re-enabled in the CMS code by setting the
8768 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8769 an MMA defence is not necessary.
8770 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8771 this issue. ([CVE-2012-0884])
8775 * Fix CVE-2011-4619: make sure we really are receiving a
8776 client hello before rejecting multiple SGC restarts. Thanks to
8777 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8781 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8783 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8784 Thanks to Antonio Martin, Enterprise Secure Access Research and
8785 Development, Cisco Systems, Inc. for discovering this bug and
8786 preparing a fix. ([CVE-2012-0050])
8790 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8792 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8793 of the Vaudenay padding oracle attack on CBC mode encryption
8794 which enables an efficient plaintext recovery attack against
8795 the OpenSSL implementation of DTLS. Their attack exploits timing
8796 differences arising during decryption processing. A research
8797 paper describing this attack can be found at:
8798 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8799 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8800 Security Group at Royal Holloway, University of London
8801 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8802 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8803 for preparing the fix. ([CVE-2011-4108])
8805 *Robin Seggelmann, Michael Tuexen*
8807 * Clear bytes used for block padding of SSL 3.0 records.
8810 *Adam Langley (Google)*
8812 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8813 Kadianakis <desnacked@gmail.com> for discovering this issue and
8814 Adam Langley for preparing the fix. ([CVE-2011-4619])
8816 *Adam Langley (Google)*
8818 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8820 *Andrey Kulikov <amdeich@gmail.com>*
8822 * Prevent malformed RFC3779 data triggering an assertion failure.
8823 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8824 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8826 *Rob Austein <sra@hactrn.net>*
8828 * Improved PRNG seeding for VOS.
8830 *Paul Green <Paul.Green@stratus.com>*
8832 * Fix ssl_ciph.c set-up race.
8834 *Adam Langley (Google)*
8836 * Fix spurious failures in ecdsatest.c.
8838 *Emilia Käsper (Google)*
8840 * Fix the BIO_f_buffer() implementation (which was mixing different
8841 interpretations of the `..._len` fields).
8843 *Adam Langley (Google)*
8845 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8846 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8847 threads won't reuse the same blinding coefficients.
8849 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8850 lock to call BN_BLINDING_invert_ex, and avoids one use of
8851 BN_BLINDING_update for each BN_BLINDING structure (previously,
8852 the last update always remained unused).
8854 *Emilia Käsper (Google)*
8856 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8858 *Bob Buckholz (Google)*
8860 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8862 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8863 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8865 *Kaspar Brand <ossl@velox.ch>*
8867 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8868 for multi-threaded use of ECDH. ([CVE-2011-3210])
8870 *Adam Langley (Google)*
8872 * Fix x509_name_ex_d2i memory leak on bad inputs.
8876 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8877 signature public key algorithm by using OID xref utilities instead.
8878 Before this you could only use some ECC ciphersuites with SHA1 only.
8882 * Add protection against ECDSA timing attacks as mentioned in the paper
8883 by Billy Bob Brumley and Nicola Tuveri, see:
8884 <http://eprint.iacr.org/2011/232.pdf>
8886 *Billy Bob Brumley and Nicola Tuveri*
8888 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8890 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8892 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8894 * Fix bug in string printing code: if *any* escaping is enabled we must
8895 escape the escape character (backslash) or the resulting string is
8900 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8902 * Disable code workaround for ancient and obsolete Netscape browsers
8903 and servers: an attacker can use it in a ciphersuite downgrade attack.
8904 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8908 * Fixed J-PAKE implementation error, originally discovered by
8909 Sebastien Martini, further info and confirmation from Stefan
8910 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8914 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8916 * Fix extension code to avoid race conditions which can result in a buffer
8917 overrun vulnerability: resumed sessions must not be modified as they can
8918 be shared by multiple threads. CVE-2010-3864
8922 * Fix WIN32 build system to correctly link an ENGINE directory into
8927 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8929 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8932 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8934 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8936 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8937 context. The operation can be customised via the ctrl mechanism in
8938 case ENGINEs want to include additional functionality.
8942 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8946 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8947 output hashes compatible with older versions of OpenSSL.
8949 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8951 * Fix compression algorithm handling: if resuming a session use the
8952 compression algorithm of the resumed session instead of determining
8953 it from client hello again. Don't allow server to change algorithm.
8957 * Add load_crls() function to commands tidying load_certs() too. Add option
8958 to verify utility to allow additional CRLs to be included.
8962 * Update OCSP request code to permit adding custom headers to the request:
8963 some responders need this.
8967 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8970 *Julia Lawall <julia@diku.dk>*
8972 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8973 needlessly dereferenced structures, used obsolete functions and
8974 didn't handle all updated verify codes correctly.
8978 * Disable MD2 in the default configuration.
8982 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8983 indicate the initial BIO being pushed or popped. This makes it possible
8984 to determine whether the BIO is the one explicitly called or as a result
8985 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8986 it handles reference counts correctly and doesn't zero out the I/O bio
8987 when it is not being explicitly popped. WARNING: applications which
8988 included workarounds for the old buggy behaviour will need to be modified
8989 or they could free up already freed BIOs.
8993 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8994 renaming to all platforms (within the 0.9.8 branch, this was
8995 done conditionally on Netware platforms to avoid a name clash).
8997 *Guenter <lists@gknw.net>*
8999 * Add ECDHE and PSK support to DTLS.
9001 *Michael Tuexen <tuexen@fh-muenster.de>*
9003 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9008 * Add "missing" function EVP_MD_flags() (without this the only way to
9009 retrieve a digest flags is by accessing the structure directly. Update
9010 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9011 or cipher is registered as in the "from" argument. Print out all
9012 registered digests in the dgst usage message instead of manually
9013 attempting to work them out.
9017 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9018 this allows the use of compression and extensions. Change default cipher
9019 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9020 by default unless an application cipher string requests it.
9024 * Alter match criteria in PKCS12_parse(). It used to try to use local
9025 key ids to find matching certificates and keys but some PKCS#12 files
9026 don't follow the (somewhat unwritten) rules and this strategy fails.
9027 Now just gather all certificates together and the first private key
9028 then look for the first certificate that matches the key.
9032 * Support use of registered digest and cipher names for dgst and cipher
9033 commands instead of having to add each one as a special case. So now
9040 openssl dgst -sha256 foo
9042 and this works for ENGINE based algorithms too.
9046 * Update Gost ENGINE to support parameter files.
9048 *Victor B. Wagner <vitus@cryptocom.ru>*
9050 * Support GeneralizedTime in ca utility.
9052 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9054 * Enhance the hash format used for certificate directory links. The new
9055 form uses the canonical encoding (meaning equivalent names will work
9056 even if they aren't identical) and uses SHA1 instead of MD5. This form
9057 is incompatible with the older format and as a result c_rehash should
9058 be used to rebuild symbolic links.
9062 * Make PKCS#8 the default write format for private keys, replacing the
9063 traditional format. This form is standardised, more secure and doesn't
9064 include an implicit MD5 dependency.
9068 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9069 committed to OpenSSL should pass this lot as a minimum.
9073 * Add session ticket override functionality for use by EAP-FAST.
9075 *Jouni Malinen <j@w1.fi>*
9077 * Modify HMAC functions to return a value. Since these can be implemented
9078 in an ENGINE errors can occur.
9082 * Type-checked OBJ_bsearch_ex.
9086 * Type-checked OBJ_bsearch. Also some constification necessitated
9087 by type-checking. Still to come: TXT_DB, bsearch(?),
9088 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9093 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9094 seconds to a tm structure directly, instead of going through OS
9095 specific date routines. This avoids any issues with OS routines such
9096 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9097 and X509_time_adj_ex() to cover the extended range. The existing
9098 X509_time_adj() is still usable and will no longer have any date issues.
9102 * Delta CRL support. New use deltas option which will attempt to locate
9103 and search any appropriate delta CRLs available.
9105 This work was sponsored by Google.
9109 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9110 code and add additional score elements. Validate alternate CRL paths
9111 as part of the CRL checking and indicate a new error "CRL path validation
9112 error" in this case. Applications wanting additional details can use
9113 the verify callback and check the new "parent" field. If this is not
9114 NULL CRL path validation is taking place. Existing applications won't
9115 see this because it requires extended CRL support which is off by
9118 This work was sponsored by Google.
9122 * Support for freshest CRL extension.
9124 This work was sponsored by Google.
9128 * Initial indirect CRL support. Currently only supported in the CRLs
9129 passed directly and not via lookup. Process certificate issuer
9130 CRL entry extension and lookup CRL entries by bother issuer name
9131 and serial number. Check and process CRL issuer entry in IDP extension.
9133 This work was sponsored by Google.
9137 * Add support for distinct certificate and CRL paths. The CRL issuer
9138 certificate is validated separately in this case. Only enabled if
9139 an extended CRL support flag is set: this flag will enable additional
9140 CRL functionality in future.
9142 This work was sponsored by Google.
9146 * Add support for policy mappings extension.
9148 This work was sponsored by Google.
9152 * Fixes to pathlength constraint, self issued certificate handling,
9153 policy processing to align with RFC3280 and PKITS tests.
9155 This work was sponsored by Google.
9159 * Support for name constraints certificate extension. DN, email, DNS
9160 and URI types are currently supported.
9162 This work was sponsored by Google.
9166 * To cater for systems that provide a pointer-based thread ID rather
9167 than numeric, deprecate the current numeric thread ID mechanism and
9168 replace it with a structure and associated callback type. This
9169 mechanism allows a numeric "hash" to be extracted from a thread ID in
9170 either case, and on platforms where pointers are larger than 'long',
9171 mixing is done to help ensure the numeric 'hash' is usable even if it
9172 can't be guaranteed unique. The default mechanism is to use "&errno"
9173 as a pointer-based thread ID to distinguish between threads.
9175 Applications that want to provide their own thread IDs should now use
9176 CRYPTO_THREADID_set_callback() to register a callback that will call
9177 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9179 Note that ERR_remove_state() is now deprecated, because it is tied
9180 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9181 to free the current thread's error state should be replaced by
9182 ERR_remove_thread_state(NULL).
9184 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9185 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9186 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9187 application was previously providing a numeric thread callback that
9188 was inappropriate for distinguishing threads, then uniqueness might
9189 have been obtained with &errno that happened immediately in the
9190 intermediate development versions of OpenSSL; this is no longer the
9191 case, the numeric thread callback will now override the automatic use
9194 *Geoff Thorpe, with help from Bodo Moeller*
9196 * Initial support for different CRL issuing certificates. This covers a
9197 simple case where the self issued certificates in the chain exist and
9198 the real CRL issuer is higher in the existing chain.
9200 This work was sponsored by Google.
9204 * Removed effectively defunct crypto/store from the build.
9208 * Revamp of STACK to provide stronger type-checking. Still to come:
9209 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9210 ASN1_STRING, CONF_VALUE.
9214 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9215 RAM on SSL connections. This option can save about 34k per idle SSL.
9219 * Revamp of LHASH to provide stronger type-checking. Still to come:
9220 STACK, TXT_DB, bsearch, qsort.
9224 * Initial support for Cryptographic Message Syntax (aka CMS) based
9225 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9226 support for data, signedData, compressedData, digestedData and
9227 encryptedData, envelopedData types included. Scripts to check against
9228 RFC4134 examples draft and interop and consistency checks of many
9229 content types and variants.
9233 * Add options to enc utility to support use of zlib compression BIO.
9237 * Extend mk1mf to support importing of options and assembly language
9238 files from Configure script, currently only included in VC-WIN32.
9239 The assembly language rules can now optionally generate the source
9240 files from the associated perl scripts.
9244 * Implement remaining functionality needed to support GOST ciphersuites.
9245 Interop testing has been performed using CryptoPro implementations.
9247 *Victor B. Wagner <vitus@cryptocom.ru>*
9249 * s390x assembler pack.
9253 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9258 * Implement Opaque PRF Input TLS extension as specified in
9259 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9260 official specification yet and no extension type assignment by
9261 IANA exists, this extension (for now) will have to be explicitly
9262 enabled when building OpenSSL by providing the extension number
9263 to use. For example, specify an option
9265 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9267 to the "config" or "Configure" script to enable the extension,
9268 assuming extension number 0x9527 (which is a completely arbitrary
9269 and unofficial assignment based on the MD5 hash of the Internet
9270 Draft). Note that by doing so, you potentially lose
9271 interoperability with other TLS implementations since these might
9272 be using the same extension number for other purposes.
9274 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9275 opaque PRF input value to use in the handshake. This will create
9276 an internal copy of the length-'len' string at 'src', and will
9277 return non-zero for success.
9279 To get more control and flexibility, provide a callback function
9282 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9283 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9287 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9290 Callback function 'cb' will be called in handshakes, and is
9291 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9292 Argument 'arg' is for application purposes (the value as given to
9293 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9294 be provided to the callback function). The callback function
9295 has to return non-zero to report success: usually 1 to use opaque
9296 PRF input just if possible, or 2 to enforce use of the opaque PRF
9297 input. In the latter case, the library will abort the handshake
9298 if opaque PRF input is not successfully negotiated.
9300 Arguments 'peerinput' and 'len' given to the callback function
9301 will always be NULL and 0 in the case of a client. A server will
9302 see the client's opaque PRF input through these variables if
9303 available (NULL and 0 otherwise). Note that if the server
9304 provides an opaque PRF input, the length must be the same as the
9305 length of the client's opaque PRF input.
9307 Note that the callback function will only be called when creating
9308 a new session (session resumption can resume whatever was
9309 previously negotiated), and will not be called in SSL 2.0
9310 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9311 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9312 for applications that need to enforce opaque PRF input.
9316 * Update ssl code to support digests other than SHA1+MD5 for handshake
9319 *Victor B. Wagner <vitus@cryptocom.ru>*
9321 * Add RFC4507 support to OpenSSL. This includes the corrections in
9322 RFC4507bis. The encrypted ticket format is an encrypted encoded
9323 SSL_SESSION structure, that way new session features are automatically
9326 If a client application caches session in an SSL_SESSION structure
9327 support is transparent because tickets are now stored in the encoded
9330 The SSL_CTX structure automatically generates keys for ticket
9331 protection in servers so again support should be possible
9332 with no application modification.
9334 If a client or server wishes to disable RFC4507 support then the option
9335 SSL_OP_NO_TICKET can be set.
9337 Add a TLS extension debugging callback to allow the contents of any client
9338 or server extensions to be examined.
9340 This work was sponsored by Google.
9344 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9345 OpenSSL should now compile cleanly on gcc 4.2
9347 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9349 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9350 support including streaming MAC support: this is required for GOST
9351 ciphersuite support.
9353 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9355 * Add option -stream to use PKCS#7 streaming in smime utility. New
9356 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9357 to output in BER and PEM format.
9361 * Experimental support for use of HMAC via EVP_PKEY interface. This
9362 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9363 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9364 ENGINE support for HMAC keys which are unextractable. New -mac and
9365 -macopt options to dgst utility.
9369 * New option -sigopt to dgst utility. Update dgst to use
9370 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9371 alternative signing parameters such as X9.31 or PSS in the dgst
9376 * Change ssl_cipher_apply_rule(), the internal function that does
9377 the work each time a ciphersuite string requests enabling
9378 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9379 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9380 the order of disabled ciphersuites such that those ciphersuites
9381 that most recently went from enabled to disabled not only stay
9382 in order with respect to each other, but also have higher priority
9383 than other disabled ciphersuites the next time ciphersuites are
9386 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9387 the same ciphersuites as with "HIGH" alone, but in a specific
9388 order where the PSK ciphersuites come first (since they are the
9389 most recently disabled ciphersuites when "HIGH" is parsed).
9391 Also, change ssl_create_cipher_list() (using this new
9392 functionality) such that between otherwise identical
9393 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9398 * Change ssl_create_cipher_list() so that it automatically
9399 arranges the ciphersuites in reasonable order before starting
9400 to process the rule string. Thus, the definition for "DEFAULT"
9401 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9402 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9403 This makes it much easier to arrive at a reasonable default order
9404 in applications for which anonymous ciphers are OK (meaning
9405 that you can't actually use DEFAULT).
9407 *Bodo Moeller; suggested by Victor Duchovni*
9409 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9410 processing) into multiple integers instead of setting
9411 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9412 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9413 (These masks as well as the individual bit definitions are hidden
9414 away into the non-exported interface ssl/ssl_locl.h, so this
9415 change to the definition of the SSL_CIPHER structure shouldn't
9416 affect applications.) This give us more bits for each of these
9417 categories, so there is no longer a need to coagulate AES128 and
9418 AES256 into a single algorithm bit, and to coagulate Camellia128
9419 and Camellia256 into a single algorithm bit, which has led to all
9422 Thus, among other things, the kludge introduced in 0.9.7m and
9423 0.9.8e for masking out AES256 independently of AES128 or masking
9424 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9426 With the change, we also introduce new ciphersuite aliases that
9427 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9432 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9433 Use the leftmost N bytes of the signature input if the input is
9434 larger than the prime q (with N being the size in bytes of q).
9438 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9439 it yet and it is largely untested.
9443 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9447 * Initial incomplete changes to avoid need for function casts in OpenSSL
9448 some compilers (gcc 4.2 and later) reject their use. Safestack is
9449 reimplemented. Update ASN1 to avoid use of legacy functions.
9453 * Win32/64 targets are linked with Winsock2.
9457 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9458 to external functions. This can be used to increase CRL handling
9459 efficiency especially when CRLs are very large by (for example) storing
9460 the CRL revoked certificates in a database.
9464 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9465 new CRLs added to a directory can be used. New command line option
9466 -verify_return_error to s_client and s_server. This causes real errors
9467 to be returned by the verify callback instead of carrying on no matter
9468 what. This reflects the way a "real world" verify callback would behave.
9472 * GOST engine, supporting several GOST algorithms and public key formats.
9473 Kindly donated by Cryptocom.
9477 * Partial support for Issuing Distribution Point CRL extension. CRLs
9478 partitioned by DP are handled but no indirect CRL or reason partitioning
9479 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9480 selected via a scoring technique which handles IDP and AKID in CRLs.
9484 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9485 will ultimately be used for all verify operations: this will remove the
9486 X509_STORE dependency on certificate verification and allow alternative
9487 lookup methods. X509_STORE based implementations of these two callbacks.
9491 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9492 Modify get_crl() to find a valid (unexpired) CRL if possible.
9496 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9497 this would be called X509_CRL_cmp() but that name is already used by
9498 a function that just compares CRL issuer names. Cache several CRL
9499 extensions in X509_CRL structure and cache CRLDP in X509.
9503 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9504 this maps equivalent X509_NAME structures into a consistent structure.
9505 Name comparison can then be performed rapidly using memcmp().
9509 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9514 * Allow digests to supply their own micalg string for S/MIME type using
9515 the ctrl EVP_MD_CTRL_MICALG.
9519 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9520 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9521 ctrl. It can then customise the structure before and/or after signing
9526 * New function OBJ_add_sigid() to allow application defined signature OIDs
9527 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9528 to free up any added signature OIDs.
9532 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9533 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9534 digest and cipher tables. New options added to openssl utility:
9535 list-message-digest-algorithms and list-cipher-algorithms.
9539 * Change the array representation of binary polynomials: the list
9540 of degrees of non-zero coefficients is now terminated with -1.
9541 Previously it was terminated with 0, which was also part of the
9542 value; thus, the array representation was not applicable to
9543 polynomials where t^0 has coefficient zero. This change makes
9544 the array representation useful in a more general context.
9548 * Various modifications and fixes to SSL/TLS cipher string
9549 handling. For ECC, the code now distinguishes between fixed ECDH
9550 with RSA certificates on the one hand and with ECDSA certificates
9551 on the other hand, since these are separate ciphersuites. The
9552 unused code for Fortezza ciphersuites has been removed.
9554 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9555 (not "ECDHE"). For consistency with the code for DH
9556 certificates, use of ECDH certificates is now considered ECDH
9557 authentication, not RSA or ECDSA authentication (the latter is
9558 merely the CA's signing algorithm and not actively used in the
9561 The temporary ciphersuite alias "ECCdraft" is no longer
9562 available, and ECC ciphersuites are no longer excluded from "ALL"
9563 and "DEFAULT". The following aliases now exist for RFC 4492
9564 ciphersuites, most of these by analogy with the DH case:
9566 kECDHr - ECDH cert, signed with RSA
9567 kECDHe - ECDH cert, signed with ECDSA
9568 kECDH - ECDH cert (signed with either RSA or ECDSA)
9569 kEECDH - ephemeral ECDH
9570 ECDH - ECDH cert or ephemeral ECDH
9576 AECDH - anonymous ECDH
9577 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9581 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9582 Use correct micalg parameters depending on digest(s) in signed message.
9586 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9587 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9591 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9592 an engine to register a method. Add ENGINE lookups for methods and
9593 functional reference processing.
9597 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9598 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9603 * New -resign option to smime utility. This adds one or more signers
9604 to an existing PKCS#7 signedData structure. Also -md option to use an
9605 alternative message digest algorithm for signing.
9609 * Tidy up PKCS#7 routines and add new functions to make it easier to
9610 create PKCS7 structures containing multiple signers. Update smime
9611 application to support multiple signers.
9615 * New -macalg option to pkcs12 utility to allow setting of an alternative
9620 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9621 Reorganize PBE internals to lookup from a static table using NIDs,
9622 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9623 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9624 PRF which will be automatically used with PBES2.
9628 * Replace the algorithm specific calls to generate keys in "req" with the
9633 * Update PKCS#7 enveloped data routines to use new API. This is now
9634 supported by any public key method supporting the encrypt operation. A
9635 ctrl is added to allow the public key algorithm to examine or modify
9636 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9641 * Add a ctrl to asn1 method to allow a public key algorithm to express
9642 a default digest type to use. In most cases this will be SHA1 but some
9643 algorithms (such as GOST) need to specify an alternative digest. The
9644 return value indicates how strong the preference is 1 means optional and
9645 2 is mandatory (that is it is the only supported type). Modify
9646 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9647 use the default md. Update openssl utilities to use the default digest
9648 type for signing if it is not explicitly indicated.
9652 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9653 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9654 signing method from the key type. This effectively removes the link
9655 between digests and public key types.
9659 * Add an OID cross reference table and utility functions. Its purpose is to
9660 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9661 rsaEncryption. This will allow some of the algorithm specific hackery
9662 needed to use the correct OID to be removed.
9666 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9667 structures for PKCS7_sign(). They are now set up by the relevant public
9672 * Add provisional EC pkey method with support for ECDSA and ECDH.
9676 * Add support for key derivation (agreement) in the API, DH method and
9681 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9682 public and private key formats. As a side effect these add additional
9683 command line functionality not previously available: DSA signatures can be
9684 generated and verified using pkeyutl and DH key support and generation in
9691 *Oliver Tappe <zooey@hirschkaefer.de>*
9693 * New make target "install_html_docs" installs HTML renditions of the
9696 *Oliver Tappe <zooey@hirschkaefer.de>*
9698 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9699 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9700 support key and parameter generation and add initial key generation
9701 functionality for RSA.
9705 * Add functions for main EVP_PKEY_method operations. The undocumented
9706 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9707 `EVP_PKEY_{encrypt,decrypt}_old`.
9711 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9712 key API, doesn't do much yet.
9716 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9717 public key algorithms. New option to openssl utility:
9718 "list-public-key-algorithms" to print out info.
9722 * Implement the Supported Elliptic Curves Extension for
9723 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9727 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9728 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9732 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9733 utilities such as rsa, dsa, dsaparam etc except they process any key
9738 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9739 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9740 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9745 * Initial support for pluggable public key ASN1.
9746 De-spaghettify the public key ASN1 handling. Move public and private
9747 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9748 algorithm specific handling to a single module within the relevant
9749 algorithm directory. Add functions to allow (near) opaque processing
9750 of public and private key structures.
9754 * Implement the Supported Point Formats Extension for
9755 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9759 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9760 for the psk identity [hint] and the psk callback functions to the
9761 SSL_SESSION, SSL and SSL_CTX structure.
9764 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9768 SSL_CTX_use_psk_identity_hint
9769 SSL_get_psk_identity_hint
9770 SSL_get_psk_identity
9771 SSL_use_psk_identity_hint
9773 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9775 * Add RFC 3161 compliant time stamp request creation, response generation
9776 and response verification functionality.
9778 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9780 * Add initial support for TLS extensions, specifically for the server_name
9781 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9782 have new members for a hostname. The SSL data structure has an
9783 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9784 stored in that context to allow for session resumption, even after the
9785 SSL has been switched to a new SSL_CTX in reaction to a client's
9786 server_name extension.
9788 New functions (subject to change):
9790 SSL_get_servername()
9791 SSL_get_servername_type()
9794 New CTRL codes and macros (subject to change):
9796 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9797 - SSL_CTX_set_tlsext_servername_callback()
9798 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9799 - SSL_CTX_set_tlsext_servername_arg()
9800 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9802 openssl s_client has a new '-servername ...' option.
9804 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9805 '-key2 ...', '-servername_fatal' (subject to change). This allows
9806 testing the HostName extension for a specific single hostname ('-cert'
9807 and '-key' remain fallbacks for handshakes without HostName
9808 negotiation). If the unrecognized_name alert has to be sent, this by
9809 default is a warning; it becomes fatal with the '-servername_fatal'
9812 *Peter Sylvester, Remy Allais, Christophe Renou*
9814 * Whirlpool hash implementation is added.
9818 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9819 bn(64,32). Because of instruction set limitations it doesn't have
9820 any negative impact on performance. This was done mostly in order
9821 to make it possible to share assembler modules, such as bn_mul_mont
9822 implementations, between 32- and 64-bit builds without hassle.
9826 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9827 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9832 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9833 dedicated Montgomery multiplication procedure, is introduced.
9834 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9835 "64-bit" performance on certain 32-bit targets.
9839 * New option SSL_OP_NO_COMP to disable use of compression selectively
9840 in SSL structures. New SSL ctrl to set maximum send fragment size.
9841 Save memory by setting the I/O buffer sizes dynamically instead of
9842 using the maximum available value.
9846 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9847 in addition to the text details.
9851 * Very, very preliminary EXPERIMENTAL support for printing of general
9852 ASN1 structures. This currently produces rather ugly output and doesn't
9853 handle several customised structures at all.
9857 * Integrated support for PVK file format and some related formats such
9858 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9859 these in the 'rsa' and 'dsa' utilities.
9863 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9867 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9868 place for the (very old) "NETSCAPE" format certificates which are now
9869 handled using new ASN1 code equivalents.
9873 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9874 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9875 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9879 * Modify CRL distribution points extension code to print out previously
9880 unsupported fields. Enhance extension setting code to allow setting of
9885 * Add print and set support for Issuing Distribution Point CRL extension.
9889 * Change 'Configure' script to enable Camellia by default.
9896 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9898 * When rejecting SSL/TLS records due to an incorrect version number, never
9899 update s->server with a new major version number. As of
9900 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9901 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9902 the previous behavior could result in a read attempt at NULL when
9903 receiving specific incorrect SSL/TLS records once record payload
9904 protection is active. ([CVE-2010-0740])
9906 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9908 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9909 could be crashed if the relevant tables were not present (e.g. chrooted).
9911 *Tomas Hoger <thoger@redhat.com>*
9913 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9915 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9917 *Martin Olsson, Neel Mehta*
9919 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9920 accommodate for stack sorting, always a write lock!).
9924 * On some versions of WIN32 Heap32Next is very slow. This can cause
9925 excessive delays in the RAND_poll(): over a minute. As a workaround
9926 include a time check in the inner Heap32Next loop too.
9930 * The code that handled flushing of data in SSL/TLS originally used the
9931 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9932 the problem outlined in PR#1949. The fix suggested there however can
9933 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9934 of Apache). So instead simplify the code to flush unconditionally.
9935 This should be fine since flushing with no data to flush is a no op.
9939 * Handle TLS versions 2.0 and later properly and correctly use the
9940 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9941 off ancient servers have a habit of sticking around for a while...
9945 * Modify compression code so it frees up structures without using the
9946 ex_data callbacks. This works around a problem where some applications
9947 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9948 restarting) then use compression (e.g. SSL with compression) later.
9949 This results in significant per-connection memory leaks and
9950 has caused some security issues including CVE-2008-1678 and
9955 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9956 change when encrypting or decrypting.
9960 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9961 connect and renegotiate with servers which do not support RI.
9962 Until RI is more widely deployed this option is enabled by default.
9966 * Add "missing" ssl ctrls to clear options and mode.
9970 * If client attempts to renegotiate and doesn't support RI respond with
9971 a no_renegotiation alert as required by RFC5746. Some renegotiating
9972 TLS clients will continue a connection gracefully when they receive
9973 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9974 waiting for a server hello which it will never receive. Now we treat a
9975 received no_renegotiation alert as a fatal error. This is because
9976 applications requesting a renegotiation might well expect it to succeed
9977 and would have no code in place to handle the server denying it so the
9978 only safe thing to do is to terminate the connection.
9982 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9983 peer supports secure renegotiation and 0 otherwise. Print out peer
9984 renegotiation support in s_client/s_server.
9988 * Replace the highly broken and deprecated SPKAC certification method with
9989 the updated NID creation version. This should correctly handle UTF8.
9993 * Implement RFC5746. Re-enable renegotiation but require the extension
9994 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9995 turns out to be a bad idea. It has been replaced by
9996 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9997 SSL_CTX_set_options(). This is really not recommended unless you
9998 know what you are doing.
10000 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10002 * Fixes to stateless session resumption handling. Use initial_ctx when
10003 issuing and attempting to decrypt tickets in case it has changed during
10004 servername handling. Use a non-zero length session ID when attempting
10005 stateless session resumption: this makes it possible to determine if
10006 a resumption has occurred immediately after receiving server hello
10007 (several places in OpenSSL subtly assume this) instead of later in
10012 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10013 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10014 fixes for a few places where the return code is not checked
10017 *Julia Lawall <julia@diku.dk>*
10019 * Add --strict-warnings option to Configure script to include devteam
10020 warnings in other configurations.
10024 * Add support for --libdir option and LIBDIR variable in makefiles. This
10025 makes it possible to install openssl libraries in locations which
10026 have names other than "lib", for example "/usr/lib64" which some
10029 *Steve Henson, based on patch from Jeremy Utley*
10031 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10032 X690 8.9.12 and can produce some misleading textual output of OIDs.
10034 *Steve Henson, reported by Dan Kaminsky*
10036 * Delete MD2 from algorithm tables. This follows the recommendation in
10037 several standards that it is not used in new applications due to
10038 several cryptographic weaknesses. For binary compatibility reasons
10039 the MD2 API is still compiled in by default.
10043 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10048 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10049 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10052 *Guenter <lists@gknw.net>*
10054 * Fix the server certificate chain building code to use X509_verify_cert(),
10055 it used to have an ad-hoc builder which was unable to cope with anything
10056 other than a simple chain.
10058 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10060 * Don't check self signed certificate signatures in X509_verify_cert()
10061 by default (a flag can override this): it just wastes time without
10062 adding any security. As a useful side effect self signed root CAs
10063 with non-FIPS digests are now usable in FIPS mode.
10067 * In dtls1_process_out_of_seq_message() the check if the current message
10068 is already buffered was missing. For every new message was memory
10069 allocated, allowing an attacker to perform an denial of service attack
10070 with sending out of seq handshake messages until there is no memory
10071 left. Additionally every future message was buffered, even if the
10072 sequence number made no sense and would be part of another handshake.
10073 So only messages with sequence numbers less than 10 in advance will be
10074 buffered. ([CVE-2009-1378])
10076 *Robin Seggelmann, discovered by Daniel Mentz*
10078 * Records are buffered if they arrive with a future epoch to be
10079 processed after finishing the corresponding handshake. There is
10080 currently no limitation to this buffer allowing an attacker to perform
10081 a DOS attack with sending records with future epochs until there is no
10082 memory left. This patch adds the pqueue_size() function to determine
10083 the size of a buffer and limits the record buffer to 100 entries.
10086 *Robin Seggelmann, discovered by Daniel Mentz*
10088 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10089 parent structure is freed. ([CVE-2009-1379])
10093 * Handle non-blocking I/O properly in SSL_shutdown() call.
10095 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10097 * Add `2.5.4.*` OIDs
10099 *Ilya O. <vrghost@gmail.com>*
10101 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10103 * Disable renegotiation completely - this fixes a severe security
10104 problem ([CVE-2009-3555]) at the cost of breaking all
10105 renegotiation. Renegotiation can be re-enabled by setting
10106 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10107 run-time. This is really not recommended unless you know what
10112 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10114 * Don't set val to NULL when freeing up structures, it is freed up by
10115 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10116 zeroing past the valid field. ([CVE-2009-0789])
10118 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10120 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10121 checked correctly. This would allow some invalid signed attributes to
10122 appear to verify correctly. ([CVE-2009-0591])
10124 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10126 * Reject UniversalString and BMPString types with invalid lengths. This
10127 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10128 a legal length. ([CVE-2009-0590])
10132 * Set S/MIME signing as the default purpose rather than setting it
10133 unconditionally. This allows applications to override it at the store
10138 * Permit restricted recursion of ASN1 strings. This is needed in practice
10139 to handle some structures.
10143 * Improve efficiency of mem_gets: don't search whole buffer each time
10146 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10148 * New -hex option for openssl rand.
10152 * Print out UTF8String and NumericString when parsing ASN1.
10156 * Support NumericString type for name components.
10160 * Allow CC in the environment to override the automatically chosen
10161 compiler. Note that nothing is done to ensure flags work with the
10166 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10168 * Properly check EVP_VerifyFinal() and similar return values
10171 *Ben Laurie, Bodo Moeller, Google Security Team*
10173 * Enable TLS extensions by default.
10177 * Allow the CHIL engine to be loaded, whether the application is
10178 multithreaded or not. (This does not release the developer from the
10179 obligation to set up the dynamic locking callbacks.)
10181 *Sander Temme <sander@temme.net>*
10183 * Use correct exit code if there is an error in dgst command.
10185 *Steve Henson; problem pointed out by Roland Dirlewanger*
10187 * Tweak Configure so that you need to say "experimental-jpake" to enable
10188 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10192 * Add experimental JPAKE support, including demo authentication in
10193 s_client and s_server.
10197 * Set the comparison function in v3_addr_canonize().
10199 *Rob Austein <sra@hactrn.net>*
10201 * Add support for XMPP STARTTLS in s_client.
10203 *Philip Paeps <philip@freebsd.org>*
10205 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10206 to ensure that even with this option, only ciphersuites in the
10207 server's preference list will be accepted. (Note that the option
10208 applies only when resuming a session, so the earlier behavior was
10209 just about the algorithm choice for symmetric cryptography.)
10213 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10215 * Fix NULL pointer dereference if a DTLS server received
10216 ChangeCipherSpec as first record ([CVE-2009-1386]).
10220 * Fix a state transition in s3_srvr.c and d1_srvr.c
10221 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10223 *Nagendra Modadugu*
10225 * The fix in 0.9.8c that supposedly got rid of unsafe
10226 double-checked locking was incomplete for RSA blinding,
10227 addressing just one layer of what turns out to have been
10228 doubly unsafe triple-checked locking.
10230 So now fix this for real by retiring the MONT_HELPER macro
10231 in crypto/rsa/rsa_eay.c.
10233 *Bodo Moeller; problem pointed out by Marius Schilder*
10235 * Various precautionary measures:
10237 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10239 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10240 (NB: This would require knowledge of the secret session ticket key
10241 to exploit, in which case you'd be SOL either way.)
10243 - Change bn_nist.c so that it will properly handle input BIGNUMs
10244 outside the expected range.
10246 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10249 *Neel Mehta, Bodo Moeller*
10251 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10252 the load fails. Useful for distros.
10254 *Ben Laurie and the FreeBSD team*
10256 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10260 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10264 * Expand ENGINE to support engine supplied SSL client certificate functions.
10266 This work was sponsored by Logica.
10270 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10271 keystores. Support for SSL/TLS client authentication too.
10272 Not compiled unless enable-capieng specified to Configure.
10274 This work was sponsored by Logica.
10278 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10279 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10280 attribute creation routines such as certificate requests and PKCS#12
10285 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10287 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10288 handshake which could lead to a client crash as found using the
10289 Codenomicon TLS test suite ([CVE-2008-1672])
10291 *Steve Henson, Mark Cox*
10293 * Fix double free in TLS server name extensions which could lead to
10294 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10298 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10300 Clear the error queue to ensure that error entries left from
10301 older function calls do not interfere with the correct operation.
10303 *Lutz Jaenicke, Erik de Castro Lopo*
10305 * Remove root CA certificates of commercial CAs:
10307 The OpenSSL project does not recommend any specific CA and does not
10308 have any policy with respect to including or excluding any CA.
10309 Therefore, it does not make any sense to ship an arbitrary selection
10310 of root CA certificates with the OpenSSL software.
10314 * RSA OAEP patches to fix two separate invalid memory reads.
10315 The first one involves inputs when 'lzero' is greater than
10316 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10317 before the beginning of from). The second one involves inputs where
10318 the 'db' section contains nothing but zeroes (there is a one-byte
10319 invalid read after the end of 'db').
10321 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10323 * Partial backport from 0.9.9-dev:
10325 Introduce bn_mul_mont (dedicated Montgomery multiplication
10326 procedure) as a candidate for BIGNUM assembler implementation.
10327 While 0.9.9-dev uses assembler for various architectures, only
10328 x86_64 is available by default here in the 0.9.8 branch, and
10329 32-bit x86 is available through a compile-time setting.
10331 To try the 32-bit x86 assembler implementation, use Configure
10332 option "enable-montasm" (which exists only for this backport).
10334 As "enable-montasm" for 32-bit x86 disclaims code stability
10335 anyway, in this constellation we activate additional code
10336 backported from 0.9.9-dev for further performance improvements,
10337 namely BN_from_montgomery_word. (To enable this otherwise,
10338 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10340 *Andy Polyakov (backport partially by Bodo Moeller)*
10342 * Add TLS session ticket callback. This allows an application to set
10343 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10344 values. This is useful for key rollover for example where several key
10345 sets may exist with different names.
10349 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10350 This was broken until now in 0.9.8 releases, such that the only way
10351 a registered ENGINE could be used (assuming it initialises
10352 successfully on the host) was to explicitly set it as the default
10353 for the relevant algorithms. This is in contradiction with 0.9.7
10354 behaviour and the documentation. With this fix, when an ENGINE is
10355 registered into a given algorithm's table of implementations, the
10356 'uptodate' flag is reset so that auto-discovery will be used next
10357 time a new context for that algorithm attempts to select an
10360 *Ian Lister (tweaked by Geoff Thorpe)*
10362 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10363 implementation in the following ways:
10365 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10368 Lack of BER streaming support means one pass streaming processing is
10369 only supported if data is detached: setting the streaming flag is
10370 ignored for embedded content.
10372 CMS support is disabled by default and must be explicitly enabled
10373 with the enable-cms configuration option.
10377 * Update the GMP engine glue to do direct copies between BIGNUM and
10378 mpz_t when openssl and GMP use the same limb size. Otherwise the
10379 existing "conversion via a text string export" trick is still used.
10381 *Paul Sheer <paulsheer@gmail.com>*
10383 * Zlib compression BIO. This is a filter BIO which compressed and
10384 uncompresses any data passed through it.
10388 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10389 RFC3394 compatible AES key wrapping.
10393 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10394 sets string data without copying. X509_ALGOR_set0() and
10395 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10396 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10397 from an X509_ATTRIBUTE structure optionally checking it occurs only
10398 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10403 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10404 to get the expected BN_FLG_CONSTTIME behavior.
10406 *Bodo Moeller (Google)*
10410 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10411 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10412 - added some more tests to do_tests.pl
10413 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10414 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10415 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10416 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10417 - various changes to netware.pl to enable gcc-cross builds on Win32
10419 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10420 - various changes to fix missing prototype warnings
10421 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10422 - added AES, WHIRLPOOL and CPUID assembler code to build files
10423 - added missing AES assembler make rules to mk1mf.pl
10424 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10426 *Guenter Knauf <eflash@gmx.net>*
10428 * Implement certificate status request TLS extension defined in RFC3546.
10429 A client can set the appropriate parameters and receive the encoded
10430 OCSP response via a callback. A server can query the supplied parameters
10431 and set the encoded OCSP response in the callback. Add simplified examples
10432 to s_client and s_server.
10436 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10438 * Fix various bugs:
10439 + Binary incompatibility of ssl_ctx_st structure
10440 + DTLS interoperation with non-compliant servers
10441 + Don't call get_session_cb() without proposed session
10442 + Fix ia64 assembler code
10444 *Andy Polyakov, Steve Henson*
10446 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10448 * DTLS Handshake overhaul. There were longstanding issues with
10449 OpenSSL DTLS implementation, which were making it impossible for
10450 RFC 4347 compliant client to communicate with OpenSSL server.
10451 Unfortunately just fixing these incompatibilities would "cut off"
10452 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10453 server keeps tolerating non RFC compliant syntax. The opposite is
10454 not true, 0.9.8f client can not communicate with earlier server.
10455 This update even addresses CVE-2007-4995.
10459 * Changes to avoid need for function casts in OpenSSL: some compilers
10460 (gcc 4.2 and later) reject their use.
10461 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10464 * Add RFC4507 support to OpenSSL. This includes the corrections in
10465 RFC4507bis. The encrypted ticket format is an encrypted encoded
10466 SSL_SESSION structure, that way new session features are automatically
10469 If a client application caches session in an SSL_SESSION structure
10470 support is transparent because tickets are now stored in the encoded
10473 The SSL_CTX structure automatically generates keys for ticket
10474 protection in servers so again support should be possible
10475 with no application modification.
10477 If a client or server wishes to disable RFC4507 support then the option
10478 SSL_OP_NO_TICKET can be set.
10480 Add a TLS extension debugging callback to allow the contents of any client
10481 or server extensions to be examined.
10483 This work was sponsored by Google.
10487 * Add initial support for TLS extensions, specifically for the server_name
10488 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10489 have new members for a hostname. The SSL data structure has an
10490 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10491 stored in that context to allow for session resumption, even after the
10492 SSL has been switched to a new SSL_CTX in reaction to a client's
10493 server_name extension.
10495 New functions (subject to change):
10497 SSL_get_servername()
10498 SSL_get_servername_type()
10501 New CTRL codes and macros (subject to change):
10503 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10504 - SSL_CTX_set_tlsext_servername_callback()
10505 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10506 - SSL_CTX_set_tlsext_servername_arg()
10507 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10509 openssl s_client has a new '-servername ...' option.
10511 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10512 '-key2 ...', '-servername_fatal' (subject to change). This allows
10513 testing the HostName extension for a specific single hostname ('-cert'
10514 and '-key' remain fallbacks for handshakes without HostName
10515 negotiation). If the unrecognized_name alert has to be sent, this by
10516 default is a warning; it becomes fatal with the '-servername_fatal'
10519 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10521 * Add AES and SSE2 assembly language support to VC++ build.
10525 * Mitigate attack on final subtraction in Montgomery reduction.
10529 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10530 (which previously caused an internal error).
10534 * Squeeze another 10% out of IGE mode when in != out.
10538 * AES IGE mode speedup.
10540 *Dean Gaudet (Google)*
10542 * Add the Korean symmetric 128-bit cipher SEED (see
10543 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10544 add SEED ciphersuites from RFC 4162:
10546 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10547 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10548 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10549 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10551 To minimize changes between patchlevels in the OpenSSL 0.9.8
10552 series, SEED remains excluded from compilation unless OpenSSL
10553 is configured with 'enable-seed'.
10555 *KISA, Bodo Moeller*
10557 * Mitigate branch prediction attacks, which can be practical if a
10558 single processor is shared, allowing a spy process to extract
10559 information. For detailed background information, see
10560 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10561 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10562 and Necessary Software Countermeasures"). The core of the change
10563 are new versions BN_div_no_branch() and
10564 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10565 respectively, which are slower, but avoid the security-relevant
10566 conditional branches. These are automatically called by BN_div()
10567 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10568 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10569 remove a conditional branch.
10571 BN_FLG_CONSTTIME is the new name for the previous
10572 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10573 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10574 in the exponent causes BN_mod_exp_mont() to use the alternative
10575 implementation in BN_mod_exp_mont_consttime().) The old name
10576 remains as a deprecated alias.
10578 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10579 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10580 constant-time implementations for more than just exponentiation.
10581 Here too the old name is kept as a deprecated alias.
10583 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10584 the BN_BLINDING structure gets an independent copy of the
10585 modulus. This means that the previous `BIGNUM *m` argument to
10586 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10587 essentially becomes `const BIGNUM *m`, although we can't actually
10588 change this in the header file before 0.9.9. It allows
10589 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10590 enable BN_FLG_CONSTTIME.
10592 *Matthew D Wood (Intel Corp)*
10594 * In the SSL/TLS server implementation, be strict about session ID
10595 context matching (which matters if an application uses a single
10596 external cache for different purposes). Previously,
10597 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10598 set. This did ensure strict client verification, but meant that,
10599 with applications using a single external cache for quite
10600 different requirements, clients could circumvent ciphersuite
10601 restrictions for a given session ID context by starting a session
10602 in a different context.
10606 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10607 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10608 authentication-only ciphersuites.
10612 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10613 not complete and could lead to a possible single byte overflow
10614 ([CVE-2007-5135]) [Ben Laurie]
10616 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10618 * Since AES128 and AES256 (and similarly Camellia128 and
10619 Camellia256) share a single mask bit in the logic of
10620 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10621 kludge to work properly if AES128 is available and AES256 isn't
10622 (or if Camellia128 is available and Camellia256 isn't).
10626 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10627 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10628 When a point or a seed is encoded in a BIT STRING, we need to
10629 prevent the removal of trailing zero bits to get the proper DER
10630 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10631 of a NamedBitList, for which trailing 0 bits need to be removed.)
10635 * Have SSL/TLS server implementation tolerate "mismatched" record
10636 protocol version while receiving ClientHello even if the
10637 ClientHello is fragmented. (The server can't insist on the
10638 particular protocol version it has chosen before the ServerHello
10639 message has informed the client about his choice.)
10643 * Add RFC 3779 support.
10645 *Rob Austein for ARIN, Ben Laurie*
10647 * Load error codes if they are not already present instead of using a
10648 static variable. This allows them to be cleanly unloaded and reloaded.
10649 Improve header file function name parsing.
10653 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10654 or CAPABILITY handshake as required by RFCs.
10656 *Goetz Babin-Ebell*
10658 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10660 * Introduce limits to prevent malicious keys being able to
10661 cause a denial of service. ([CVE-2006-2940])
10663 *Steve Henson, Bodo Moeller*
10665 * Fix ASN.1 parsing of certain invalid structures that can result
10666 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10668 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10669 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10671 * Fix SSL client code which could crash if connecting to a
10672 malicious SSLv2 server. ([CVE-2006-4343])
10674 *Tavis Ormandy and Will Drewry, Google Security Team*
10676 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10677 match only those. Before that, "AES256-SHA" would be interpreted
10678 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10679 the same strength classification in 0.9.7h) as we currently only
10680 have a single AES bit in the ciphersuite description bitmap.
10681 That change, however, also applied to ciphersuite strings such as
10682 "RC4-MD5" that intentionally matched multiple ciphersuites --
10683 namely, SSL 2.0 ciphersuites in addition to the more common ones
10684 from SSL 3.0/TLS 1.0.
10686 So we change the selection algorithm again: Naming an explicit
10687 ciphersuite selects this one ciphersuite, and any other similar
10688 ciphersuite (same bitmap) from *other* protocol versions.
10689 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10690 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10692 Since SSL 2.0 does not have any ciphersuites for which the
10693 128/256 bit distinction would be relevant, this works for now.
10694 The proper fix will be to use different bits for AES128 and
10695 AES256, which would have avoided the problems from the beginning;
10696 however, bits are scarce, so we can only do this in a new release
10697 (not just a patchlevel) when we can change the SSL_CIPHER
10698 definition to split the single 'unsigned long mask' bitmap into
10699 multiple values to extend the available space.
10703 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10705 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10706 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10708 * Add AES IGE and biIGE modes.
10712 * Change the Unix randomness entropy gathering to use poll() when
10713 possible instead of select(), since the latter has some
10714 undesirable limitations.
10716 *Darryl Miles via Richard Levitte and Bodo Moeller*
10718 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10719 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10720 cannot be implicitly activated as part of, e.g., the "AES" alias.
10721 However, please upgrade to OpenSSL 0.9.9[-dev] for
10722 non-experimental use of the ECC ciphersuites to get TLS extension
10723 support, which is required for curve and point format negotiation
10724 to avoid potential handshake problems.
10728 * Disable rogue ciphersuites:
10730 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10731 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10732 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10734 The latter two were purportedly from
10735 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10738 Also deactivate the remaining ciphersuites from
10739 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10740 unofficial, and the ID has long expired.
10744 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10745 dual-core machines) and other potential thread-safety issues.
10749 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10750 versions), which is now available for royalty-free use
10751 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10752 Also, add Camellia TLS ciphersuites from RFC 4132.
10754 To minimize changes between patchlevels in the OpenSSL 0.9.8
10755 series, Camellia remains excluded from compilation unless OpenSSL
10756 is configured with 'enable-camellia'.
10760 * Disable the padding bug check when compression is in use. The padding
10761 bug check assumes the first packet is of even length, this is not
10762 necessarily true if compression is enabled and can result in false
10763 positives causing handshake failure. The actual bug test is ancient
10764 code so it is hoped that implementations will either have fixed it by
10765 now or any which still have the bug do not support compression.
10769 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10771 * When applying a cipher rule check to see if string match is an explicit
10772 cipher suite and only match that one cipher suite if it is.
10776 * Link in manifests for VC++ if needed.
10778 *Austin Ziegler <halostatue@gmail.com>*
10780 * Update support for ECC-based TLS ciphersuites according to
10781 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10782 TLS extensions, which are supported starting with the 0.9.9
10783 branch, not in the OpenSSL 0.9.8 branch).
10787 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10788 opaque EVP_CIPHER_CTX handling.
10792 * Fixes and enhancements to zlib compression code. We now only use
10793 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10794 to conform with the standards mentioned here:
10795 <http://www.zlib.net/DLL_FAQ.txt>
10796 Static zlib linking now works on Windows and the new --with-zlib-include
10797 --with-zlib-lib options to Configure can be used to supply the location
10798 of the headers and library. Gracefully handle case where zlib library
10803 * Several fixes and enhancements to the OID generation code. The old code
10804 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10805 handle numbers larger than ULONG_MAX, truncated printing and had a
10806 non standard OBJ_obj2txt() behaviour.
10810 * Add support for building of engines under engine/ as shared libraries
10811 under VC++ build system.
10815 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10816 Hopefully, we will not see any false combination of paths any more.
10820 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10822 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10823 (part of SSL_OP_ALL). This option used to disable the
10824 countermeasure against man-in-the-middle protocol-version
10825 rollback in the SSL 2.0 server implementation, which is a bad
10826 idea. ([CVE-2005-2969])
10828 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10829 for Information Security, National Institute of Advanced Industrial
10830 Science and Technology [AIST], Japan)*
10832 * Add two function to clear and return the verify parameter flags.
10836 * Keep cipherlists sorted in the source instead of sorting them at
10837 runtime, thus removing the need for a lock.
10841 * Avoid some small subgroup attacks in Diffie-Hellman.
10843 *Nick Mathewson and Ben Laurie*
10845 * Add functions for well-known primes.
10849 * Extended Windows CE support.
10851 *Satoshi Nakamura and Andy Polyakov*
10853 * Initialize SSL_METHOD structures at compile time instead of during
10854 runtime, thus removing the need for a lock.
10858 * Make PKCS7_decrypt() work even if no certificate is supplied by
10859 attempting to decrypt each encrypted key in turn. Add support to
10864 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10866 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10869 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10873 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10874 key into the same file any more.
10878 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10882 * Add -utf8 command line and config file option to 'ca'.
10884 *Stefan <stf@udoma.org*
10886 * Removed the macro des_crypt(), as it seems to conflict with some
10887 libraries. Use DES_crypt().
10891 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10892 involves renaming the source and generated shared-libs for
10893 both. The engines will accept the corrected or legacy ids
10894 ('ncipher' and '4758_cca' respectively) when binding. NB,
10895 this only applies when building 'shared'.
10897 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10899 * Add attribute functions to EVP_PKEY structure. Modify
10900 PKCS12_create() to recognize a CSP name attribute and
10901 use it. Make -CSP option work again in pkcs12 utility.
10905 * Add new functionality to the bn blinding code:
10906 - automatic re-creation of the BN_BLINDING parameters after
10907 a fixed number of uses (currently 32)
10908 - add new function for parameter creation
10909 - introduce flags to control the update behaviour of the
10910 BN_BLINDING parameters
10911 - hide BN_BLINDING structure
10912 Add a second BN_BLINDING slot to the RSA structure to improve
10913 performance when a single RSA object is shared among several
10918 * Add support for DTLS.
10920 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10922 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10923 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10927 * Remove buggy and incomplete DH cert support from
10928 ssl/ssl_rsa.c and ssl/s3_both.c
10932 * Use SHA-1 instead of MD5 as the default digest algorithm for
10933 the `apps/openssl` commands.
10937 * Compile clean with "-Wall -Wmissing-prototypes
10938 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10939 DEBUG_SAFESTACK must also be set.
10943 * Change ./Configure so that certain algorithms can be disabled by default.
10944 The new counterpiece to "no-xxx" is "enable-xxx".
10946 The patented RC5 and MDC2 algorithms will now be disabled unless
10947 "enable-rc5" and "enable-mdc2", respectively, are specified.
10949 (IDEA remains enabled despite being patented. This is because IDEA
10950 is frequently required for interoperability, and there is no license
10951 fee for non-commercial use. As before, "no-idea" can be used to
10952 avoid this algorithm.)
10956 * Add processing of proxy certificates (see RFC 3820). This work was
10957 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10958 EGEE (Enabling Grids for E-science in Europe).
10962 * RC4 performance overhaul on modern architectures/implementations, such
10963 as Intel P4, IA-64 and AMD64.
10967 * New utility extract-section.pl. This can be used specify an alternative
10968 section number in a pod file instead of having to treat each file as
10969 a separate case in Makefile. This can be done by adding two lines to the
10972 =for comment openssl_section:XXX
10974 The blank line is mandatory.
10978 * New arguments -certform, -keyform and -pass for s_client and s_server
10979 to allow alternative format key and certificate files and passphrase
10984 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10985 update associated structures and add various utility functions.
10987 Add new policy related verify parameters, include policy checking in
10988 standard verify code. Enhance 'smime' application with extra parameters
10989 to support policy checking and print out.
10993 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10994 Nehemiah processors. These extensions support AES encryption in hardware
10995 as well as RNG (though RNG support is currently disabled).
10997 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10999 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11003 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11005 *Andy Polyakov and a number of other people*
11007 * Improved PowerPC platform support. Most notably BIGNUM assembler
11008 implementation contributed by IBM.
11010 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11012 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11013 exponent rather than 'unsigned long'. There is a corresponding change to
11014 the new 'rsa_keygen' element of the RSA_METHOD structure.
11016 *Jelte Jansen, Geoff Thorpe*
11018 * Functionality for creating the initial serial number file is now
11019 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11021 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11022 number file to 1, which is bound to cause problems. To avoid
11023 the problems while respecting compatibility between different 0.9.7
11024 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11025 CA.pl for serial number initialization. With the new release 0.9.8,
11026 we can fix the problem directly in the 'ca' utility.)
11030 * Reduced header interdependencies by declaring more opaque objects in
11031 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11032 give fewer recursive includes, which could break lazy source code - so
11033 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11034 developers should define this symbol when building and using openssl to
11035 ensure they track the recommended behaviour, interfaces, [etc], but
11036 backwards-compatible behaviour prevails when this isn't defined.
11040 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11044 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11045 This will generate a random key of the appropriate length based on the
11046 cipher context. The EVP_CIPHER can provide its own random key generation
11047 routine to support keys of a specific form. This is used in the des and
11048 3des routines to generate a key of the correct parity. Update S/MIME
11049 code to use new functions and hence generate correct parity DES keys.
11050 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11051 valid (weak or incorrect parity).
11055 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11056 as looking them up. This is useful when the verified structure may contain
11057 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11058 present unless the new PKCS7_NO_CRL flag is asserted.
11062 * Extend ASN1 oid configuration module. It now additionally accepts the
11065 shortName = some long name, 1.2.3.4
11069 * Reimplemented the BN_CTX implementation. There is now no more static
11070 limitation on the number of variables it can handle nor the depth of the
11071 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11072 information can now expand as required, and rather than having a single
11073 static array of bignums, BN_CTX now uses a linked-list of such arrays
11074 allowing it to expand on demand whilst maintaining the usefulness of
11075 BN_CTX's "bundling".
11079 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11080 to allow all RSA operations to function using a single BN_CTX.
11084 * Preliminary support for certificate policy evaluation and checking. This
11085 is initially intended to pass the tests outlined in "Conformance Testing
11086 of Relying Party Client Certificate Path Processing Logic" v1.07.
11090 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11091 remained unused and not that useful. A variety of other little bignum
11092 tweaks and fixes have also been made continuing on from the audit (see
11097 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11098 associated ASN1, EVP and SSL functions and old ASN1 macros.
11102 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11103 and this should never fail. So the return value from the use of
11104 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11105 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11109 * BN_CTX_get() should return zero-valued bignums, providing the same
11110 initialised value as BN_new().
11112 *Geoff Thorpe, suggested by Ulf Möller*
11114 * Support for inhibitAnyPolicy certificate extension.
11118 * An audit of the BIGNUM code is underway, for which debugging code is
11119 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11120 is considered valid when processing BIGNUMs, and causes execution to
11121 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11122 further steps are taken to deliberately pollute unused data in BIGNUM
11123 structures to try and expose faulty code further on. For now, openssl will
11124 (in its default mode of operation) continue to tolerate the inconsistent
11125 forms that it has tolerated in the past, but authors and packagers should
11126 consider trying openssl and their own applications when compiled with
11127 these debugging symbols defined. It will help highlight potential bugs in
11128 their own code, and will improve the test coverage for OpenSSL itself. At
11129 some point, these tighter rules will become openssl's default to improve
11130 maintainability, though the assert()s and other overheads will remain only
11131 in debugging configurations. See bn.h for more details.
11133 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11135 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11136 that can only be obtained through BN_CTX_new() (which implicitly
11137 initialises it). The presence of this function only made it possible
11138 to overwrite an existing structure (and cause memory leaks).
11142 * Because of the callback-based approach for implementing LHASH as a
11143 template type, lh_insert() adds opaque objects to hash-tables and
11144 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11145 to clean up those corresponding objects before destroying the hash table
11146 (and losing the object pointers). So some over-zealous constifications in
11147 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11148 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11149 prototyped to have "const" restrictions on the object pointers they are
11150 given (and so aren't required to cast them away any more).
11154 * The tmdiff.h API was so ugly and minimal that our own timing utility
11155 (speed) prefers to use its own implementation. The two implementations
11156 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11157 its object type properly exposed (MS_TM) instead of casting to/from
11158 `char *`. This may still change yet if someone realises MS_TM and
11160 aren't necessarily the greatest nomenclatures - but this is what was used
11161 internally to the implementation so I've used that for now.
11165 * Ensure that deprecated functions do not get compiled when
11166 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11167 the self-tests were still using deprecated key-generation functions so
11168 these have been updated also.
11172 * Reorganise PKCS#7 code to separate the digest location functionality
11173 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11174 New function PKCS7_set_digest() to set the digest type for PKCS#7
11175 digestedData type. Add additional code to correctly generate the
11176 digestedData type and add support for this type in PKCS7 initialization
11181 * New function PKCS7_set0_type_other() this initializes a PKCS7
11182 structure of type "other".
11186 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11187 sure the loop does correctly stop and breaking ("division by zero")
11188 modulus operations are not performed. The (pre-generated) prime
11189 table crypto/bn/bn_prime.h was already correct, but it could not be
11190 re-generated on some platforms because of the "division by zero"
11191 situation in the script.
11193 *Ralf S. Engelschall*
11195 * Update support for ECC-based TLS ciphersuites according to
11196 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11197 SHA-1 now is only used for "small" curves (where the
11198 representation of a field element takes up to 24 bytes); for
11199 larger curves, the field element resulting from ECDH is directly
11200 used as premaster secret.
11202 *Douglas Stebila (Sun Microsystems Laboratories)*
11204 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11205 curve secp160r1 to the tests.
11207 *Douglas Stebila (Sun Microsystems Laboratories)*
11209 * Add the possibility to load symbols globally with DSO.
11211 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11213 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11214 control of the error stack.
11218 * Add support for STORE in ENGINE.
11222 * Add the STORE type. The intention is to provide a common interface
11223 to certificate and key stores, be they simple file-based stores, or
11224 HSM-type store, or LDAP stores, or...
11225 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11229 * Add a generic structure called OPENSSL_ITEM. This can be used to
11230 pass a list of arguments to any function as well as provide a way
11231 for a function to pass data back to the caller.
11235 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11236 works like BUF_strdup() but can be used to duplicate a portion of
11237 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11242 * Add the function sk_find_ex() which works like sk_find(), but will
11243 return an index to an element even if an exact match couldn't be
11244 found. The index is guaranteed to point at the element where the
11245 searched-for key would be inserted to preserve sorting order.
11249 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11250 takes an extra flags argument for optional functionality. Currently,
11251 the following flags are defined:
11253 OBJ_BSEARCH_VALUE_ON_NOMATCH
11254 This one gets OBJ_bsearch_ex() to return a pointer to the first
11255 element where the comparing function returns a negative or zero
11258 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11259 This one gets OBJ_bsearch_ex() to return a pointer to the first
11260 element where the comparing function returns zero. This is useful
11261 if there are more than one element where the comparing function
11266 * Make it possible to create self-signed certificates with 'openssl ca'
11267 in such a way that the self-signed certificate becomes part of the
11268 CA database and uses the same mechanisms for serial number generation
11269 as all other certificate signing. The new flag '-selfsign' enables
11270 this functionality. Adapt CA.sh and CA.pl.in.
11274 * Add functionality to check the public key of a certificate request
11275 against a given private. This is useful to check that a certificate
11276 request can be signed by that key (self-signing).
11280 * Make it possible to have multiple active certificates with the same
11281 subject in the CA index file. This is done only if the keyword
11282 'unique_subject' is set to 'no' in the main CA section (default
11283 if 'CA_default') of the configuration file. The value is saved
11284 with the database itself in a separate index attribute file,
11285 named like the index file with '.attr' appended to the name.
11289 * Generate multi-valued AVAs using '+' notation in config files for
11294 * Support for nameConstraints certificate extension.
11298 * Support for policyConstraints certificate extension.
11302 * Support for policyMappings certificate extension.
11306 * Make sure the default DSA_METHOD implementation only uses its
11307 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11308 and change its own handlers to be NULL so as to remove unnecessary
11309 indirection. This lets alternative implementations fallback to the
11310 default implementation more easily.
11314 * Support for directoryName in GeneralName related extensions
11319 * Make it possible to link applications using Makefile.shared.
11320 Make that possible even when linking against static libraries!
11324 * Support for single pass processing for S/MIME signing. This now
11325 means that S/MIME signing can be done from a pipe, in addition
11326 cleartext signing (multipart/signed type) is effectively streaming
11327 and the signed data does not need to be all held in memory.
11329 This is done with a new flag PKCS7_STREAM. When this flag is set
11330 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11331 is done after the data is output (and digests calculated) in
11332 SMIME_write_PKCS7().
11336 * Add full support for -rpath/-R, both in shared libraries and
11337 applications, at least on the platforms where it's known how
11342 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11343 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11344 will now compute a table of multiples of the generator that
11345 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11346 faster (notably in the case of a single point multiplication,
11347 scalar * generator).
11349 *Nils Larsch, Bodo Moeller*
11351 * IPv6 support for certificate extensions. The various extensions
11352 which use the IP:a.b.c.d can now take IPv6 addresses using the
11353 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11358 * Added an ENGINE that implements RSA by performing private key
11359 exponentiations with the GMP library. The conversions to and from
11360 GMP's mpz_t format aren't optimised nor are any montgomery forms
11361 cached, and on x86 it appears OpenSSL's own performance has caught up.
11362 However there are likely to be other architectures where GMP could
11363 provide a boost. This ENGINE is not built in by default, but it can be
11364 specified at Configure time and should be accompanied by the necessary
11365 linker additions, eg;
11366 ./config -DOPENSSL_USE_GMP -lgmp
11370 * "openssl engine" will not display ENGINE/DSO load failure errors when
11371 testing availability of engines with "-t" - the old behaviour is
11372 produced by increasing the feature's verbosity with "-tt".
11376 * ECDSA routines: under certain error conditions uninitialized BN objects
11377 could be freed. Solution: make sure initialization is performed early
11378 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11383 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11384 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11385 software implementations. For DSA and DH, parameter generation can
11386 also be overridden by providing the appropriate method callbacks.
11390 * Change the "progress" mechanism used in key-generation and
11391 primality testing to functions that take a new BN_GENCB pointer in
11392 place of callback/argument pairs. The new API functions have `_ex`
11393 postfixes and the older functions are reimplemented as wrappers for
11394 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11395 declarations of the old functions to help (graceful) attempts to
11396 migrate to the new functions. Also, the new key-generation API
11397 functions operate on a caller-supplied key-structure and return
11398 success/failure rather than returning a key or NULL - this is to
11399 help make "keygen" another member function of RSA_METHOD etc.
11401 Example for using the new callback interface:
11403 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11404 void *my_arg = ...;
11407 BN_GENCB_set(&my_cb, my_callback, my_arg);
11409 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11410 /* For the meaning of a, b in calls to my_callback(), see the
11411 * documentation of the function that calls the callback.
11412 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11413 * my_callback should return 1 if it wants BN_is_prime_ex()
11414 * to continue, or 0 to stop.
11419 * Change the ZLIB compression method to be stateful, and make it
11420 available to TLS with the number defined in
11421 draft-ietf-tls-compression-04.txt.
11425 * Add the ASN.1 structures and functions for CertificatePair, which
11426 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11428 CertificatePair ::= SEQUENCE {
11429 forward [0] Certificate OPTIONAL,
11430 reverse [1] Certificate OPTIONAL,
11431 -- at least one of the pair shall be present -- }
11433 Also implement the PEM functions to read and write certificate
11434 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11436 This needed to be defined, mostly for the sake of the LDAP
11437 attribute crossCertificatePair, but may prove useful elsewhere as
11442 * Make it possible to inhibit symlinking of shared libraries in
11443 Makefile.shared, for Cygwin's sake.
11447 * Extend the BIGNUM API by creating a function
11448 void BN_set_negative(BIGNUM *a, int neg);
11449 and a macro that behave like
11450 int BN_is_negative(const BIGNUM *a);
11452 to avoid the need to access 'a->neg' directly in applications.
11456 * Implement fast modular reduction for pseudo-Mersenne primes
11457 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11458 EC_GROUP_new_curve_GFp() will now automatically use this
11461 *Nils Larsch <nla@trustcenter.de>*
11463 * Add new lock type (CRYPTO_LOCK_BN).
11467 * Change the ENGINE framework to automatically load engines
11468 dynamically from specific directories unless they could be
11469 found to already be built in or loaded. Move all the
11470 current engines except for the cryptodev one to a new
11471 directory engines/.
11472 The engines in engines/ are built as shared libraries if
11473 the "shared" options was given to ./Configure or ./config.
11474 Otherwise, they are inserted in libcrypto.a.
11475 /usr/local/ssl/engines is the default directory for dynamic
11476 engines, but that can be overridden at configure time through
11477 the usual use of --prefix and/or --openssldir, and at run
11478 time with the environment variable OPENSSL_ENGINES.
11480 *Geoff Thorpe and Richard Levitte*
11482 * Add Makefile.shared, a helper makefile to build shared
11483 libraries. Adapt Makefile.org.
11487 * Add version info to Win32 DLLs.
11489 *Peter 'Luna' Runestig" <peter@runestig.com>*
11491 * Add new 'medium level' PKCS#12 API. Certificates and keys
11492 can be added using this API to created arbitrary PKCS#12
11493 files while avoiding the low-level API.
11495 New options to PKCS12_create(), key or cert can be NULL and
11496 will then be omitted from the output file. The encryption
11497 algorithm NIDs can be set to -1 for no encryption, the mac
11498 iteration count can be set to 0 to omit the mac.
11500 Enhance pkcs12 utility by making the -nokeys and -nocerts
11501 options work when creating a PKCS#12 file. New option -nomac
11502 to omit the mac, NONE can be set for an encryption algorithm.
11503 New code is modified to use the enhanced PKCS12_create()
11504 instead of the low-level API.
11508 * Extend ASN1 encoder to support indefinite length constructed
11509 encoding. This can output sequences tags and octet strings in
11510 this form. Modify pk7_asn1.c to support indefinite length
11511 encoding. This is experimental and needs additional code to
11512 be useful, such as an ASN1 bio and some enhanced streaming
11515 Extend template encode functionality so that tagging is passed
11516 down to the template encoder.
11520 * Let 'openssl req' fail if an argument to '-newkey' is not
11521 recognized instead of using RSA as a default.
11525 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11526 As these are not official, they are not included in "ALL";
11527 the "ECCdraft" ciphersuite group alias can be used to select them.
11529 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11531 * Add ECDH engine support.
11533 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11535 * Add ECDH in new directory crypto/ecdh/.
11537 *Douglas Stebila (Sun Microsystems Laboratories)*
11539 * Let BN_rand_range() abort with an error after 100 iterations
11540 without success (which indicates a broken PRNG).
11544 * Change BN_mod_sqrt() so that it verifies that the input value
11545 is really the square of the return value. (Previously,
11546 BN_mod_sqrt would show GIGO behaviour.)
11550 * Add named elliptic curves over binary fields from X9.62, SECG,
11551 and WAP/WTLS; add OIDs that were still missing.
11553 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11555 * Extend the EC library for elliptic curves over binary fields
11556 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11559 EC_GF2m_simple_method
11563 EC_GROUP_new_curve_GF2m
11564 EC_GROUP_set_curve_GF2m
11565 EC_GROUP_get_curve_GF2m
11566 EC_POINT_set_affine_coordinates_GF2m
11567 EC_POINT_get_affine_coordinates_GF2m
11568 EC_POINT_set_compressed_coordinates_GF2m
11570 Point compression for binary fields is disabled by default for
11571 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11574 As binary polynomials are represented as BIGNUMs, various members
11575 of the EC_GROUP and EC_POINT data structures can be shared
11576 between the implementations for prime fields and binary fields;
11577 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11578 are essentially identical to their `..._GFp` counterparts.
11579 (For simplicity, the `..._GFp` prefix has been dropped from
11580 various internal method names.)
11582 An internal 'field_div' method (similar to 'field_mul' and
11583 'field_sqr') has been added; this is used only for binary fields.
11585 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11587 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11588 through methods ('mul', 'precompute_mult').
11590 The generic implementations (now internally called 'ec_wNAF_mul'
11591 and 'ec_wNAF_precomputed_mult') remain the default if these
11592 methods are undefined.
11594 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11596 * New function EC_GROUP_get_degree, which is defined through
11597 EC_METHOD. For curves over prime fields, this returns the bit
11598 length of the modulus.
11600 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11602 * New functions EC_GROUP_dup, EC_POINT_dup.
11603 (These simply call ..._new and ..._copy).
11605 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11607 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11608 Polynomials are represented as BIGNUMs (where the sign bit is not
11609 used) in the following functions [macros]:
11612 BN_GF2m_sub [= BN_GF2m_add]
11613 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11614 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11615 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11617 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11618 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11619 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11620 BN_GF2m_cmp [= BN_ucmp]
11622 (Note that only the 'mod' functions are actually for fields GF(2^m).
11623 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11625 For some functions, an the irreducible polynomial defining a
11626 field can be given as an 'unsigned int[]' with strictly
11627 decreasing elements giving the indices of those bits that are set;
11628 i.e., p[] represents the polynomial
11629 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11631 p[0] > p[1] > ... > p[k] = 0.
11632 This applies to the following functions:
11635 BN_GF2m_mod_mul_arr
11636 BN_GF2m_mod_sqr_arr
11637 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11638 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11639 BN_GF2m_mod_exp_arr
11640 BN_GF2m_mod_sqrt_arr
11641 BN_GF2m_mod_solve_quad_arr
11645 Conversion can be performed by the following functions:
11650 bntest.c has additional tests for binary polynomial arithmetic.
11652 Two implementations for BN_GF2m_mod_div() are available.
11653 The default algorithm simply uses BN_GF2m_mod_inv() and
11654 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11655 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11656 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11658 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11660 * Add new error code 'ERR_R_DISABLED' that can be used when some
11661 functionality is disabled at compile-time.
11663 *Douglas Stebila <douglas.stebila@sun.com>*
11665 * Change default behaviour of 'openssl asn1parse' so that more
11666 information is visible when viewing, e.g., a certificate:
11668 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11669 mode the content of non-printable OCTET STRINGs is output in a
11670 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11671 avoid the appearance of a printable string.
11673 *Nils Larsch <nla@trustcenter.de>*
11675 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11677 EC_GROUP_set_asn1_flag()
11678 EC_GROUP_get_asn1_flag()
11679 EC_GROUP_set_point_conversion_form()
11680 EC_GROUP_get_point_conversion_form()
11681 These control ASN1 encoding details:
11682 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11683 has been set to OPENSSL_EC_NAMED_CURVE.
11684 - Points are encoded in uncompressed form by default; options for
11685 asn1_for are as for point2oct, namely
11686 POINT_CONVERSION_COMPRESSED
11687 POINT_CONVERSION_UNCOMPRESSED
11688 POINT_CONVERSION_HYBRID
11690 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11692 EC_GROUP_set_seed()
11693 EC_GROUP_get0_seed()
11694 EC_GROUP_get_seed_len()
11695 This is used only for ASN1 purposes (so far).
11697 *Nils Larsch <nla@trustcenter.de>*
11699 * Add 'field_type' member to EC_METHOD, which holds the NID
11700 of the appropriate field type OID. The new function
11701 EC_METHOD_get_field_type() returns this value.
11703 *Nils Larsch <nla@trustcenter.de>*
11706 EC_POINT_point2bn()
11707 EC_POINT_bn2point()
11708 EC_POINT_point2hex()
11709 EC_POINT_hex2point()
11710 providing useful interfaces to EC_POINT_point2oct() and
11711 EC_POINT_oct2point().
11713 *Nils Larsch <nla@trustcenter.de>*
11715 * Change internals of the EC library so that the functions
11716 EC_GROUP_set_generator()
11717 EC_GROUP_get_generator()
11718 EC_GROUP_get_order()
11719 EC_GROUP_get_cofactor()
11720 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11721 to methods, which would lead to unnecessary code duplication when
11722 adding different types of curves.
11724 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11726 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11727 arithmetic, and such that modified wNAFs are generated
11728 (which avoid length expansion in many cases).
11732 * Add a function EC_GROUP_check_discriminant() (defined via
11733 EC_METHOD) that verifies that the curve discriminant is non-zero.
11735 Add a function EC_GROUP_check() that makes some sanity tests
11736 on a EC_GROUP, its generator and order. This includes
11737 EC_GROUP_check_discriminant().
11739 *Nils Larsch <nla@trustcenter.de>*
11741 * Add ECDSA in new directory crypto/ecdsa/.
11743 Add applications 'openssl ecparam' and 'openssl ecdsa'
11744 (these are based on 'openssl dsaparam' and 'openssl dsa').
11746 ECDSA support is also included in various other files across the
11747 library. Most notably,
11748 - 'openssl req' now has a '-newkey ecdsa:file' option;
11749 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11750 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11751 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11752 them suitable for ECDSA where domain parameters must be
11753 extracted before the specific public key;
11754 - ECDSA engine support has been added.
11756 *Nils Larsch <nla@trustcenter.de>*
11758 * Include some named elliptic curves, and add OIDs from X9.62,
11759 SECG, and WAP/WTLS. Each curve can be obtained from the new
11761 EC_GROUP_new_by_curve_name(),
11762 and the list of available named curves can be obtained with
11763 EC_get_builtin_curves().
11764 Also add a 'curve_name' member to EC_GROUP objects, which can be
11766 EC_GROUP_set_curve_name()
11767 EC_GROUP_get_curve_name()
11769 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11771 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11772 was actually never needed) and in BN_mul(). The removal in BN_mul()
11773 required a small change in bn_mul_part_recursive() and the addition
11774 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11775 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11776 bn_sub_words() and bn_add_words() except they take arrays with
11781 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11783 * Cleanse PEM buffers before freeing them since they may contain
11786 *Benjamin Bennett <ben@psc.edu>*
11788 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11789 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11790 authentication-only ciphersuites.
11794 * Since AES128 and AES256 share a single mask bit in the logic of
11795 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11796 kludge to work properly if AES128 is available and AES256 isn't.
11800 * Expand security boundary to match 1.1.1 module.
11804 * Remove redundant features: hash file source, editing of test vectors
11805 modify fipsld to use external fips_premain.c signature.
11809 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11810 run algorithm test programs.
11814 * Make algorithm test programs more tolerant of whitespace.
11818 * Have SSL/TLS server implementation tolerate "mismatched" record
11819 protocol version while receiving ClientHello even if the
11820 ClientHello is fragmented. (The server can't insist on the
11821 particular protocol version it has chosen before the ServerHello
11822 message has informed the client about his choice.)
11826 * Load error codes if they are not already present instead of using a
11827 static variable. This allows them to be cleanly unloaded and reloaded.
11831 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11833 * Introduce limits to prevent malicious keys being able to
11834 cause a denial of service. ([CVE-2006-2940])
11836 *Steve Henson, Bodo Moeller*
11838 * Fix ASN.1 parsing of certain invalid structures that can result
11839 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11841 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11842 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11844 * Fix SSL client code which could crash if connecting to a
11845 malicious SSLv2 server. ([CVE-2006-4343])
11847 *Tavis Ormandy and Will Drewry, Google Security Team*
11849 * Change ciphersuite string processing so that an explicit
11850 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11851 will no longer include "AES128-SHA"), and any other similar
11852 ciphersuite (same bitmap) from *other* protocol versions (so that
11853 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11854 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11855 changes from 0.9.8b and 0.9.8d.
11859 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11861 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11862 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11864 * Change the Unix randomness entropy gathering to use poll() when
11865 possible instead of select(), since the latter has some
11866 undesirable limitations.
11868 *Darryl Miles via Richard Levitte and Bodo Moeller*
11870 * Disable rogue ciphersuites:
11872 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11873 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11874 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11876 The latter two were purportedly from
11877 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11880 Also deactivate the remaining ciphersuites from
11881 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11882 unofficial, and the ID has long expired.
11886 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11887 dual-core machines) and other potential thread-safety issues.
11891 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11893 * Adapt fipsld and the build system to link against the validated FIPS
11894 module in FIPS mode.
11898 * Fixes for VC++ 2005 build under Windows.
11902 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11903 from a Windows bash shell such as MSYS. It is autodetected from the
11904 "config" script when run from a VC++ environment. Modify standard VC++
11905 build to use fipscanister.o from the GNU make build.
11909 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11911 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11912 The value now differs depending on if you build for FIPS or not.
11913 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11914 safely run with a non-FIPSed libcrypto, as it may crash because of
11915 the difference induced by this change.
11919 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11921 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11922 (part of SSL_OP_ALL). This option used to disable the
11923 countermeasure against man-in-the-middle protocol-version
11924 rollback in the SSL 2.0 server implementation, which is a bad
11925 idea. ([CVE-2005-2969])
11927 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11928 for Information Security, National Institute of Advanced Industrial
11929 Science and Technology [AIST, Japan)]*
11931 * Minimal support for X9.31 signatures and PSS padding modes. This is
11932 mainly for FIPS compliance and not fully integrated at this stage.
11936 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11937 the exponentiation using a fixed-length exponent. (Otherwise,
11938 the information leaked through timing could expose the secret key
11939 after many signatures; cf. Bleichenbacher's attack on DSA with
11944 * Make a new fixed-window mod_exp implementation the default for
11945 RSA, DSA, and DH private-key operations so that the sequence of
11946 squares and multiplies and the memory access pattern are
11947 independent of the particular secret key. This will mitigate
11948 cache-timing and potential related attacks.
11950 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11951 and this is automatically used by BN_mod_exp_mont() if the new flag
11952 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11953 will use this BN flag for private exponents unless the flag
11954 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11955 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11957 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11959 * Change the client implementation for SSLv23_method() and
11960 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11961 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11962 (Previously, the SSL 2.0 backwards compatible Client Hello
11963 message format would be used even with SSL_OP_NO_SSLv2.)
11967 * Add support for smime-type MIME parameter in S/MIME messages which some
11972 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11973 a threadsafe manner. Modify rsa code to use new function and add calls
11974 to dsa and dh code (which had race conditions before).
11978 * Include the fixed error library code in the C error file definitions
11979 instead of fixing them up at runtime. This keeps the error code
11980 structures constant.
11984 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11986 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11989 * Fixes for newer kerberos headers. NB: the casts are needed because
11990 the 'length' field is signed on one version and unsigned on another
11991 with no (?) obvious way to tell the difference, without these VC++
11992 complains. Also the "definition" of FAR (blank) is no longer included
11993 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11994 some needed definitions.
11998 * Undo Cygwin change.
12002 * Added support for proxy certificates according to RFC 3820.
12003 Because they may be a security thread to unaware applications,
12004 they must be explicitly allowed in run-time. See
12005 docs/HOWTO/proxy_certificates.txt for further information.
12009 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12011 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12012 server and client random values. Previously
12013 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12014 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12016 This change has negligible security impact because:
12018 1. Server and client random values still have 24 bytes of pseudo random
12021 2. Server and client random values are sent in the clear in the initial
12024 3. The master secret is derived using the premaster secret (48 bytes in
12025 size for static RSA ciphersuites) as well as client server and random
12028 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12031 *Stephen Henson, reported by UK NISCC*
12033 * Use Windows randomness collection on Cygwin.
12037 * Fix hang in EGD/PRNGD query when communication socket is closed
12038 prematurely by EGD/PRNGD.
12040 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12042 * Prompt for pass phrases when appropriate for PKCS12 input format.
12046 * Back-port of selected performance improvements from development
12047 branch, as well as improved support for PowerPC platforms.
12051 * Add lots of checks for memory allocation failure, error codes to indicate
12052 failure and freeing up memory if a failure occurs.
12054 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12056 * Add new -passin argument to dgst.
12060 * Perform some character comparisons of different types in X509_NAME_cmp:
12061 this is needed for some certificates that re-encode DNs into UTF8Strings
12062 (in violation of RFC3280) and can't or won't issue name rollover
12067 * Make an explicit check during certificate validation to see that
12068 the CA setting in each certificate on the chain is correct. As a
12069 side effect always do the following basic checks on extensions,
12070 not just when there's an associated purpose to the check:
12072 - if there is an unhandled critical extension (unless the user
12073 has chosen to ignore this fault)
12074 - if the path length has been exceeded (if one is set at all)
12075 - that certain extensions fit the associated purpose (if one has
12080 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12082 * Avoid a race condition when CRLs are checked in a multi threaded
12083 environment. This would happen due to the reordering of the revoked
12084 entries during signature checking and serial number lookup. Now the
12085 encoding is cached and the serial number sort performed under a lock.
12086 Add new STACK function sk_is_sorted().
12090 * Add Delta CRL to the extension code.
12094 * Various fixes to s3_pkt.c so alerts are sent properly.
12096 *David Holmes <d.holmes@f5.com>*
12098 * Reduce the chances of duplicate issuer name and serial numbers (in
12099 violation of RFC3280) using the OpenSSL certificate creation utilities.
12100 This is done by creating a random 64 bit value for the initial serial
12101 number when a serial number file is created or when a self signed
12102 certificate is created using 'openssl req -x509'. The initial serial
12103 number file is created using 'openssl x509 -next_serial' in CA.pl
12104 rather than being initialized to 1.
12108 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12110 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12111 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12113 *Joe Orton, Steve Henson*
12115 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12118 *Joe Orton, Steve Henson*
12120 * Make it possible to have multiple active certificates with the same
12121 subject in the CA index file. This is done only if the keyword
12122 'unique_subject' is set to 'no' in the main CA section (default
12123 if 'CA_default') of the configuration file. The value is saved
12124 with the database itself in a separate index attribute file,
12125 named like the index file with '.attr' appended to the name.
12129 * X509 verify fixes. Disable broken certificate workarounds when
12130 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12131 keyUsage extension present. Don't accept CRLs with unhandled critical
12132 extensions: since verify currently doesn't process CRL extensions this
12133 rejects a CRL with *any* critical extensions. Add new verify error codes
12138 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12139 A clarification of RFC2560 will require the use of OCTET STRINGs and
12140 some implementations cannot handle the current raw format. Since OpenSSL
12141 copies and compares OCSP nonces as opaque blobs without any attempt at
12142 parsing them this should not create any compatibility issues.
12146 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12147 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12148 this HMAC (and other) operations are several times slower than OpenSSL
12153 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12155 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12157 * Use the correct content when signing type "other".
12161 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12163 * Fix various bugs revealed by running the NISCC test suite:
12165 Stop out of bounds reads in the ASN1 code when presented with
12166 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12168 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12170 If verify callback ignores invalid public key errors don't try to check
12171 certificate signature with the NULL public key.
12175 * New -ignore_err option in ocsp application to stop the server
12176 exiting on the first error in a request.
12180 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12181 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12186 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12187 extra data after the compression methods not only for TLS 1.0
12188 but also for SSL 3.0 (as required by the specification).
12190 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12192 * Change X509_certificate_type() to mark the key as exported/exportable
12193 when it's 512 *bits* long, not 512 bytes.
12197 * Change AES_cbc_encrypt() so it outputs exact multiple of
12198 blocks during encryption.
12202 * Various fixes to base64 BIO and non blocking I/O. On write
12203 flushes were not handled properly if the BIO retried. On read
12204 data was not being buffered properly and had various logic bugs.
12205 This also affects blocking I/O when the data being decoded is a
12210 * Various S/MIME bugfixes and compatibility changes:
12211 output correct application/pkcs7 MIME type if
12212 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12213 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12214 of files as .eml work). Correctly handle very long lines in MIME
12219 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12221 * Countermeasure against the Klima-Pokorny-Rosa extension of
12222 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12223 a protocol version number mismatch like a decryption error
12224 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12228 * Turn on RSA blinding by default in the default implementation
12229 to avoid a timing attack. Applications that don't want it can call
12230 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12231 They would be ill-advised to do so in most cases.
12233 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12235 * Change RSA blinding code so that it works when the PRNG is not
12236 seeded (in this case, the secret RSA exponent is abused as
12237 an unpredictable seed -- if it is not unpredictable, there
12238 is no point in blinding anyway). Make RSA blinding thread-safe
12239 by remembering the creator's thread ID in rsa->blinding and
12240 having all other threads use local one-time blinding factors
12241 (this requires more computation than sharing rsa->blinding, but
12242 avoids excessive locking; and if an RSA object is not shared
12243 between threads, blinding will still be very fast).
12247 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12248 ENGINE as defaults for all supported algorithms irrespective of
12249 the 'flags' parameter. 'flags' is now honoured, so applications
12250 should make sure they are passing it correctly.
12254 * Target "mingw" now allows native Windows code to be generated in
12255 the Cygwin environment as well as with the MinGW compiler.
12259 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12261 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12262 via timing by performing a MAC computation even if incorrect
12263 block cipher padding has been found. This is a countermeasure
12264 against active attacks where the attacker has to distinguish
12265 between bad padding and a MAC verification error. ([CVE-2003-0078])
12267 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12268 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12269 Martin Vuagnoux (EPFL, Ilion)*
12271 * Make the no-err option work as intended. The intention with no-err
12272 is not to have the whole error stack handling routines removed from
12273 libcrypto, it's only intended to remove all the function name and
12274 reason texts, thereby removing some of the footprint that may not
12275 be interesting if those errors aren't displayed anyway.
12277 NOTE: it's still possible for any application or module to have its
12278 own set of error texts inserted. The routines are there, just not
12279 used by default when no-err is given.
12283 * Add support for FreeBSD on IA64.
12285 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12287 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12288 Kerberos function mit_des_cbc_cksum(). Before this change,
12289 the value returned by DES_cbc_cksum() was like the one from
12290 mit_des_cbc_cksum(), except the bytes were swapped.
12292 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12294 * Allow an application to disable the automatic SSL chain building.
12295 Before this a rather primitive chain build was always performed in
12296 ssl3_output_cert_chain(): an application had no way to send the
12297 correct chain if the automatic operation produced an incorrect result.
12299 Now the chain builder is disabled if either:
12301 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12303 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12305 The reasoning behind this is that an application would not want the
12306 auto chain building to take place if extra chain certificates are
12307 present and it might also want a means of sending no additional
12308 certificates (for example the chain has two certificates and the
12313 * Add the possibility to build without the ENGINE framework.
12315 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12317 * Under Win32 gmtime() can return NULL: check return value in
12318 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12322 * DSA routines: under certain error conditions uninitialized BN objects
12323 could be freed. Solution: make sure initialization is performed early
12324 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12325 Nils Larsch <nla@trustcenter.de> via PR#459)
12329 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12330 checked on reconnect on the client side, therefore session resumption
12331 could still fail with a "ssl session id is different" error. This
12332 behaviour is masked when SSL_OP_ALL is used due to
12333 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12334 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12335 followup to PR #377.
12339 * IA-32 assembler support enhancements: unified ELF targets, support
12340 for SCO/Caldera platforms, fix for Cygwin shared build.
12344 * Add support for FreeBSD on sparc64. As a consequence, support for
12345 FreeBSD on non-x86 processors is separate from x86 processors on
12346 the config script, much like the NetBSD support.
12348 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12350 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12352 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12355 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12356 code (06) was taken as the first octet of the session ID and the last
12357 octet was ignored consequently. As a result SSLv2 client side session
12358 caching could not have worked due to the session ID mismatch between
12360 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12365 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12366 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12371 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12372 seems that in spite of existing for more than a year, many application
12373 author have done nothing to provide the necessary callbacks, which
12374 means that this particular engine will not work properly anywhere.
12375 This is a very unfortunate situation which forces us, in the name
12376 of usability, to give the hw_ncipher.c a static lock, which is part
12378 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12379 appear in 0.9.8 or later. We EXPECT application authors to have
12380 dealt properly with this when 0.9.8 is released (unless we actually
12381 make such changes in the libcrypto locking code that changes will
12382 have to be made anyway).
12386 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12387 octets have been read, EOF or an error occurs. Without this change
12388 some truncated ASN1 structures will not produce an error.
12392 * Disable Heimdal support, since it hasn't been fully implemented.
12393 Still give the possibility to force the use of Heimdal, but with
12394 warnings and a request that patches get sent to openssl-dev.
12398 * Add the VC-CE target, introduce the WINCE sysname, and add
12399 INSTALL.WCE and appropriate conditionals to make it build.
12401 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12403 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12404 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12405 edit numbers of the version.
12407 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12409 * Introduce safe string copy and catenation functions
12410 (BUF_strlcpy() and BUF_strlcat()).
12412 *Ben Laurie (CHATS) and Richard Levitte*
12414 * Avoid using fixed-size buffers for one-line DNs.
12416 *Ben Laurie (CHATS)*
12418 * Add BUF_MEM_grow_clean() to avoid information leakage when
12419 resizing buffers containing secrets, and use where appropriate.
12421 *Ben Laurie (CHATS)*
12423 * Avoid using fixed size buffers for configuration file location.
12425 *Ben Laurie (CHATS)*
12427 * Avoid filename truncation for various CA files.
12429 *Ben Laurie (CHATS)*
12431 * Use sizeof in preference to magic numbers.
12433 *Ben Laurie (CHATS)*
12435 * Avoid filename truncation in cert requests.
12437 *Ben Laurie (CHATS)*
12439 * Add assertions to check for (supposedly impossible) buffer
12442 *Ben Laurie (CHATS)*
12444 * Don't cache truncated DNS entries in the local cache (this could
12445 potentially lead to a spoofing attack).
12447 *Ben Laurie (CHATS)*
12449 * Fix various buffers to be large enough for hex/decimal
12450 representations in a platform independent manner.
12452 *Ben Laurie (CHATS)*
12454 * Add CRYPTO_realloc_clean() to avoid information leakage when
12455 resizing buffers containing secrets, and use where appropriate.
12457 *Ben Laurie (CHATS)*
12459 * Add BIO_indent() to avoid much slightly worrying code to do
12462 *Ben Laurie (CHATS)*
12464 * Convert sprintf()/BIO_puts() to BIO_printf().
12466 *Ben Laurie (CHATS)*
12468 * buffer_gets() could terminate with the buffer only half
12471 *Ben Laurie (CHATS)*
12473 * Add assertions to prevent user-supplied crypto functions from
12474 overflowing internal buffers by having large block sizes, etc.
12476 *Ben Laurie (CHATS)*
12478 * New OPENSSL_assert() macro (similar to assert(), but enabled
12481 *Ben Laurie (CHATS)*
12483 * Eliminate unused copy of key in RC4.
12485 *Ben Laurie (CHATS)*
12487 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12489 *Ben Laurie (CHATS)*
12491 * Fix off-by-one error in EGD path.
12493 *Ben Laurie (CHATS)*
12495 * If RANDFILE path is too long, ignore instead of truncating.
12497 *Ben Laurie (CHATS)*
12499 * Eliminate unused and incorrectly sized X.509 structure
12502 *Ben Laurie (CHATS)*
12504 * Eliminate unused and dangerous function knumber().
12506 *Ben Laurie (CHATS)*
12508 * Eliminate unused and dangerous structure, KSSL_ERR.
12510 *Ben Laurie (CHATS)*
12512 * Protect against overlong session ID context length in an encoded
12513 session object. Since these are local, this does not appear to be
12516 *Ben Laurie (CHATS)*
12518 * Change from security patch (see 0.9.6e below) that did not affect
12519 the 0.9.6 release series:
12521 Remote buffer overflow in SSL3 protocol - an attacker could
12522 supply an oversized master key in Kerberos-enabled versions.
12525 *Ben Laurie (CHATS)*
12527 * Change the SSL kerb5 codes to match RFC 2712.
12531 * Make -nameopt work fully for req and add -reqopt switch.
12533 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12535 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12537 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12539 * Make sure tests can be performed even if the corresponding algorithms
12540 have been removed entirely. This was also the last step to make
12541 OpenSSL compilable with DJGPP under all reasonable conditions.
12543 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12545 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12546 to allow version independent disabling of normally unselected ciphers,
12547 which may be activated as a side-effect of selecting a single cipher.
12549 (E.g., cipher list string "RSA" enables ciphersuites that are left
12550 out of "ALL" because they do not provide symmetric encryption.
12551 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12553 *Lutz Jaenicke, Bodo Moeller*
12555 * Add appropriate support for separate platform-dependent build
12556 directories. The recommended way to make a platform-dependent
12557 build directory is the following (tested on Linux), maybe with
12560 # Place yourself outside of the OpenSSL source tree. In
12561 # this example, the environment variable OPENSSL_SOURCE
12562 # is assumed to contain the absolute OpenSSL source directory.
12563 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12564 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12565 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12566 mkdir -p `dirname $F`
12567 ln -s $OPENSSL_SOURCE/$F $F
12570 To be absolutely sure not to disturb the source tree, a "make clean"
12571 is a good thing. If it isn't successful, don't worry about it,
12572 it probably means the source directory is very clean.
12576 * Make sure any ENGINE control commands make local copies of string
12577 pointers passed to them whenever necessary. Otherwise it is possible
12578 the caller may have overwritten (or deallocated) the original string
12579 data when a later ENGINE operation tries to use the stored values.
12581 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12583 * Improve diagnostics in file reading and command-line digests.
12585 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12587 * Add AES modes CFB and OFB to the object database. Correct an
12588 error in AES-CFB decryption.
12592 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12593 allows existing EVP_CIPHER_CTX structures to be reused after
12594 calling `EVP_*Final()`. This behaviour is used by encryption
12595 BIOs and some applications. This has the side effect that
12596 applications must explicitly clean up cipher contexts with
12597 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12601 * Check the values of dna and dnb in bn_mul_recursive before calling
12602 bn_mul_comba (a non zero value means the a or b arrays do not contain
12603 n2 elements) and fallback to bn_mul_normal if either is not zero.
12607 * Fix escaping of non-ASCII characters when using the -subj option
12608 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12612 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12613 form for "surname", serialNumber has no short form.
12614 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12615 therefore remove "mail" short name for "internet 7".
12616 The OID for unique identifiers in X509 certificates is
12617 x500UniqueIdentifier, not uniqueIdentifier.
12618 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12622 * Add an "init" command to the ENGINE config module and auto initialize
12623 ENGINEs. Without any "init" command the ENGINE will be initialized
12624 after all ctrl commands have been executed on it. If init=1 the
12625 ENGINE is initialized at that point (ctrls before that point are run
12626 on the uninitialized ENGINE and after on the initialized one). If
12627 init=0 then the ENGINE will not be initialized at all.
12631 * Fix the 'app_verify_callback' interface so that the user-defined
12632 argument is actually passed to the callback: In the
12633 SSL_CTX_set_cert_verify_callback() prototype, the callback
12634 declaration has been changed from
12637 int (*cb)(X509_STORE_CTX *,void *);
12638 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12639 i=s->ctx->app_verify_callback(&ctx)
12640 has been changed into
12641 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12643 To update applications using SSL_CTX_set_cert_verify_callback(),
12644 a dummy argument can be added to their callback functions.
12646 *D. K. Smetters <smetters@parc.xerox.com>*
12648 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12650 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12652 * Add and OPENSSL_LOAD_CONF define which will cause
12653 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12654 This allows older applications to transparently support certain
12655 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12656 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12657 load the config file and OPENSSL_add_all_algorithms_conf() which will
12658 always load it have also been added.
12662 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12663 Adjust NIDs and EVP layer.
12665 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12667 * Config modules support in openssl utility.
12669 Most commands now load modules from the config file,
12670 though in a few (such as version) this isn't done
12671 because it couldn't be used for anything.
12673 In the case of ca and req the config file used is
12674 the same as the utility itself: that is the -config
12675 command line option can be used to specify an
12680 * Move default behaviour from OPENSSL_config(). If appname is NULL
12681 use "openssl_conf" if filename is NULL use default openssl config file.
12685 * Add an argument to OPENSSL_config() to allow the use of an alternative
12686 config section name. Add a new flag to tolerate a missing config file
12687 and move code to CONF_modules_load_file().
12691 * Support for crypto accelerator cards from Accelerated Encryption
12692 Processing, www.aep.ie. (Use engine 'aep')
12693 The support was copied from 0.9.6c [engine] and adapted/corrected
12694 to work with the new engine framework.
12696 *AEP Inc. and Richard Levitte*
12698 * Support for SureWare crypto accelerator cards from Baltimore
12699 Technologies. (Use engine 'sureware')
12700 The support was copied from 0.9.6c [engine] and adapted
12701 to work with the new engine framework.
12705 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12706 make the newer ENGINE framework commands for the CHIL engine work.
12708 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12710 * Make it possible to produce shared libraries on ReliantUNIX.
12712 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12714 * Add the configuration target debug-linux-ppro.
12715 Make 'openssl rsa' use the general key loading routines
12716 implemented in `apps.c`, and make those routines able to
12717 handle the key format FORMAT_NETSCAPE and the variant
12720 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12722 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12724 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12726 * Add -keyform to rsautl, and document -engine.
12728 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12730 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12731 BIO_R_NO_SUCH_FILE error code rather than the generic
12732 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12736 * Add new functions
12737 ERR_peek_last_error
12738 ERR_peek_last_error_line
12739 ERR_peek_last_error_line_data.
12740 These are similar to
12742 ERR_peek_error_line
12743 ERR_peek_error_line_data,
12744 but report on the latest error recorded rather than the first one
12745 still in the error queue.
12747 *Ben Laurie, Bodo Moeller*
12749 * default_algorithms option in ENGINE config module. This allows things
12751 default_algorithms = ALL
12752 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12756 * Preliminary ENGINE config module.
12760 * New experimental application configuration code.
12764 * Change the AES code to follow the same name structure as all other
12765 symmetric ciphers, and behave the same way. Move everything to
12766 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12768 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12770 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12772 *Ben Laurie and Theo de Raadt*
12774 * Add option to output public keys in req command.
12776 *Massimiliano Pala madwolf@openca.org*
12778 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12779 (up to about 10% better than before for P-192 and P-224).
12783 * New functions/macros
12785 SSL_CTX_set_msg_callback(ctx, cb)
12786 SSL_CTX_set_msg_callback_arg(ctx, arg)
12787 SSL_set_msg_callback(ssl, cb)
12788 SSL_set_msg_callback_arg(ssl, arg)
12790 to request calling a callback function
12792 void cb(int write_p, int version, int content_type,
12793 const void *buf, size_t len, SSL *ssl, void *arg)
12795 whenever a protocol message has been completely received
12796 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12797 protocol version according to which the SSL library interprets
12798 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12799 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12800 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12801 specification (change_cipher_spec(20), alert(21), handshake(22)).
12802 'buf' and 'len' point to the actual message, 'ssl' to the
12803 SSL object, and 'arg' is the application-defined value set by
12804 SSL[_CTX]_set_msg_callback_arg().
12806 'openssl s_client' and 'openssl s_server' have new '-msg' options
12807 to enable a callback that displays all protocol messages.
12811 * Change the shared library support so shared libraries are built as
12812 soon as the corresponding static library is finished, and thereby get
12813 openssl and the test programs linked against the shared library.
12814 This still only happens when the keyword "shard" has been given to
12815 the configuration scripts.
12817 NOTE: shared library support is still an experimental thing, and
12818 backward binary compatibility is still not guaranteed.
12820 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12822 * Add support for Subject Information Access extension.
12824 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12826 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12827 additional bytes when new memory had to be allocated, not just
12828 when reusing an existing buffer.
12832 * New command line and configuration option 'utf8' for the req command.
12833 This allows field values to be specified as UTF8 strings.
12837 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12838 runs for the former and machine-readable output for the latter.
12842 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12843 of the e-mail address in the DN (i.e., it will go into a certificate
12844 extension only). The new configuration file option 'email_in_dn = no'
12845 has the same effect.
12847 *Massimiliano Pala madwolf@openca.org*
12849 * Change all functions with names starting with `des_` to be starting
12850 with `DES_` instead. Add wrappers that are compatible with libdes,
12851 but are named `_ossl_old_des_*`. Finally, add macros that map the
12852 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12853 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12854 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12857 Since we provide two compatibility mappings, the user needs to
12858 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12859 compatibility is desired. The default (i.e., when that macro
12860 isn't defined) is OpenSSL 0.9.6c compatibility.
12862 There are also macros that enable and disable the support of old
12863 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12864 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12865 are defined, the default will apply: to support the old des routines.
12867 In either case, one must include openssl/des.h to get the correct
12868 definitions. Do not try to just include openssl/des_old.h, that
12871 NOTE: This is a major break of an old API into a new one. Software
12872 authors are encouraged to switch to the `DES_` style functions. Some
12873 time in the future, des_old.h and the libdes compatibility functions
12874 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12875 default), and then completely removed.
12879 * Test for certificates which contain unsupported critical extensions.
12880 If such a certificate is found during a verify operation it is
12881 rejected by default: this behaviour can be overridden by either
12882 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12883 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12884 X509_supported_extension() has also been added which returns 1 if a
12885 particular extension is supported.
12889 * Modify the behaviour of EVP cipher functions in similar way to digests
12890 to retain compatibility with existing code.
12894 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12895 compatibility with existing code. In particular the 'ctx' parameter does
12896 not have to be to be initialized before the call to EVP_DigestInit() and
12897 it is tidied up after a call to EVP_DigestFinal(). New function
12898 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12899 EVP_MD_CTX_copy() changed to not require the destination to be
12900 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12901 requires the destination to be valid.
12903 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12904 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12908 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12909 so that complete 'Handshake' protocol structures are kept in memory
12910 instead of overwriting 'msg_type' and 'length' with 'body' data.
12914 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12916 *Massimo Santin via Richard Levitte*
12918 * Major restructuring to the underlying ENGINE code. This includes
12919 reduction of linker bloat, separation of pure "ENGINE" manipulation
12920 (initialisation, etc) from functionality dealing with implementations
12921 of specific crypto interfaces. This change also introduces integrated
12922 support for symmetric ciphers and digest implementations - so ENGINEs
12923 can now accelerate these by providing EVP_CIPHER and EVP_MD
12924 implementations of their own. This is detailed in
12925 [crypto/engine/README.md](crypto/engine/README.md)
12926 as it couldn't be adequately described here. However, there are a few
12927 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12928 were changed in the original introduction of ENGINE code have now
12929 reverted back - the hooking from this code to ENGINE is now a good
12930 deal more passive and at run-time, operations deal directly with
12931 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12932 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12933 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12934 they were not being used by the framework as there is no concept of a
12935 BIGNUM_METHOD and they could not be generalised to the new
12936 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12937 ENGINE_cpy() has been removed as it cannot be consistently defined in
12942 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12946 * Change mkdef.pl to sort symbols that get the same entry number,
12947 and make sure the automatically generated functions `ERR_load_*`
12948 become part of libeay.num as well.
12952 * New function SSL_renegotiate_pending(). This returns true once
12953 renegotiation has been requested (either SSL_renegotiate() call
12954 or HelloRequest/ClientHello received from the peer) and becomes
12955 false once a handshake has been completed.
12956 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12957 sends a HelloRequest, but does not ensure that a handshake takes
12958 place. SSL_renegotiate_pending() is useful for checking if the
12959 client has followed the request.)
12963 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12964 By default, clients may request session resumption even during
12965 renegotiation (if session ID contexts permit); with this option,
12966 session resumption is possible only in the first handshake.
12968 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12969 more bits available for options that should not be part of
12970 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12974 * Add some demos for certificate and certificate request creation.
12978 * Make maximum certificate chain size accepted from the peer application
12979 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12980 "Douglas E. Engert" <deengert@anl.gov>.
12984 * Add support for shared libraries for Unixware-7
12985 (Boyd Lynn Gerber <gerberb@zenez.com>).
12989 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12990 be done prior to destruction. Use this to unload error strings from
12991 ENGINEs that load their own error strings. NB: This adds two new API
12992 functions to "get" and "set" this destroy handler in an ENGINE.
12996 * Alter all existing ENGINE implementations (except "openssl" and
12997 "openbsd") to dynamically instantiate their own error strings. This
12998 makes them more flexible to be built both as statically-linked ENGINEs
12999 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13000 Also, add stub code to each that makes building them as self-contained
13001 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13005 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13006 implementations into applications that are completely implemented in
13007 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13008 commands that can be used to configure what shared-library to load and
13009 to control aspects of the way it is handled. Also, made an update to
13010 the [README-Engine.md](README-Engine.md) file
13011 that brings its information up-to-date and
13012 provides some information and instructions on the "dynamic" ENGINE
13013 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13017 * Make it possible to unload ranges of ERR strings with a new
13018 "ERR_unload_strings" function.
13022 * Add a copy() function to EVP_MD.
13026 * Make EVP_MD routines take a context pointer instead of just the
13027 md_data void pointer.
13031 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13032 that the digest can only process a single chunk of data
13033 (typically because it is provided by a piece of
13034 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13035 is only going to provide a single chunk of data, and hence the
13036 framework needn't accumulate the data for oneshot drivers.
13040 * As with "ERR", make it possible to replace the underlying "ex_data"
13041 functions. This change also alters the storage and management of global
13042 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13043 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13044 index counters. The API functions that use this state have been changed
13045 to take a "class_index" rather than pointers to the class's local STACK
13046 and counter, and there is now an API function to dynamically create new
13047 classes. This centralisation allows us to (a) plug a lot of the
13048 thread-safety problems that existed, and (b) makes it possible to clean
13049 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13050 such data would previously have always leaked in application code and
13051 workarounds were in place to make the memory debugging turn a blind eye
13052 to it. Application code that doesn't use this new function will still
13053 leak as before, but their memory debugging output will announce it now
13054 rather than letting it slide.
13056 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13057 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13058 has a return value to indicate success or failure.
13062 * Make it possible to replace the underlying "ERR" functions such that the
13063 global state (2 LHASH tables and 2 locks) is only used by the "default"
13064 implementation. This change also adds two functions to "get" and "set"
13065 the implementation prior to it being automatically set the first time
13066 any other ERR function takes place. Ie. an application can call "get",
13067 pass the return value to a module it has just loaded, and that module
13068 can call its own "set" function using that value. This means the
13069 module's "ERR" operations will use (and modify) the error state in the
13070 application and not in its own statically linked copy of OpenSSL code.
13074 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13075 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13076 the operation, and provides a more encapsulated way for external code
13077 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13078 to use these functions rather than manually incrementing the counts.
13080 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13084 * Add EVP test program.
13088 * Add symmetric cipher support to ENGINE. Expect the API to change!
13092 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13093 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13094 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13095 These allow a CRL to be built without having to access X509_CRL fields
13096 directly. Modify 'ca' application to use new functions.
13100 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13101 bug workarounds. Rollback attack detection is a security feature.
13102 The problem will only arise on OpenSSL servers when TLSv1 is not
13103 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13104 Software authors not wanting to support TLSv1 will have special reasons
13105 for their choice and can explicitly enable this option.
13107 *Bodo Moeller, Lutz Jaenicke*
13109 * Rationalise EVP so it can be extended: don't include a union of
13110 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13111 (similar to those existing for EVP_CIPHER_CTX).
13116 EVP_MD_CTX_init(&md); /* new function call */
13117 EVP_DigestInit(&md, EVP_sha1());
13118 EVP_DigestUpdate(&md, in, len);
13119 EVP_DigestFinal(&md, out, NULL);
13120 EVP_MD_CTX_cleanup(&md); /* new function call */
13124 * Make DES key schedule conform to the usual scheme, as well as
13125 correcting its structure. This means that calls to DES functions
13126 now have to pass a pointer to a des_key_schedule instead of a
13127 plain des_key_schedule (which was actually always a pointer
13130 des_key_schedule ks;
13132 des_set_key_checked(..., &ks);
13133 des_ncbc_encrypt(..., &ks, ...);
13135 (Note that a later change renames 'des_...' into 'DES_...'.)
13139 * Initial reduction of linker bloat: the use of some functions, such as
13140 PEM causes large amounts of unused functions to be linked in due to
13141 poor organisation. For example pem_all.c contains every PEM function
13142 which has a knock on effect of linking in large amounts of (unused)
13143 ASN1 code. Grouping together similar functions and splitting unrelated
13144 functions prevents this.
13148 * Cleanup of EVP macros.
13152 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13153 correct `_ecb suffix`.
13157 * Add initial OCSP responder support to ocsp application. The
13158 revocation information is handled using the text based index
13159 use by the ca application. The responder can either handle
13160 requests generated internally, supplied in files (for example
13161 via a CGI script) or using an internal minimal server.
13165 * Add configuration choices to get zlib compression for TLS.
13169 * Changes to Kerberos SSL for RFC 2712 compliance:
13170 1. Implemented real KerberosWrapper, instead of just using
13171 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13172 2. Implemented optional authenticator field of KerberosWrapper.
13174 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13175 and authenticator structs; see crypto/krb5/.
13177 Generalized Kerberos calls to support multiple Kerberos libraries.
13178 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13179 via Richard Levitte*
13181 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13182 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13183 values for each of the key sizes rather than having just
13184 parameters (and 'speed' generating keys each time).
13188 * Speed up EVP routines.
13191 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13192 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13193 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13194 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13196 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13197 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13198 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13201 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13203 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13207 * Added the OS2-EMX target.
13209 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13211 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13212 New functions to support `NCONF` routines in extension code.
13213 New function `CONF_set_nconf()`
13214 to allow functions which take an `NCONF` to also handle the old `LHASH`
13215 structure: this means that the old `CONF` compatible routines can be
13216 retained (in particular w.rt. extensions) without having to duplicate the
13217 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13221 * Enhance the general user interface with mechanisms for inner control
13222 and with possibilities to have yes/no kind of prompts.
13226 * Change all calls to low-level digest routines in the library and
13227 applications to use EVP. Add missing calls to HMAC_cleanup() and
13228 don't assume HMAC_CTX can be copied using memcpy().
13230 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13232 * Add the possibility to control engines through control names but with
13233 arbitrary arguments instead of just a string.
13234 Change the key loaders to take a UI_METHOD instead of a callback
13235 function pointer. NOTE: this breaks binary compatibility with earlier
13236 versions of OpenSSL [engine].
13237 Adapt the nCipher code for these new conditions and add a card insertion
13242 * Enhance the general user interface with mechanisms to better support
13243 dialog box interfaces, application-defined prompts, the possibility
13244 to use defaults (for example default passwords from somewhere else)
13245 and interrupts/cancellations.
13249 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13250 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13254 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13255 tidy up some unnecessarily weird code in 'sk_new()').
13257 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13259 * Change the key loading routines for ENGINEs to use the same kind
13260 callback (pem_password_cb) as all other routines that need this
13265 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13266 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13267 than this minimum value is recommended.
13271 * New random seeder for OpenVMS, using the system process statistics
13272 that are easily reachable.
13276 * Windows apparently can't transparently handle global
13277 variables defined in DLLs. Initialisations such as:
13279 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13281 won't compile. This is used by the any applications that need to
13282 declare their own ASN1 modules. This was fixed by adding the option
13283 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13284 needed for static libraries under Win32.
13288 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13289 setting of purpose and trust fields. New X509_STORE trust and
13290 purpose functions and tidy up setting in other SSL functions.
13294 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13295 structure. These are inherited by X509_STORE_CTX when it is
13296 initialised. This allows various defaults to be set in the
13297 X509_STORE structure (such as flags for CRL checking and custom
13298 purpose or trust settings) for functions which only use X509_STORE_CTX
13299 internally such as S/MIME.
13301 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13302 trust settings if they are not set in X509_STORE. This allows X509_STORE
13303 purposes and trust (in S/MIME for example) to override any set by default.
13305 Add command line options for CRL checking to smime, s_client and s_server
13310 * Initial CRL based revocation checking. If the CRL checking flag(s)
13311 are set then the CRL is looked up in the X509_STORE structure and
13312 its validity and signature checked, then if the certificate is found
13313 in the CRL the verify fails with a revoked error.
13315 Various new CRL related callbacks added to X509_STORE_CTX structure.
13317 Command line options added to 'verify' application to support this.
13319 This needs some additional work, such as being able to handle multiple
13320 CRLs with different times, extension based lookup (rather than just
13321 by subject name) and ultimately more complete V2 CRL extension
13326 * Add a general user interface API (crypto/ui/). This is designed
13327 to replace things like des_read_password and friends (backward
13328 compatibility functions using this new API are provided).
13329 The purpose is to remove prompting functions from the DES code
13330 section as well as provide for prompting through dialog boxes in
13331 a window system and the like.
13335 * Add "ex_data" support to ENGINE so implementations can add state at a
13336 per-structure level rather than having to store it globally.
13340 * Make it possible for ENGINE structures to be copied when retrieved by
13341 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13342 This causes the "original" ENGINE structure to act like a template,
13343 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13344 operational state can be localised to each ENGINE structure, despite the
13345 fact they all share the same "methods". New ENGINE structures returned in
13346 this case have no functional references and the return value is the single
13347 structural reference. This matches the single structural reference returned
13348 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13353 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13354 needs to match any other type at all we need to manually clear the
13359 * Changes to the "openssl engine" utility to include;
13360 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13361 about an ENGINE's available control commands.
13362 - executing control commands from command line arguments using the
13363 '-pre' and '-post' switches. '-post' is only used if '-t' is
13364 specified and the ENGINE is successfully initialised. The syntax for
13365 the individual commands are colon-separated, for example;
13366 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13370 * New dynamic control command support for ENGINEs. ENGINEs can now
13371 declare their own commands (numbers), names (strings), descriptions,
13372 and input types for run-time discovery by calling applications. A
13373 subset of these commands are implicitly classed as "executable"
13374 depending on their input type, and only these can be invoked through
13375 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13376 can be based on user input, config files, etc). The distinction is
13377 that "executable" commands cannot return anything other than a boolean
13378 result and can only support numeric or string input, whereas some
13379 discoverable commands may only be for direct use through
13380 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13381 pointers, or other custom uses. The "executable" commands are to
13382 support parameterisations of ENGINE behaviour that can be
13383 unambiguously defined by ENGINEs and used consistently across any
13384 OpenSSL-based application. Commands have been added to all the
13385 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13386 control over shared-library paths without source code alterations.
13390 * Changed all ENGINE implementations to dynamically allocate their
13391 ENGINEs rather than declaring them statically. Apart from this being
13392 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13393 this also allows the implementations to compile without using the
13394 internal engine_int.h header.
13398 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13399 'const' value. Any code that should be able to modify a RAND_METHOD
13400 should already have non-const pointers to it (ie. they should only
13401 modify their own ones).
13405 * Made a variety of little tweaks to the ENGINE code.
13406 - "atalla" and "ubsec" string definitions were moved from header files
13407 to C code. "nuron" string definitions were placed in variables
13408 rather than hard-coded - allowing parameterisation of these values
13409 later on via ctrl() commands.
13410 - Removed unused "#if 0"'d code.
13411 - Fixed engine list iteration code so it uses ENGINE_free() to release
13412 structural references.
13413 - Constified the RAND_METHOD element of ENGINE structures.
13414 - Constified various get/set functions as appropriate and added
13415 missing functions (including a catch-all ENGINE_cpy that duplicates
13416 all ENGINE values onto a new ENGINE except reference counts/state).
13417 - Removed NULL parameter checks in get/set functions. Setting a method
13418 or function to NULL is a way of cancelling out a previously set
13419 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13420 and doesn't justify the extra error symbols and code.
13421 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13422 flags from engine_int.h to engine.h.
13423 - Changed prototypes for ENGINE handler functions (init(), finish(),
13424 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13428 * Implement binary inversion algorithm for BN_mod_inverse in addition
13429 to the algorithm using long division. The binary algorithm can be
13430 used only if the modulus is odd. On 32-bit systems, it is faster
13431 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13432 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13433 up to 450 bits. In 64-bit environments, the binary algorithm
13434 appears to be advantageous for much longer moduli; here we use it
13435 for moduli up to 2048 bits.
13439 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13440 could not support the combine flag in choice fields.
13444 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13445 extensions from a certificate request to the certificate.
13449 * Allow multiple 'certopt' and 'nameopt' options to be separated
13450 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13451 file: this allows the display of the certificate about to be
13452 signed to be customised, to allow certain fields to be included
13453 or excluded and extension details. The old system didn't display
13454 multicharacter strings properly, omitted fields not in the policy
13455 and couldn't display additional details such as extensions.
13459 * Function EC_POINTs_mul for multiple scalar multiplication
13460 of an arbitrary number of elliptic curve points
13461 \sum scalars[i]*points[i],
13462 optionally including the generator defined for the EC_GROUP:
13463 scalar*generator + \sum scalars[i]*points[i].
13465 EC_POINT_mul is a simple wrapper function for the typical case
13466 that the point list has just one item (besides the optional
13471 * First EC_METHODs for curves over GF(p):
13473 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13474 operations and provides various method functions that can also
13475 operate with faster implementations of modular arithmetic.
13477 EC_GFp_mont_method() reuses most functions that are part of
13478 EC_GFp_simple_method, but uses Montgomery arithmetic.
13480 *Bodo Moeller; point addition and point doubling
13481 implementation directly derived from source code provided by
13482 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13484 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13485 crypto/ec/ec_lib.c):
13487 Curves are EC_GROUP objects (with an optional group generator)
13488 based on EC_METHODs that are built into the library.
13490 Points are EC_POINT objects based on EC_GROUP objects.
13492 Most of the framework would be able to handle curves over arbitrary
13493 finite fields, but as there are no obvious types for fields other
13494 than GF(p), some functions are limited to that for now.
13498 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13499 that the file contains a complete HTTP response.
13503 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13504 change the def and num file printf format specifier from "%-40sXXX"
13505 to "%-39s XXX". The latter will always guarantee a space after the
13506 field while the former will cause them to run together if the field
13507 is 40 of more characters long.
13511 * Constify the cipher and digest 'method' functions and structures
13512 and modify related functions to take constant EVP_MD and EVP_CIPHER
13517 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13518 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13522 * Modify `EVP_Digest*()` routines so they now return values. Although the
13523 internal software routines can never fail additional hardware versions
13528 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13530 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13531 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13534 ERR_R_NESTED_ASN1_ERROR
13536 ERR_R_MISSING_ASN1_EOS
13537 were 4 .. 9, conflicting with
13538 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13540 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13541 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13543 Add new error code 'ERR_R_INTERNAL_ERROR'.
13547 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13552 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13553 sets the subject name for a new request or supersedes the
13554 subject name in a given request. Formats that can be parsed are
13555 'CN=Some Name, OU=myOU, C=IT'
13557 'CN=Some Name/OU=myOU/C=IT'.
13559 Add options '-batch' and '-verbose' to 'openssl req'.
13561 *Massimiliano Pala <madwolf@hackmasters.net>*
13563 * Introduce the possibility to access global variables through
13564 functions on platform were that's the best way to handle exporting
13565 global variables in shared libraries. To enable this functionality,
13566 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13567 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13568 is normally done by Configure or something similar).
13570 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13571 in the source file (foo.c) like this:
13573 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13574 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13576 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13577 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13579 OPENSSL_DECLARE_GLOBAL(int,foo);
13580 #define foo OPENSSL_GLOBAL_REF(foo)
13581 OPENSSL_DECLARE_GLOBAL(double,bar);
13582 #define bar OPENSSL_GLOBAL_REF(bar)
13584 The #defines are very important, and therefore so is including the
13585 header file everywhere where the defined globals are used.
13587 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13588 of ASN.1 items, but that structure is a bit different.
13590 The largest change is in util/mkdef.pl which has been enhanced with
13591 better and easier to understand logic to choose which symbols should
13592 go into the Windows .def files as well as a number of fixes and code
13593 cleanup (among others, algorithm keywords are now sorted
13594 lexicographically to avoid constant rewrites).
13598 * In BN_div() keep a copy of the sign of 'num' before writing the
13599 result to 'rm' because if rm==num the value will be overwritten
13600 and produce the wrong result if 'num' is negative: this caused
13601 problems with BN_mod() and BN_nnmod().
13605 * Function OCSP_request_verify(). This checks the signature on an
13606 OCSP request and verifies the signer certificate. The signer
13607 certificate is just checked for a generic purpose and OCSP request
13612 * Add OCSP_check_validity() function to check the validity of OCSP
13613 responses. OCSP responses are prepared in real time and may only
13614 be a few seconds old. Simply checking that the current time lies
13615 between thisUpdate and nextUpdate max reject otherwise valid responses
13616 caused by either OCSP responder or client clock inaccuracy. Instead
13617 we allow thisUpdate and nextUpdate to fall within a certain period of
13618 the current time. The age of the response can also optionally be
13619 checked. Two new options -validity_period and -status_age added to
13624 * If signature or public key algorithm is unrecognized print out its
13625 OID rather that just UNKNOWN.
13629 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13630 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13631 ID to be generated from the issuer certificate alone which can then be
13632 passed to OCSP_id_issuer_cmp().
13636 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13637 ASN1 modules to export functions returning ASN1_ITEM pointers
13638 instead of the ASN1_ITEM structures themselves. This adds several
13639 new macros which allow the underlying ASN1 function/structure to
13640 be accessed transparently. As a result code should not use ASN1_ITEM
13641 references directly (such as &X509_it) but instead use the relevant
13642 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13643 use of the new ASN1 code on platforms where exporting structures
13644 is problematical (for example in shared libraries) but exporting
13645 functions returning pointers to structures is not.
13649 * Add support for overriding the generation of SSL/TLS session IDs.
13650 These callbacks can be registered either in an SSL_CTX or per SSL.
13651 The purpose of this is to allow applications to control, if they wish,
13652 the arbitrary values chosen for use as session IDs, particularly as it
13653 can be useful for session caching in multiple-server environments. A
13654 command-line switch for testing this (and any client code that wishes
13655 to use such a feature) has been added to "s_server".
13657 *Geoff Thorpe, Lutz Jaenicke*
13659 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13660 of the form `#if defined(...) || defined(...) || ...` and
13661 `#if !defined(...) && !defined(...) && ...`. This also avoids
13662 the growing number of special cases it was previously handling.
13666 * Make all configuration macros available for application by making
13667 sure they are available in opensslconf.h, by giving them names starting
13668 with `OPENSSL_` to avoid conflicts with other packages and by making
13669 sure e_os2.h will cover all platform-specific cases together with
13671 Additionally, it is now possible to define configuration/platform-
13672 specific names (called "system identities"). In the C code, these
13673 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13674 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13675 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13680 * New option -set_serial to 'req' and 'x509' this allows the serial
13681 number to use to be specified on the command line. Previously self
13682 signed certificates were hard coded with serial number 0 and the
13683 CA options of 'x509' had to use a serial number in a file which was
13688 * New options to 'ca' utility to support V2 CRL entry extensions.
13689 Currently CRL reason, invalidity date and hold instruction are
13690 supported. Add new CRL extensions to V3 code and some new objects.
13694 * New function EVP_CIPHER_CTX_set_padding() this is used to
13695 disable standard block padding (aka PKCS#5 padding) in the EVP
13696 API, which was previously mandatory. This means that the data is
13697 not padded in any way and so the total length much be a multiple
13698 of the block size, otherwise an error occurs.
13702 * Initial (incomplete) OCSP SSL support.
13706 * New function OCSP_parse_url(). This splits up a URL into its host,
13707 port and path components: primarily to parse OCSP URLs. New -url
13708 option to ocsp utility.
13712 * New nonce behavior. The return value of OCSP_check_nonce() now
13713 reflects the various checks performed. Applications can decide
13714 whether to tolerate certain situations such as an absent nonce
13715 in a response when one was present in a request: the ocsp application
13716 just prints out a warning. New function OCSP_add1_basic_nonce()
13717 this is to allow responders to include a nonce in a response even if
13718 the request is nonce-less.
13722 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13723 skipped when using openssl x509 multiple times on a single input file,
13724 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13728 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13729 set string type: to handle setting ASN1_TIME structures. Fix ca
13730 utility to correctly initialize revocation date of CRLs.
13734 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13735 the clients preferred ciphersuites and rather use its own preferences.
13736 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13737 Internet Explorer by ensuring unchanged hash method during stepup.
13738 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13742 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13743 to aes and add a new 'exist' option to print out symbols that don't
13748 * Additional options to ocsp utility to allow flags to be set and
13749 additional certificates supplied.
13753 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13754 OCSP client a number of certificate to only verify the response
13759 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13760 handle the new API. Currently only ECB, CBC modes supported. Add new
13763 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13764 Encryption Standard (AES) Ciphersuites for Transport Layer
13765 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13766 not enabled by default and were not part of the "ALL" ciphersuite
13767 alias because they were not yet official; they could be
13768 explicitly requested by specifying the "AESdraft" ciphersuite
13769 group alias. In the final release of OpenSSL 0.9.7, the group
13770 alias is called "AES" and is part of "ALL".)
13772 *Ben Laurie, Steve Henson, Bodo Moeller*
13774 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13775 request to response.
13779 * Functions for OCSP responders. OCSP_request_onereq_count(),
13780 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13781 extract information from a certificate request. OCSP_response_create()
13782 creates a response and optionally adds a basic response structure.
13783 OCSP_basic_add1_status() adds a complete single response to a basic
13784 response and returns the OCSP_SINGLERESP structure just added (to allow
13785 extensions to be included for example). OCSP_basic_add1_cert() adds a
13786 certificate to a basic response and OCSP_basic_sign() signs a basic
13787 response with various flags. New helper functions ASN1_TIME_check()
13788 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13789 (converts ASN1_TIME to GeneralizedTime).
13793 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13794 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13795 structure from a certificate. X509_pubkey_digest() digests the public_key
13796 contents: this is used in various key identifiers.
13800 * Make sk_sort() tolerate a NULL argument.
13802 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13804 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13805 passed by the function are trusted implicitly. If any of them signed the
13806 response then it is assumed to be valid and is not verified.
13810 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13811 to data. This was previously part of the PKCS7 ASN1 code. This
13812 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13813 *Steve Henson, reported by Kenneth R. Robinette
13814 <support@securenetterm.com>*
13816 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13817 routines: without these tracing memory leaks is very painful.
13818 Fix leaks in PKCS12 and PKCS7 routines.
13822 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13823 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13824 effectively meant GeneralizedTime would never be used. Now it
13825 is initialised to -1 but X509_time_adj() now has to check the value
13826 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13827 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13828 *Steve Henson, reported by Kenneth R. Robinette
13829 <support@securenetterm.com>*
13831 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13832 result in a zero length in the ASN1_INTEGER structure which was
13833 not consistent with the structure when d2i_ASN1_INTEGER() was used
13834 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13835 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13836 where it did not print out a minus for negative ASN1_INTEGER.
13840 * Add summary printout to ocsp utility. The various functions which
13841 convert status values to strings have been renamed to:
13842 OCSP_response_status_str(), OCSP_cert_status_str() and
13843 OCSP_crl_reason_str() and are no longer static. New options
13844 to verify nonce values and to disable verification. OCSP response
13845 printout format cleaned up.
13849 * Add additional OCSP certificate checks. These are those specified
13850 in RFC2560. This consists of two separate checks: the CA of the
13851 certificate being checked must either be the OCSP signer certificate
13852 or the issuer of the OCSP signer certificate. In the latter case the
13853 OCSP signer certificate must contain the OCSP signing extended key
13854 usage. This check is performed by attempting to match the OCSP
13855 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13856 in the OCSP_CERTID structures of the response.
13860 * Initial OCSP certificate verification added to OCSP_basic_verify()
13861 and related routines. This uses the standard OpenSSL certificate
13862 verify routines to perform initial checks (just CA validity) and
13863 to obtain the certificate chain. Then additional checks will be
13864 performed on the chain. Currently the root CA is checked to see
13865 if it is explicitly trusted for OCSP signing. This is used to set
13866 a root CA as a global signing root: that is any certificate that
13867 chains to that CA is an acceptable OCSP signing certificate.
13871 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13872 extensions from a separate configuration file.
13873 As when reading extensions from the main configuration file,
13874 the '-extensions ...' option may be used for specifying the
13877 *Massimiliano Pala <madwolf@comune.modena.it>*
13879 * New OCSP utility. Allows OCSP requests to be generated or
13880 read. The request can be sent to a responder and the output
13881 parsed, outputted or printed in text form. Not complete yet:
13882 still needs to check the OCSP response validity.
13886 * New subcommands for 'openssl ca':
13887 `openssl ca -status <serial>` prints the status of the cert with
13888 the given serial number (according to the index file).
13889 `openssl ca -updatedb` updates the expiry status of certificates
13892 *Massimiliano Pala <madwolf@comune.modena.it>*
13894 * New '-newreq-nodes' command option to CA.pl. This is like
13895 '-newreq', but calls 'openssl req' with the '-nodes' option
13896 so that the resulting key is not encrypted.
13898 *Damien Miller <djm@mindrot.org>*
13900 * New configuration for the GNU Hurd.
13902 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13904 * Initial code to implement OCSP basic response verify. This
13905 is currently incomplete. Currently just finds the signer's
13906 certificate and verifies the signature on the response.
13910 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13911 value of OPENSSLDIR. This is available via the new '-d' option
13912 to 'openssl version', and is also included in 'openssl version -a'.
13916 * Allowing defining memory allocation callbacks that will be given
13917 file name and line number information in additional arguments
13918 (a `const char*` and an int). The basic functionality remains, as
13919 well as the original possibility to just replace malloc(),
13920 realloc() and free() by functions that do not know about these
13921 additional arguments. To register and find out the current
13922 settings for extended allocation functions, the following
13923 functions are provided:
13925 CRYPTO_set_mem_ex_functions
13926 CRYPTO_set_locked_mem_ex_functions
13927 CRYPTO_get_mem_ex_functions
13928 CRYPTO_get_locked_mem_ex_functions
13930 These work the same way as CRYPTO_set_mem_functions and friends.
13931 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13932 extended allocation function is enabled.
13933 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13934 a conventional allocation function is enabled.
13936 *Richard Levitte, Bodo Moeller*
13938 * Finish off removing the remaining LHASH function pointer casts.
13939 There should no longer be any prototype-casting required when using
13940 the LHASH abstraction, and any casts that remain are "bugs". See
13941 the callback types and macros at the head of lhash.h for details
13942 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13946 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13947 If /dev/[u]random devices are not available or do not return enough
13948 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13950 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13951 /etc/entropy will be queried once each in this sequence, querying stops
13952 when enough entropy was collected without querying more sockets.
13956 * Change the Unix RAND_poll() variant to be able to poll several
13957 random devices, as specified by DEVRANDOM, until a sufficient amount
13958 of data has been collected. We spend at most 10 ms on each file
13959 (select timeout) and read in non-blocking mode. DEVRANDOM now
13960 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13961 (previously it was just the string "/dev/urandom"), so on typical
13962 platforms the 10 ms delay will never occur.
13963 Also separate out the Unix variant to its own file, rand_unix.c.
13964 For VMS, there's a currently-empty rand_vms.c.
13968 * Move OCSP client related routines to ocsp_cl.c. These
13969 provide utility functions which an application needing
13970 to issue a request to an OCSP responder and analyse the
13971 response will typically need: as opposed to those which an
13972 OCSP responder itself would need which will be added later.
13974 OCSP_request_sign() signs an OCSP request with an API similar
13975 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13976 response. OCSP_response_get1_basic() extracts basic response
13977 from response. OCSP_resp_find_status(): finds and extracts status
13978 information from an OCSP_CERTID structure (which will be created
13979 when the request structure is built). These are built from lower
13980 level functions which work on OCSP_SINGLERESP structures but
13981 won't normally be used unless the application wishes to examine
13982 extensions in the OCSP response for example.
13984 Replace nonce routines with a pair of functions.
13985 OCSP_request_add1_nonce() adds a nonce value and optionally
13986 generates a random value. OCSP_check_nonce() checks the
13987 validity of the nonce in an OCSP response.
13991 * Change function OCSP_request_add() to OCSP_request_add0_id().
13992 This doesn't copy the supplied OCSP_CERTID and avoids the
13993 need to free up the newly created id. Change return type
13994 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13995 This can then be used to add extensions to the request.
13996 Deleted OCSP_request_new(), since most of its functionality
13997 is now in OCSP_REQUEST_new() (and the case insensitive name
13998 clash) apart from the ability to set the request name which
13999 will be added elsewhere.
14003 * Update OCSP API. Remove obsolete extensions argument from
14004 various functions. Extensions are now handled using the new
14005 OCSP extension code. New simple OCSP HTTP function which
14006 can be used to send requests and parse the response.
14010 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14011 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14012 uses the special reorder version of SET OF to sort the attributes
14013 and reorder them to match the encoded order. This resolves a long
14014 standing problem: a verify on a PKCS7 structure just after signing
14015 it used to fail because the attribute order did not match the
14016 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14017 it uses the received order. This is necessary to tolerate some broken
14018 software that does not order SET OF. This is handled by encoding
14019 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14020 to produce the required SET OF.
14024 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14025 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14026 files to get correct declarations of the ASN.1 item variables.
14030 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14031 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14032 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14033 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14034 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14035 ASN1_ITEM and no wrapper functions.
14039 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14040 replace the old function pointer based I/O routines. Change most of
14041 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14045 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14046 lines, recognize more "algorithms" that can be deselected, and make
14047 it complain about algorithm deselection that isn't recognised.
14051 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14052 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14053 to use new functions. Add NO_ASN1_OLD which can be set to remove
14054 some old style ASN1 functions: this can be used to determine if old
14055 code will still work when these eventually go away.
14059 * New extension functions for OCSP structures, these follow the
14060 same conventions as certificates and CRLs.
14064 * New function X509V3_add1_i2d(). This automatically encodes and
14065 adds an extension. Its behaviour can be customised with various
14066 flags to append, replace or delete. Various wrappers added for
14067 certificates and CRLs.
14071 * Fix to avoid calling the underlying ASN1 print routine when
14072 an extension cannot be parsed. Correct a typo in the
14073 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14077 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14078 entries for variables.
14082 * Add functionality to `apps/openssl.c` for detecting locking
14083 problems: As the program is single-threaded, all we have
14084 to do is register a locking callback using an array for
14085 storing which locks are currently held by the program.
14089 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14090 SSL_get_ex_data_X509_STORE_idx(), which is used in
14091 ssl_verify_cert_chain() and thus can be called at any time
14092 during TLS/SSL handshakes so that thread-safety is essential.
14093 Unfortunately, the ex_data design is not at all suited
14094 for multi-threaded use, so it probably should be abolished.
14098 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14100 *Broadcom, tweaked and integrated by Geoff Thorpe*
14102 * Move common extension printing code to new function
14103 X509V3_print_extensions(). Reorganise OCSP print routines and
14104 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14108 * New function X509_signature_print() to remove duplication in some
14113 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14114 set (this was treated exactly the same as SET OF previously). This
14115 is used to reorder the STACK representing the structure to match the
14116 encoding. This will be used to get round a problem where a PKCS7
14117 structure which was signed could not be verified because the STACK
14118 order did not reflect the encoded order.
14122 * Reimplement the OCSP ASN1 module using the new code.
14126 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14127 for its ASN1 operations. The old style function pointers still exist
14128 for now but they will eventually go away.
14132 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14133 completely replaces the old ASN1 functionality with a table driven
14134 encoder and decoder which interprets an ASN1_ITEM structure describing
14135 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14136 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14137 has also been converted to the new form.
14141 * Change BN_mod_exp_recp so that negative moduli are tolerated
14142 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14143 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14144 for negative moduli.
14148 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14149 of not touching the result's sign bit.
14153 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14158 * Changed the LHASH code to use prototypes for callbacks, and created
14159 macros to declare and implement thin (optionally static) functions
14160 that provide type-safety and avoid function pointer casting for the
14161 type-specific callbacks.
14165 * Added Kerberos Cipher Suites to be used with TLS, as written in
14167 *Veers Staats <staatsvr@asc.hpc.mil>,
14168 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14170 * Reformat the FAQ so the different questions and answers can be divided
14171 in sections depending on the subject.
14175 * Have the zlib compression code load ZLIB.DLL dynamically under
14180 * New function BN_mod_sqrt for computing square roots modulo a prime
14181 (using the probabilistic Tonelli-Shanks algorithm unless
14182 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14183 be handled deterministically).
14185 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14187 * Make BN_mod_inverse faster by explicitly handling small quotients
14188 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14189 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14193 * New function BN_kronecker.
14197 * Fix BN_gcd so that it works on negative inputs; the result is
14198 positive unless both parameters are zero.
14199 Previously something reasonably close to an infinite loop was
14200 possible because numbers could be growing instead of shrinking
14201 in the implementation of Euclid's algorithm.
14205 * Fix BN_is_word() and BN_is_one() macros to take into account the
14206 sign of the number in question.
14208 Fix BN_is_word(a,w) to work correctly for w == 0.
14210 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14211 because its test if the absolute value of 'a' equals 'w'.
14212 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14213 it exists mostly for use in the implementations of BN_is_zero(),
14214 BN_is_one(), and BN_is_word().
14218 * New function BN_swap.
14222 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14223 the exponentiation functions are more likely to produce reasonable
14224 results on negative inputs.
14228 * Change BN_mod_mul so that the result is always non-negative.
14229 Previously, it could be negative if one of the factors was negative;
14230 I don't think anyone really wanted that behaviour.
14234 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14235 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14236 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14237 and add new functions:
14246 BN_mod_lshift1_quick
14248 BN_mod_lshift_quick
14250 These functions always generate non-negative results.
14252 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14253 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14255 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14256 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14257 be reduced modulo `m`.
14259 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14262 The following entry accidentally appeared in the CHANGES file
14263 distributed with OpenSSL 0.9.7. The modifications described in
14264 it do *not* apply to OpenSSL 0.9.7.
14266 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14267 was actually never needed) and in BN_mul(). The removal in BN_mul()
14268 required a small change in bn_mul_part_recursive() and the addition
14269 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14270 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14271 bn_sub_words() and bn_add_words() except they take arrays with
14277 * In 'openssl passwd', verify passwords read from the terminal
14278 unless the '-salt' option is used (which usually means that
14279 verification would just waste user's time since the resulting
14280 hash is going to be compared with some given password hash)
14281 or the new '-noverify' option is used.
14283 This is an incompatible change, but it does not affect
14284 non-interactive use of 'openssl passwd' (passwords on the command
14285 line, '-stdin' option, '-in ...' option) and thus should not
14286 cause any problems.
14290 * Remove all references to RSAref, since there's no more need for it.
14294 * Make DSO load along a path given through an environment variable
14295 (SHLIB_PATH) with shl_load().
14299 * Constify the ENGINE code as a result of BIGNUM constification.
14300 Also constify the RSA code and most things related to it. In a
14301 few places, most notable in the depth of the ASN.1 code, ugly
14302 casts back to non-const were required (to be solved at a later
14307 * Make it so the openssl application has all engines loaded by default.
14311 * Constify the BIGNUM routines a little more.
14315 * Add the following functions:
14317 ENGINE_load_cswift()
14319 ENGINE_load_atalla()
14320 ENGINE_load_nuron()
14321 ENGINE_load_builtin_engines()
14323 That way, an application can itself choose if external engines that
14324 are built-in in OpenSSL shall ever be used or not. The benefit is
14325 that applications won't have to be linked with libdl or other dso
14326 libraries unless it's really needed.
14328 Changed 'openssl engine' to load all engines on demand.
14329 Changed the engine header files to avoid the duplication of some
14330 declarations (they differed!).
14334 * 'openssl engine' can now list capabilities.
14338 * Better error reporting in 'openssl engine'.
14342 * Never call load_dh_param(NULL) in s_server.
14346 * Add engine application. It can currently list engines by name and
14347 identity, and test if they are actually available.
14351 * Improve RPM specification file by forcing symbolic linking and making
14352 sure the installed documentation is also owned by root.root.
14354 *Damien Miller <djm@mindrot.org>*
14356 * Give the OpenSSL applications more possibilities to make use of
14357 keys (public as well as private) handled by engines.
14361 * Add OCSP code that comes from CertCo.
14365 * Add VMS support for the Rijndael code.
14369 * Added untested support for Nuron crypto accelerator.
14373 * Add support for external cryptographic devices. This code was
14374 previously distributed separately as the "engine" branch.
14376 *Geoff Thorpe, Richard Levitte*
14378 * Rework the filename-translation in the DSO code. It is now possible to
14379 have far greater control over how a "name" is turned into a filename
14380 depending on the operating environment and any oddities about the
14381 different shared library filenames on each system.
14385 * Support threads on FreeBSD-elf in Configure.
14389 * Fix for SHA1 assembly problem with MASM: it produces
14390 warnings about corrupt line number information when assembling
14391 with debugging information. This is caused by the overlapping
14394 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14397 NCONF_get_number() has no error checking at all. As a replacement,
14398 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14399 promoted strongly. The old NCONF_get_number is kept around for
14400 binary backward compatibility.
14401 Make it possible for methods to load from something other than a BIO,
14402 by providing a function pointer that is given a name instead of a BIO.
14403 For example, this could be used to load configuration data from an
14408 * Fix for non blocking accept BIOs. Added new I/O special reason
14409 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14410 with non blocking I/O was not possible because no retry code was
14411 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14416 * Added the beginnings of Rijndael support.
14420 * Fix for bug in DirectoryString mask setting. Add support for
14421 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14422 to allow certificate printing to more controllable, additional
14423 'certopt' option to 'x509' to allow new printing options to be
14428 * Clean old EAY MD5 hack from e_os.h.
14432 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14434 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14435 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14437 *Joe Orton, Steve Henson*
14439 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14441 * Fix additional bug revealed by the NISCC test suite:
14443 Stop bug triggering large recursion when presented with
14444 certain ASN.1 tags ([CVE-2003-0851])
14448 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14450 * Fix various bugs revealed by running the NISCC test suite:
14452 Stop out of bounds reads in the ASN1 code when presented with
14453 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14455 If verify callback ignores invalid public key errors don't try to check
14456 certificate signature with the NULL public key.
14460 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14461 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14466 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14467 extra data after the compression methods not only for TLS 1.0
14468 but also for SSL 3.0 (as required by the specification).
14470 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14472 * Change X509_certificate_type() to mark the key as exported/exportable
14473 when it's 512 *bits* long, not 512 bytes.
14477 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14479 * Countermeasure against the Klima-Pokorny-Rosa extension of
14480 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14481 a protocol version number mismatch like a decryption error
14482 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14486 * Turn on RSA blinding by default in the default implementation
14487 to avoid a timing attack. Applications that don't want it can call
14488 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14489 They would be ill-advised to do so in most cases.
14491 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14493 * Change RSA blinding code so that it works when the PRNG is not
14494 seeded (in this case, the secret RSA exponent is abused as
14495 an unpredictable seed -- if it is not unpredictable, there
14496 is no point in blinding anyway). Make RSA blinding thread-safe
14497 by remembering the creator's thread ID in rsa->blinding and
14498 having all other threads use local one-time blinding factors
14499 (this requires more computation than sharing rsa->blinding, but
14500 avoids excessive locking; and if an RSA object is not shared
14501 between threads, blinding will still be very fast).
14505 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14507 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14508 via timing by performing a MAC computation even if incorrect
14509 block cipher padding has been found. This is a countermeasure
14510 against active attacks where the attacker has to distinguish
14511 between bad padding and a MAC verification error. ([CVE-2003-0078])
14513 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14514 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14515 Martin Vuagnoux (EPFL, Ilion)*
14517 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14519 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14520 memory from its contents. This is done with a counter that will
14521 place alternating values in each byte. This can be used to solve
14522 two issues: 1) the removal of calls to memset() by highly optimizing
14523 compilers, and 2) cleansing with other values than 0, since those can
14524 be read through on certain media, for example a swap space on disk.
14528 * Bugfix: client side session caching did not work with external caching,
14529 because the session->cipher setting was not restored when reloading
14530 from the external cache. This problem was masked, when
14531 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14532 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14536 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14537 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14539 *Zeev Lieber <zeev-l@yahoo.com>*
14541 * Undo an undocumented change introduced in 0.9.6e which caused
14542 repeated calls to OpenSSL_add_all_ciphers() and
14543 OpenSSL_add_all_digests() to be ignored, even after calling
14548 * Change the default configuration reader to deal with last line not
14549 being properly terminated.
14553 * Change X509_NAME_cmp() so it applies the special rules on handling
14554 DN values that are of type PrintableString, as well as RDNs of type
14555 emailAddress where the value has the type ia5String.
14557 *stefank@valicert.com via Richard Levitte*
14559 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14560 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14561 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14562 the bitwise-OR of the two for use by the majority of applications
14563 wanting this behaviour, and update the docs. The documented
14564 behaviour and actual behaviour were inconsistent and had been
14565 changing anyway, so this is more a bug-fix than a behavioural
14568 *Geoff Thorpe, diagnosed by Nadav Har'El*
14570 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14571 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14575 * Fix initialization code race conditions in
14576 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14577 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14578 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14579 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14580 ssl2_get_cipher_by_char(),
14581 ssl3_get_cipher_by_char().
14583 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14585 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14586 the cached sessions are flushed, as the remove_cb() might use ex_data
14587 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14588 (see [openssl.org #212]).
14590 *Geoff Thorpe, Lutz Jaenicke*
14592 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14593 length, instead of the encoding length to d2i_ASN1_OBJECT.
14597 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14599 * [In 0.9.6g-engine release:]
14600 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14602 *Lynn Gazis <lgazis@rainbow.com>*
14604 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14606 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14607 and get fix the header length calculation.
14608 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14609 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14611 * Use proper error handling instead of 'assertions' in buffer
14612 overflow checks added in 0.9.6e. This prevents DoS (the
14613 assertions could call abort()).
14615 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14617 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14619 * Add various sanity checks to asn1_get_length() to reject
14620 the ASN1 length bytes if they exceed sizeof(long), will appear
14621 negative or the content length exceeds the length of the
14624 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14626 * Fix cipher selection routines: ciphers without encryption had no flags
14627 for the cipher strength set and where therefore not handled correctly
14628 by the selection routines (PR #130).
14632 * Fix EVP_dsa_sha macro.
14637 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14638 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14639 that was added in OpenSSL 0.9.6d.
14641 As the countermeasure turned out to be incompatible with some
14642 broken SSL implementations, the new option is part of SSL_OP_ALL.
14643 SSL_OP_ALL is usually employed when compatibility with weird SSL
14644 implementations is desired (e.g. '-bugs' option to 's_client' and
14645 's_server'), so the new option is automatically set in many
14650 * Changes in security patch:
14652 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14653 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14654 Air Force Materiel Command, USAF, under agreement number
14657 * Add various sanity checks to asn1_get_length() to reject
14658 the ASN1 length bytes if they exceed sizeof(long), will appear
14659 negative or the content length exceeds the length of the
14660 supplied buffer. ([CVE-2002-0659])
14662 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14664 * Assertions for various potential buffer overflows, not known to
14665 happen in practice.
14667 *Ben Laurie (CHATS)*
14669 * Various temporary buffers to hold ASCII versions of integers were
14670 too small for 64 bit platforms. ([CVE-2002-0655])
14671 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14673 * Remote buffer overflow in SSL3 protocol - an attacker could
14674 supply an oversized session ID to a client. ([CVE-2002-0656])
14676 *Ben Laurie (CHATS)*
14678 * Remote buffer overflow in SSL2 protocol - an attacker could
14679 supply an oversized client master key. ([CVE-2002-0656])
14681 *Ben Laurie (CHATS)*
14683 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14685 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14686 encoded as NULL) with id-dsa-with-sha1.
14688 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14690 * Check various `X509_...()` return values in `apps/req.c`.
14692 *Nils Larsch <nla@trustcenter.de>*
14694 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14695 an end-of-file condition would erroneously be flagged, when the CRLF
14696 was just at the end of a processed block. The bug was discovered when
14697 processing data through a buffering memory BIO handing the data to a
14698 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14699 <ptsekov@syntrex.com> and Nedelcho Stanev.
14703 * Implement a countermeasure against a vulnerability recently found
14704 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14705 before application data chunks to avoid the use of known IVs
14706 with data potentially chosen by the attacker.
14710 * Fix length checks in ssl3_get_client_hello().
14714 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14715 to prevent ssl3_read_internal() from incorrectly assuming that
14716 ssl3_read_bytes() found application data while handshake
14717 processing was enabled when in fact s->s3->in_read_app_data was
14718 merely automatically cleared during the initial handshake.
14720 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14722 * Fix object definitions for Private and Enterprise: they were not
14723 recognized in their shortname (=lowercase) representation. Extend
14724 obj_dat.pl to issue an error when using undefined keywords instead
14725 of silently ignoring the problem (Svenning Sorensen
14726 <sss@sss.dnsalias.net>).
14730 * Fix DH_generate_parameters() so that it works for 'non-standard'
14731 generators, i.e. generators other than 2 and 5. (Previously, the
14732 code did not properly initialise the 'add' and 'rem' values to
14733 BN_generate_prime().)
14735 In the new general case, we do not insist that 'generator' is
14736 actually a primitive root: This requirement is rather pointless;
14737 a generator of the order-q subgroup is just as good, if not
14742 * Map new X509 verification errors to alerts. Discovered and submitted by
14743 Tom Wu <tom@arcot.com>.
14747 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14748 returning non-zero before the data has been completely received
14749 when using non-blocking I/O.
14751 *Bodo Moeller; problem pointed out by John Hughes*
14753 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14755 *Ben Laurie, Lutz Jaenicke*
14757 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14758 Yoram Zahavi <YoramZ@gilian.com>).
14762 * Add information about CygWin 1.3 and on, and preserve proper
14763 configuration for the versions before that.
14765 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14767 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14768 check whether we deal with a copy of a session and do not delete from
14769 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14770 <izhar@checkpoint.com>.
14774 * Do not store session data into the internal session cache, if it
14775 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14776 flag is set). Proposed by Aslam <aslam@funk.com>.
14780 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14785 * [In 0.9.6d-engine release:]
14786 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14788 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14790 * Add the configuration target linux-s390x.
14792 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14794 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14795 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14796 variable as an indication that a ClientHello message has been
14797 received. As the flag value will be lost between multiple
14798 invocations of ssl3_accept when using non-blocking I/O, the
14799 function may not be aware that a handshake has actually taken
14800 place, thus preventing a new session from being added to the
14803 To avoid this problem, we now set s->new_session to 2 instead of
14804 using a local variable.
14806 *Lutz Jaenicke, Bodo Moeller*
14808 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14809 if the SSL_R_LENGTH_MISMATCH error is detected.
14811 *Geoff Thorpe, Bodo Moeller*
14813 * New 'shared_ldflag' column in Configure platform table.
14817 * Fix EVP_CIPHER_mode macro.
14819 *"Dan S. Camper" <dan@bti.net>*
14821 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14822 type, we must throw them away by setting rr->length to 0.
14824 *D P Chang <dpc@qualys.com>*
14826 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14828 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14829 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14830 worked incorrectly for those cases where range = `10..._2` and
14831 `3*range` is two bits longer than range.)
14835 * Only add signing time to PKCS7 structures if it is not already
14840 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14841 OBJ_ld_ce should be OBJ_id_ce.
14842 Also some ip-pda OIDs in crypto/objects/objects.txt were
14843 incorrect (cf. RFC 3039).
14845 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14847 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14848 returns early because it has nothing to do.
14850 *Andy Schneider <andy.schneider@bjss.co.uk>*
14852 * [In 0.9.6c-engine release:]
14853 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14855 *Andy Schneider <andy.schneider@bjss.co.uk>*
14857 * [In 0.9.6c-engine release:]
14858 Add support for Cryptographic Appliance's keyserver technology.
14859 (Use engine 'keyclient')
14861 *Cryptographic Appliances and Geoff Thorpe*
14863 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14864 is called via tools/c89.sh because arguments have to be
14865 rearranged (all '-L' options must appear before the first object
14868 *Richard Shapiro <rshapiro@abinitio.com>*
14870 * [In 0.9.6c-engine release:]
14871 Add support for Broadcom crypto accelerator cards, backported
14874 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14876 * [In 0.9.6c-engine release:]
14877 Add support for SureWare crypto accelerator cards from
14878 Baltimore Technologies. (Use engine 'sureware')
14880 *Baltimore Technologies and Mark Cox*
14882 * [In 0.9.6c-engine release:]
14883 Add support for crypto accelerator cards from Accelerated
14884 Encryption Processing, www.aep.ie. (Use engine 'aep')
14886 *AEP Inc. and Mark Cox*
14888 * Add a configuration entry for gcc on UnixWare.
14890 *Gary Benson <gbenson@redhat.com>*
14892 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14893 messages are stored in a single piece (fixed-length part and
14894 variable-length part combined) and fix various bugs found on the way.
14898 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14899 instead. BIO_gethostbyname() does not know what timeouts are
14900 appropriate, so entries would stay in cache even when they have
14902 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14904 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14905 faced with a pathologically small ClientHello fragment that does
14906 not contain client_version: Instead of aborting with an error,
14907 simply choose the highest available protocol version (i.e.,
14908 TLS 1.0 unless it is disabled). In practice, ClientHello
14909 messages are never sent like this, but this change gives us
14910 strictly correct behaviour at least for TLS.
14914 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14915 never resets s->method to s->ctx->method when called from within
14916 one of the SSL handshake functions.
14918 *Bodo Moeller; problem pointed out by Niko Baric*
14920 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14921 (sent using the client's version number) if client_version is
14922 smaller than the protocol version in use. Also change
14923 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14924 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14925 the client will at least see that alert.
14929 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14934 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14935 client receives HelloRequest while in a handshake.
14937 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14939 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14940 should end in 'break', not 'goto end' which circumvents various
14941 cleanups done in state SSL_ST_OK. But session related stuff
14942 must be disabled for SSL_ST_OK in the case that we just sent a
14945 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14946 before just sending a HelloRequest.
14948 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14950 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14951 reveal whether illegal block cipher padding was found or a MAC
14952 verification error occurred. (Neither SSLerr() codes nor alerts
14953 are directly visible to potential attackers, but the information
14954 may leak via logfiles.)
14956 Similar changes are not required for the SSL 2.0 implementation
14957 because the number of padding bytes is sent in clear for SSL 2.0,
14958 and the extra bytes are just ignored. However ssl/s2_pkt.c
14959 failed to verify that the purported number of padding bytes is in
14964 * Add OpenUNIX-8 support including shared libraries
14965 (Boyd Lynn Gerber <gerberb@zenez.com>).
14969 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14970 'wristwatch attack' using huge encoding parameters (cf.
14971 James H. Manger's CRYPTO 2001 paper). Note that the
14972 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14973 encoding parameters and hence was not vulnerable.
14977 * BN_sqr() bug fix.
14979 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14981 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14982 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14983 followed by modular reduction.
14985 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14987 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14988 equivalent based on BN_pseudo_rand() instead of BN_rand().
14992 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14993 This function was broken, as the check for a new client hello message
14994 to handle SGC did not allow these large messages.
14995 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14999 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15003 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15004 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15008 * Rework the configuration and shared library support for Tru64 Unix.
15009 The configuration part makes use of modern compiler features and
15010 still retains old compiler behavior for those that run older versions
15011 of the OS. The shared library support part includes a variant that
15012 uses the RPATH feature, and is available through the special
15013 configuration target "alpha-cc-rpath", which will never be selected
15016 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15018 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15019 with the same message size as in ssl3_get_certificate_request().
15020 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15021 messages might inadvertently be reject as too long.
15023 *Petr Lampa <lampa@fee.vutbr.cz>*
15025 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15029 * Modified SSL library such that the verify_callback that has been set
15030 specifically for an SSL object with SSL_set_verify() is actually being
15031 used. Before the change, a verify_callback set with this function was
15032 ignored and the verify_callback() set in the SSL_CTX at the time of
15033 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15034 to allow the necessary settings.
15038 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15039 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15040 done automatically (in contradiction to the requirements of the C
15041 standard). This made problems when used from OpenSSH.
15045 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15046 dh->length and always used
15048 BN_rand_range(priv_key, dh->p).
15050 BN_rand_range() is not necessary for Diffie-Hellman, and this
15051 specific range makes Diffie-Hellman unnecessarily inefficient if
15052 dh->length (recommended exponent length) is much smaller than the
15053 length of dh->p. We could use BN_rand_range() if the order of
15054 the subgroup was stored in the DH structure, but we only have
15059 BN_rand(priv_key, l, ...)
15061 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15068 RSA_eay_public_encrypt
15069 RSA_eay_private_decrypt
15070 RSA_eay_private_encrypt (signing)
15071 RSA_eay_public_decrypt (signature verification)
15073 (default implementations for RSA_public_encrypt,
15074 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15075 always reject numbers >= n.
15079 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15080 to synchronize access to 'locking_thread'. This is necessary on
15081 systems where access to 'locking_thread' (an 'unsigned long'
15082 variable) is not atomic.
15086 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15087 *before* setting the 'crypto_lock_rand' flag. The previous code had
15088 a race condition if 0 is a valid thread ID.
15090 *Travis Vitek <vitek@roguewave.com>*
15092 * Add support for shared libraries under Irix.
15094 *Albert Chin-A-Young <china@thewrittenword.com>*
15096 * Add configuration option to build on Linux on both big-endian and
15097 little-endian MIPS.
15099 *Ralf Baechle <ralf@uni-koblenz.de>*
15101 * Add the possibility to create shared libraries on HP-UX.
15105 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15107 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15108 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15109 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15110 PRNG state recovery was possible based on the output of
15111 one PRNG request appropriately sized to gain knowledge on
15112 'md' followed by enough consecutive 1-byte PRNG requests
15113 to traverse all of 'state'.
15115 1. When updating 'md_local' (the current thread's copy of 'md')
15116 during PRNG output generation, hash all of the previous
15117 'md_local' value, not just the half used for PRNG output.
15119 2. Make the number of bytes from 'state' included into the hash
15120 independent from the number of PRNG bytes requested.
15122 The first measure alone would be sufficient to avoid
15123 Markku-Juhani's attack. (Actually it had never occurred
15124 to me that the half of 'md_local' used for chaining was the
15125 half from which PRNG output bytes were taken -- I had always
15126 assumed that the secret half would be used.) The second
15127 measure makes sure that additional data from 'state' is never
15128 mixed into 'md_local' in small portions; this heuristically
15129 further strengthens the PRNG.
15133 * Fix crypto/bn/asm/mips3.s.
15137 * When only the key is given to "enc", the IV is undefined. Print out
15138 an error message in this case.
15142 * Handle special case when X509_NAME is empty in X509 printing routines.
15146 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15147 positive and less than q.
15151 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15152 used: it isn't thread safe and the add_lock_callback should handle
15155 *Paul Rose <Paul.Rose@bridge.com>*
15157 * Verify that incoming data obeys the block size in
15158 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15164 *Ulf Möller, Bodo Möller*
15166 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15167 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15168 when fixing the server behaviour for backwards-compatible 'client
15169 hello' messages. (Note that the attack is impractical against
15170 SSL 3.0 and TLS 1.0 anyway because length and version checking
15171 means that the probability of guessing a valid ciphertext is
15172 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15175 Before 0.9.5, the countermeasure (hide the error by generating a
15176 random 'decryption result') did not work properly because
15177 ERR_clear_error() was missing, meaning that SSL_get_error() would
15178 detect the supposedly ignored error.
15180 Both problems are now fixed.
15184 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15185 (previously it was 1024).
15189 * Fix for compatibility mode trust settings: ignore trust settings
15190 unless some valid trust or reject settings are present.
15194 * Fix for blowfish EVP: its a variable length cipher.
15198 * Fix various bugs related to DSA S/MIME verification. Handle missing
15199 parameters in DSA public key structures and return an error in the
15200 DSA routines if parameters are absent.
15204 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15205 in the current directory if neither $RANDFILE nor $HOME was set.
15206 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15207 caused some confusion to Windows users who haven't defined $HOME.
15208 Thus RAND_file_name() is changed again: e_os.h can define a
15209 DEFAULT_HOME, which will be used if $HOME is not set.
15210 For Windows, we use "C:"; on other platforms, we still require
15211 environment variables.
15213 * Move 'if (!initialized) RAND_poll()' into regions protected by
15214 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15215 having multiple threads call RAND_poll() concurrently.
15219 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15220 combination of a flag and a thread ID variable.
15221 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15222 flag), *other* threads can enter ssleay_add_bytes without obeying
15223 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15224 that they do not hold after the first thread unsets add_do_not_lock).
15228 * Change bctest again: '-x' expressions are not available in all
15229 versions of 'test'.
15233 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15235 * Fix a couple of memory leaks in PKCS7_dataDecode()
15237 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15239 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15240 the default extension for executables, if any. Also, make the perl
15241 scripts that use symlink() to test if it really exists and use "cp"
15242 if it doesn't. All this made OpenSSL compilable and installable in
15247 * Fix for asn1_GetSequence() for indefinite length constructed data.
15248 If SEQUENCE is length is indefinite just set c->slen to the total
15249 amount of data available.
15251 *Steve Henson, reported by shige@FreeBSD.org*
15253 *This change does not apply to 0.9.7.*
15255 * Change bctest to avoid here-documents inside command substitution
15256 (workaround for FreeBSD /bin/sh bug).
15257 For compatibility with Ultrix, avoid shell functions (introduced
15258 in the bctest version that searches along $PATH).
15262 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15263 with des_encrypt() defined on some operating systems, like Solaris
15268 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15269 On the Importance of Eliminating Errors in Cryptographic
15270 Computations, J. Cryptology 14 (2001) 2, 101-119,
15271 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15275 * MIPS assembler BIGNUM division bug fix.
15279 * Disabled incorrect Alpha assembler code.
15283 * Fix PKCS#7 decode routines so they correctly update the length
15284 after reading an EOC for the EXPLICIT tag.
15288 *This change does not apply to 0.9.7.*
15290 * Fix bug in PKCS#12 key generation routines. This was triggered
15291 if a 3DES key was generated with a 0 initial byte. Include
15292 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15293 (but broken) behaviour.
15297 * Enhance bctest to search for a working bc along $PATH and print
15300 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15302 * Fix memory leaks in err.c: free err_data string if necessary;
15303 don't write to the wrong index in ERR_set_error_data.
15307 * Implement ssl23_peek (analogous to ssl23_read), which previously
15312 * Replace rdtsc with `_emit` statements for VC++ version 5.
15314 *Jeremy Cooper <jeremy@baymoo.org>*
15316 * Make it possible to reuse SSLv2 sessions.
15320 * In copy_email() check for >= 0 as a return value for
15321 X509_NAME_get_index_by_NID() since 0 is a valid index.
15323 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15325 * Avoid coredump with unsupported or invalid public keys by checking if
15326 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15327 PKCS7_verify() fails with non detached data.
15331 * Don't use getenv in library functions when run as setuid/setgid.
15332 New function OPENSSL_issetugid().
15336 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15337 due to incorrect handling of multi-threading:
15339 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15341 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15343 3. Count how many times MemCheck_off() has been called so that
15344 nested use can be treated correctly. This also avoids
15345 inband-signalling in the previous code (which relied on the
15346 assumption that thread ID 0 is impossible).
15350 * Add "-rand" option also to s_client and s_server.
15354 * Fix CPU detection on Irix 6.x.
15355 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15356 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15358 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15363 *This change does not apply to 0.9.7.*
15365 * Use the cached encoding of an X509_NAME structure rather than
15366 copying it. This is apparently the reason for the libsafe "errors"
15367 but the code is actually correct.
15371 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15372 Bleichenbacher's DSA attack.
15373 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15374 to be set and top=0 forces the highest bit to be set; top=-1 is new
15375 and leaves the highest bit random.
15377 *Ulf Moeller, Bodo Moeller*
15379 * In the `NCONF_...`-based implementations for `CONF_...` queries
15380 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15381 a temporary CONF structure with the data component set to NULL
15382 (which gives segmentation faults in lh_retrieve).
15383 Instead, use NULL for the CONF pointer in CONF_get_string and
15384 CONF_get_number (which may use environment variables) and directly
15385 return NULL from CONF_get_section.
15389 * Fix potential buffer overrun for EBCDIC.
15393 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15394 keyUsage if basicConstraints absent for a CA.
15398 * Make SMIME_write_PKCS7() write mail header values with a format that
15399 is more generally accepted (no spaces before the semicolon), since
15400 some programs can't parse those values properly otherwise. Also make
15401 sure BIO's that break lines after each write do not create invalid
15406 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15407 macros previously used would not encode an empty SEQUENCE OF
15408 and break the signature.
15412 *This change does not apply to 0.9.7.*
15414 * Zero the premaster secret after deriving the master secret in
15419 * Add some EVP_add_digest_alias registrations (as found in
15420 OpenSSL_add_all_digests()) to SSL_library_init()
15421 aka OpenSSL_add_ssl_algorithms(). This provides improved
15422 compatibility with peers using X.509 certificates
15423 with unconventional AlgorithmIdentifier OIDs.
15427 * Fix for Irix with NO_ASM.
15429 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15431 * ./config script fixes.
15433 *Ulf Moeller, Richard Levitte*
15435 * Fix 'openssl passwd -1'.
15439 * Change PKCS12_key_gen_asc() so it can cope with non null
15440 terminated strings whose length is passed in the passlen
15441 parameter, for example from PEM callbacks. This was done
15442 by adding an extra length parameter to asc2uni().
15444 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15446 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15447 call failed, free the DSA structure.
15451 * Fix to uni2asc() to cope with zero length Unicode strings.
15452 These are present in some PKCS#12 files.
15456 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15457 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15458 when writing a 32767 byte record.
15460 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15462 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15463 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15465 (RSA objects have a reference count access to which is protected
15466 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15467 so they are meant to be shared between threads.)
15468 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15469 "Reddie, Steven" <Steven.Reddie@ca.com>*
15471 * Fix a deadlock in CRYPTO_mem_leaks().
15475 * Use better test patterns in bntest.
15479 * rand_win.c fix for Borland C.
15483 * BN_rshift bugfix for n == 0.
15487 * Add a 'bctest' script that checks for some known 'bc' bugs
15488 so that 'make test' does not abort just because 'bc' is broken.
15492 * Store verify_result within SSL_SESSION also for client side to
15493 avoid potential security hole. (Re-used sessions on the client side
15494 always resulted in verify_result==X509_V_OK, not using the original
15495 result of the server certificate verification.)
15499 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15500 SSL3_RT_APPLICATION_DATA, return 0.
15501 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15506 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15507 releases, have been re-implemented by renaming the previous
15508 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15509 and ssl3_read_internal, respectively, and adding 'peek' parameters
15510 to them. The new ssl[23]_{read,peek} functions are calls to
15511 ssl[23]_read_internal with the 'peek' flag set appropriately.
15512 A 'peek' parameter has also been added to ssl3_read_bytes, which
15513 does the actual work for ssl3_read_internal.
15517 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15518 the method-specific "init()" handler. Also clean up ex_data after
15519 calling the method-specific "finish()" handler. Previously, this was
15520 happening the other way round.
15524 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15525 The previous value, 12, was not always sufficient for BN_mod_exp().
15529 * Make sure that shared libraries get the internal name engine with
15530 the full version number and not just 0. This should mark the
15531 shared libraries as not backward compatible. Of course, this should
15532 be changed again when we can guarantee backward binary compatibility.
15536 * Fix typo in get_cert_by_subject() in by_dir.c
15538 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15540 * Rework the system to generate shared libraries:
15542 - Make note of the expected extension for the shared libraries and
15543 if there is a need for symbolic links from for example libcrypto.so.0
15544 to libcrypto.so.0.9.7. There is extended info in Configure for
15547 - Make as few rebuilds of the shared libraries as possible.
15549 - Still avoid linking the OpenSSL programs with the shared libraries.
15551 - When installing, install the shared libraries separately from the
15556 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15558 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15559 and not in SSL_clear because the latter is also used by the
15560 accept/connect functions; previously, the settings made by
15561 SSL_set_read_ahead would be lost during the handshake.
15563 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15565 * Correct util/mkdef.pl to be selective about disabled algorithms.
15566 Previously, it would create entries for disabled algorithms no
15571 * Added several new manual pages for SSL_* function.
15575 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15577 * In ssl23_get_client_hello, generate an error message when faced
15578 with an initial SSL 3.0/TLS record that is too small to contain the
15579 first two bytes of the ClientHello message, i.e. client_version.
15580 (Note that this is a pathologic case that probably has never happened
15581 in real life.) The previous approach was to use the version number
15582 from the record header as a substitute; but our protocol choice
15583 should not depend on that one because it is not authenticated
15584 by the Finished messages.
15588 * More robust randomness gathering functions for Windows.
15590 *Jeffrey Altman <jaltman@columbia.edu>*
15592 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15593 not set then we don't setup the error code for issuer check errors
15594 to avoid possibly overwriting other errors which the callback does
15595 handle. If an application does set the flag then we assume it knows
15596 what it is doing and can handle the new informational codes
15601 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15602 a general "ANY" type, as such it should be able to decode anything
15603 including tagged types. However it didn't check the class so it would
15604 wrongly interpret tagged types in the same way as their universal
15605 counterpart and unknown types were just rejected. Changed so that the
15606 tagged and unknown types are handled in the same way as a SEQUENCE:
15607 that is the encoding is stored intact. There is also a new type
15608 "V_ASN1_OTHER" which is used when the class is not universal, in this
15609 case we have no idea what the actual type is so we just lump them all
15614 * On VMS, stdout may very well lead to a file that is written to
15615 in a record-oriented fashion. That means that every write() will
15616 write a separate record, which will be read separately by the
15617 programs trying to read from it. This can be very confusing.
15619 The solution is to put a BIO filter in the way that will buffer
15620 text until a linefeed is reached, and then write everything a
15621 line at a time, so every record written will be an actual line,
15622 not chunks of lines and not (usually doesn't happen, but I've
15623 seen it once) several lines in one record. BIO_f_linebuffer() is
15626 Currently, it's a VMS-only method, because that's where it has
15627 been tested well enough.
15631 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15632 it can return incorrect results.
15633 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15634 but it was in 0.9.6-beta[12].)
15638 * Disable the check for content being present when verifying detached
15639 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15640 include zero length content when signing messages.
15644 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15645 BIO_ctrl (for BIO pairs).
15649 * Add DSO method for VMS.
15653 * Bug fix: Montgomery multiplication could produce results with the
15658 * Add RPM specification openssl.spec and modify it to build three
15659 packages. The default package contains applications, application
15660 documentation and run-time libraries. The devel package contains
15661 include files, static libraries and function documentation. The
15662 doc package contains the contents of the doc directory. The original
15663 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15667 * Add a large number of documentation files for many SSL routines.
15669 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15671 * Add a configuration entry for Sony News 4.
15673 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15675 * Don't set the two most significant bits to one when generating a
15676 random number < q in the DSA library.
15680 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15681 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15682 the underlying transport is blocking) if a handshake took place.
15683 (The default behaviour is needed by applications such as s_client
15684 and s_server that use select() to determine when to use SSL_read;
15685 but for applications that know in advance when to expect data, it
15686 just makes things more complicated.)
15690 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15695 * Add a few more EBCDIC conditionals that make `req` and `x509`
15696 work better on such systems.
15698 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15700 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15701 Update PKCS12_parse() so it copies the friendlyName and the
15702 keyid to the certificates aux info.
15706 * Fix bug in PKCS7_verify() which caused an infinite loop
15707 if there was more than one signature.
15709 *Sven Uszpelkat <su@celocom.de>*
15711 * Major change in util/mkdef.pl to include extra information
15712 about each symbol, as well as presenting variables as well
15713 as functions. This change means that there's n more need
15714 to rebuild the .num files when some algorithms are excluded.
15718 * Allow the verify time to be set by an application,
15719 rather than always using the current time.
15723 * Phase 2 verify code reorganisation. The certificate
15724 verify code now looks up an issuer certificate by a
15725 number of criteria: subject name, authority key id
15726 and key usage. It also verifies self signed certificates
15727 by the same criteria. The main comparison function is
15728 X509_check_issued() which performs these checks.
15730 Lot of changes were necessary in order to support this
15731 without completely rewriting the lookup code.
15733 Authority and subject key identifier are now cached.
15735 The LHASH 'certs' is X509_STORE has now been replaced
15736 by a STACK_OF(X509_OBJECT). This is mainly because an
15737 LHASH can't store or retrieve multiple objects with
15738 the same hash value.
15740 As a result various functions (which were all internal
15741 use only) have changed to handle the new X509_STORE
15742 structure. This will break anything that messed round
15743 with X509_STORE internally.
15745 The functions X509_STORE_add_cert() now checks for an
15746 exact match, rather than just subject name.
15748 The X509_STORE API doesn't directly support the retrieval
15749 of multiple certificates matching a given criteria, however
15750 this can be worked round by performing a lookup first
15751 (which will fill the cache with candidate certificates)
15752 and then examining the cache for matches. This is probably
15753 the best we can do without throwing out X509_LOOKUP
15754 entirely (maybe later...).
15756 The X509_VERIFY_CTX structure has been enhanced considerably.
15758 All certificate lookup operations now go via a get_issuer()
15759 callback. Although this currently uses an X509_STORE it
15760 can be replaced by custom lookups. This is a simple way
15761 to bypass the X509_STORE hackery necessary to make this
15762 work and makes it possible to use more efficient techniques
15763 in future. A very simple version which uses a simple
15764 STACK for its trusted certificate store is also provided
15765 using X509_STORE_CTX_trusted_stack().
15767 The verify_cb() and verify() callbacks now have equivalents
15768 in the X509_STORE_CTX structure.
15770 X509_STORE_CTX also has a 'flags' field which can be used
15771 to customise the verify behaviour.
15775 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15776 excludes S/MIME capabilities.
15780 * When a certificate request is read in keep a copy of the
15781 original encoding of the signed data and use it when outputting
15782 again. Signatures then use the original encoding rather than
15783 a decoded, encoded version which may cause problems if the
15784 request is improperly encoded.
15788 * For consistency with other BIO_puts implementations, call
15789 buffer_write(b, ...) directly in buffer_puts instead of calling
15792 In BIO_puts, increment b->num_write as in BIO_write.
15794 *Peter.Sylvester@EdelWeb.fr*
15796 * Fix BN_mul_word for the case where the word is 0. (We have to use
15797 BN_zero, we may not return a BIGNUM with an array consisting of
15798 words set to zero.)
15802 * Avoid calling abort() from within the library when problems are
15803 detected, except if preprocessor symbols have been defined
15804 (such as REF_CHECK, BN_DEBUG etc.).
15808 * New openssl application 'rsautl'. This utility can be
15809 used for low-level RSA operations. DER public key
15810 BIO/fp routines also added.
15814 * New Configure entry and patches for compiling on QNX 4.
15816 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15818 * A demo state-machine implementation was sponsored by
15819 Nuron (<http://www.nuron.com/>) and is now available in
15820 demos/state_machine.
15824 * New options added to the 'dgst' utility for signature
15825 generation and verification.
15829 * Unrecognized PKCS#7 content types are now handled via a
15830 catch all ASN1_TYPE structure. This allows unsupported
15831 types to be stored as a "blob" and an application can
15832 encode and decode it manually.
15836 * Fix various signed/unsigned issues to make a_strex.c
15837 compile under VC++.
15839 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15841 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15842 length if passed a buffer. ASN1_INTEGER_to_BN failed
15843 if passed a NULL BN and its argument was negative.
15845 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15847 * Modification to PKCS#7 encoding routines to output definite
15848 length encoding. Since currently the whole structures are in
15849 memory there's not real point in using indefinite length
15850 constructed encoding. However if OpenSSL is compiled with
15851 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15855 * Added BIO_vprintf() and BIO_vsnprintf().
15859 * Added more prefixes to parse for in the strings written
15860 through a logging bio, to cover all the levels that are available
15861 through syslog. The prefixes are now:
15863 PANIC, EMERG, EMR => LOG_EMERG
15864 ALERT, ALR => LOG_ALERT
15865 CRIT, CRI => LOG_CRIT
15866 ERROR, ERR => LOG_ERR
15867 WARNING, WARN, WAR => LOG_WARNING
15868 NOTICE, NOTE, NOT => LOG_NOTICE
15869 INFO, INF => LOG_INFO
15870 DEBUG, DBG => LOG_DEBUG
15872 and as before, if none of those prefixes are present at the
15873 beginning of the string, LOG_ERR is chosen.
15875 On Win32, the `LOG_*` levels are mapped according to this:
15877 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15878 LOG_WARNING => EVENTLOG_WARNING_TYPE
15879 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15883 * Made it possible to reconfigure with just the configuration
15884 argument "reconf" or "reconfigure". The command line arguments
15885 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15886 and are retrieved from there when reconfiguring.
15892 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15894 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15898 * The obj_dat.pl script was messing up the sorting of object
15899 names. The reason was that it compared the quoted version
15900 of strings as a result "OCSP" > "OCSP Signing" because
15901 " > SPACE. Changed script to store unquoted versions of
15902 names and add quotes on output. It was also omitting some
15903 names from the lookup table if they were given a default
15904 value (that is if SN is missing it is given the same
15905 value as LN and vice versa), these are now added on the
15906 grounds that if an object has a name we should be able to
15907 look it up. Finally added warning output when duplicate
15908 short or long names are found.
15912 * Changes needed for Tandem NSK.
15914 *Scott Uroff <scott@xypro.com>*
15916 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15917 RSA_padding_check_SSLv23(), special padding was never detected
15918 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15919 version rollback attacks was not effective.
15921 In s23_clnt.c, don't use special rollback-attack detection padding
15922 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15923 client; similarly, in s23_srvr.c, don't do the rollback check if
15924 SSL 2.0 is the only protocol enabled in the server.
15928 * Make it possible to get hexdumps of unprintable data with 'openssl
15929 asn1parse'. By implication, the functions ASN1_parse_dump() and
15930 BIO_dump_indent() are added.
15934 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15935 these print out strings and name structures based on various
15936 flags including RFC2253 support and proper handling of
15937 multibyte characters. Added options to the 'x509' utility
15938 to allow the various flags to be set.
15942 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15943 Also change the functions X509_cmp_current_time() and
15944 X509_gmtime_adj() work with an ASN1_TIME structure,
15945 this will enable certificates using GeneralizedTime in validity
15946 dates to be checked.
15950 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15951 negative public key encodings) on by default,
15952 NO_NEG_PUBKEY_BUG can be set to disable it.
15956 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15957 content octets. An i2c_ASN1_OBJECT is unnecessary because
15958 the encoding can be trivially obtained from the structure.
15962 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15963 not read locks (`CRYPTO_r_[un]lock`).
15967 * A first attempt at creating official support for shared
15968 libraries through configuration. I've kept it so the
15969 default is static libraries only, and the OpenSSL programs
15970 are always statically linked for now, but there are
15971 preparations for dynamic linking in place.
15972 This has been tested on Linux and Tru64.
15976 * Randomness polling function for Win9x, as described in:
15977 Peter Gutmann, Software Generation of Practically Strong
15982 * Fix so PRNG is seeded in req if using an already existing
15987 * New options to smime application. -inform and -outform
15988 allow alternative formats for the S/MIME message including
15989 PEM and DER. The -content option allows the content to be
15990 specified separately. This should allow things like Netscape
15991 form signing output easier to verify.
15995 * Fix the ASN1 encoding of tags using the 'long form'.
15999 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16000 STRING types. These convert content octets to and from the
16001 underlying type. The actual tag and length octets are
16002 already assumed to have been read in and checked. These
16003 are needed because all other string types have virtually
16004 identical handling apart from the tag. By having versions
16005 of the ASN1 functions that just operate on content octets
16006 IMPLICIT tagging can be handled properly. It also allows
16007 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16008 and ASN1_INTEGER are identical apart from the tag.
16012 * Change the handling of OID objects as follows:
16014 - New object identifiers are inserted in objects.txt, following
16015 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16016 - objects.pl is used to process obj_mac.num and create a new
16018 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16021 This is currently kind of a hack, and the perl code in objects.pl
16022 isn't very elegant, but it works as I intended. The simplest way
16023 to check that it worked correctly is to look in obj_dat.h and
16024 check the array nid_objs and make sure the objects haven't moved
16025 around (this is important!). Additions are OK, as well as
16026 consistent name changes.
16030 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16034 * Addition of the command line parameter '-rand file' to 'openssl req'.
16035 The given file adds to whatever has already been seeded into the
16036 random pool through the RANDFILE configuration file option or
16037 environment variable, or the default random state file.
16041 * mkstack.pl now sorts each macro group into lexical order.
16042 Previously the output order depended on the order the files
16043 appeared in the directory, resulting in needless rewriting
16048 * Patches to make OpenSSL compile under Win32 again. Mostly
16049 work arounds for the VC++ problem that it treats func() as
16050 func(void). Also stripped out the parts of mkdef.pl that
16051 added extra typesafe functions: these no longer exist.
16055 * Reorganisation of the stack code. The macros are now all
16056 collected in safestack.h . Each macro is defined in terms of
16057 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16058 DEBUG_SAFESTACK is now handled in terms of function casts,
16059 this has the advantage of retaining type safety without the
16060 use of additional functions. If DEBUG_SAFESTACK is not defined
16061 then the non typesafe macros are used instead. Also modified the
16062 mkstack.pl script to handle the new form. Needs testing to see
16063 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16064 the default if no major problems. Similar behaviour for ASN1_SET_OF
16065 and PKCS12_STACK_OF.
16069 * When some versions of IIS use the 'NET' form of private key the
16070 key derivation algorithm is different. Normally MD5(password) is
16071 used as a 128 bit RC4 key. In the modified case
16072 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16073 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16074 as the old Netscape_RSA functions except they have an additional
16075 'sgckey' parameter which uses the modified algorithm. Also added
16076 an -sgckey command line option to the rsa utility. Thanks to
16077 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16078 algorithm to openssl-dev.
16082 * The evp_local.h macros were using 'c.##kname' which resulted in
16083 invalid expansion on some systems (SCO 5.0.5 for example).
16084 Corrected to 'c.kname'.
16086 *Phillip Porch <root@theporch.com>*
16088 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16089 a STACK of email addresses from a certificate or request, these look
16090 in the subject name and the subject alternative name extensions and
16091 omit any duplicate addresses.
16095 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16096 This makes DSA verification about 2 % faster.
16100 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16101 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16102 plus overhead for 1024 bit moduli).
16103 This makes exponentiations about 0.5 % faster for 1024 bit
16104 exponents (as measured by "openssl speed rsa2048").
16108 * Rename memory handling macros to avoid conflicts with other
16110 Malloc => OPENSSL_malloc
16111 Malloc_locked => OPENSSL_malloc_locked
16112 Realloc => OPENSSL_realloc
16113 Free => OPENSSL_free
16117 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16118 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16122 * CygWin32 support.
16124 *John Jarvie <jjarvie@newsguy.com>*
16126 * The type-safe stack code has been rejigged. It is now only compiled
16127 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16128 by default all type-specific stack functions are "#define"d back to
16129 standard stack functions. This results in more streamlined output
16130 but retains the type-safety checking possibilities of the original
16135 * The STACK code has been cleaned up, and certain type declarations
16136 that didn't make a lot of sense have been brought in line. This has
16137 also involved a cleanup of sorts in safestack.h to more correctly
16138 map type-safe stack functions onto their plain stack counterparts.
16139 This work has also resulted in a variety of "const"ifications of
16140 lots of the code, especially `_cmp` operations which should normally
16141 be prototyped with "const" parameters anyway.
16145 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16146 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16147 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16148 where all of 'md' is used each time the PRNG is used, but 'state'
16149 is used only indexed by a cyclic counter. As entropy may not be
16150 well distributed from the beginning, 'md' is important as a
16151 chaining variable. However, the output function chains only half
16152 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16153 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16154 in all of 'state' being rewritten, with the new values depending
16155 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16159 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16160 the handshake is continued after ssl_verify_cert_chain();
16161 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16162 can lead to 'unexplainable' connection aborts later.
16164 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16166 * Major EVP API cipher revision.
16167 Add hooks for extra EVP features. This allows various cipher
16168 parameters to be set in the EVP interface. Support added for variable
16169 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16170 setting of RC2 and RC5 parameters.
16172 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16175 Remove lots of duplicated code from the EVP library. For example *every*
16176 cipher init() function handles the 'iv' in the same way according to the
16177 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16178 for CFB and OFB modes they zero ctx->num.
16180 New functionality allows removal of S/MIME code RC2 hack.
16182 Most of the routines have the same form and so can be declared in terms
16185 By shifting this to the top level EVP_CipherInit() it can be removed from
16186 all individual ciphers. If the cipher wants to handle IVs or keys
16187 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16190 Change lots of functions like EVP_EncryptUpdate() to now return a
16191 value: although software versions of the algorithms cannot fail
16192 any installed hardware versions can.
16196 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16197 this option is set, tolerate broken clients that send the negotiated
16198 protocol version number instead of the requested protocol version
16203 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16204 i.e. non-zero for export ciphersuites, zero otherwise.
16205 Previous versions had this flag inverted, inconsistent with
16206 rsa_tmp_cb (..._TMP_RSA_CB).
16208 *Bodo Moeller; problem reported by Amit Chopra*
16210 * Add missing DSA library text string. Work around for some IIS
16211 key files with invalid SEQUENCE encoding.
16215 * Add a document (doc/standards.txt) that list all kinds of standards
16216 and so on that are implemented in OpenSSL.
16220 * Enhance c_rehash script. Old version would mishandle certificates
16221 with the same subject name hash and wouldn't handle CRLs at all.
16222 Added -fingerprint option to crl utility, to support new c_rehash
16227 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16231 * Fix for SSL server purpose checking. Server checking was
16232 rejecting certificates which had extended key usage present
16233 but no ssl client purpose.
16235 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16237 * Make PKCS#12 code work with no password. The PKCS#12 spec
16238 is a little unclear about how a blank password is handled.
16239 Since the password in encoded as a BMPString with terminating
16240 double NULL a zero length password would end up as just the
16241 double NULL. However no password at all is different and is
16242 handled differently in the PKCS#12 key generation code. NS
16243 treats a blank password as zero length. MSIE treats it as no
16244 password on export: but it will try both on import. We now do
16245 the same: PKCS12_parse() tries zero length and no password if
16246 the password is set to "" or NULL (NULL is now a valid password:
16247 it wasn't before) as does the pkcs12 application.
16251 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16252 perror when PEM_read_bio_X509_REQ fails, the error message must
16253 be obtained from the error queue.
16257 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16258 it in ERR_remove_state if appropriate, and change ERR_get_state
16259 accordingly to avoid race conditions (this is necessary because
16260 thread_hash is no longer constant once set).
16264 * Bugfix for linux-elf makefile.one.
16268 * RSA_get_default_method() will now cause a default
16269 RSA_METHOD to be chosen if one doesn't exist already.
16270 Previously this was only set during a call to RSA_new()
16271 or RSA_new_method(NULL) meaning it was possible for
16272 RSA_get_default_method() to return NULL.
16276 * Added native name translation to the existing DSO code
16277 that will convert (if the flag to do so is set) filenames
16278 that are sufficiently small and have no path information
16279 into a canonical native form. Eg. "blah" converted to
16280 "libblah.so" or "blah.dll" etc.
16284 * New function ERR_error_string_n(e, buf, len) which is like
16285 ERR_error_string(e, buf), but writes at most 'len' bytes
16286 including the 0 terminator. For ERR_error_string_n, 'buf'
16289 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16291 * CONF library reworked to become more general. A new CONF
16292 configuration file reader "class" is implemented as well as a
16293 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16294 old `CONF_*` functions are still there, but are reimplemented to
16295 work in terms of the new functions. Also, a set of functions
16296 to handle the internal storage of the configuration data is
16297 provided to make it easier to write new configuration file
16298 reader "classes" (I can definitely see something reading a
16299 configuration file in XML format, for example), called `_CONF_*`,
16300 or "the configuration storage API"...
16302 The new configuration file reading functions are:
16304 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16305 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16307 NCONF_default, NCONF_WIN32
16309 NCONF_dump_fp, NCONF_dump_bio
16311 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16312 NCONF_new creates a new CONF object. This works in the same way
16313 as other interfaces in OpenSSL, like the BIO interface.
16314 `NCONF_dump_*` dump the internal storage of the configuration file,
16315 which is useful for debugging. All other functions take the same
16316 arguments as the old `CONF_*` functions with the exception of the
16317 first that must be a `CONF *` instead of a `LHASH *`.
16319 To make it easier to use the new classes with the old `CONF_*` functions,
16320 the function CONF_set_default_method is provided.
16324 * Add '-tls1' option to 'openssl ciphers', which was already
16325 mentioned in the documentation but had not been implemented.
16326 (This option is not yet really useful because even the additional
16327 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16331 * Initial DSO code added into libcrypto for letting OpenSSL (and
16332 OpenSSL-based applications) load shared libraries and bind to
16333 them in a portable way.
16335 *Geoff Thorpe, with contributions from Richard Levitte*
16337 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16339 * Make sure _lrotl and _lrotr are only used with MSVC.
16341 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16342 (the default implementation of RAND_status).
16344 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16345 to '-clrext' (= clear extensions), as intended and documented.
16346 *Bodo Moeller; inconsistency pointed out by Michael Attili
16347 <attili@amaxo.com>*
16349 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16350 was larger than the MD block size.
16352 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16354 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16355 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16356 using the passed key: if the passed key was a private key the result
16357 of X509_print(), for example, would be to print out all the private key
16362 * des_quad_cksum() byte order bug fix.
16363 *Ulf Möller, using the problem description in krb4-0.9.7, where
16364 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16366 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16369 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16371 * For easily testing in shell scripts whether some command
16372 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16373 returns with exit code 0 iff no command of the given name is available.
16374 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16375 the output goes to stdout and nothing is printed to stderr.
16376 Additional arguments are always ignored.
16378 Since for each cipher there is a command of the same name,
16379 the 'no-cipher' compilation switches can be tested this way.
16381 ('openssl no-XXX' is not able to detect pseudo-commands such
16382 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16386 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16390 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16391 is set; it will be thrown away anyway because each handshake creates
16393 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16394 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16395 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16396 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16400 * New s_client option -ign_eof: EOF at stdin is ignored, and
16401 'Q' and 'R' lose their special meanings (quit/renegotiate).
16402 This is part of what -quiet does; unlike -quiet, -ign_eof
16403 does not suppress any output.
16407 * Add compatibility options to the purpose and trust code. The
16408 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16409 accepts a certificate or CA, this was the previous behaviour,
16410 with all the associated security issues.
16412 X509_TRUST_COMPAT is the old trust behaviour: only and
16413 automatically trust self signed roots in certificate store. A
16414 new trust setting X509_TRUST_DEFAULT is used to specify that
16415 a purpose has no associated trust setting and it should instead
16416 use the value in the default purpose.
16420 * Fix the PKCS#8 DSA private key code so it decodes keys again
16421 and fix a memory leak.
16425 * In util/mkerr.pl (which implements 'make errors'), preserve
16426 reason strings from the previous version of the .c file, as
16427 the default to have only downcase letters (and digits) in
16428 automatically generated reasons codes is not always appropriate.
16432 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16433 using strerror. Previously, ERR_reason_error_string() returned
16434 library names as reason strings for SYSerr; but SYSerr is a special
16435 case where small numbers are errno values, not library numbers.
16439 * Add '-dsaparam' option to 'openssl dhparam' application. This
16440 converts DSA parameters into DH parameters. (When creating parameters,
16441 DSA_generate_parameters is used.)
16445 * Include 'length' (recommended exponent length) in C code generated
16446 by 'openssl dhparam -C'.
16450 * The second argument to set_label in perlasm was already being used
16451 so couldn't be used as a "file scope" flag. Moved to third argument
16456 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16457 instead of RAND_bytes for encryption IVs and salts.
16461 * Include RAND_status() into RAND_METHOD instead of implementing
16462 it only for md_rand.c Otherwise replacing the PRNG by calling
16463 RAND_set_rand_method would be impossible.
16467 * Don't let DSA_generate_key() enter an infinite loop if the random
16468 number generation fails.
16472 * New 'rand' application for creating pseudo-random output.
16476 * Added configuration support for Linux/IA64
16478 *Rolf Haberrecker <rolf@suse.de>*
16480 * Assembler module support for Mingw32.
16484 * Shared library support for HPUX (in shlib/).
16486 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16488 * Shared library support for Solaris gcc.
16490 *Lutz Behnke <behnke@trustcenter.de>*
16492 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16494 * PKCS7_encrypt() was adding text MIME headers twice because they
16495 were added manually and by SMIME_crlf_copy().
16499 * In bntest.c don't call BN_rand with zero bits argument.
16501 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16503 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16504 case was implemented. This caused BN_div_recp() to fail occasionally.
16508 * Add an optional second argument to the set_label() in the perl
16509 assembly language builder. If this argument exists and is set
16510 to 1 it signals that the assembler should use a symbol whose
16511 scope is the entire file, not just the current function. This
16512 is needed with MASM which uses the format label:: for this scope.
16514 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16516 * Change the ASN1 types so they are typedefs by default. Before
16517 almost all types were #define'd to ASN1_STRING which was causing
16518 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16523 * Change names of new functions to the new get1/get0 naming
16524 convention: After 'get1', the caller owns a reference count
16525 and has to call `..._free`; 'get0' returns a pointer to some
16526 data structure without incrementing reference counters.
16527 (Some of the existing 'get' functions increment a reference
16528 counter, some don't.)
16529 Similarly, 'set1' and 'add1' functions increase reference
16530 counters or duplicate objects.
16534 * Allow for the possibility of temp RSA key generation failure:
16535 the code used to assume it always worked and crashed on failure.
16539 * Fix potential buffer overrun problem in BIO_printf().
16540 *Ulf Möller, using public domain code by Patrick Powell; problem
16541 pointed out by David Sacerdote <das33@cornell.edu>*
16543 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16544 RAND_egd() and RAND_status(). In the command line application,
16545 the EGD socket can be specified like a seed file using RANDFILE
16550 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16551 Some CAs (e.g. Verisign) distribute certificates in this form.
16555 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16556 list to exclude them. This means that no special compilation option
16557 is needed to use anonymous DH: it just needs to be included in the
16562 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16563 EVP_MD_type. The old functionality is available in a new macro called
16564 EVP_MD_md(). Change code that uses it and update docs.
16568 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16569 where the `void *` argument is replaced by a function pointer argument.
16570 Previously `void *` was abused to point to functions, which works on
16571 many platforms, but is not correct. As these functions are usually
16572 called by macros defined in OpenSSL header files, most source code
16573 should work without changes.
16577 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16578 sections with information on -D... compiler switches used for
16579 compiling the library so that applications can see them. To enable
16580 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16581 must be defined. E.g.,
16582 #define OPENSSL_ALGORITHM_DEFINES
16583 #include <openssl/opensslconf.h>
16584 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16586 *Richard Levitte, Ulf and Bodo Möller*
16588 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16593 * Change the 'other' type in certificate aux info to a STACK_OF
16594 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16595 the required ASN1 format: arbitrary types determined by an OID.
16599 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16600 argument to 'req'. This is not because the function is newer or
16601 better than others it just uses the work 'NEW' in the certificate
16602 request header lines. Some software needs this.
16606 * Reorganise password command line arguments: now passwords can be
16607 obtained from various sources. Delete the PEM_cb function and make
16608 it the default behaviour: i.e. if the callback is NULL and the
16609 usrdata argument is not NULL interpret it as a null terminated pass
16610 phrase. If usrdata and the callback are NULL then the pass phrase
16611 is prompted for as usual.
16615 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16616 the support is automatically enabled. The resulting binaries will
16617 autodetect the card and use it if present.
16619 *Ben Laurie and Compaq Inc.*
16621 * Work around for Netscape hang bug. This sends certificate request
16622 and server done in one record. Since this is perfectly legal in the
16623 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16624 the bugs/SSLv3 entry for more info.
16628 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16632 * Add -rand argument to smime and pkcs12 applications and read/write
16637 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16641 * Add command line password options to the remaining applications.
16645 * Bug fix for BN_div_recp() for numerators with an even number of
16650 * More tests in bntest.c, and changed test_bn output.
16654 * ./config recognizes MacOS X now.
16658 * Bug fix for BN_div() when the first words of num and divisor are
16659 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16663 * Add support for various broken PKCS#8 formats, and command line
16664 options to produce them.
16668 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16669 get temporary BIGNUMs from a BN_CTX.
16673 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16678 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16679 include a #define from the old name to the new. The original intent
16680 was that statically linked binaries could for example just call
16681 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16682 link with digests. This never worked because SSLeay_add_all_digests()
16683 and SSLeay_add_all_ciphers() were in the same source file so calling
16684 one would link with the other. They are now in separate source files.
16688 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16692 * Use a less unusual form of the Miller-Rabin primality test (it used
16693 a binary algorithm for exponentiation integrated into the Miller-Rabin
16694 loop, our standard modexp algorithms are faster).
16698 * Support for the EBCDIC character set completed.
16700 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16702 * Source code cleanups: use const where appropriate, eliminate casts,
16703 use `void *` instead of `char *` in lhash.
16707 * Bugfix: ssl3_send_server_key_exchange was not restartable
16708 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16709 this the server could overwrite ephemeral keys that the client
16714 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16715 using 50 iterations of the Rabin-Miller test.
16717 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16718 iterations of the Rabin-Miller test as required by the appendix
16719 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16720 As BN_is_prime_fasttest includes trial division, DSA parameter
16721 generation becomes much faster.
16723 This implies a change for the callback functions in DSA_is_prime
16724 and DSA_generate_parameters: The callback function is called once
16725 for each positive witness in the Rabin-Miller test, not just
16726 occasionally in the inner loop; and the parameters to the
16727 callback function now provide an iteration count for the outer
16728 loop rather than for the current invocation of the inner loop.
16729 DSA_generate_parameters additionally can call the callback
16730 function with an 'iteration count' of -1, meaning that a
16731 candidate has passed the trial division test (when q is generated
16732 from an application-provided seed, trial division is skipped).
16736 * New function BN_is_prime_fasttest that optionally does trial
16737 division before starting the Rabin-Miller test and has
16738 an additional BN_CTX * argument (whereas BN_is_prime always
16739 has to allocate at least one BN_CTX).
16740 'callback(1, -1, cb_arg)' is called when a number has passed the
16741 trial division stage.
16745 * Fix for bug in CRL encoding. The validity dates weren't being handled
16750 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16754 * New function BN_pseudo_rand().
16758 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16759 bignum version of BN_from_montgomery() with the working code from
16760 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16765 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16766 made it impossible to use the same SSL_SESSION data structure in
16767 SSL2 clients in multiple threads.
16771 * The return value of RAND_load_file() no longer counts bytes obtained
16772 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16773 to seed the PRNG (previously an explicit byte count was required).
16775 *Ulf Möller, Bodo Möller*
16777 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16778 used `char *` instead of `void *` and had casts all over the place.
16782 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16786 * Retain source code compatibility for BN_prime_checks macro:
16787 BN_is_prime(..., BN_prime_checks, ...) now uses
16788 BN_prime_checks_for_size to determine the appropriate number of
16789 Rabin-Miller iterations.
16793 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16794 DH_CHECK_P_NOT_SAFE_PRIME.
16795 (Check if this is true? OpenPGP calls them "strong".)
16799 * Merge the functionality of "dh" and "gendh" programs into a new program
16800 "dhparam". The old programs are retained for now but will handle DH keys
16801 (instead of parameters) in future.
16805 * Make the ciphers, s_server and s_client programs check the return values
16806 when a new cipher list is set.
16810 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16811 ciphers. Before when the 56bit ciphers were enabled the sorting was
16814 The syntax for the cipher sorting has been extended to support sorting by
16815 cipher-strength (using the strength_bits hard coded in the tables).
16816 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16818 Fix a bug in the cipher-command parser: when supplying a cipher command
16819 string with an "undefined" symbol (neither command nor alphanumeric
16820 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16821 an error is flagged.
16823 Due to the strength-sorting extension, the code of the
16824 ssl_create_cipher_list() function was completely rearranged. I hope that
16825 the readability was also increased :-)
16827 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16829 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16830 for the first serial number and places 2 in the serial number file. This
16831 avoids problems when the root CA is created with serial number zero and
16832 the first user certificate has the same issuer name and serial number
16837 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16838 the new code. Add documentation for this stuff.
16842 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16843 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16844 structures and behave in an analogous way to the X509v3 functions:
16845 they shouldn't be called directly but wrapper functions should be used
16848 So we also now have some wrapper functions that call the X509at functions
16849 when passed certificate requests. (TO DO: similar things can be done with
16850 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16851 things. Some of these need some d2i or i2d and print functionality
16852 because they handle more complex structures.)
16856 * Add missing #ifndefs that caused missing symbols when building libssl
16857 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16858 NO_RSA in `ssl/s2*.c`.
16860 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16862 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16863 has a return value which indicates the quality of the random data
16864 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16865 error queue. New function RAND_pseudo_bytes() generates output that is
16866 guaranteed to be unique but not unpredictable. RAND_add is like
16867 RAND_seed, but takes an extra argument for an entropy estimate
16868 (RAND_seed always assumes full entropy).
16872 * Do more iterations of Rabin-Miller probable prime test (specifically,
16873 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16874 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16875 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16876 false-positive rate of at most 2^-80 for random input.
16880 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16884 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16885 in the 0.9.5 release), this returns the chain
16886 from an X509_CTX structure with a dup of the stack and all
16887 the X509 reference counts upped: so the stack will exist
16888 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16891 Also make SSL_SESSION_print() print out the verify return
16896 * Add manpage for the pkcs12 command. Also change the default
16897 behaviour so MAC iteration counts are used unless the new
16898 -nomaciter option is used. This improves file security and
16899 only older versions of MSIE (4.0 for example) need it.
16903 * Honor the no-xxx Configure options when creating .DEF files.
16907 * Add PKCS#10 attributes to field table: challengePassword,
16908 unstructuredName and unstructuredAddress. These are taken from
16909 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16910 international characters are used.
16912 More changes to X509_ATTRIBUTE code: allow the setting of types
16913 based on strings. Remove the 'loc' parameter when adding
16914 attributes because these will be a SET OF encoding which is sorted
16919 * Initial changes to the 'req' utility to allow request generation
16920 automation. This will allow an application to just generate a template
16921 file containing all the field values and have req construct the
16924 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16925 used all over the place including certificate requests and PKCS#7
16926 structures. They are currently handled manually where necessary with
16927 some primitive wrappers for PKCS#7. The new functions behave in a
16928 manner analogous to the X509 extension functions: they allow
16929 attributes to be looked up by NID and added.
16931 Later something similar to the X509V3 code would be desirable to
16932 automatically handle the encoding, decoding and printing of the
16933 more complex types. The string types like challengePassword can
16934 be handled by the string table functions.
16936 Also modified the multi byte string table handling. Now there is
16937 a 'global mask' which masks out certain types. The table itself
16938 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16939 is useful when for example there is only one permissible type
16940 (as in countryName) and using the mask might result in no valid
16945 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16946 SSL_get_peer_finished to allow applications to obtain the latest
16947 Finished messages sent to the peer or expected from the peer,
16948 respectively. (SSL_get_peer_finished is usually the Finished message
16949 actually received from the peer, otherwise the protocol will be aborted.)
16951 As the Finished message are message digests of the complete handshake
16952 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16953 be used for external authentication procedures when the authentication
16954 provided by SSL/TLS is not desired or is not enough.
16958 * Enhanced support for Alpha Linux is added. Now ./config checks if
16959 the host supports BWX extension and if Compaq C is present on the
16960 $PATH. Just exploiting of the BWX extension results in 20-30%
16961 performance kick for some algorithms, e.g. DES and RC4 to mention
16962 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16967 * Add support for MS "fast SGC". This is arguably a violation of the
16968 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16969 weak crypto and after checking the certificate is SGC a second one
16970 with strong crypto. MS SGC stops the first handshake after receiving
16971 the server certificate message and sends a second client hello. Since
16972 a server will typically do all the time consuming operations before
16973 expecting any further messages from the client (server key exchange
16974 is the most expensive) there is little difference between the two.
16976 To get OpenSSL to support MS SGC we have to permit a second client
16977 hello message after we have sent server done. In addition we have to
16978 reset the MAC if we do get this second client hello.
16982 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16983 if a DER encoded private key is RSA or DSA traditional format. Changed
16984 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16985 format DER encoded private key. Newer code should use PKCS#8 format which
16986 has the key type encoded in the ASN1 structure. Added DER private key
16987 support to pkcs8 application.
16991 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16992 ciphersuites has been selected (as required by the SSL 3/TLS 1
16993 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16994 is set, we interpret this as a request to violate the specification
16995 (the worst that can happen is a handshake failure, and 'correct'
16996 behaviour would result in a handshake failure anyway).
17000 * In SSL_CTX_add_session, take into account that there might be multiple
17001 SSL_SESSION structures with the same session ID (e.g. when two threads
17002 concurrently obtain them from an external cache).
17003 The internal cache can handle only one SSL_SESSION with a given ID,
17004 so if there's a conflict, we now throw out the old one to achieve
17009 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17010 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17011 some routines that use cipher OIDs: some ciphers do not have OIDs
17012 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17017 * Simplify the trust setting structure and code. Now we just have
17018 two sequences of OIDs for trusted and rejected settings. These will
17019 typically have values the same as the extended key usage extension
17020 and any application specific purposes.
17022 The trust checking code now has a default behaviour: it will just
17023 check for an object with the same NID as the passed id. Functions can
17024 be provided to override either the default behaviour or the behaviour
17025 for a given id. SSL client, server and email already have functions
17026 in place for compatibility: they check the NID and also return "trusted"
17027 if the certificate is self signed.
17031 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17032 traditional format into an EVP_PKEY structure.
17036 * Add a password callback function PEM_cb() which either prompts for
17037 a password if usr_data is NULL or otherwise assumes it is a null
17038 terminated password. Allow passwords to be passed on command line
17039 environment or config files in a few more utilities.
17043 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17044 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17045 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17046 Update documentation.
17050 * Support for ASN1 "NULL" type. This could be handled before by using
17051 ASN1_TYPE but there wasn't any function that would try to read a NULL
17052 and produce an error if it couldn't. For compatibility we also have
17053 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17054 don't allocate anything because they don't need to.
17058 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17061 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17063 * Rebuild of the memory allocation routines used by OpenSSL code and
17064 possibly others as well. The purpose is to make an interface that
17065 provide hooks so anyone can build a separate set of allocation and
17066 deallocation routines to be used by OpenSSL, for example memory
17067 pool implementations, or something else, which was previously hard
17068 since Malloc(), Realloc() and Free() were defined as macros having
17069 the values malloc, realloc and free, respectively (except for Win32
17070 compilations). The same is provided for memory debugging code.
17071 OpenSSL already comes with functionality to find memory leaks, but
17072 this gives people a chance to debug other memory problems.
17074 With these changes, a new set of functions and macros have appeared:
17076 CRYPTO_set_mem_debug_functions() [F]
17077 CRYPTO_get_mem_debug_functions() [F]
17078 CRYPTO_dbg_set_options() [F]
17079 CRYPTO_dbg_get_options() [F]
17080 CRYPTO_malloc_debug_init() [M]
17082 The memory debug functions are NULL by default, unless the library
17083 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17084 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17085 gives the standard debugging functions that come with OpenSSL) or
17086 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17087 provided by the library user) must be used. When the standard
17088 debugging functions are used, CRYPTO_dbg_set_options can be used to
17089 request additional information:
17090 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17091 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17093 Also, things like CRYPTO_set_mem_functions will always give the
17094 expected result (the new set of functions is used for allocation
17095 and deallocation) at all times, regardless of platform and compiler
17098 To finish it up, some functions that were never use in any other
17099 way than through macros have a new API and new semantic:
17101 CRYPTO_dbg_malloc()
17102 CRYPTO_dbg_realloc()
17105 All macros of value have retained their old syntax.
17107 *Richard Levitte and Bodo Moeller*
17109 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17110 ordering of SMIMECapabilities wasn't in "strength order" and there
17111 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17116 * Some ASN1 types with illegal zero length encoding (INTEGER,
17117 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17119 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17121 * Merge in my S/MIME library for OpenSSL. This provides a simple
17122 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17123 functionality to handle multipart/signed properly) and a utility
17124 called 'smime' to call all this stuff. This is based on code I
17125 originally wrote for Celo who have kindly allowed it to be
17126 included in OpenSSL.
17130 * Add variants des_set_key_checked and des_set_key_unchecked of
17131 des_set_key (aka des_key_sched). Global variable des_check_key
17132 decides which of these is called by des_set_key; this way
17133 des_check_key behaves as it always did, but applications and
17134 the library itself, which was buggy for des_check_key == 1,
17135 have a cleaner way to pick the version they need.
17139 * New function PKCS12_newpass() which changes the password of a
17144 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17145 dynamic mix. In both cases the ids can be used as an index into the
17146 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17147 functions so they accept a list of the field values and the
17148 application doesn't need to directly manipulate the X509_TRUST
17153 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17158 * Modify the way the V3 extension code looks up extensions. This now
17159 works in a similar way to the object code: we have some "standard"
17160 extensions in a static table which is searched with OBJ_bsearch()
17161 and the application can add dynamic ones if needed. The file
17162 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17163 updated whenever a new extension is added to the core code and kept
17164 in ext_nid order. There is a simple program 'tabtest.c' which checks
17165 this. New extensions are not added too often so this file can readily
17166 be maintained manually.
17168 There are two big advantages in doing things this way. The extensions
17169 can be looked up immediately and no longer need to be "added" using
17170 X509V3_add_standard_extensions(): this function now does nothing.
17171 Side note: I get *lots* of email saying the extension code doesn't
17172 work because people forget to call this function.
17173 Also no dynamic allocation is done unless new extensions are added:
17174 so if we don't add custom extensions there is no need to call
17175 X509V3_EXT_cleanup().
17179 * Modify enc utility's salting as follows: make salting the default. Add a
17180 magic header, so unsalted files fail gracefully instead of just decrypting
17181 to garbage. This is because not salting is a big security hole, so people
17182 should be discouraged from doing it.
17186 * Fixes and enhancements to the 'x509' utility. It allowed a message
17187 digest to be passed on the command line but it only used this
17188 parameter when signing a certificate. Modified so all relevant
17189 operations are affected by the digest parameter including the
17190 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17191 DSA key was used because it didn't fix the digest.
17195 * Initial certificate chain verify code. Currently tests the untrusted
17196 certificates for consistency with the verify purpose (which is set
17197 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17199 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17200 this is because it will reject chains with invalid extensions whereas
17201 every previous version of OpenSSL and SSLeay made no checks at all.
17203 Trust code: checks the root CA for the relevant trust settings. Trust
17204 settings have an initial value consistent with the verify purpose: e.g.
17205 if the verify purpose is for SSL client use it expects the CA to be
17206 trusted for SSL client use. However the default value can be changed to
17207 permit custom trust settings: one example of this would be to only trust
17208 certificates from a specific "secure" set of CAs.
17210 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17211 which should be used for version portability: especially since the
17212 verify structure is likely to change more often now.
17214 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17215 to set them. If not set then assume SSL clients will verify SSL servers
17218 Two new options to the verify program: -untrusted allows a set of
17219 untrusted certificates to be passed in and -purpose which sets the
17220 intended purpose of the certificate. If a purpose is set then the
17221 new chain verify code is used to check extension consistency.
17225 * Support for the authority information access extension.
17229 * Modify RSA and DSA PEM read routines to transparently handle
17230 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17231 public keys in a format compatible with certificate
17232 SubjectPublicKeyInfo structures. Unfortunately there were already
17233 functions called *_PublicKey_* which used various odd formats so
17234 these are retained for compatibility: however the DSA variants were
17235 never in a public release so they have been deleted. Changed dsa/rsa
17236 utilities to handle the new format: note no releases ever handled public
17237 keys so we should be OK.
17239 The primary motivation for this change is to avoid the same fiasco
17240 that dogs private keys: there are several incompatible private key
17241 formats some of which are standard and some OpenSSL specific and
17242 require various evil hacks to allow partial transparent handling and
17243 even then it doesn't work with DER formats. Given the option anything
17244 other than PKCS#8 should be dumped: but the other formats have to
17245 stay in the name of compatibility.
17247 With public keys and the benefit of hindsight one standard format
17248 is used which works with EVP_PKEY, RSA or DSA structures: though
17249 it clearly returns an error if you try to read the wrong kind of key.
17251 Added a -pubkey option to the 'x509' utility to output the public key.
17252 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17253 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17254 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17255 that do the same as the `EVP_PKEY_assign_*()` except they up the
17256 reference count of the added key (they don't "swallow" the
17261 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17262 CRLs would fail if the file contained no certificates or no CRLs:
17263 added a new function to read in both types and return the number
17264 read: this means that if none are read it will be an error. The
17265 DER versions of the certificate and CRL reader would always fail
17266 because it isn't possible to mix certificates and CRLs in DER format
17267 without choking one or the other routine. Changed this to just read
17268 a certificate: this is the best we can do. Also modified the code
17269 in `apps/verify.c` to take notice of return codes: it was previously
17270 attempting to read in certificates from NULL pointers and ignoring
17271 any errors: this is one reason why the cert and CRL reader seemed
17272 to work. It doesn't check return codes from the default certificate
17273 routines: these may well fail if the certificates aren't installed.
17277 * Code to support otherName option in GeneralName.
17281 * First update to verify code. Change the verify utility
17282 so it warns if it is passed a self signed certificate:
17283 for consistency with the normal behaviour. X509_verify
17284 has been modified to it will now verify a self signed
17285 certificate if *exactly* the same certificate appears
17286 in the store: it was previously impossible to trust a
17287 single self signed certificate. This means that:
17288 openssl verify ss.pem
17289 now gives a warning about a self signed certificate but
17290 openssl verify -CAfile ss.pem ss.pem
17295 * For servers, store verify_result in SSL_SESSION data structure
17296 (and add it to external session representation).
17297 This is needed when client certificate verifications fails,
17298 but an application-provided verification callback (set by
17299 SSL_CTX_set_cert_verify_callback) allows accepting the session
17300 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17301 but returns 1): When the session is reused, we have to set
17302 ssl->verify_result to the appropriate error code to avoid
17305 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17307 * Fix a bug in the new PKCS#7 code: it didn't consider the
17308 case in PKCS7_dataInit() where the signed PKCS7 structure
17309 didn't contain any existing data because it was being created.
17311 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17313 * Add a salt to the key derivation routines in enc.c. This
17314 forms the first 8 bytes of the encrypted file. Also add a
17315 -S option to allow a salt to be input on the command line.
17319 * New function X509_cmp(). Oddly enough there wasn't a function
17320 to compare two certificates. We do this by working out the SHA1
17321 hash and comparing that. X509_cmp() will be needed by the trust
17326 * SSL_get1_session() is like SSL_get_session(), but increments
17327 the reference count in the SSL_SESSION returned.
17329 *Geoff Thorpe <geoff@eu.c2.net>*
17331 * Fix for 'req': it was adding a null to request attributes.
17332 Also change the X509_LOOKUP and X509_INFO code to handle
17333 certificate auxiliary information.
17337 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17342 * Add the possibility to add extra information to the memory leak
17343 detecting output, to form tracebacks, showing from where each
17344 allocation was originated: CRYPTO_push_info("constant string") adds
17345 the string plus current file name and line number to a per-thread
17346 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17347 is like calling CYRPTO_pop_info() until the stack is empty.
17348 Also updated memory leak detection code to be multi-thread-safe.
17352 * Add options -text and -noout to pkcs7 utility and delete the
17353 encryption options which never did anything. Update docs.
17357 * Add options to some of the utilities to allow the pass phrase
17358 to be included on either the command line (not recommended on
17359 OSes like Unix) or read from the environment. Update the
17360 manpages and fix a few bugs.
17364 * Add a few manpages for some of the openssl commands.
17368 * Fix the -revoke option in ca. It was freeing up memory twice,
17369 leaking and not finding already revoked certificates.
17373 * Extensive changes to support certificate auxiliary information.
17374 This involves the use of X509_CERT_AUX structure and X509_AUX
17375 functions. An X509_AUX function such as PEM_read_X509_AUX()
17376 can still read in a certificate file in the usual way but it
17377 will also read in any additional "auxiliary information". By
17378 doing things this way a fair degree of compatibility can be
17379 retained: existing certificates can have this information added
17380 using the new 'x509' options.
17382 Current auxiliary information includes an "alias" and some trust
17383 settings. The trust settings will ultimately be used in enhanced
17384 certificate chain verification routines: currently a certificate
17385 can only be trusted if it is self signed and then it is trusted
17390 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17391 The problem was that one of the replacement routines had not been working
17392 since SSLeay releases. For now the offending routine has been replaced
17393 with non-optimised assembler. Even so, this now gives around 95%
17394 performance improvement for 1024 bit RSA signs.
17398 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17399 handling. Most clients have the effective key size in bits equal to
17400 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17401 A few however don't do this and instead use the size of the decrypted key
17402 to determine the RC2 key length and the AlgorithmIdentifier to determine
17403 the effective key length. In this case the effective key length can still
17404 be 40 bits but the key length can be 168 bits for example. This is fixed
17405 by manually forcing an RC2 key into the EVP_PKEY structure because the
17406 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17407 the key length and effective key length are equal.
17411 * Add a bunch of functions that should simplify the creation of
17412 X509_NAME structures. Now you should be able to do:
17413 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17414 and have it automatically work out the correct field type and fill in
17415 the structures. The more adventurous can try:
17416 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17417 and it will (hopefully) work out the correct multibyte encoding.
17421 * Change the 'req' utility to use the new field handling and multibyte
17422 copy routines. Before the DN field creation was handled in an ad hoc
17423 way in req, ca, and x509 which was rather broken and didn't support
17424 BMPStrings or UTF8Strings. Since some software doesn't implement
17425 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17426 using the dirstring_type option. See the new comment in the default
17427 openssl.cnf for more info.
17431 * Make crypto/rand/md_rand.c more robust:
17432 - Assure unique random numbers after fork().
17433 - Make sure that concurrent threads access the global counter and
17434 md serializably so that we never lose entropy in them
17435 or use exactly the same state in multiple threads.
17436 Access to the large state is not always serializable because
17437 the additional locking could be a performance killer, and
17438 md should be large enough anyway.
17442 * New file `apps/app_rand.c` with commonly needed functionality
17443 for handling the random seed file.
17445 Use the random seed file in some applications that previously did not:
17447 dsaparam -genkey (which also ignored its '-rand' option),
17450 x509 (when signing).
17451 Except on systems with /dev/urandom, it is crucial to have a random
17452 seed file at least for key creation, DSA signing, and for DH exchanges;
17453 for RSA signatures we could do without one.
17455 gendh and gendsa (unlike genrsa) used to read only the first byte
17456 of each file listed in the '-rand' option. The function as previously
17457 found in genrsa is now in app_rand.c and is used by all programs
17458 that support '-rand'.
17462 * In RAND_write_file, use mode 0600 for creating files;
17463 don't just chmod when it may be too late.
17467 * Report an error from X509_STORE_load_locations
17468 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17472 * New function ASN1_mbstring_copy() this copies a string in either
17473 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17474 into an ASN1_STRING type. A mask of permissible types is passed
17475 and it chooses the "minimal" type to use or an error if not type
17480 * Add function equivalents to the various macros in asn1.h. The old
17481 macros are retained with an `M_` prefix. Code inside the library can
17482 use the `M_` macros. External code (including the openssl utility)
17483 should *NOT* in order to be "shared library friendly".
17487 * Add various functions that can check a certificate's extensions
17488 to see if it usable for various purposes such as SSL client,
17489 server or S/MIME and CAs of these types. This is currently
17490 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17491 verification. Also added a -purpose flag to x509 utility to
17492 print out all the purposes.
17496 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17501 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17502 for, obtain and decode and extension and obtain its critical flag.
17503 This allows all the necessary extension code to be handled in a
17504 single function call.
17508 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17509 platforms. See crypto/rc4/rc4_enc.c for further details.
17513 * New -noout option to asn1parse. This causes no output to be produced
17514 its main use is when combined with -strparse and -out to extract data
17515 from a file (which may not be in ASN.1 format).
17519 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17520 when producing the local key id.
17522 *Richard Levitte <levitte@stacken.kth.se>*
17524 * New option -dhparam in s_server. This allows a DH parameter file to be
17525 stated explicitly. If it is not stated then it tries the first server
17526 certificate file. The previous behaviour hard coded the filename
17531 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17532 a public key to be input or output. For example:
17533 openssl rsa -in key.pem -pubout -out pubkey.pem
17534 Also added necessary DSA public key functions to handle this.
17538 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17539 in the message. This was handled by allowing
17540 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17542 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17544 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17545 to the end of the strings whereas this didn't. This would cause problems
17546 if strings read with d2i_ASN1_bytes() were later modified.
17548 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17550 * Fix for base64 decode bug. When a base64 bio reads only one line of
17551 data and it contains EOF it will end up returning an error. This is
17552 caused by input 46 bytes long. The cause is due to the way base64
17553 BIOs find the start of base64 encoded data. They do this by trying a
17554 trial decode on each line until they find one that works. When they
17555 do a flag is set and it starts again knowing it can pass all the
17556 data directly through the decoder. Unfortunately it doesn't reset
17557 the context it uses. This means that if EOF is reached an attempt
17558 is made to pass two EOFs through the context and this causes the
17559 resulting error. This can also cause other problems as well. As is
17560 usual with these problems it takes *ages* to find and the fix is
17561 trivial: move one line.
17563 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17565 * Ugly workaround to get s_client and s_server working under Windows. The
17566 old code wouldn't work because it needed to select() on sockets and the
17567 tty (for keypresses and to see if data could be written). Win32 only
17568 supports select() on sockets so we select() with a 1s timeout on the
17569 sockets and then see if any characters are waiting to be read, if none
17570 are present then we retry, we also assume we can always write data to
17571 the tty. This isn't nice because the code then blocks until we've
17572 received a complete line of data and it is effectively polling the
17573 keyboard at 1s intervals: however it's quite a bit better than not
17574 working at all :-) A dedicated Windows application might handle this
17575 with an event loop for example.
17579 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17580 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17581 will be called when RSA_sign() and RSA_verify() are used. This is useful
17582 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17583 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17584 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17585 This necessitated the support of an extra signature type NID_md5_sha1
17586 for SSL signatures and modifications to the SSL library to use it instead
17587 of calling RSA_public_decrypt() and RSA_private_encrypt().
17591 * Add new -verify -CAfile and -CApath options to the crl program, these
17592 will lookup a CRL issuers certificate and verify the signature in a
17593 similar way to the verify program. Tidy up the crl program so it
17594 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17595 less strict. It will now permit CRL extensions even if it is not
17596 a V2 CRL: this will allow it to tolerate some broken CRLs.
17600 * Initialize all non-automatic variables each time one of the openssl
17601 sub-programs is started (this is necessary as they may be started
17602 multiple times from the "OpenSSL>" prompt).
17604 *Lennart Bang, Bodo Moeller*
17606 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17607 removing all other RSA functionality (this is what NO_RSA does). This
17608 is so (for example) those in the US can disable those operations covered
17609 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17614 * Non-copying interface to BIO pairs.
17615 (still largely untested)
17619 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17620 ASCII string. This was handled independently in various places before.
17624 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17625 UTF8 strings a character at a time.
17629 * Use client_version from client hello to select the protocol
17630 (s23_srvr.c) and for RSA client key exchange verification
17631 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17635 * Add various utility functions to handle SPKACs, these were previously
17636 handled by poking round in the structure internals. Added new function
17637 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17638 print, verify and generate SPKACs. Based on an original idea from
17639 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17643 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17647 * Allow the config file extension section to be overwritten on the
17648 command line. Based on an original idea from Massimiliano Pala
17649 <madwolf@comune.modena.it>. The new option is called -extensions
17650 and can be applied to ca, req and x509. Also -reqexts to override
17651 the request extensions in req and -crlexts to override the crl extensions
17656 * Add new feature to the SPKAC handling in ca. Now you can include
17657 the same field multiple times by preceding it by "XXXX." for example:
17660 this is the same syntax as used in the req config file.
17664 * Allow certificate extensions to be added to certificate requests. These
17665 are specified in a 'req_extensions' option of the req section of the
17666 config file. They can be printed out with the -text option to req but
17667 are otherwise ignored at present.
17671 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17672 data read consists of only the final block it would not decrypted because
17673 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17674 A misplaced 'break' also meant the decrypted final block might not be
17675 copied until the next read.
17679 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17680 a few extra parameters to the DH structure: these will be useful if
17681 for example we want the value of 'q' or implement X9.42 DH.
17685 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17686 provides hooks that allow the default DSA functions or functions on a
17687 "per key" basis to be replaced. This allows hardware acceleration and
17688 hardware key storage to be handled without major modification to the
17689 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17690 associated functions.
17694 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17695 as "read only": it can't be written to and the buffer it points to will
17696 not be freed. Reading from a read only BIO is much more efficient than
17697 a normal memory BIO. This was added because there are several times when
17698 an area of memory needs to be read from a BIO. The previous method was
17699 to create a memory BIO and write the data to it, this results in two
17700 copies of the data and an O(n^2) reading algorithm. There is a new
17701 function BIO_new_mem_buf() which creates a read only memory BIO from
17702 an area of memory. Also modified the PKCS#7 routines to use read only
17707 * Bugfix: ssl23_get_client_hello did not work properly when called in
17708 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17709 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17710 but a retry condition occurred while trying to read the rest.
17714 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17715 NID_pkcs7_encrypted by default: this was wrong since this should almost
17716 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17717 the encrypted data type: this is a more sensible place to put it and it
17718 allows the PKCS#12 code to be tidied up that duplicated this
17723 * Changed obj_dat.pl script so it takes its input and output files on
17724 the command line. This should avoid shell escape redirection problems
17729 * Initial support for certificate extension requests, these are included
17730 in things like Xenroll certificate requests. Included functions to allow
17731 extensions to be obtained and added.
17735 * -crlf option to s_client and s_server for sending newlines as
17736 CRLF (as required by many protocols).
17740 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17742 * Install libRSAglue.a when OpenSSL is built with RSAref.
17744 *Ralf S. Engelschall*
17746 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17748 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17750 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17755 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17756 DH parameters/keys (q is lost during that conversion, but the resulting
17757 DH parameters contain its length).
17759 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17760 much faster than DH_generate_parameters (which creates parameters
17761 where `p = 2*q + 1`), and also the smaller q makes DH computations
17762 much more efficient (160-bit exponentiation instead of 1024-bit
17763 exponentiation); so this provides a convenient way to support DHE
17764 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17765 utter importance to use
17766 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17768 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17769 when such DH parameters are used, because otherwise small subgroup
17770 attacks may become possible!
17774 * Avoid memory leak in i2d_DHparams.
17778 * Allow the -k option to be used more than once in the enc program:
17779 this allows the same encrypted message to be read by multiple recipients.
17783 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17784 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17785 it will always use the numerical form of the OID, even if it has a short
17790 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17791 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17792 otherwise bn_mod_exp was called. In the case of hardware keys for example
17793 no private key components need be present and it might store extra data
17794 in the RSA structure, which cannot be accessed from bn_mod_exp.
17795 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17796 private key operations.
17800 * Added support for SPARC Linux.
17804 * pem_password_cb function type incompatibly changed from
17805 typedef int pem_password_cb(char *buf, int size, int rwflag);
17807 ....(char *buf, int size, int rwflag, void *userdata);
17808 so that applications can pass data to their callbacks:
17809 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17810 additional void * argument, which is just handed through whenever
17811 the password callback is called.
17813 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17815 New function SSL_CTX_set_default_passwd_cb_userdata.
17817 Compatibility note: As many C implementations push function arguments
17818 onto the stack in reverse order, the new library version is likely to
17819 interoperate with programs that have been compiled with the old
17820 pem_password_cb definition (PEM_whatever takes some data that
17821 happens to be on the stack as its last argument, and the callback
17822 just ignores this garbage); but there is no guarantee whatsoever that
17825 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17826 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17827 problems not only on Windows, but also on some Unix platforms.
17828 To avoid problematic command lines, these definitions are now in an
17829 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17830 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17834 * MIPS III/IV assembler module is reimplemented.
17838 * More DES library cleanups: remove references to srand/rand and
17839 delete an unused file.
17843 * Add support for the free Netwide assembler (NASM) under Win32,
17844 since not many people have MASM (ml) and it can be hard to obtain.
17845 This is currently experimental but it seems to work OK and pass all
17846 the tests. Check out INSTALL.W32 for info.
17850 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17851 without temporary keys kept an extra copy of the server key,
17852 and connections with temporary keys did not free everything in case
17857 * New function RSA_check_key and new openssl rsa option -check
17858 for verifying the consistency of RSA keys.
17860 *Ulf Moeller, Bodo Moeller*
17862 * Various changes to make Win32 compile work:
17863 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17864 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17865 comparison" warnings.
17866 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17870 * Add a debugging option to PKCS#5 v2 key generation function: when
17871 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17872 derived keys are printed to stderr.
17876 * Copy the flags in ASN1_STRING_dup().
17878 *Roman E. Pavlov <pre@mo.msk.ru>*
17880 * The x509 application mishandled signing requests containing DSA
17881 keys when the signing key was also DSA and the parameters didn't match.
17883 It was supposed to omit the parameters when they matched the signing key:
17884 the verifying software was then supposed to automatically use the CA's
17885 parameters if they were absent from the end user certificate.
17887 Omitting parameters is no longer recommended. The test was also
17888 the wrong way round! This was probably due to unusual behaviour in
17889 EVP_cmp_parameters() which returns 1 if the parameters match.
17890 This meant that parameters were omitted when they *didn't* match and
17891 the certificate was useless. Certificates signed with 'ca' didn't have
17894 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17896 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17897 The interface is as follows:
17898 Applications can use
17899 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17900 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17901 "off" is now the default.
17902 The library internally uses
17903 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17904 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17905 to disable memory-checking temporarily.
17907 Some inconsistent states that previously were possible (and were
17908 even the default) are now avoided.
17910 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17911 with each memory chunk allocated; this is occasionally more helpful
17912 than just having a counter.
17914 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17916 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17921 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17922 which largely parallels "options", but is for changing API behaviour,
17923 whereas "options" are about protocol behaviour.
17924 Initial "mode" flags are:
17926 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17927 a single record has been written.
17928 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17929 retries use the same buffer location.
17930 (But all of the contents must be
17935 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17938 * Fix problems with no-hmac etc.
17940 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17942 * New functions RSA_get_default_method(), RSA_set_method() and
17943 RSA_get_method(). These allows replacement of RSA_METHODs without having
17944 to mess around with the internals of an RSA structure.
17948 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17949 Also really enable memory leak checks in openssl.c and in some
17952 *Chad C. Mulligan, Bodo Moeller*
17954 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17955 up the length of negative integers. This has now been simplified to just
17956 store the length when it is first determined and use it later, rather
17957 than trying to keep track of where data is copied and updating it to
17959 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17961 * Add a new function PKCS7_signatureVerify. This allows the verification
17962 of a PKCS#7 signature but with the signing certificate passed to the
17963 function itself. This contrasts with PKCS7_dataVerify which assumes the
17964 certificate is present in the PKCS#7 structure. This isn't always the
17965 case: certificates can be omitted from a PKCS#7 structure and be
17966 distributed by "out of band" means (such as a certificate database).
17970 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17971 function prototypes in pem.h, also change util/mkdef.pl to add the
17972 necessary function names.
17976 * mk1mf.pl (used by Windows builds) did not properly read the
17977 options set by Configure in the top level Makefile, and Configure
17978 was not even able to write more than one option correctly.
17979 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17983 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17984 file to be loaded from a BIO or FILE pointer. The BIO version will
17985 for example allow memory BIOs to contain config info.
17989 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17990 Whoever hopes to achieve shared-library compatibility across versions
17991 must use this, not the compile-time macro.
17992 (Exercise 0.9.4: Which is the minimum library version required by
17994 Note: All this applies only to multi-threaded programs, others don't
17999 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18000 through a BIO pair triggered the default case, i.e.
18001 SSLerr(...,SSL_R_UNKNOWN_STATE).
18005 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18006 can use the SSL library even if none of the specific BIOs is
18011 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18012 for the encoded length.
18014 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18016 * Add initial documentation of the X509V3 functions.
18020 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18021 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18022 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18023 secure PKCS#8 private key format with a high iteration count.
18027 * Fix determination of Perl interpreter: A perl or perl5
18028 *directory* in $PATH was also accepted as the interpreter.
18030 *Ralf S. Engelschall*
18032 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18033 wrong with it but it was very old and did things like calling
18034 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18035 unusual formatting.
18039 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18040 to use the new extension code.
18044 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18045 with macros. This should make it easier to change their form, add extra
18046 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18051 * Add to configuration table a new entry that can specify an alternative
18052 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18053 according to Mark Crispin <MRC@Panda.COM>.
18057 * DES CBC did not update the IV. Weird.
18061 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18062 Changing the behaviour of the former might break existing programs --
18063 where IV updating is needed, des_ncbc_encrypt can be used.
18066 * When bntest is run from "make test" it drives bc to check its
18067 calculations, as well as internally checking them. If an internal check
18068 fails, it needs to cause bc to give a non-zero result or make test carries
18069 on without noticing the failure. Fixed.
18073 * DES library cleanups.
18077 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18078 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18079 ciphers. NOTE: although the key derivation function has been verified
18080 against some published test vectors it has not been extensively tested
18081 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18086 * Instead of "mkdir -p", which is not fully portable, use new
18087 Perl script "util/mkdir-p.pl".
18091 * Rewrite the way password based encryption (PBE) is handled. It used to
18092 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18093 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18094 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18095 the 'parameter' field of the AlgorithmIdentifier is passed to the
18096 underlying key generation function so it must do its own ASN1 parsing.
18097 This has also changed the EVP_PBE_CipherInit() function which now has a
18098 'parameter' argument instead of literal salt and iteration count values
18099 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18103 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18104 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18105 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18106 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18107 value was just used as a "magic string" and not used directly its
18108 value doesn't matter.
18112 * Introduce some semblance of const correctness to BN. Shame C doesn't
18117 * "linux-sparc64" configuration (ultrapenguin).
18119 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18120 "linux-sparc" configuration.
18122 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18124 * config now generates no-xxx options for missing ciphers.
18128 * Support the EBCDIC character set (work in progress).
18129 File ebcdic.c not yet included because it has a different license.
18131 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18133 * Support BS2000/OSD-POSIX.
18135 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18137 * Make callbacks for key generation use `void *` instead of `char *`.
18141 * Make S/MIME samples compile (not yet tested).
18145 * Additional typesafe stacks.
18149 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18153 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18155 * New configuration variant "sco5-gcc".
18157 * Updated some demos.
18159 *Sean O Riordain, Wade Scholine*
18161 * Add missing BIO_free at exit of pkcs12 application.
18165 * Fix memory leak in conf.c.
18169 * Updates for Win32 to assembler version of MD5.
18173 * Set #! path to perl in `apps/der_chop` to where we found it
18174 instead of using a fixed path.
18178 * SHA library changes for irix64-mips4-cc.
18182 * Improvements for VMS support.
18186 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18188 * Bignum library bug fix. IRIX 6 passes "make test" now!
18189 This also avoids the problems with SC4.2 and unpatched SC5.
18191 *Andy Polyakov <appro@fy.chalmers.se>*
18193 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18194 These are required because of the typesafe stack would otherwise break
18195 existing code. If old code used a structure member which used to be STACK
18196 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18197 sk_num or sk_value it would produce an error because the num, data members
18198 are not present in STACK_OF. Now it just produces a warning. sk_set
18199 replaces the old method of assigning a value to sk_value
18200 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18201 that does this will no longer work (and should use sk_set instead) but
18202 this could be regarded as a "questionable" behaviour anyway.
18206 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18207 correctly handle encrypted S/MIME data.
18211 * Change type of various DES function arguments from des_cblock
18212 (which means, in function argument declarations, pointer to char)
18213 to des_cblock * (meaning pointer to array with 8 char elements),
18214 which allows the compiler to do more typechecking; it was like
18215 that back in SSLeay, but with lots of ugly casts.
18217 Introduce new type const_des_cblock.
18221 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18222 problems: find RecipientInfo structure that matches recipient certificate
18223 and initialise the ASN1 structures properly based on passed cipher.
18227 * Belatedly make the BN tests actually check the results.
18231 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18232 to and from BNs: it was completely broken. New compilation option
18233 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18234 key elements as negative integers.
18238 * Reorganize and speed up MD5.
18240 *Andy Polyakov <appro@fy.chalmers.se>*
18244 *Richard Levitte <richard@levitte.org>*
18246 * New option -out to asn1parse to allow the parsed structure to be
18247 output to a file. This is most useful when combined with the -strparse
18248 option to examine the output of things like OCTET STRINGS.
18252 * Make SSL library a little more fool-proof by not requiring any longer
18253 that `SSL_set_{accept,connect}_state` be called before
18254 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18255 in many applications because usually everything *appeared* to work as
18256 intended anyway -- now it really works as intended).
18260 * Move openssl.cnf out of lib/.
18264 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18265 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18266 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18268 *Ralf S. Engelschall*
18270 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18271 handle PKCS#7 enveloped data properly.
18273 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18275 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18276 copying pointers. The cert_st handling is changed by this in
18277 various ways (and thus what used to be known as ctx->default_cert
18278 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18279 any longer when s->cert does not give us what we need).
18280 ssl_cert_instantiate becomes obsolete by this change.
18281 As soon as we've got the new code right (possibly it already is?),
18282 we have solved a couple of bugs of the earlier code where s->cert
18283 was used as if it could not have been shared with other SSL structures.
18285 Note that using the SSL API in certain dirty ways now will result
18286 in different behaviour than observed with earlier library versions:
18287 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18288 does not influence s as it used to.
18290 In order to clean up things more thoroughly, inside SSL_SESSION
18291 we don't use CERT any longer, but a new structure SESS_CERT
18292 that holds per-session data (if available); currently, this is
18293 the peer's certificate chain and, for clients, the server's certificate
18294 and temporary key. CERT holds only those values that can have
18295 meaningful defaults in an SSL_CTX.
18299 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18300 from the internal representation. Various PKCS#7 fixes: remove some
18301 evil casts and set the enc_dig_alg field properly based on the signing
18306 * Allow PKCS#12 password to be set from the command line or the
18307 environment. Let 'ca' get its config file name from the environment
18308 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18313 * Allow certificate policies extension to use an IA5STRING for the
18314 organization field. This is contrary to the PKIX definition but
18315 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18320 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18321 without disallowing inline assembler and the like for non-pedantic builds.
18325 * Support Borland C++ builder.
18327 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18333 * SHA-1 cleanups and performance enhancements.
18335 *Andy Polyakov <appro@fy.chalmers.se>*
18337 * Sparc v8plus assembler for the bignum library.
18339 *Andy Polyakov <appro@fy.chalmers.se>*
18341 * Accept any -xxx and +xxx compiler options in Configure.
18345 * Update HPUX configuration.
18349 * Add missing `sk_<type>_unshift()` function to safestack.h
18351 *Ralf S. Engelschall*
18353 * New function SSL_CTX_use_certificate_chain_file that sets the
18354 "extra_cert"s in addition to the certificate. (This makes sense
18355 only for "PEM" format files, as chains as a whole are not
18360 * Support verify_depth from the SSL API.
18361 x509_vfy.c had what can be considered an off-by-one-error:
18362 Its depth (which was not part of the external interface)
18363 was actually counting the number of certificates in a chain;
18364 now it really counts the depth.
18368 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18369 instead of X509err, which often resulted in confusing error
18370 messages since the error codes are not globally unique
18371 (e.g. an alleged error in ssl3_accept when a certificate
18372 didn't match the private key).
18374 * New function SSL_CTX_set_session_id_context that allows to set a default
18375 value (so that you don't need SSL_set_session_id_context for each
18376 connection using the SSL_CTX).
18380 * OAEP decoding bug fix.
18384 * Support INSTALL_PREFIX for package builders, as proposed by
18389 * New Configure options "threads" and "no-threads". For systems
18390 where the proper compiler options are known (currently Solaris
18391 and Linux), "threads" is the default.
18395 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18399 * Install various scripts to $(OPENSSLDIR)/misc, not to
18400 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18401 such as /usr/local/bin.
18405 * "make linux-shared" to build shared libraries.
18407 *Niels Poppe <niels@netbox.org>*
18409 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18413 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18414 extension adding in x509 utility.
18418 * Remove NOPROTO sections and error code comments.
18422 * Partial rewrite of the DEF file generator to now parse the ANSI
18427 * New Configure options --prefix=DIR and --openssldir=DIR.
18431 * Complete rewrite of the error code script(s). It is all now handled
18432 by one script at the top level which handles error code gathering,
18433 header rewriting and C source file generation. It should be much better
18434 than the old method: it now uses a modified version of Ulf's parser to
18435 read the ANSI prototypes in all header files (thus the old K&R definitions
18436 aren't needed for error creation any more) and do a better job of
18437 translating function codes into names. The old 'ASN1 error code embedded
18438 in a comment' is no longer necessary and it doesn't use .err files which
18439 have now been deleted. Also the error code call doesn't have to appear all
18440 on one line (which resulted in some large lines...).
18444 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18448 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18449 0 (which usually indicates a closed connection), but continue reading.
18453 * Fix some race conditions.
18457 * Add support for CRL distribution points extension. Add Certificate
18458 Policies and CRL distribution points documentation.
18462 * Move the autogenerated header file parts to crypto/opensslconf.h.
18466 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18467 8 of keying material. Merlin has also confirmed interop with this fix
18468 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18470 *Merlin Hughes <merlin@baltimore.ie>*
18472 * Fix lots of warnings.
18474 *Richard Levitte <levitte@stacken.kth.se>*
18476 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18477 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18479 *Richard Levitte <levitte@stacken.kth.se>*
18481 * Fix problems with sizeof(long) == 8.
18483 *Andy Polyakov <appro@fy.chalmers.se>*
18485 * Change functions to ANSI C.
18489 * Fix typos in error codes.
18491 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18493 * Remove defunct assembler files from Configure.
18497 * SPARC v8 assembler BIGNUM implementation.
18499 *Andy Polyakov <appro@fy.chalmers.se>*
18501 * Support for Certificate Policies extension: both print and set.
18502 Various additions to support the r2i method this uses.
18506 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18507 return a const string when you are expecting an allocated buffer.
18511 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18512 types DirectoryString and DisplayText.
18516 * Add code to allow r2i extensions to access the configuration database,
18517 add an LHASH database driver and add several ctx helper functions.
18521 * Fix an evil bug in bn_expand2() which caused various BN functions to
18522 fail when they extended the size of a BIGNUM.
18526 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18527 support typesafe stack.
18531 * Fix typo in SSL_[gs]et_options().
18533 *Nils Frostberg <nils@medcom.se>*
18535 * Delete various functions and files that belonged to the (now obsolete)
18536 old X509V3 handling code.
18540 * New Configure option "rsaref".
18544 * Don't auto-generate pem.h.
18548 * Introduce type-safe ASN.1 SETs.
18552 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18554 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18556 * Introduce type-safe STACKs. This will almost certainly break lots of code
18557 that links with OpenSSL (well at least cause lots of warnings), but fear
18558 not: the conversion is trivial, and it eliminates loads of evil casts. A
18559 few STACKed things have been converted already. Feel free to convert more.
18560 In the fullness of time, I'll do away with the STACK type altogether.
18564 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18565 specified in `<certfile>` by updating the entry in the index.txt file.
18566 This way one no longer has to edit the index.txt file manually for
18567 revoking a certificate. The -revoke option does the gory details now.
18569 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18571 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18572 `-text` option at all and this way the `-noout -text` combination was
18573 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18575 *Ralf S. Engelschall*
18577 * Make sure a corresponding plain text error message exists for the
18578 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18579 verify callback function determined that a certificate was revoked.
18581 *Ralf S. Engelschall*
18583 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18584 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18585 all available ciphers including rc5, which was forgotten until now.
18586 In order to let the testing shell script know which algorithms
18587 are available, a new (up to now undocumented) command
18588 `openssl list-cipher-commands` is used.
18592 * Bugfix: s_client occasionally would sleep in select() when
18593 it should have checked SSL_pending() first.
18597 * New functions DSA_do_sign and DSA_do_verify to provide access to
18598 the raw DSA values prior to ASN.1 encoding.
18602 * Tweaks to Configure
18604 *Niels Poppe <niels@netbox.org>*
18606 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18611 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18615 * New config option to avoid instructions that are illegal on the 80386.
18616 The default code is faster, but requires at least a 486.
18620 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18621 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18622 same as SSL2_VERSION anyway.
18626 * New "-showcerts" option for s_client.
18630 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18631 application. Various cleanups and fixes.
18635 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18636 modify error routines to work internally. Add error codes and PBE init
18637 to library startup routines.
18641 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18642 packing functions to asn1 and evp. Changed function names and error
18643 codes along the way.
18647 * PKCS12 integration: and so it begins... First of several patches to
18648 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18649 objects to objects.h
18653 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18654 and display support for Thawte strong extranet extension.
18658 * Add LinuxPPC support.
18660 *Jeff Dubrule <igor@pobox.org>*
18662 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18663 bn_div_words in alpha.s.
18665 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18667 * Make sure the RSA OAEP test is skipped under -DRSAref because
18668 OAEP isn't supported when OpenSSL is built with RSAref.
18670 *Ulf Moeller <ulf@fitug.de>*
18672 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18673 so they no longer are missing under -DNOPROTO.
18675 *Soren S. Jorvang <soren@t.dk>*
18677 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18679 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18680 doesn't work when the session is reused. Coming soon!
18684 * Fix a security hole, that allows sessions to be reused in the wrong
18685 context thus bypassing client cert protection! All software that uses
18686 client certs and session caches in multiple contexts NEEDS PATCHING to
18687 allow session reuse! A fuller solution is in the works.
18689 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18691 * Some more source tree cleanups (removed obsolete files
18692 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18693 permission on "config" script to be executable) and a fix for the INSTALL
18696 *Ulf Moeller <ulf@fitug.de>*
18698 * Remove some legacy and erroneous uses of malloc, free instead of
18701 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18703 * Make rsa_oaep_test return non-zero on error.
18705 *Ulf Moeller <ulf@fitug.de>*
18707 * Add support for native Solaris shared libraries. Configure
18708 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18709 if someone would make that last step automatic.
18711 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18713 * ctx_size was not built with the right compiler during "make links". Fixed.
18717 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18718 except NULL ciphers". This means the default cipher list will no longer
18719 enable NULL ciphers. They need to be specifically enabled e.g. with
18720 the string "DEFAULT:eNULL".
18724 * Fix to RSA private encryption routines: if p < q then it would
18725 occasionally produce an invalid result. This will only happen with
18726 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18730 * Be less restrictive and allow also `perl util/perlpath.pl
18731 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18732 because this way one can also use an interpreter named `perl5` (which is
18733 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18734 installed as `perl`).
18736 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18738 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18740 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18742 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18743 advapi32.lib to Win32 build and change the pem test comparison
18744 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18745 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18746 and crypto/des/ede_cbcm_enc.c.
18750 * DES quad checksum was broken on big-endian architectures. Fixed.
18754 * Comment out two functions in bio.h that aren't implemented. Fix up the
18755 Win32 test batch file so it (might) work again. The Win32 test batch file
18756 is horrible: I feel ill....
18760 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18761 in e_os.h. Audit of header files to check ANSI and non ANSI
18762 sections: 10 functions were absent from non ANSI section and not exported
18763 from Windows DLLs. Fixed up libeay.num for new functions.
18767 * Make `openssl version` output lines consistent.
18769 *Ralf S. Engelschall*
18771 * Fix Win32 symbol export lists for BIO functions: Added
18772 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18773 to ms/libeay{16,32}.def.
18775 *Ralf S. Engelschall*
18777 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18778 fine under Unix and passes some trivial tests I've now added. But the
18779 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18780 added to make sure no one expects that this stuff really works in the
18781 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18782 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18785 *Ralf S. Engelschall*
18787 * Fix the generation of two part addresses in perl.
18789 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18791 * Add config entry for Linux on MIPS.
18793 *John Tobey <jtobey@channel1.com>*
18795 * Make links whenever Configure is run, unless we are on Windoze.
18799 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18800 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18805 * Add a useful kludge to allow package maintainers to specify compiler and
18806 other platforms details on the command line without having to patch the
18807 Configure script every time: One now can use
18808 `perl Configure <id>:<details>`,
18809 i.e. platform ids are allowed to have details appended
18810 to them (separated by colons). This is treated as there would be a static
18811 pre-configured entry in Configure's %table under key `<id>` with value
18812 `<details>` and `perl Configure <id>` is called. So, when you want to
18813 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18814 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18815 now, which overrides the FreeBSD-elf entry on-the-fly.
18817 *Ralf S. Engelschall*
18819 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18823 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18824 on the `perl Configure ...` command line. This way one can compile
18825 OpenSSL libraries with Position Independent Code (PIC) which is needed
18826 for linking it into DSOs.
18828 *Ralf S. Engelschall*
18830 * Remarkably, export ciphers were totally broken and no-one had noticed!
18835 * Cleaned up the LICENSE document: The official contact for any license
18836 questions now is the OpenSSL core team under openssl-core@openssl.org.
18837 And add a paragraph about the dual-license situation to make sure people
18838 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18839 to the OpenSSL toolkit.
18841 *Ralf S. Engelschall*
18843 * General source tree makefile cleanups: Made `making xxx in yyy...`
18844 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18845 Additionally cleaned up the `make links` target: Remove unnecessary
18846 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18847 to speed processing and no longer clutter the display with confusing
18848 stuff. Instead only the actually done links are displayed.
18850 *Ralf S. Engelschall*
18852 * Permit null encryption ciphersuites, used for authentication only. It used
18853 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18854 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18859 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18860 signed attributes when verifying signatures (this would break them),
18861 the detached data encoding was wrong and public keys obtained using
18862 X509_get_pubkey() weren't freed.
18866 * Add text documentation for the BUFFER functions. Also added a work around
18867 to a Win95 console bug. This was triggered by the password read stuff: the
18868 last character typed gets carried over to the next fread(). If you were
18869 generating a new cert request using 'req' for example then the last
18870 character of the passphrase would be CR which would then enter the first
18875 * Added the new 'Includes OpenSSL Cryptography Software' button as
18876 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18877 button and can be used by applications based on OpenSSL to show the
18878 relationship to the OpenSSL project.
18880 *Ralf S. Engelschall*
18882 * Remove confusing variables in function signatures in files
18883 ssl/ssl_lib.c and ssl/ssl.h.
18885 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18887 * Don't install bss_file.c under PREFIX/include/
18889 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18891 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18892 functions that return function pointers and has support for NT specific
18893 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18894 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18895 unsigned to signed types: this was killing the Win32 compile.
18899 * Add new certificate file to stack functions,
18900 SSL_add_dir_cert_subjects_to_stack() and
18901 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18902 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18903 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18904 This means that Apache-SSL and similar packages don't have to mess around
18905 to add as many CAs as they want to the preferred list.
18909 * Experiment with doxygen documentation. Currently only partially applied to
18911 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18912 openssl.doxy as the configuration file.
18916 * Get rid of remaining C++-style comments which strict C compilers hate.
18918 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18920 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18921 compiled in by default: it has problems with large keys.
18925 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18926 DH private keys and/or callback functions which directly correspond to
18927 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18928 is needed for applications which have to configure certificates on a
18929 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18931 For the RSA certificate situation is makes no difference, but
18932 for the DSA certificate situation this fixes the "no shared cipher"
18933 problem where the OpenSSL cipher selection procedure failed because the
18934 temporary keys were not overtaken from the context and the API provided
18935 no way to reconfigure them.
18936 The new functions now let applications reconfigure the stuff and they
18937 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18938 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18939 non-public-API function ssl_cert_instantiate() is used as a helper
18940 function and also to reduce code redundancy inside ssl_rsa.c.
18942 *Ralf S. Engelschall*
18944 * Move s_server -dcert and -dkey options out of the undocumented feature
18945 area because they are useful for the DSA situation and should be
18946 recognized by the users.
18948 *Ralf S. Engelschall*
18950 * Fix the cipher decision scheme for export ciphers: the export bits are
18951 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18952 SSL_EXP_MASK. So, the original variable has to be used instead of the
18953 already masked variable.
18955 *Richard Levitte <levitte@stacken.kth.se>*
18957 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18959 *Richard Levitte <levitte@stacken.kth.se>*
18961 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18962 from `int` to `unsigned int` because it is a length and initialized by
18963 EVP_DigestFinal() which expects an `unsigned int *`.
18965 *Richard Levitte <levitte@stacken.kth.se>*
18967 * Don't hard-code path to Perl interpreter on shebang line of Configure
18968 script. Instead use the usual Shell->Perl transition trick.
18970 *Ralf S. Engelschall*
18972 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18973 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18974 -noout -modulus` as it's already the case for `openssl rsa -noout
18975 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18976 currently the public key is printed (a decision which was already done by
18977 `openssl dsa -modulus` in the past) which serves a similar purpose.
18978 Additionally the NO_RSA no longer completely removes the whole -modulus
18979 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18982 *Ralf S. Engelschall*
18984 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18985 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18987 *Arne Ansper <arne@ats.cyber.ee>*
18989 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18990 to be added. Now both 'req' and 'ca' can use new objects defined in the
18995 * Add cool BIO that does syslog (or event log on NT).
18997 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18999 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19000 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19001 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19002 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19006 * Add preliminary config info for new extension code.
19010 * Make RSA_NO_PADDING really use no padding.
19012 *Ulf Moeller <ulf@fitug.de>*
19014 * Generate errors when private/public key check is done.
19018 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19019 for some CRL extensions and new objects added.
19023 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19024 key usage extension and fuller support for authority key id.
19028 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19029 padding method for RSA, which is recommended for new applications in PKCS
19030 #1 v2.0 (RFC 2437, October 1998).
19031 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19032 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19033 against Bleichbacher's attack on RSA.
19034 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19037 * Updates to the new SSL compression code
19039 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19041 * Fix so that the version number in the master secret, when passed
19042 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19043 (because the server will not accept higher), that the version number
19044 is 0x03,0x01, not 0x03,0x00
19046 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19048 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19049 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19050 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19054 * Support for RAW extensions where an arbitrary extension can be
19055 created by including its DER encoding. See `apps/openssl.cnf` for
19060 * Make sure latest Perl versions don't interpret some generated C array
19061 code as Perl array code in the crypto/err/err_genc.pl script.
19063 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19065 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19066 not many people have the assembler. Various Win32 compilation fixes and
19067 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19068 build instructions.
19072 * Modify configure script 'Configure' to automatically create crypto/date.h
19073 file under Win32 and also build pem.h from pem.org. New script
19074 util/mkfiles.pl to create the MINFO file on environments that can't do a
19075 'make files': perl util/mkfiles.pl >MINFO should work.
19079 * Major rework of DES function declarations, in the pursuit of correctness
19080 and purity. As a result, many evil casts evaporated, and some weirdness,
19081 too. You may find this causes warnings in your code. Zapping your evil
19082 casts will probably fix them. Mostly.
19086 * Fix for a typo in asn1.h. Bug fix to object creation script
19087 obj_dat.pl. It considered a zero in an object definition to mean
19088 "end of object": none of the objects in objects.h have any zeros
19089 so it wasn't spotted.
19091 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19093 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19094 Masking (CBCM). In the absence of test vectors, the best I have been able
19095 to do is check that the decrypt undoes the encrypt, so far. Send me test
19096 vectors if you have them.
19100 * Correct calculation of key length for export ciphers (too much space was
19101 allocated for null ciphers). This has not been tested!
19105 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19106 message is now correct (it understands "crypto" and "ssl" on its
19107 command line). There is also now an "update" option. This will update
19108 the util/ssleay.num and util/libeay.num files with any new functions.
19110 perl util/mkdef.pl crypto ssl update
19111 it will update them.
19115 * Overhauled the Perl interface:
19116 - ported BN stuff to OpenSSL's different BN library
19117 - made the perl/ source tree CVS-aware
19118 - renamed the package from SSLeay to OpenSSL (the files still contain
19119 their history because I've copied them in the repository)
19120 - removed obsolete files (the test scripts will be replaced
19121 by better Test::Harness variants in the future)
19123 *Ralf S. Engelschall*
19125 * First cut for a very conservative source tree cleanup:
19126 1. merge various obsolete readme texts into doc/ssleay.txt
19127 where we collect the old documents and readme texts.
19128 2. remove the first part of files where I'm already sure that we no
19129 longer need them because of three reasons: either they are just temporary
19130 files which were left by Eric or they are preserved original files where
19131 I've verified that the diff is also available in the CVS via "cvs diff
19132 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19133 the crypto/md/ stuff).
19135 *Ralf S. Engelschall*
19137 * More extension code. Incomplete support for subject and issuer alt
19138 name, issuer and authority key id. Change the i2v function parameters
19139 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19140 what that's for :-) Fix to ASN1 macro which messed up
19141 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19145 * Preliminary support for ENUMERATED type. This is largely copied from the
19150 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19152 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19154 * Make sure `make rehash` target really finds the `openssl` program.
19156 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19158 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19159 like to hear about it if this slows down other processors.
19163 * Add CygWin32 platform information to Configure script.
19165 *Alan Batie <batie@aahz.jf.intel.com>*
19167 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19169 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19171 * New program nseq to manipulate netscape certificate sequences
19175 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19180 * Fixes to BN code. Previously the default was to define BN_RECURSION
19181 but the BN code had some problems that would cause failures when
19182 doing certificate verification and some other functions.
19184 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19186 * Add ASN1 and PEM code to support netscape certificate sequences.
19190 * Add ASN1 and PEM code to support netscape certificate sequences.
19194 * Add several PKIX and private extended key usage OIDs.
19198 * Modify the 'ca' program to handle the new extension code. Modify
19199 openssl.cnf for new extension format, add comments.
19203 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19204 and add a sample to openssl.cnf so req -x509 now adds appropriate
19209 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19210 error code, add initial support to X509_print() and x509 application.
19214 * Takes a deep breath and start adding X509 V3 extension support code. Add
19215 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19216 stuff is currently isolated and isn't even compiled yet.
19220 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19221 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19222 Removed the versions check from X509 routines when loading extensions:
19223 this allows certain broken certificates that don't set the version
19224 properly to be processed.
19228 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19229 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19230 can still be regenerated with "make depend".
19234 * Spelling mistake in C version of CAST-128.
19236 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19238 * Changes to the error generation code. The perl script err-code.pl
19239 now reads in the old error codes and retains the old numbers, only
19240 adding new ones if necessary. It also only changes the .err files if new
19241 codes are added. The makefiles have been modified to only insert errors
19242 when needed (to avoid needlessly modifying header files). This is done
19243 by only inserting errors if the .err file is newer than the auto generated
19244 C file. To rebuild all the error codes from scratch (the old behaviour)
19245 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19246 or delete all the .err files.
19250 * CAST-128 was incorrectly implemented for short keys. The C version has
19251 been fixed, but is untested. The assembler versions are also fixed, but
19252 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19253 to regenerate it if needed.
19254 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19255 Hagino <itojun@kame.net>*
19257 * File was opened incorrectly in randfile.c.
19259 *Ulf Möller <ulf@fitug.de>*
19261 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19262 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19263 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19264 al: it's just almost always a UTCTime. Note this patch adds new error
19265 codes so do a "make errors" if there are problems.
19269 * Correct Linux 1 recognition in config.
19271 *Ulf Möller <ulf@fitug.de>*
19273 * Remove pointless MD5 hash when using DSA keys in ca.
19275 *Anonymous <nobody@replay.com>*
19277 * Generate an error if given an empty string as a cert directory. Also
19278 generate an error if handed NULL (previously returned 0 to indicate an
19279 error, but didn't set one).
19281 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19283 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19287 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19288 parameters. This was causing a warning which killed off the Win32 compile.
19292 * Remove C++ style comments from crypto/bn/bn_local.h.
19294 *Neil Costigan <neil.costigan@celocom.com>*
19296 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19297 based on a text string, looking up short and long names and finally
19298 "dot" format. The "dot" format stuff didn't work. Added new function
19299 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19300 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19301 OID is not part of the table.
19305 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19306 X509_LOOKUP_by_alias().
19310 * Sort openssl functions by name.
19314 * Get the `gendsa` command working and add it to the `list` command. Remove
19315 encryption from sample DSA keys (in case anyone is interested the password
19320 * Make *all* `*_free` functions accept a NULL pointer.
19322 *Frans Heymans <fheymans@isaserver.be>*
19324 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19327 *Anonymous <nobody@replay.com>*
19329 * s_server should send the CAfile as acceptable CAs, not its own cert.
19331 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19333 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19335 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19337 * Temp key "for export" tests were wrong in s3_srvr.c.
19339 *Anonymous <nobody@replay.com>*
19341 * Add prototype for temp key callback functions
19342 SSL_CTX_set_tmp_{rsa,dh}_callback().
19346 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19347 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19351 * X509_name_add_entry() freed the wrong thing after an error.
19353 *Arne Ansper <arne@ats.cyber.ee>*
19355 * rsa_eay.c would attempt to free a NULL context.
19357 *Arne Ansper <arne@ats.cyber.ee>*
19359 * BIO_s_socket() had a broken should_retry() on Windoze.
19361 *Arne Ansper <arne@ats.cyber.ee>*
19363 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19365 *Arne Ansper <arne@ats.cyber.ee>*
19367 * Make sure the already existing X509_STORE->depth variable is initialized
19368 in X509_STORE_new(), but document the fact that this variable is still
19369 unused in the certificate verification process.
19371 *Ralf S. Engelschall*
19373 * Fix the various library and `apps/` files to free up pkeys obtained from
19374 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19378 * Fix reference counting in X509_PUBKEY_get(). This makes
19379 demos/maurice/example2.c work, amongst others, probably.
19381 *Steve Henson and Ben Laurie*
19383 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19384 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19385 are no longer created. This way we have a single and consistent command
19386 line interface `openssl <command>`, similar to `cvs <command>`.
19388 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19390 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19391 BIT STRING wrapper always have zero unused bits.
19395 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19399 * Make the top-level INSTALL documentation easier to understand.
19403 * Makefiles updated to exit if an error occurs in a sub-directory
19404 make (including if user presses ^C) [Paul Sutton]
19406 * Make Montgomery context stuff explicit in RSA data structure.
19410 * Fix build order of pem and err to allow for generated pem.h.
19414 * Fix renumbering bug in X509_NAME_delete_entry().
19418 * Enhanced the err-ins.pl script so it makes the error library number
19419 global and can add a library name. This is needed for external ASN1 and
19420 other error libraries.
19424 * Fixed sk_insert which never worked properly.
19428 * Fix ASN1 macros so they can handle indefinite length constructed
19429 EXPLICIT tags. Some non standard certificates use these: they can now
19434 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19435 into a single doc/ssleay.txt bundle. This way the information is still
19436 preserved but no longer messes up this directory. Now it's new room for
19437 the new set of documentation files.
19439 *Ralf S. Engelschall*
19441 * SETs were incorrectly DER encoded. This was a major pain, because they
19442 shared code with SEQUENCEs, which aren't coded the same. This means that
19443 almost everything to do with SETs or SEQUENCEs has either changed name or
19444 number of arguments.
19446 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19448 * Fix test data to work with the above.
19452 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19453 was already fixed by Eric for 0.9.1 it seems.
19455 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19457 * Autodetect FreeBSD3.
19461 * Fix various bugs in Configure. This affects the following platforms:
19465 unixware-2.0-pentium
19470 * Eliminate generated files from CVS. Reorder tests to regenerate files
19471 before they are needed.
19475 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19479 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19481 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19482 changed SSLeay to OpenSSL in version strings.
19484 *Ralf S. Engelschall*
19486 * Some fixups to the top-level documents.
19490 * Fixed the nasty bug where rsaref.h was not found under compile-time
19491 because the symlink to include/ was missing.
19493 *Ralf S. Engelschall*
19495 * Incorporated the popular no-RSA/DSA-only patches
19496 which allow to compile an RSA-free SSLeay.
19498 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19500 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19501 when "ssleay" is still not found.
19503 *Ralf S. Engelschall*
19505 * Added more platforms to Configure: Cray T3E, HPUX 11,
19507 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19509 * Updated the README file.
19511 *Ralf S. Engelschall*
19513 * Added various .cvsignore files in the CVS repository subdirs
19514 to make a "cvs update" really silent.
19516 *Ralf S. Engelschall*
19518 * Recompiled the error-definition header files and added
19519 missing symbols to the Win32 linker tables.
19521 *Ralf S. Engelschall*
19523 * Cleaned up the top-level documents;
19524 o new files: CHANGES and LICENSE
19525 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19526 o merged COPYRIGHT into LICENSE
19527 o removed obsolete TODO file
19528 o renamed MICROSOFT to INSTALL.W32
19530 *Ralf S. Engelschall*
19532 * Removed dummy files from the 0.9.1b source tree:
19533 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19534 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19535 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19536 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19537 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19539 *Ralf S. Engelschall*
19541 * Added various platform portability fixes.
19545 * The Genesis of the OpenSSL rpject:
19546 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19547 Young and Tim J. Hudson created while they were working for C2Net until
19550 *The OpenSSL Project*
19552 ### Changes between 0.9.0b and 0.9.1b [not released]
19554 * Updated a few CA certificates under certs/
19558 * Changed some BIGNUM api stuff.
19562 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19563 DGUX x86, Linux Alpha, etc.
19567 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19568 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19573 * Add -strparse option to asn1pars program which parses nested
19576 *Dr Stephen Henson <shenson@bigfoot.com>*
19578 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19582 * DSA fix for "ca" program.
19586 * Added "-genkey" option to "dsaparam" program.
19590 * Added RIPE MD160 (rmd160) message digest.
19594 * Added -a (all) option to "ssleay version" command.
19598 * Added PLATFORM define which is the id given to Configure.
19602 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19606 * Extended the ASN.1 parser routines.
19610 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19614 * Added a BN_CTX to the BN library.
19618 * Fixed the weak key values in DES library
19622 * Changed API in EVP library for cipher aliases.
19626 * Added support for RC2/64bit cipher.
19630 * Converted the lhash library to the crypto/mem.c functions.
19634 * Added more recognized ASN.1 object ids.
19638 * Added more RSA padding checks for SSL/TLS.
19642 * Added BIO proxy/filter functionality.
19646 * Added extra_certs to SSL_CTX which can be used
19647 send extra CA certificates to the client in the CA cert chain sending
19648 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19652 * Now Fortezza is denied in the authentication phase because
19653 this is key exchange mechanism is not supported by SSLeay at all.
19657 * Additional PKCS1 checks.
19661 * Support the string "TLSv1" for all TLS v1 ciphers.
19665 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19666 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19670 * Fixed a few memory leaks.
19674 * Fixed various code and comment typos.
19678 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19679 bytes sent in the client random.
19681 *Edward Bishop <ebishop@spyglass.com>*
19685 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19686 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19687 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19688 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19689 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19690 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19691 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19692 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19693 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19694 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19695 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19696 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19697 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19698 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19699 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19700 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19701 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19702 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19703 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19704 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19705 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19706 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19707 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19708 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19709 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19710 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19711 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19712 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19713 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19714 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19715 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19716 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19717 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19718 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19719 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19720 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19721 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19722 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19723 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19724 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19725 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19726 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19727 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19728 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19729 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19730 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19731 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19732 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19733 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19734 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19735 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19736 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19737 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19738 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19739 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19740 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19741 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19742 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19743 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19744 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19745 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19746 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19747 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19748 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19749 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19750 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19751 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19752 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19753 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19754 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19755 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19756 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19757 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19758 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19759 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19760 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19761 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19762 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19763 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19764 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19765 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19766 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19767 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19768 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19769 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19770 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19771 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19772 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19773 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19774 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19775 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19776 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19777 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19778 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19779 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19780 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19781 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19782 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19783 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19784 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19785 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19786 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19787 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19788 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19789 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19790 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19791 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19792 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19793 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19794 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19795 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19796 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19797 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19798 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19799 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19800 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19801 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19802 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19803 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19804 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19805 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19806 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19807 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19808 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19809 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19810 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19811 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19812 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19813 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19814 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19815 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19816 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19817 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19818 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19819 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19820 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19821 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19822 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19823 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19824 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19825 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19826 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19827 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19828 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19829 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19830 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19831 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19832 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19833 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19834 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19835 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19836 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19837 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19838 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19839 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19840 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19841 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19842 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19843 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19844 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19845 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19846 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655