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 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
27 * The FIPS provider includes a few non-approved algorithms for
28 backward compatibility purposes and the "fips=yes" property query
29 must be used for all algorithm fetches to ensure FIPS compliance.
31 The algorithms that are included but not approved are Triple DES ECB
36 * Added support for KMAC in KBKDF.
40 * RNDR and RNDRRS support in provider functions to provide
41 random number generation for Arm CPUs (aarch64).
45 * s_client and s_server apps now explicitly say when the TLS version
46 does not include the renegotiation mechanism. This avoids confusion
47 between that scenario versus when the TLS version includes secure
48 renegotiation but the peer lacks support for it.
52 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
54 *Tomasz Kantecki, Andrey Matyukov*
56 * The various OBJ_* functions have been made thread safe.
60 * Parallel dual-prime 1536/2048-bit modular exponentiation for
61 AVX512_IFMA capable processors.
63 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
65 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
66 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
67 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
68 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
69 `OPENSSL_NO_DEPRECATED_3_1`.
71 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
72 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
73 definitions for these functions regardless of whether
74 `OPENSSL_NO_DEPRECATED_3_1` is defined.
76 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
77 functions regardless of whether they are using them. It is recommended that
78 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
82 * When generating safe-prime DH parameters set the recommended private key
83 length equivalent to minimum key lengths as in RFC 7919.
87 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
88 maximum size that is smaller or equal to the digest length to comply with
89 FIPS 186-4 section 5. This is implemented by a new option
90 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
91 `rsa_pss_saltlen` parameter, which is now the default. Signature
92 verification is not affected by this change and continues to work as before.
99 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
100 listed here are only a brief description.
101 The migration guide contains more detailed information related to new features,
102 breaking changes, and mappings for the large list of deprecated functions.
104 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
106 ### Changes between 3.0.7 and 3.0.8 [xx XXX xxxx]
108 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
109 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
110 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
111 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
112 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
113 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
114 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
115 for legacy EC and SM2 keys is also changed similarly to honor the
116 equivalent conversion format flag as specified in the underlying
117 `EC_KEY` object being exported to a provider, when this function is
118 called through `EVP_PKEY_export()`.
122 * Fixed a type confusion vulnerability relating to X.400 address processing
123 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
124 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
125 vulnerability may allow an attacker who can provide a certificate chain and
126 CRL (neither of which need have a valid signature) to pass arbitrary pointers
127 to a `memcmp` call, creating a possible read primitive, subject to some
128 constraints. Refer to the advisory for more information. Thanks to David
129 Benjamin for discovering this issue. ([CVE-2023-0286])
131 This issue has been fixed by changing the public header file definition of
132 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
133 possible for any existing application to successfully use the existing
134 definition; however, if any application references the `x400Address` field
135 (e.g. in dead code), note that the type of this field has changed. There is
140 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
142 * Fixed two buffer overflows in punycode decoding functions.
144 A buffer overrun can be triggered in X.509 certificate verification,
145 specifically in name constraint checking. Note that this occurs after
146 certificate chain signature verification and requires either a CA to
147 have signed the malicious certificate or for the application to continue
148 certificate verification despite failure to construct a path to a trusted
151 In a TLS client, this can be triggered by connecting to a malicious
152 server. In a TLS server, this can be triggered if the server requests
153 client authentication and a malicious client connects.
155 An attacker can craft a malicious email address to overflow
156 an arbitrary number of bytes containing the `.` character (decimal 46)
157 on the stack. This buffer overflow could result in a crash (causing a
161 An attacker can craft a malicious email address to overflow four
162 attacker-controlled bytes on the stack. This buffer overflow could
163 result in a crash (causing a denial of service) or potentially remote code
164 execution depending on stack layout for any given platform/compiler.
169 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
170 parameters in OpenSSL code.
171 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
172 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
173 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
174 Using these invalid names may cause algorithms to use slower methods
175 that ignore the CRT parameters.
179 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
184 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
185 data to be signed before signing the certificate.
189 * Added RIPEMD160 to the default provider.
193 * Ensured that the key share group sent or accepted for the key exchange
194 is allowed for the protocol version.
198 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
200 * OpenSSL supports creating a custom cipher via the legacy
201 EVP_CIPHER_meth_new() function and associated function calls. This function
202 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
203 to use the new provider mechanism in order to implement custom ciphers.
205 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
206 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
207 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
208 and decryption initialisation functions). Instead of using the custom cipher
209 directly it incorrectly tries to fetch an equivalent cipher from the
210 available providers. An equivalent cipher is found based on the NID passed to
211 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
212 given cipher. However it is possible for an application to incorrectly pass
213 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
214 is used in this way the OpenSSL encryption/decryption initialisation function
215 will match the NULL cipher as being equivalent and will fetch this from the
216 available providers. This will succeed if the default provider has been
217 loaded (or if a third party provider has been loaded that offers this
218 cipher). Using the NULL cipher means that the plaintext is emitted as the
221 Applications are only affected by this issue if they call
222 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
223 encryption/decryption initialisation function. Applications that only use
224 SSL/TLS are not impacted by this issue.
229 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
234 * Fixed the linux-mips64 Configure target which was missing the
235 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
240 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
245 * Correctly handle a retransmitted ClientHello in DTLS
249 * Fixed detection of ktls support in cross-compile environment on Linux
253 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
258 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
259 report correct results in some cases
263 * Fix UWP builds by defining VirtualLock
267 * For known safe primes use the minimum key length according to RFC 7919.
268 Longer private key sizes unnecessarily raise the cycles needed to compute the
269 shared secret without any increase of the real security. This fixes a
270 regression from 1.1.1 where these shorter keys were generated for the known
275 * Added the loongarch64 target
279 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
280 only passed to the FIPS provider and not to the default or legacy provider.
284 * Fixed reported performance degradation on aarch64. Restored the
285 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
286 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
287 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
288 The new algorithm is still used for 32 bit targets.
292 * Added a missing header for memcmp that caused compilation failure on some
297 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
299 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
300 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
301 This issue makes the RSA implementation with 2048 bit private keys
302 incorrect on such machines and memory corruption will happen during
303 the computation. As a consequence of the memory corruption an attacker
304 may be able to trigger a remote code execution on the machine performing
307 SSL/TLS servers or other servers using 2048 bit RSA private keys running
308 on machines supporting AVX512IFMA instructions of the X86_64 architecture
309 are affected by this issue.
314 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
315 implementation would not encrypt the entirety of the data under some
316 circumstances. This could reveal sixteen bytes of data that was
317 preexisting in the memory that wasn't written. In the special case of
318 "in place" encryption, sixteen bytes of the plaintext would be revealed.
320 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
321 they are both unaffected.
324 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
326 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
328 * In addition to the c_rehash shell command injection identified in
329 CVE-2022-1292, further bugs where the c_rehash script does not
330 properly sanitise shell metacharacters to prevent command injection have been
333 When the CVE-2022-1292 was fixed it was not discovered that there
334 are other places in the script where the file names of certificates
335 being hashed were possibly passed to a command executed through the shell.
337 This script is distributed by some operating systems in a manner where
338 it is automatically executed. On such operating systems, an attacker
339 could execute arbitrary commands with the privileges of the script.
341 Use of the c_rehash script is considered obsolete and should be replaced
342 by the OpenSSL rehash command line tool.
345 *Daniel Fiala, Tomáš Mráz*
347 * Case insensitive string comparison no longer uses locales. It has instead
348 been directly implemented.
352 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
354 * Case insensitive string comparison is reimplemented via new locale-agnostic
355 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
356 comparison. The previous implementation had problems when the Turkish locale
361 * Fixed a bug in the c_rehash script which was not properly sanitising shell
362 metacharacters to prevent command injection. This script is distributed by
363 some operating systems in a manner where it is automatically executed. On
364 such operating systems, an attacker could execute arbitrary commands with the
365 privileges of the script.
367 Use of the c_rehash script is considered obsolete and should be replaced
368 by the OpenSSL rehash command line tool.
373 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
374 certificate on an OCSP response. The bug caused the function in the case
375 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
376 response (meaning a successful verification) even in the case where the
377 response signing certificate fails to verify.
379 It is anticipated that most users of `OCSP_basic_verify` will not use the
380 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
381 a negative value (indicating a fatal error) in the case of a certificate
382 verification failure. The normal expected return value in this case would be
385 This issue also impacts the command line OpenSSL "ocsp" application. When
386 verifying an ocsp response with the "-no_cert_checks" option the command line
387 application will report that the verification is successful even though it
388 has in fact failed. In this case the incorrect successful response will also
389 be accompanied by error messages showing the failure and contradicting the
390 apparently successful result.
395 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
396 AAD data as the MAC key. This made the MAC key trivially predictable.
398 An attacker could exploit this issue by performing a man-in-the-middle attack
399 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
400 that the modified data would still pass the MAC integrity check.
402 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
403 endpoint will always be rejected by the recipient and the connection will
404 fail at that point. Many application protocols require data to be sent from
405 the client to the server first. Therefore, in such a case, only an OpenSSL
406 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
408 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
409 sent in both directions. In this case both clients and servers could be
410 affected, regardless of the application protocol.
412 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
413 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
414 the handshake when using this ciphersuite.
416 The confidentiality of data is not impacted by this issue, i.e. an attacker
417 cannot decrypt data that has been encrypted using this ciphersuite - they can
420 In order for this attack to work both endpoints must legitimately negotiate
421 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
422 OpenSSL 3.0, and is not available within the default provider or the default
423 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
424 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
425 following must have occurred:
427 1) OpenSSL must have been compiled with the (non-default) compile time option
428 enable-weak-ssl-ciphers
430 2) OpenSSL must have had the legacy provider explicitly loaded (either
431 through application code or via configuration)
433 3) The ciphersuite must have been explicitly added to the ciphersuite list
435 4) The libssl security level must have been set to 0 (default is 1)
437 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
439 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
440 others that both endpoints have in common
445 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
446 occupied by the removed hash table entries.
448 This function is used when decoding certificates or keys. If a long lived
449 process periodically decodes certificates or keys its memory usage will
450 expand without bounds and the process might be terminated by the operating
451 system causing a denial of service. Also traversing the empty hash table
452 entries will take increasingly more time.
454 Typically such long lived processes might be TLS clients or TLS servers
455 configured to accept client certificate authentication.
458 *Hugo Landau, Aliaksei Levin*
460 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
461 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
462 statistics are no longer supported. For compatibility, these statistics are
463 still listed in the output but are now always reported as zero.
467 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
469 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
470 for non-prime moduli.
472 Internally this function is used when parsing certificates that contain
473 elliptic curve public keys in compressed form or explicit elliptic curve
474 parameters with a base point encoded in compressed form.
476 It is possible to trigger the infinite loop by crafting a certificate that
477 has invalid explicit curve parameters.
479 Since certificate parsing happens prior to verification of the certificate
480 signature, any process that parses an externally supplied certificate may thus
481 be subject to a denial of service attack. The infinite loop can also be
482 reached when parsing crafted private keys as they can contain explicit
483 elliptic curve parameters.
485 Thus vulnerable situations include:
487 - TLS clients consuming server certificates
488 - TLS servers consuming client certificates
489 - Hosting providers taking certificates or private keys from customers
490 - Certificate authorities parsing certification requests from subscribers
491 - Anything else which parses ASN.1 elliptic curve parameters
493 Also any other applications that use the BN_mod_sqrt() where the attacker
494 can control the parameter values are vulnerable to this DoS issue.
499 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
500 to the list of ciphersuites providing Perfect Forward Secrecy as
501 required by SECLEVEL >= 3.
503 *Dmitry Belyavskiy, Nicola Tuveri*
505 * Made the AES constant time code for no-asm configurations
506 optional due to the resulting 95% performance degradation.
507 The AES constant time code can be enabled, for no assembly
508 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
512 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
517 * The negative return value handling of the certificate verification callback
518 was reverted. The replacement is to set the verification retry state with
519 the SSL_set_retry_verify() function.
523 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
525 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
526 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
527 verify a certificate supplied by a server. That function may return a
528 negative return value to indicate an internal error (for example out of
529 memory). Such a negative return value is mishandled by OpenSSL and will cause
530 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
531 success and a subsequent call to SSL_get_error() to return the value
532 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
533 returned by OpenSSL if the application has previously called
534 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
535 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
536 totally unexpected and applications may not behave correctly as a result. The
537 exact behaviour will depend on the application but it could result in
538 crashes, infinite loops or other similar incorrect responses.
540 This issue is made more serious in combination with a separate bug in OpenSSL
541 3.0 that will cause X509_verify_cert() to indicate an internal error when
542 processing a certificate chain. This will occur where a certificate does not
543 include the Subject Alternative Name extension but where a Certificate
544 Authority has enforced name constraints. This issue can occur even with valid
550 * Corrected a few file name and file reference bugs in the build,
551 installation and setup scripts, which lead to installation verification
552 failures. Slightly enhanced the installation verification script.
556 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
561 * Fixed PVK encoder to properly query for the passphrase.
565 * Multiple fixes in the OSSL_HTTP API functions.
569 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
570 OSSL_PARAM_INTEGER data type and return error on negative numbers
571 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
572 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
576 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
580 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
584 * Multiple threading fixes.
588 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
592 * Allow fetching an operation from the provider that owns an unexportable key
593 as a fallback if that is still allowed by the property query.
597 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
599 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
604 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
605 S390X capability vector to zero. This simplifies testing of different code
606 paths on S390X architecture.
610 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
611 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
612 SP 800-38D". The communication will fail at this point.
616 * The EC_GROUP_clear_free() function is deprecated as there is nothing
617 confidential in EC_GROUP data.
621 * The byte order mark (BOM) character is ignored if encountered at the
622 beginning of a PEM-formatted file.
626 * Added CMS support for the Russian GOST algorithms.
630 * Due to move of the implementation of cryptographic operations
631 to the providers, validation of various operation parameters can
632 be postponed until the actual operation is executed where previously
633 it happened immediately when an operation parameter was set.
635 For example when setting an unsupported curve with
636 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
637 fail but later keygen operations with the EVP_PKEY_CTX will fail.
639 *OpenSSL team members and many third party contributors*
641 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
642 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
643 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
644 instead to retrieve these algorithms from a provider.
648 * On build targets where the multilib postfix is set in the build
649 configuration the libdir directory was changing based on whether
650 the lib directory with the multilib postfix exists on the system
651 or not. This unpredictable behavior was removed and eventual
652 multilib postfix is now always added to the default libdir. Use
653 `--libdir=lib` to override the libdir if adding the postfix is
658 * The triple DES key wrap functionality now conforms to RFC 3217 but is
659 no longer interoperable with OpenSSL 1.1.1.
663 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
664 function codes, this function can only cause problems for calling
669 * Add a configurable flag to output date formats as ISO 8601. Does not
670 change the default date format.
674 * Version of MSVC earlier than 1300 could get link warnings, which could
675 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
676 Support for this flag has been removed.
680 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
681 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
682 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
683 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
684 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
688 * The signatures of the functions to get and set options on SSL and
689 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
690 Some source code changes may be required.
694 * The public definitions of conf_method_st and conf_st have been
695 deprecated. They will be made opaque in a future release.
697 *Rich Salz and Tomáš Mráz*
699 * Client-initiated renegotiation is disabled by default. To allow it, use
700 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
701 flag, or the "ClientRenegotiation" config parameter as appropriate.
705 * Add "abspath" and "includedir" pragma's to config files, to prevent,
706 or modify relative pathname inclusion.
710 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
711 validated. Please consult the README-FIPS and
712 README-PROVIDERS files, as well as the migration guide.
714 *OpenSSL team members and many third party contributors*
716 * For the key types DH and DHX the allowed settable parameters are now different.
720 * The openssl commands that read keys, certificates, and CRLs now
721 automatically detect the PEM or DER format of the input files.
723 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
725 * Added enhanced PKCS#12 APIs which accept a library context.
729 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
733 * Added support for Kernel TLS (KTLS).
735 *Boris Pismenny, John Baldwin and Andrew Gallatin*
737 * Support for RFC 5746 secure renegotiation is now required by default for
738 SSL or TLS connections to succeed.
742 * The signature of the `copy` functional parameter of the
743 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
744 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
745 the signature of the `pub_decode` functional parameter of the
746 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
747 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
751 * The error return values from some control calls (ctrl) have changed.
755 * A public key check is now performed during EVP_PKEY_derive_set_peer().
759 * Many functions in the EVP_ namespace that are getters of values from
760 implementations or contexts were renamed to include get or get0 in their
761 names. Old names are provided as macro aliases for compatibility and
766 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
767 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
768 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
773 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
776 * The output from the command line applications may have minor
781 * The output from numerous "printing" may have minor changes.
785 * Windows thread synchronization uses read/write primitives (SRWLock) when
786 supported by the OS, otherwise CriticalSection continues to be used.
790 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
791 work on read only BIO source/sinks that do not support these functions.
792 This allows piping or redirection of a file BIO using stdin to be buffered
793 into memory. This is used internally in OSSL_DECODER_from_bio().
797 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
798 this function would return one of the values OSSL_STORE_INFO_NAME,
799 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
800 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
801 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
802 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
803 using this function should be amended to handle the changed return value.
807 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
808 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
809 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
810 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
811 contains more than one certificate identifier: This means that all
812 certificates referenced there MUST be part of the validation chain.
816 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
817 RC5, DESX and DES have been moved to the legacy provider.
821 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
822 RIPEMD-160 have been moved to the legacy provider.
826 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
831 * The deprecated functions EVP_PKEY_get0_RSA(),
832 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
833 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
834 well as the similarly named "get1" functions behave differently in
839 * A number of functions handling low-level keys or engines were deprecated
840 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
841 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
842 EVP_PKEY_get0_siphash().
846 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
847 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
848 will need to load the legacy crypto provider. This includes these PBE
849 algorithms which use this KDF:
850 - NID_pbeWithMD2AndDES_CBC
851 - NID_pbeWithMD5AndDES_CBC
852 - NID_pbeWithSHA1AndRC2_CBC
853 - NID_pbeWithMD2AndRC2_CBC
854 - NID_pbeWithMD5AndRC2_CBC
855 - NID_pbeWithSHA1AndDES_CBC
859 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
860 BIO_debug_callback() functions.
864 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
865 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
869 * The RAND_METHOD APIs have been deprecated.
873 * The SRP APIs have been deprecated.
877 * Add a compile time option to prevent the caching of provider fetched
878 algorithms. This is enabled by including the no-cached-fetch option
879 at configuration time.
883 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
884 count of PKCS12_DEFAULT_ITER.
886 *Tomáš Mráz and Sahana Prasad*
888 * The openssl speed command does not use low-level API calls anymore.
892 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
895 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
897 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
901 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
902 providers may supply their own group implementations (using either the "key
903 exchange" or the "key encapsulation" methods) which will automatically be
904 detected and used by libssl.
906 *Matt Caswell, Nicola Tuveri*
908 * The undocumented function X509_certificate_type() has been deprecated;
912 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
916 * Removed RSA padding mode for SSLv23 (which was only used for
917 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
918 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
923 * Deprecated the obsolete X9.31 RSA key generation related functions.
925 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
926 is not allowed to return a value > 1, this is no more taken as failure.
928 *Viktor Dukhovni and David von Oheimb*
930 * Deprecated the obsolete X9.31 RSA key generation related functions
931 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
932 BN_X931_generate_prime_ex().
936 * The default key generation method for the regular 2-prime RSA keys was
937 changed to the FIPS 186-4 B.3.6 method.
941 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
945 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
949 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
950 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
952 *Rich Salz, Richard Levitte, and David von Oheimb*
954 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
958 * Deprecated `OCSP_parse_url()`.
962 * Validation of SM2 keys has been separated from the validation of regular EC
967 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
968 switches: a validation failure triggers an early exit, returning a failure
969 exit status to the parent process.
973 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
974 to ignore unknown ciphers.
978 * The `-cipher-commands` and `-digest-commands` options
979 of the command line utility `list` have been deprecated.
980 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
984 * Added convenience functions for generating asymmetric key pairs:
985 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
986 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
990 * All of the low-level EC_KEY functions have been deprecated.
992 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
994 * Deprecated all the libcrypto and libssl error string loading
999 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1000 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1005 * The `-crypt` option to the `passwd` command line tool has been removed.
1009 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1014 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1018 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1019 EVP_PKEY_get1_tls_encodedpoint().
1023 * The security callback, which can be customised by application code, supports
1024 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1025 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1029 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1030 interface. Their functionality remains unchanged.
1034 * Added new option for 'openssl list', '-providers', which will display the
1035 list of loaded providers, their names, version and status. It optionally
1036 displays their gettable parameters.
1040 * Removed EVP_PKEY_set_alias_type().
1044 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1045 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1049 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1050 parameters are still checked for type safety at compile time via helper
1055 * Remove the RAND_DRBG API
1057 *Paul Dale and Matthias St. Pierre*
1059 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1060 as well as actual hostnames.
1064 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1065 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1066 conversely, silently ignore DTLS protocol version bounds when configuring
1067 TLS-based contexts. The commands can be repeated to set bounds of both
1068 types. The same applies with the corresponding "min_protocol" and
1069 "max_protocol" command-line switches, in case some application uses both TLS
1072 SSL_CTX instances that are created for a fixed protocol version (e.g.
1073 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1074 attempts to apply bounds to these protocol versions would result in an
1075 error. Now only the "version-flexible" SSL_CTX instances are subject to
1076 limits in configuration files in command-line options.
1080 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1085 * Reworked the recorded ERR codes to make better space for system errors.
1086 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1087 given code is a system error (true) or an OpenSSL error (false).
1091 * Reworked the test perl framework to better allow parallel testing.
1093 *Nicola Tuveri and David von Oheimb*
1095 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1096 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1100 * 'Configure' has been changed to figure out the configuration target if
1101 none is given on the command line. Consequently, the 'config' script is
1102 now only a mere wrapper. All documentation is changed to only mention
1105 *Rich Salz and Richard Levitte*
1107 * Added a library context `OSSL_LIB_CTX` that applications as well as
1108 other libraries can use to form a separate context within which
1109 libcrypto operations are performed.
1113 * Added various `_ex` functions to the OpenSSL API that support using
1114 a non-default `OSSL_LIB_CTX`.
1118 * Handshake now fails if Extended Master Secret extension is dropped
1123 * Dropped interactive mode from the `openssl` program.
1127 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1129 *David von Oheimb and Shane Lontis*
1131 * Deprecated `EC_METHOD_get_field_type()`.
1135 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1136 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1137 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1141 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1145 * Add CAdES-BES signature verification support, mostly derived
1146 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1148 *Filipe Raimundo da Silva*
1150 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1154 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1155 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1159 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1163 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1164 EC_KEY_precompute_mult().
1168 * Deprecated EC_POINTs_mul().
1172 * Removed FIPS_mode() and FIPS_mode_set().
1176 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1180 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1181 EC_POINT_get_Jprojective_coordinates_GFp().
1185 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1186 arrays to be more easily constructed via a series of utility functions.
1187 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1188 the various push functions and finally convert to a passable OSSL_PARAM
1189 array using OSSL_PARAM_BLD_to_param().
1193 * The security strength of SHA1 and MD5 based signatures in TLS has been
1198 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1199 contain a provider side internal key.
1203 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1207 * Project text documents not yet having a proper file name extension
1208 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1209 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1213 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1214 have been converted to Markdown with the goal to produce documents
1215 which not only look pretty when viewed online in the browser, but
1216 remain well readable inside a plain text editor.
1218 To achieve this goal, a 'minimalistic' Markdown style has been applied
1219 which avoids formatting elements that interfere too much with the
1220 reading flow in the text file. For example, it
1222 * avoids [ATX headings][] and uses [setext headings][] instead
1223 (which works for `<h1>` and `<h2>` headings only).
1224 * avoids [inline links][] and uses [reference links][] instead.
1225 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1227 [ATX headings]: https://github.github.com/gfm/#atx-headings
1228 [setext headings]: https://github.github.com/gfm/#setext-headings
1229 [inline links]: https://github.github.com/gfm/#inline-link
1230 [reference links]: https://github.github.com/gfm/#reference-link
1231 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1232 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1234 *Matthias St. Pierre*
1236 * The test suite is changed to preserve results of each test recipe.
1237 A new directory test-runs/ with subdirectories named like the
1238 test recipes are created in the build tree for this purpose.
1242 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1243 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1244 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1246 *David von Oheimb, Martin Peylo*
1248 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1249 It supports arbitrary request and response content types, GET redirection,
1250 TLS, connections via HTTP(S) proxies, connections and exchange via
1251 user-defined BIOs (allowing implicit connections), persistent connections,
1252 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1253 The legacy OCSP-focused (and only partly documented) API
1254 is retained for backward compatibility, while most of it is deprecated.
1258 * Added `util/check-format.pl`, a tool for checking adherence to the
1259 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1260 The checks performed are incomplete and yield some false positives.
1261 Still the tool should be useful for detecting most typical glitches.
1265 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1266 If domain name resolution yields multiple IP addresses all of them are tried
1267 after `connect()` failures.
1271 * All of the low-level RSA functions have been deprecated.
1275 * X509 certificates signed using SHA1 are no longer allowed at security
1280 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1281 modified to use PKEY APIs. These commands are now in maintenance mode
1282 and no new features will be added to them.
1286 * The command line utility rsautl has been deprecated.
1290 * The command line utilities genrsa and rsa have been modified to use PKEY
1291 APIs. They now write PKCS#8 keys by default. These commands are now in
1292 maintenance mode and no new features will be added to them.
1296 * All of the low-level DH functions have been deprecated.
1298 *Paul Dale and Matt Caswell*
1300 * All of the low-level DSA functions have been deprecated.
1304 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1305 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1309 * Deprecated low-level ECDH and ECDSA functions.
1313 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1317 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1318 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1319 a new formulation to include all the things it can be used for,
1320 as well as words of caution.
1324 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1328 * All of the low-level HMAC functions have been deprecated.
1330 *Paul Dale and David von Oheimb*
1332 * Over two thousand fixes were made to the documentation, including:
1333 - Common options (such as -rand/-writerand, TLS version control, etc)
1334 were refactored and point to newly-enhanced descriptions in openssl.pod.
1335 - Added style conformance for all options (with help from Richard Levitte),
1336 documented all reported missing options, added a CI build to check
1337 that all options are documented and that no unimplemented options
1339 - Documented some internals, such as all use of environment variables.
1340 - Addressed all internal broken L<> references.
1344 * All of the low-level CMAC functions have been deprecated.
1348 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1349 functions have been deprecated.
1351 *Paul Dale and David von Oheimb*
1353 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1354 set of functions. The documentation mentioned negative values for some
1355 errors, but this was never the case, so the mention of negative values
1358 Code that followed the documentation and thereby check with something
1359 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1363 * All of the low-level cipher functions have been deprecated.
1365 *Matt Caswell and Paul Dale*
1367 * Removed include/openssl/opensslconf.h.in and replaced it with
1368 include/openssl/configuration.h.in, which differs in not including
1369 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1370 was added to include both.
1372 This allows internal hacks where one might need to modify the set
1373 of configured macros, for example this if deprecated symbols are
1374 still supposed to be available internally:
1376 #include <openssl/configuration.h>
1378 #undef OPENSSL_NO_DEPRECATED
1379 #define OPENSSL_SUPPRESS_DEPRECATED
1381 #include <openssl/macros.h>
1383 This should not be used by applications that use the exported
1384 symbols, as that will lead to linking errors.
1388 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1389 used in exponentiation with 512-bit moduli. No EC algorithms are
1390 affected. Analysis suggests that attacks against 2-prime RSA1024,
1391 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1392 difficult to perform and are not believed likely. Attacks against DH512
1393 are considered just feasible. However, for an attack the target would
1394 have to re-use the DH512 private key, which is not recommended anyway.
1395 Also applications directly using the low-level API BN_mod_exp may be
1396 affected if they use BN_FLG_CONSTTIME.
1401 * Most memory-debug features have been deprecated, and the functionality
1402 replaced with no-ops.
1406 * Added documentation for the STACK API.
1410 * Introduced a new method type and API, OSSL_ENCODER, to represent
1411 generic encoders. These do the same sort of job that PEM writers
1412 and d2i functions do, but with support for methods supplied by
1413 providers, and the possibility for providers to support other
1418 * Introduced a new method type and API, OSSL_DECODER, to represent
1419 generic decoders. These do the same sort of job that PEM readers
1420 and i2d functions do, but with support for methods supplied by
1421 providers, and the possibility for providers to support other
1426 * Added a .pragma directive to the syntax of configuration files, to
1427 allow varying behavior in a supported and predictable manner.
1428 Currently added pragma:
1432 This allows dollar signs to be a keyword character unless it's
1433 followed by a opening brace or parenthesis. This is useful for
1434 platforms where dollar signs are commonly used in names, such as
1435 volume names and system directory names on VMS.
1439 * Added functionality to create an EVP_PKEY from user data.
1443 * Change the interpretation of the '--api' configuration option to
1444 mean that this is a desired API compatibility level with no
1445 further meaning. The previous interpretation, that this would
1446 also mean to remove all deprecated symbols up to and including
1447 the given version, no requires that 'no-deprecated' is also used
1448 in the configuration.
1450 When building applications, the desired API compatibility level
1451 can be set with the OPENSSL_API_COMPAT macro like before. For
1452 API compatibility version below 3.0, the old style numerical
1453 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1454 For version 3.0 and on, the value is expected to be the decimal
1455 value calculated from the major and minor version like this:
1457 MAJOR * 10000 + MINOR * 100
1461 -DOPENSSL_API_COMPAT=30000 For 3.0
1462 -DOPENSSL_API_COMPAT=30200 For 3.2
1464 To hide declarations that are deprecated up to and including the
1465 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1466 given when building the application as well.
1470 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1471 access to certificate and CRL stores via URIs and OSSL_STORE
1474 This adds the following functions:
1476 - X509_LOOKUP_store()
1477 - X509_STORE_load_file()
1478 - X509_STORE_load_path()
1479 - X509_STORE_load_store()
1480 - SSL_add_store_cert_subjects_to_stack()
1481 - SSL_CTX_set_default_verify_store()
1482 - SSL_CTX_load_verify_file()
1483 - SSL_CTX_load_verify_dir()
1484 - SSL_CTX_load_verify_store()
1488 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1489 The presence of this system service is determined at run-time.
1493 * Added functionality to create an EVP_PKEY context based on data
1494 for methods from providers. This takes an algorithm name and a
1495 property query string and simply stores them, with the intent
1496 that any operation that uses this context will use those strings
1497 to fetch the needed methods implicitly, thereby making the port
1498 of application written for pre-3.0 OpenSSL easier.
1502 * The undocumented function NCONF_WIN32() has been deprecated; for
1503 conversion details see the HISTORY section of doc/man5/config.pod
1507 * Introduced the new functions EVP_DigestSignInit_ex() and
1508 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1509 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1510 pages for further details.
1514 * Over two thousand fixes were made to the documentation, including:
1515 adding missing command flags, better style conformance, documentation
1518 *Rich Salz, Richard Levitte*
1520 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1521 X25519, X448, Ed25519 and Ed448.
1525 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1530 * Deprecated the public definition of `ERR_STATE` as well as the function
1531 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1536 * Added ERR functionality to give callers access to the stored function
1537 names that have replaced the older function code based functions.
1539 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1540 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1541 ERR_peek_error_all() and ERR_peek_last_error_all().
1543 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1544 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1545 ERR_func_error_string().
1549 * Extended testing to be verbose for failing tests only. The make variables
1550 VERBOSE_FAILURE or VF can be used to enable this:
1552 $ make VF=1 test # Unix
1553 $ mms /macro=(VF=1) test ! OpenVMS
1554 $ nmake VF=1 test # Windows
1558 * Added the `-copy_extensions` option to the `x509` command for use with
1559 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1560 all extensions in the request are copied to the certificate or vice versa.
1562 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1564 * Added the `-copy_extensions` option to the `req` command for use with
1565 `-x509`. When given with the `copy` or `copyall` argument,
1566 all extensions in the certification request are copied to the certificate.
1570 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1571 they generate are by default RFC 5280 compliant in the following sense:
1572 There is a subjectKeyIdentifier extension with a hash value of the public key
1573 and for not self-signed certs there is an authorityKeyIdentifier extension
1574 with a keyIdentifier field or issuer information identifying the signing key.
1575 This is done unless some configuration overrides the new default behavior,
1576 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1580 * Added several checks to `X509_verify_cert()` according to requirements in
1581 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1582 (which may be done by using the CLI option `-x509_strict`):
1583 * The basicConstraints of CA certificates must be marked critical.
1584 * CA certificates must explicitly include the keyUsage extension.
1585 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1586 * The issuer name of any certificate must not be empty.
1587 * The subject name of CA certs, certs with keyUsage crlSign,
1588 and certs without subjectAlternativeName must not be empty.
1589 * If a subjectAlternativeName extension is given it must not be empty.
1590 * The signatureAlgorithm field and the cert signature must be consistent.
1591 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1592 must not be marked critical.
1593 * The authorityKeyIdentifier must be given for X.509v3 certs
1594 unless they are self-signed.
1595 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1599 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1600 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1604 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1605 used even when parsing explicit parameters, when loading a encoded key
1606 or calling `EC_GROUP_new_from_ecpkparameters()`/
1607 `EC_GROUP_new_from_ecparameters()`.
1608 This prevents bypass of security hardening and performance gains,
1609 especially for curves with specialized EC_METHODs.
1610 By default, if a key encoded with explicit parameters is loaded and later
1611 encoded, the output is still encoded with explicit parameters, even if
1612 internally a "named" EC_GROUP is used for computation.
1616 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1617 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1618 NULL. After this change, only the cofactor parameter can be NULL. It also
1619 does some minimal sanity checks on the passed order.
1624 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1625 An attack is simple, if the first CMS_recipientInfo is valid but the
1626 second CMS_recipientInfo is chosen ciphertext. If the second
1627 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1628 encryption key will be replaced by garbage, and the message cannot be
1629 decoded, but if the RSA decryption fails, the correct encryption key is
1630 used and the recipient will not notice the attack.
1631 As a work around for this potential attack the length of the decrypted
1632 key must be equal to the cipher default key length, in case the
1633 certificate is not given and all recipientInfo are tried out.
1634 The old behaviour can be re-enabled in the CMS code by setting the
1635 CMS_DEBUG_DECRYPT flag.
1639 * Early start up entropy quality from the DEVRANDOM seed source has been
1640 improved for older Linux systems. The RAND subsystem will wait for
1641 /dev/random to be producing output before seeding from /dev/urandom.
1642 The seeded state is stored for future library initialisations using
1643 a system global shared memory segment. The shared memory identifier
1644 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1645 the desired value. The default identifier is 114.
1649 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1650 when primes for RSA keys are computed.
1651 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1652 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1653 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1654 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1655 This avoids possible fingerprinting of newly generated RSA modules.
1659 * Correct the extended master secret constant on EBCDIC systems. Without this
1660 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1661 negotiate EMS will fail. Unfortunately this also means that TLS connections
1662 between EBCDIC systems with this fix, and EBCDIC systems without this
1663 fix will fail if they negotiate EMS.
1667 * Changed the library initialisation so that the config file is now loaded
1668 by default. This was already the case for libssl. It now occurs for both
1669 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1670 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1674 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1675 where the former acts as a replacement for `ERR_put_error()`, and the
1676 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1677 `ERR_raise_data()` adds more flexibility by taking a format string and
1678 an arbitrary number of arguments following it, to be processed with
1683 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1684 to check if a named provider is loaded and available. When called, it
1685 will also activate all fallback providers if such are still present.
1689 * Enforce a minimum DH modulus size of 512 bits.
1693 * Changed DH parameters to generate the order q subgroup instead of 2q.
1694 Previously generated DH parameters are still accepted by DH_check
1695 but DH_generate_key works around that by clearing bit 0 of the
1696 private key for those. This avoids leaking bit 0 of the private key.
1700 * Significantly reduce secure memory usage by the randomness pools.
1704 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1709 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1710 algorithms. An implementation of a key exchange algorithm can be obtained
1711 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1712 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1713 the older EVP_PKEY_derive_init() function. See the man pages for the new
1714 functions for further details.
1718 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1722 * Removed the function names from error messages and deprecated the
1727 * Removed NextStep support and the macro OPENSSL_UNISTD
1731 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1732 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1733 Also removed "export var as function" capability; we do not export
1734 variables, only functions.
1738 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1739 an error and 1 indicating success. In previous versions of OpenSSL this
1740 was a void type. If a key was set longer than the maximum possible this
1745 * Support SM2 signing and verification schemes with X509 certificate.
1749 * Use SHA256 as the default digest for TS query in the `ts` app.
1753 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1757 * Default cipher lists/suites are now available via a function, the
1758 #defines are deprecated.
1762 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1763 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1764 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1768 * Join the directories crypto/x509 and crypto/x509v3
1772 * Added command 'openssl kdf' that uses the EVP_KDF API.
1776 * Added command 'openssl mac' that uses the EVP_MAC API.
1780 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1781 as default directories. Also added the command 'openssl info'
1782 for scripting purposes.
1786 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1791 * Add prediction resistance to the DRBG reseeding process.
1795 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1796 mandated by IEEE Std 1619-2018.
1800 * Added newline escaping functionality to a filename when using openssl dgst.
1801 This output format is to replicate the output format found in the `*sum`
1802 checksum programs. This aims to preserve backward compatibility.
1804 *Matt Eaton, Richard Levitte, and Paul Dale*
1806 * Removed the heartbeat message in DTLS feature, as it has very
1807 little usage and doesn't seem to fulfill a valuable purpose.
1808 The configuration option is now deprecated.
1812 * Changed the output of 'openssl {digestname} < file' to display the
1813 digest name in its output.
1817 * Added a new generic trace API which provides support for enabling
1818 instrumentation through trace output.
1820 *Richard Levitte & Matthias St. Pierre*
1822 * Added build tests for C++. These are generated files that only do one
1823 thing, to include one public OpenSSL head file each. This tests that
1824 the public header files can be usefully included in a C++ application.
1826 This test isn't enabled by default. It can be enabled with the option
1827 'enable-buildtest-c++'.
1831 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1835 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1839 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1843 * Added KMAC to EVP_MAC.
1847 * Added property based algorithm implementation selection framework to
1852 * Added SCA hardening for modular field inversion in EC_GROUP through
1853 a new dedicated field_inv() pointer in EC_METHOD.
1854 This also addresses a leakage affecting conversions from projective
1855 to affine coordinates.
1857 *Billy Bob Brumley, Nicola Tuveri*
1859 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1860 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1861 those algorithms that were already supported through the EVP_PKEY API
1862 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1863 and scrypt are now wrappers that call EVP_KDF.
1867 * Build devcrypto engine as a dynamic engine.
1869 *Eneas U de Queiroz*
1871 * Add keyed BLAKE2 to EVP_MAC.
1875 * Fix a bug in the computation of the endpoint-pair shared secret used
1876 by DTLS over SCTP. This breaks interoperability with older versions
1877 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1878 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1879 interoperability with such broken implementations. However, enabling
1880 this switch breaks interoperability with correct implementations.
1882 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1883 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1887 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1891 * Changed the license to the Apache License v2.0.
1895 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1897 - Major releases (indicated by incrementing the MAJOR release number)
1898 may introduce incompatible API/ABI changes.
1899 - Minor releases (indicated by incrementing the MINOR release number)
1900 may introduce new features but retain API/ABI compatibility.
1901 - Patch releases (indicated by incrementing the PATCH number)
1902 are intended for bug fixes and other improvements of existing
1903 features only (like improving performance or adding documentation)
1904 and retain API/ABI compatibility.
1908 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1912 * Remove the 'dist' target and add a tarball building script. The
1913 'dist' target has fallen out of use, and it shouldn't be
1914 necessary to configure just to create a source distribution.
1918 * Recreate the OS390-Unix config target. It no longer relies on a
1919 special script like it did for OpenSSL pre-1.1.0.
1923 * Instead of having the source directories listed in Configure, add
1924 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1929 * Add GMAC to EVP_MAC.
1933 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1937 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1938 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1939 to facilitate the continued use of MACs through raw private keys in
1940 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1944 * Deprecate ECDH_KDF_X9_62().
1948 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1949 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1950 are retained for backwards compatibility.
1954 * AES-XTS mode now enforces that its two keys are different to mitigate
1955 the attacked described in "Efficient Instantiations of Tweakable
1956 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1957 Details of this attack can be obtained from:
1958 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1962 * Rename the object files, i.e. give them other names than in previous
1963 versions. Their names now include the name of the final product, as
1964 well as its type mnemonic (bin, lib, shlib).
1968 * Added new option for 'openssl list', '-objects', which will display the
1969 list of built in objects, i.e. OIDs with names.
1973 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1974 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1979 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1980 improves application performance by removing data copies and providing
1981 applications with zero-copy system calls such as sendfile and splice.
1985 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1989 * `PKCS12_parse` now maintains the order of the parsed certificates
1990 when outputting them via `*ca` (rather than reversing it).
1994 * Deprecated pthread fork support methods.
1998 * Added support for FFDHE key exchange in TLS 1.3.
2002 * Added a new concept for OpenSSL plugability: providers. This
2003 functionality is designed to replace the ENGINE API and ENGINE
2004 implementations, and to be much more dynamic, allowing provider
2005 authors to introduce new algorithms among other things, as long as
2006 there's an API that supports the algorithm type.
2008 With this concept comes a new core API for interaction between
2009 libcrypto and provider implementations. Public libcrypto functions
2010 that want to use providers do so through this core API.
2012 The main documentation for this core API is found in
2013 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2014 refer to other manuals describing the API specific for supported
2015 algorithm types (also called operations).
2022 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2024 * Avoid loading of a dynamic engine twice.
2028 * Prioritise DANE TLSA issuer certs over peer certs
2032 * Fixed random API for MacOS prior to 10.12
2034 These MacOS versions don't support the CommonCrypto APIs
2038 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2040 * Fixed an SM2 Decryption Buffer Overflow.
2042 In order to decrypt SM2 encrypted data an application is expected to
2043 call the API function EVP_PKEY_decrypt(). Typically an application will
2044 call this function twice. The first time, on entry, the "out" parameter
2045 can be NULL and, on exit, the "outlen" parameter is populated with the
2046 buffer size required to hold the decrypted plaintext. The application
2047 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2048 again, but this time passing a non-NULL value for the "out" parameter.
2050 A bug in the implementation of the SM2 decryption code means that the
2051 calculation of the buffer size required to hold the plaintext returned
2052 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2053 size required by the second call. This can lead to a buffer overflow
2054 when EVP_PKEY_decrypt() is called by the application a second time with
2055 a buffer that is too small.
2057 A malicious attacker who is able present SM2 content for decryption to
2058 an application could cause attacker chosen data to overflow the buffer
2059 by up to a maximum of 62 bytes altering the contents of other data held
2060 after the buffer, possibly changing application behaviour or causing
2061 the application to crash. The location of the buffer is application
2062 dependent but is typically heap allocated.
2067 * Fixed various read buffer overruns processing ASN.1 strings
2069 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2070 structure which contains a buffer holding the string data and a field
2071 holding the buffer length. This contrasts with normal C strings which
2072 are represented as a buffer for the string data which is terminated
2073 with a NUL (0) byte.
2075 Although not a strict requirement, ASN.1 strings that are parsed using
2076 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2077 well as any string whose value has been set with the ASN1_STRING_set()
2078 function will additionally NUL terminate the byte array in the
2079 ASN1_STRING structure.
2081 However, it is possible for applications to directly construct valid
2082 ASN1_STRING structures which do not NUL terminate the byte array by
2083 directly setting the "data" and "length" fields in the ASN1_STRING
2084 array. This can also happen by using the ASN1_STRING_set0() function.
2086 Numerous OpenSSL functions that print ASN.1 data have been found to
2087 assume that the ASN1_STRING byte array will be NUL terminated, even
2088 though this is not guaranteed for strings that have been directly
2089 constructed. Where an application requests an ASN.1 structure to be
2090 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2091 been directly constructed by the application without NUL terminating
2092 the "data" field, then a read buffer overrun can occur.
2094 The same thing can also occur during name constraints processing
2095 of certificates (for example if a certificate has been directly
2096 constructed by the application instead of loading it via the OpenSSL
2097 parsing functions, and the certificate contains non NUL terminated
2098 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2099 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2101 If a malicious actor can cause an application to directly construct an
2102 ASN1_STRING and then process it through one of the affected OpenSSL
2103 functions then this issue could be hit. This might result in a crash
2104 (causing a Denial of Service attack). It could also result in the
2105 disclosure of private memory contents (such as private keys, or
2106 sensitive plaintext).
2111 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2113 * Fixed a problem with verifying a certificate chain when using the
2114 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2115 the certificates present in a certificate chain. It is not set by default.
2117 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2118 the chain that have explicitly encoded elliptic curve parameters was added
2119 as an additional strict check.
2121 An error in the implementation of this check meant that the result of a
2122 previous check to confirm that certificates in the chain are valid CA
2123 certificates was overwritten. This effectively bypasses the check
2124 that non-CA certificates must not be able to issue other certificates.
2126 If a "purpose" has been configured then there is a subsequent opportunity
2127 for checks that the certificate is a valid CA. All of the named "purpose"
2128 values implemented in libcrypto perform this check. Therefore, where
2129 a purpose is set the certificate chain will still be rejected even when the
2130 strict flag has been used. A purpose is set by default in libssl client and
2131 server certificate verification routines, but it can be overridden or
2132 removed by an application.
2134 In order to be affected, an application must explicitly set the
2135 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2136 for the certificate verification or, in the case of TLS client or server
2137 applications, override the default purpose.
2142 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2143 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2144 renegotiation ClientHello omits the signature_algorithms extension (where it
2145 was present in the initial ClientHello), but includes a
2146 signature_algorithms_cert extension then a NULL pointer dereference will
2147 result, leading to a crash and a denial of service attack.
2149 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2150 (which is the default configuration). OpenSSL TLS clients are not impacted by
2154 *Peter Kästle and Samuel Sapalski*
2156 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2158 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2159 create a unique hash value based on the issuer and serial number data
2160 contained within an X509 certificate. However, it was failing to correctly
2161 handle any errors that may occur while parsing the issuer field (which might
2162 occur if the issuer field is maliciously constructed). This may subsequently
2163 result in a NULL pointer deref and a crash leading to a potential denial of
2169 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2170 padding mode to correctly check for rollback attacks. This is considered a
2171 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2176 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2177 functions. Previously they could overflow the output length argument in some
2178 cases where the input length is close to the maximum permissible length for
2179 an integer on the platform. In such cases the return value from the function
2180 call would be 1 (indicating success), but the output length value would be
2181 negative. This could cause applications to behave incorrectly or crash.
2186 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2187 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2188 could be exploited in a side channel attack to recover the password. Since
2189 the attack is local host only this is outside of the current OpenSSL
2190 threat model and therefore no CVE is assigned.
2192 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2197 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2199 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2200 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2201 If an attacker can control both items being compared then this could lead
2202 to a possible denial of service attack. OpenSSL itself uses the
2203 GENERAL_NAME_cmp function for two purposes:
2204 1) Comparing CRL distribution point names between an available CRL and a
2205 CRL distribution point embedded in an X509 certificate
2206 2) When verifying that a timestamp response token signer matches the
2207 timestamp authority name (exposed via the API functions
2208 TS_RESP_verify_response and TS_RESP_verify_token)
2213 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2215 * Certificates with explicit curve parameters are now disallowed in
2216 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2220 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2221 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2222 conversely, silently ignore DTLS protocol version bounds when configuring
2223 TLS-based contexts. The commands can be repeated to set bounds of both
2224 types. The same applies with the corresponding "min_protocol" and
2225 "max_protocol" command-line switches, in case some application uses both TLS
2228 SSL_CTX instances that are created for a fixed protocol version (e.g.
2229 TLSv1_server_method()) also silently ignore version bounds. Previously
2230 attempts to apply bounds to these protocol versions would result in an
2231 error. Now only the "version-flexible" SSL_CTX instances are subject to
2232 limits in configuration files in command-line options.
2236 * Handshake now fails if Extended Master Secret extension is dropped
2241 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2243 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2245 * Fixed segmentation fault in SSL_check_chain()
2246 Server or client applications that call the SSL_check_chain() function
2247 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2248 dereference as a result of incorrect handling of the
2249 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2250 or unrecognised signature algorithm is received from the peer. This could
2251 be exploited by a malicious peer in a Denial of Service attack.
2256 * Added AES consttime code for no-asm configurations
2257 an optional constant time support for AES was added
2258 when building openssl for no-asm.
2259 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2260 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2261 At this time this feature is by default disabled.
2262 It will be enabled by default in 3.0.
2266 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2268 * Revert the change of EOF detection while reading in libssl to avoid
2269 regressions in applications depending on the current way of reporting
2270 the EOF. As the existing method is not fully accurate the change to
2271 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2272 branch and will be present in the 3.0 release.
2276 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2277 when primes for RSA keys are computed.
2278 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2279 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2280 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2281 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2282 This avoids possible fingerprinting of newly generated RSA modules.
2286 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2288 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2289 while reading in libssl then we would report an error back to the
2290 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2291 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2292 therefore give a hint as to what went wrong.
2296 * Check that ed25519 and ed448 are allowed by the security level. Previously
2297 signature algorithms not using an MD were not being checked that they were
2298 allowed by the security level.
2302 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2303 was not quite right. The behaviour was not consistent between resumption
2304 and normal handshakes, and also not quite consistent with historical
2305 behaviour. The behaviour in various scenarios has been clarified and
2306 it has been updated to make it match historical behaviour as closely as
2311 * *[VMS only]* The header files that the VMS compilers include automatically,
2312 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2313 that the C++ compiler doesn't understand. This is a shortcoming in the
2314 compiler, but can be worked around with `__cplusplus` guards.
2316 C++ applications that use OpenSSL libraries must be compiled using the
2317 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2318 functions. Otherwise, only functions with symbols of less than 31
2319 characters can be used, as the linker will not be able to successfully
2320 resolve symbols with longer names.
2324 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2325 The presence of this system service is determined at run-time.
2329 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2334 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2336 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2337 number generator (RNG). This was intended to include protection in the
2338 event of a fork() system call in order to ensure that the parent and child
2339 processes did not share the same RNG state. However, this protection was not
2340 being used in the default case.
2342 A partial mitigation for this issue is that the output from a high
2343 precision timer is mixed into the RNG state so the likelihood of a parent
2344 and child process sharing state is significantly reduced.
2346 If an application already calls OPENSSL_init_crypto() explicitly using
2347 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2350 *Matthias St. Pierre*
2352 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2353 used even when parsing explicit parameters, when loading a encoded key
2354 or calling `EC_GROUP_new_from_ecpkparameters()`/
2355 `EC_GROUP_new_from_ecparameters()`.
2356 This prevents bypass of security hardening and performance gains,
2357 especially for curves with specialized EC_METHODs.
2358 By default, if a key encoded with explicit parameters is loaded and later
2359 encoded, the output is still encoded with explicit parameters, even if
2360 internally a "named" EC_GROUP is used for computation.
2364 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2365 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2366 NULL. After this change, only the cofactor parameter can be NULL. It also
2367 does some minimal sanity checks on the passed order.
2372 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2373 An attack is simple, if the first CMS_recipientInfo is valid but the
2374 second CMS_recipientInfo is chosen ciphertext. If the second
2375 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2376 encryption key will be replaced by garbage, and the message cannot be
2377 decoded, but if the RSA decryption fails, the correct encryption key is
2378 used and the recipient will not notice the attack.
2379 As a work around for this potential attack the length of the decrypted
2380 key must be equal to the cipher default key length, in case the
2381 certificate is not given and all recipientInfo are tried out.
2382 The old behaviour can be re-enabled in the CMS code by setting the
2383 CMS_DEBUG_DECRYPT flag.
2388 * Early start up entropy quality from the DEVRANDOM seed source has been
2389 improved for older Linux systems. The RAND subsystem will wait for
2390 /dev/random to be producing output before seeding from /dev/urandom.
2391 The seeded state is stored for future library initialisations using
2392 a system global shared memory segment. The shared memory identifier
2393 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2394 the desired value. The default identifier is 114.
2398 * Correct the extended master secret constant on EBCDIC systems. Without this
2399 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2400 negotiate EMS will fail. Unfortunately this also means that TLS connections
2401 between EBCDIC systems with this fix, and EBCDIC systems without this
2402 fix will fail if they negotiate EMS.
2406 * Use Windows installation paths in the mingw builds
2408 Mingw isn't a POSIX environment per se, which means that Windows
2409 paths should be used for installation.
2414 * Changed DH_check to accept parameters with order q and 2q subgroups.
2415 With order 2q subgroups the bit 0 of the private key is not secret
2416 but DH_generate_key works around that by clearing bit 0 of the
2417 private key for those. This avoids leaking bit 0 of the private key.
2421 * Significantly reduce secure memory usage by the randomness pools.
2425 * Revert the DEVRANDOM_WAIT feature for Linux systems
2427 The DEVRANDOM_WAIT feature added a select() call to wait for the
2428 /dev/random device to become readable before reading from the
2429 /dev/urandom device.
2431 It turned out that this change had negative side effects on
2432 performance which were not acceptable. After some discussion it
2433 was decided to revert this feature and leave it up to the OS
2434 resp. the platform maintainer to ensure a proper initialization
2435 during early boot time.
2437 *Matthias St. Pierre*
2439 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2441 * Add build tests for C++. These are generated files that only do one
2442 thing, to include one public OpenSSL head file each. This tests that
2443 the public header files can be usefully included in a C++ application.
2445 This test isn't enabled by default. It can be enabled with the option
2446 'enable-buildtest-c++'.
2450 * Enable SHA3 pre-hashing for ECDSA and DSA.
2454 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2455 This changes the size when using the `genpkey` command when no size is given.
2456 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2457 generation commands to use 2048 bits by default.
2461 * Reorganize the manual pages to consistently have RETURN VALUES,
2462 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2463 util/fix-doc-nits accordingly.
2465 *Paul Yang, Joshua Lock*
2467 * Add the missing accessor EVP_PKEY_get0_engine()
2471 * Have commands like `s_client` and `s_server` output the signature scheme
2472 along with other cipher suite parameters when debugging.
2476 * Make OPENSSL_config() error agnostic again.
2480 * Do the error handling in RSA decryption constant time.
2484 * Prevent over long nonces in ChaCha20-Poly1305.
2486 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2487 for every encryption operation. RFC 7539 specifies that the nonce value
2488 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2489 and front pads the nonce with 0 bytes if it is less than 12
2490 bytes. However it also incorrectly allows a nonce to be set of up to 16
2491 bytes. In this case only the last 12 bytes are significant and any
2492 additional leading bytes are ignored.
2494 It is a requirement of using this cipher that nonce values are
2495 unique. Messages encrypted using a reused nonce value are susceptible to
2496 serious confidentiality and integrity attacks. If an application changes
2497 the default nonce length to be longer than 12 bytes and then makes a
2498 change to the leading bytes of the nonce expecting the new value to be a
2499 new unique nonce then such an application could inadvertently encrypt
2500 messages with a reused nonce.
2502 Additionally the ignored bytes in a long nonce are not covered by the
2503 integrity guarantee of this cipher. Any application that relies on the
2504 integrity of these ignored leading bytes of a long nonce may be further
2505 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2506 is safe because no such use sets such a long nonce value. However user
2507 applications that use this cipher directly and set a non-default nonce
2508 length to be longer than 12 bytes may be vulnerable.
2510 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2516 * Add DEVRANDOM_WAIT feature for Linux systems
2518 On older Linux systems where the getrandom() system call is not available,
2519 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2520 Contrary to getrandom(), the /dev/urandom device will not block during
2521 early boot when the kernel CSPRNG has not been seeded yet.
2523 To mitigate this known weakness, use select() to wait for /dev/random to
2524 become readable before reading from /dev/urandom.
2526 * Ensure that SM2 only uses SM3 as digest algorithm
2530 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2532 * Change the info callback signals for the start and end of a post-handshake
2533 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2534 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2535 confused by this and assume that a TLSv1.2 renegotiation has started. This
2536 can break KeyUpdate handling. Instead we no longer signal the start and end
2537 of a post handshake message exchange (although the messages themselves are
2538 still signalled). This could break some applications that were expecting
2539 the old signals. However without this KeyUpdate is not usable for many
2544 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2546 * Timing vulnerability in DSA signature generation
2548 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2549 timing side channel attack. An attacker could use variations in the signing
2550 algorithm to recover the private key.
2552 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2557 * Timing vulnerability in ECDSA signature generation
2559 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2560 timing side channel attack. An attacker could use variations in the signing
2561 algorithm to recover the private key.
2563 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2568 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2569 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2570 of two gigabytes and the error handling improved.
2572 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2573 categorized as a normal bug, not a security issue, because the DRBG reseeds
2574 automatically and is fully functional even without additional randomness
2575 provided by the application.
2577 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2579 * Add a new ClientHello callback. Provides a callback interface that gives
2580 the application the ability to adjust the nascent SSL object at the
2581 earliest stage of ClientHello processing, immediately after extensions have
2582 been collected but before they have been processed. In particular, this
2583 callback can adjust the supported TLS versions in response to the contents
2588 * Add SM2 base algorithm support.
2592 * s390x assembly pack: add (improved) hardware-support for the following
2593 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2594 aes-cfb/cfb8, aes-ecb.
2598 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2599 parameter is no longer accepted, as it leads to a corrupt table. NULL
2600 pem_str is reserved for alias entries only.
2604 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2605 step for prime curves. The new implementation is based on formulae from
2606 differential addition-and-doubling in homogeneous projective coordinates
2607 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2608 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2609 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2610 to work in projective coordinates.
2612 *Billy Bob Brumley, Nicola Tuveri*
2614 * Change generating and checking of primes so that the error rate of not
2615 being prime depends on the intended use based on the size of the input.
2616 For larger primes this will result in more rounds of Miller-Rabin.
2617 The maximal error rate for primes with more than 1080 bits is lowered
2620 *Kurt Roeckx, Annie Yousar*
2622 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2626 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2627 moving between systems, and to avoid confusion when a Windows build is
2628 done with mingw vs with MSVC. For POSIX installs, there's still a
2629 symlink or copy named 'tsget' to avoid that confusion as well.
2633 * Revert blinding in ECDSA sign and instead make problematic addition
2634 length-invariant. Switch even to fixed-length Montgomery multiplication.
2638 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2639 step for binary curves. The new implementation is based on formulae from
2640 differential addition-and-doubling in mixed Lopez-Dahab projective
2641 coordinates, modified to independently blind the operands.
2643 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2645 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2646 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2647 EC_METHODs to implement their own specialized "ladder step", to take
2648 advantage of more favorable coordinate systems or more efficient
2649 differential addition-and-doubling algorithms.
2651 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2653 * Modified the random device based seed sources to keep the relevant
2654 file descriptors open rather than reopening them on each access.
2655 This allows such sources to operate in a chroot() jail without
2656 the associated device nodes being available. This behaviour can be
2657 controlled using RAND_keep_random_devices_open().
2661 * Numerous side-channel attack mitigations have been applied. This may have
2662 performance impacts for some algorithms for the benefit of improved
2663 security. Specific changes are noted in this change log by their respective
2668 * AIX shared library support overhaul. Switch to AIX "natural" way of
2669 handling shared libraries, which means collecting shared objects of
2670 different versions and bitnesses in one common archive. This allows to
2671 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2672 doesn't affect the way 3rd party applications are linked, only how
2673 multi-version installation is managed.
2677 * Make ec_group_do_inverse_ord() more robust and available to other
2678 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2679 mitigations are applied to the fallback BN_mod_inverse().
2680 When using this function rather than BN_mod_inverse() directly, new
2681 EC cryptosystem implementations are then safer-by-default.
2685 * Add coordinate blinding for EC_POINT and implement projective
2686 coordinate blinding for generic prime curves as a countermeasure to
2687 chosen point SCA attacks.
2689 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2691 * Add blinding to ECDSA and DSA signatures to protect against side channel
2692 attacks discovered by Keegan Ryan (NCC Group).
2696 * Enforce checking in the `pkeyutl` command to ensure that the input
2697 length does not exceed the maximum supported digest length when performing
2698 a sign, verify or verifyrecover operation.
2702 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2703 I/O in combination with something like select() or poll() will hang. This
2704 can be turned off again using SSL_CTX_clear_mode().
2705 Many applications do not properly handle non-application data records, and
2706 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2707 around the problems in those applications, but can also break some.
2708 It's recommended to read the manpages about SSL_read(), SSL_write(),
2709 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2710 SSL_CTX_set_read_ahead() again.
2714 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2715 now allow empty (zero character) pass phrases.
2719 * Apply blinding to binary field modular inversion and remove patent
2720 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2724 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2725 binary and prime elliptic curves.
2729 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2730 constant time fixed point multiplication.
2734 * Revise elliptic curve scalar multiplication with timing attack
2735 defenses: ec_wNAF_mul redirects to a constant time implementation
2736 when computing fixed point and variable point multiplication (which
2737 in OpenSSL are mostly used with secret scalars in keygen, sign,
2738 ECDH derive operations).
2739 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2742 * Updated CONTRIBUTING
2746 * Updated DRBG / RAND to request nonce and additional low entropy
2747 randomness from the system.
2749 *Matthias St. Pierre*
2751 * Updated 'openssl rehash' to use OpenSSL consistent default.
2755 * Moved the load of the ssl_conf module to libcrypto, which helps
2756 loading engines that libssl uses before libssl is initialised.
2760 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2764 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2766 *Ingo Schwarze, Rich Salz*
2768 * Added output of accepting IP address and port for 'openssl s_server'
2772 * Added a new API for TLSv1.3 ciphersuites:
2773 SSL_CTX_set_ciphersuites()
2774 SSL_set_ciphersuites()
2778 * Memory allocation failures consistently add an error to the error
2783 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2784 in libcrypto when run as setuid/setgid.
2788 * Load any config file by default when libssl is used.
2792 * Added new public header file <openssl/rand_drbg.h> and documentation
2793 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2795 *Matthias St. Pierre*
2797 * QNX support removed (cannot find contributors to get their approval
2798 for the license change).
2802 * TLSv1.3 replay protection for early data has been implemented. See the
2803 SSL_read_early_data() man page for further details.
2807 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2808 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2809 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2810 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2811 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2812 configuration has been separated out. See the ciphers man page or the
2813 SSL_CTX_set_ciphersuites() man page for more information.
2817 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2818 in responder mode now supports the new "-multi" option, which
2819 spawns the specified number of child processes to handle OCSP
2820 requests. The "-timeout" option now also limits the OCSP
2821 responder's patience to wait to receive the full client request
2822 on a newly accepted connection. Child processes are respawned
2823 as needed, and the CA index file is automatically reloaded
2824 when changed. This makes it possible to run the "ocsp" responder
2825 as a long-running service, making the OpenSSL CA somewhat more
2826 feature-complete. In this mode, most diagnostic messages logged
2827 after entering the event loop are logged via syslog(3) rather than
2832 * Added support for X448 and Ed448. Heavily based on original work by
2837 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2838 objects loaded. This adds the functions OSSL_STORE_expect() and
2839 OSSL_STORE_find() as well as needed tools to construct searches and
2840 get the search data out of them.
2844 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2845 version of OpenSSL should review their configuration settings to ensure
2846 that they are still appropriate for TLSv1.3. For further information see:
2847 <https://wiki.openssl.org/index.php/TLS1.3>
2851 * Grand redesign of the OpenSSL random generator
2853 The default RAND method now utilizes an AES-CTR DRBG according to
2854 NIST standard SP 800-90Ar1. The new random generator is essentially
2855 a port of the default random generator from the OpenSSL FIPS 2.0
2856 object module. It is a hybrid deterministic random bit generator
2857 using an AES-CTR bit stream and which seeds and reseeds itself
2858 automatically using trusted system entropy sources.
2860 Some of its new features are:
2861 - Support for multiple DRBG instances with seed chaining.
2862 - The default RAND method makes use of a DRBG.
2863 - There is a public and private DRBG instance.
2864 - The DRBG instances are fork-safe.
2865 - Keep all global DRBG instances on the secure heap if it is enabled.
2866 - The public and private DRBG instance are per thread for lock free
2869 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2871 * Changed Configure so it only says what it does and doesn't dump
2872 so much data. Instead, ./configdata.pm should be used as a script
2873 to display all sorts of configuration data.
2877 * Added processing of "make variables" to Configure.
2881 * Added SHA512/224 and SHA512/256 algorithm support.
2885 * The last traces of Netware support, first removed in 1.1.0, have
2890 * Get rid of Makefile.shared, and in the process, make the processing
2891 of certain files (rc.obj, or the .def/.map/.opt files produced from
2892 the ordinal files) more visible and hopefully easier to trace and
2893 debug (or make silent).
2897 * Make it possible to have environment variable assignments as
2898 arguments to config / Configure.
2902 * Add multi-prime RSA (RFC 8017) support.
2906 * Add SM3 implemented according to GB/T 32905-2016
2907 *Jack Lloyd <jack.lloyd@ribose.com>,*
2908 *Ronald Tse <ronald.tse@ribose.com>,*
2909 *Erick Borsboom <erick.borsboom@ribose.com>*
2911 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2912 as documented in RFC6066.
2913 Based on a patch from Tomasz Moń
2915 *Filipe Raimundo da Silva*
2917 * Add SM4 implemented according to GB/T 32907-2016.
2918 *Jack Lloyd <jack.lloyd@ribose.com>,*
2919 *Ronald Tse <ronald.tse@ribose.com>,*
2920 *Erick Borsboom <erick.borsboom@ribose.com>*
2922 * Reimplement -newreq-nodes and ERR_error_string_n; the
2923 original author does not agree with the license change.
2927 * Add ARIA AEAD TLS support.
2931 * Some macro definitions to support VS6 have been removed. Visual
2932 Studio 6 has not worked since 1.1.0
2936 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2937 without clearing the errors.
2941 * Add "atfork" functions. If building on a system that without
2942 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2943 requirements. The RAND facility now uses/requires this.
2951 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2952 not possible to disable entirely. However, it's still possible to
2953 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2956 To disable, configure with 'no-ui-console'. 'no-ui' is still
2957 possible to use as an alias. Check at compile time with the
2958 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2959 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2963 * Add a STORE module, which implements a uniform and URI based reader of
2964 stores that can contain keys, certificates, CRLs and numerous other
2965 objects. The main API is loosely based on a few stdio functions,
2966 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2967 OSSL_STORE_error and OSSL_STORE_close.
2968 The implementation uses backends called "loaders" to implement arbitrary
2969 URI schemes. There is one built in "loader" for the 'file' scheme.
2973 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2974 then adjusted to work on FreeBSD 8.4 as well.
2975 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2976 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2980 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2981 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2982 error code calls like this:
2984 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2986 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2987 that can be encoded in C. For the foreseeable future, this will only
2990 *Richard Levitte and Tim Hudson*
2992 * Removed BSD cryptodev engine.
2996 * Add a build target 'build_all_generated', to build all generated files
2997 and only that. This can be used to prepare everything that requires
2998 things like perl for a system that lacks perl and then move everything
2999 to that system and do the rest of the build there.
3003 * In the UI interface, make it possible to duplicate the user data. This
3004 can be used by engines that need to retain the data for a longer time
3005 than just the call where this user data is passed.
3009 * Ignore the '-named_curve auto' value for compatibility of applications
3012 *Tomáš Mráz <tmraz@fedoraproject.org>*
3014 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3015 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3016 alerts across multiple records (some of which could be empty). In practice
3017 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3018 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3019 support this at all. Supporting it adds significant complexity to the
3020 record layer, and its removal is unlikely to cause interoperability
3025 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3026 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3027 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3032 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3033 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3035 *Richard Levitte, Andy Polyakov*
3037 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3042 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3043 platform rather than 'mingw'.
3047 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3048 success if they are asked to add an object which already exists
3049 in the store. This change cascades to other functions which load
3050 certificates and CRLs.
3054 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3055 facilitate stack unwinding even from assembly subroutines.
3059 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3060 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3064 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3065 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3066 which is the minimum version we support.
3070 * Certificate time validation (X509_cmp_time) enforces stricter
3071 compliance with RFC 5280. Fractional seconds and timezone offsets
3072 are no longer allowed.
3076 * Add support for ARIA
3080 * s_client will now send the Server Name Indication (SNI) extension by
3081 default unless the new "-noservername" option is used. The server name is
3082 based on the host provided to the "-connect" option unless overridden by
3083 using "-servername".
3087 * Add support for SipHash
3091 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3092 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3093 prevent issues where no progress is being made and the peer continually
3094 sends unrecognised record types, using up resources processing them.
3098 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3099 using the algorithm defined in
3100 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3104 * Heartbeat support has been removed; the ABI is changed for now.
3106 *Richard Levitte, Rich Salz*
3108 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3112 * The RSA "null" method, which was partially supported to avoid patent
3113 issues, has been replaced to always returns NULL.
3120 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3122 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3123 used even when parsing explicit parameters, when loading a encoded key
3124 or calling `EC_GROUP_new_from_ecpkparameters()`/
3125 `EC_GROUP_new_from_ecparameters()`.
3126 This prevents bypass of security hardening and performance gains,
3127 especially for curves with specialized EC_METHODs.
3128 By default, if a key encoded with explicit parameters is loaded and later
3129 encoded, the output is still encoded with explicit parameters, even if
3130 internally a "named" EC_GROUP is used for computation.
3134 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3135 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3136 NULL. After this change, only the cofactor parameter can be NULL. It also
3137 does some minimal sanity checks on the passed order.
3142 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3143 An attack is simple, if the first CMS_recipientInfo is valid but the
3144 second CMS_recipientInfo is chosen ciphertext. If the second
3145 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3146 encryption key will be replaced by garbage, and the message cannot be
3147 decoded, but if the RSA decryption fails, the correct encryption key is
3148 used and the recipient will not notice the attack.
3149 As a work around for this potential attack the length of the decrypted
3150 key must be equal to the cipher default key length, in case the
3151 certificate is not given and all recipientInfo are tried out.
3152 The old behaviour can be re-enabled in the CMS code by setting the
3153 CMS_DEBUG_DECRYPT flag.
3158 * Use Windows installation paths in the mingw builds
3160 Mingw isn't a POSIX environment per se, which means that Windows
3161 paths should be used for installation.
3166 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3168 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3169 This changes the size when using the `genpkey` command when no size is given.
3170 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3171 generation commands to use 2048 bits by default.
3175 * Prevent over long nonces in ChaCha20-Poly1305.
3177 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3178 for every encryption operation. RFC 7539 specifies that the nonce value
3179 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3180 and front pads the nonce with 0 bytes if it is less than 12
3181 bytes. However it also incorrectly allows a nonce to be set of up to 16
3182 bytes. In this case only the last 12 bytes are significant and any
3183 additional leading bytes are ignored.
3185 It is a requirement of using this cipher that nonce values are
3186 unique. Messages encrypted using a reused nonce value are susceptible to
3187 serious confidentiality and integrity attacks. If an application changes
3188 the default nonce length to be longer than 12 bytes and then makes a
3189 change to the leading bytes of the nonce expecting the new value to be a
3190 new unique nonce then such an application could inadvertently encrypt
3191 messages with a reused nonce.
3193 Additionally the ignored bytes in a long nonce are not covered by the
3194 integrity guarantee of this cipher. Any application that relies on the
3195 integrity of these ignored leading bytes of a long nonce may be further
3196 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3197 is safe because no such use sets such a long nonce value. However user
3198 applications that use this cipher directly and set a non-default nonce
3199 length to be longer than 12 bytes may be vulnerable.
3201 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3207 * Added SCA hardening for modular field inversion in EC_GROUP through
3208 a new dedicated field_inv() pointer in EC_METHOD.
3209 This also addresses a leakage affecting conversions from projective
3210 to affine coordinates.
3212 *Billy Bob Brumley, Nicola Tuveri*
3214 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3215 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3219 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3223 * Remove the 'dist' target and add a tarball building script. The
3224 'dist' target has fallen out of use, and it shouldn't be
3225 necessary to configure just to create a source distribution.
3229 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3231 * Timing vulnerability in DSA signature generation
3233 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3234 timing side channel attack. An attacker could use variations in the signing
3235 algorithm to recover the private key.
3237 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3242 * Timing vulnerability in ECDSA signature generation
3244 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3245 timing side channel attack. An attacker could use variations in the signing
3246 algorithm to recover the private key.
3248 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3253 * Add coordinate blinding for EC_POINT and implement projective
3254 coordinate blinding for generic prime curves as a countermeasure to
3255 chosen point SCA attacks.
3257 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3259 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3261 * Client DoS due to large DH parameter
3263 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3264 malicious server can send a very large prime value to the client. This will
3265 cause the client to spend an unreasonably long period of time generating a
3266 key for this prime resulting in a hang until the client has finished. This
3267 could be exploited in a Denial Of Service attack.
3269 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3274 * Cache timing vulnerability in RSA Key Generation
3276 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3277 a cache timing side channel attack. An attacker with sufficient access to
3278 mount cache timing attacks during the RSA key generation process could
3279 recover the private key.
3281 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3282 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3287 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3288 parameter is no longer accepted, as it leads to a corrupt table. NULL
3289 pem_str is reserved for alias entries only.
3293 * Revert blinding in ECDSA sign and instead make problematic addition
3294 length-invariant. Switch even to fixed-length Montgomery multiplication.
3298 * Change generating and checking of primes so that the error rate of not
3299 being prime depends on the intended use based on the size of the input.
3300 For larger primes this will result in more rounds of Miller-Rabin.
3301 The maximal error rate for primes with more than 1080 bits is lowered
3304 *Kurt Roeckx, Annie Yousar*
3306 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3310 * Add blinding to ECDSA and DSA signatures to protect against side channel
3311 attacks discovered by Keegan Ryan (NCC Group).
3315 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3316 now allow empty (zero character) pass phrases.
3320 * Certificate time validation (X509_cmp_time) enforces stricter
3321 compliance with RFC 5280. Fractional seconds and timezone offsets
3322 are no longer allowed.
3326 * Fixed a text canonicalisation bug in CMS
3328 Where a CMS detached signature is used with text content the text goes
3329 through a canonicalisation process first prior to signing or verifying a
3330 signature. This process strips trailing space at the end of lines, converts
3331 line terminators to CRLF and removes additional trailing line terminators
3332 at the end of a file. A bug in the canonicalisation process meant that
3333 some characters, such as form-feed, were incorrectly treated as whitespace
3334 and removed. This is contrary to the specification (RFC5485). This fix
3335 could mean that detached text data signed with an earlier version of
3336 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3337 signed with a fixed OpenSSL may fail to verify with an earlier version of
3338 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3339 and use the "-binary" flag (for the "cms" command line application) or set
3340 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3344 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3346 * Constructed ASN.1 types with a recursive definition could exceed the stack
3348 Constructed ASN.1 types with a recursive definition (such as can be found
3349 in PKCS7) could eventually exceed the stack given malicious input with
3350 excessive recursion. This could result in a Denial Of Service attack. There
3351 are no such structures used within SSL/TLS that come from untrusted sources
3352 so this is considered safe.
3354 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3360 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3362 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3363 effectively reduced to only comparing the least significant bit of each
3364 byte. This allows an attacker to forge messages that would be considered as
3365 authenticated in an amount of tries lower than that guaranteed by the
3366 security claims of the scheme. The module can only be compiled by the
3367 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3369 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3375 * Add a build target 'build_all_generated', to build all generated files
3376 and only that. This can be used to prepare everything that requires
3377 things like perl for a system that lacks perl and then move everything
3378 to that system and do the rest of the build there.
3382 * Backport SSL_OP_NO_RENGOTIATION
3384 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3385 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3386 changes this is no longer possible in 1.1.0. Therefore, the new
3387 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3388 1.1.0 to provide equivalent functionality.
3390 Note that if an application built against 1.1.0h headers (or above) is run
3391 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3392 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3396 * Removed the OS390-Unix config target. It relied on a script that doesn't
3401 * rsaz_1024_mul_avx2 overflow bug on x86_64
3403 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3404 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3405 Analysis suggests that attacks against RSA and DSA as a result of this
3406 defect would be very difficult to perform and are not believed likely.
3407 Attacks against DH1024 are considered just feasible, because most of the
3408 work necessary to deduce information about a private key may be performed
3409 offline. The amount of resources required for such an attack would be
3410 significant. However, for an attack on TLS to be meaningful, the server
3411 would have to share the DH1024 private key among multiple clients, which is
3412 no longer an option since CVE-2016-0701.
3414 This only affects processors that support the AVX2 but not ADX extensions
3415 like Intel Haswell (4th generation).
3417 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3418 was originally found via the OSS-Fuzz project.
3423 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3425 * bn_sqrx8x_internal carry bug on x86_64
3427 There is a carry propagating bug in the x86_64 Montgomery squaring
3428 procedure. No EC algorithms are affected. Analysis suggests that attacks
3429 against RSA and DSA as a result of this defect would be very difficult to
3430 perform and are not believed likely. Attacks against DH are considered just
3431 feasible (although very difficult) because most of the work necessary to
3432 deduce information about a private key may be performed offline. The amount
3433 of resources required for such an attack would be very significant and
3434 likely only accessible to a limited number of attackers. An attacker would
3435 additionally need online access to an unpatched system using the target
3436 private key in a scenario with persistent DH parameters and a private
3437 key that is shared between multiple clients.
3439 This only affects processors that support the BMI1, BMI2 and ADX extensions
3440 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3442 This issue was reported to OpenSSL by the OSS-Fuzz project.
3447 * Malformed X.509 IPAddressFamily could cause OOB read
3449 If an X.509 certificate has a malformed IPAddressFamily extension,
3450 OpenSSL could do a one-byte buffer overread. The most likely result
3451 would be an erroneous display of the certificate in text format.
3453 This issue was reported to OpenSSL by the OSS-Fuzz project.
3458 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3460 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3461 platform rather than 'mingw'.
3465 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3466 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3467 which is the minimum version we support.
3471 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3473 * Encrypt-Then-Mac renegotiation crash
3475 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3476 negotiated where it was not in the original handshake (or vice-versa) then
3477 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3478 and servers are affected.
3480 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3485 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3487 * Truncated packet could crash via OOB read
3489 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3490 cipher is being used, then a truncated packet can cause that host to
3491 perform an out-of-bounds read, usually resulting in a crash.
3493 This issue was reported to OpenSSL by Robert Święcki of Google.
3498 * Bad (EC)DHE parameters cause a client crash
3500 If a malicious server supplies bad parameters for a DHE or ECDHE key
3501 exchange then this can result in the client attempting to dereference a
3502 NULL pointer leading to a client crash. This could be exploited in a Denial
3505 This issue was reported to OpenSSL by Guido Vranken.
3510 * BN_mod_exp may produce incorrect results on x86_64
3512 There is a carry propagating bug in the x86_64 Montgomery squaring
3513 procedure. No EC algorithms are affected. Analysis suggests that attacks
3514 against RSA and DSA as a result of this defect would be very difficult to
3515 perform and are not believed likely. Attacks against DH are considered just
3516 feasible (although very difficult) because most of the work necessary to
3517 deduce information about a private key may be performed offline. The amount
3518 of resources required for such an attack would be very significant and
3519 likely only accessible to a limited number of attackers. An attacker would
3520 additionally need online access to an unpatched system using the target
3521 private key in a scenario with persistent DH parameters and a private
3522 key that is shared between multiple clients. For example this can occur by
3523 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3524 similar to CVE-2015-3193 but must be treated as a separate problem.
3526 This issue was reported to OpenSSL by the OSS-Fuzz project.
3531 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3533 * ChaCha20/Poly1305 heap-buffer-overflow
3535 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3536 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3537 crash. This issue is not considered to be exploitable beyond a DoS.
3539 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3544 * CMS Null dereference
3546 Applications parsing invalid CMS structures can crash with a NULL pointer
3547 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3548 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3549 structure callback if an attempt is made to free certain invalid encodings.
3550 Only CHOICE structures using a callback which do not handle NULL value are
3553 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3558 * Montgomery multiplication may produce incorrect results
3560 There is a carry propagating bug in the Broadwell-specific Montgomery
3561 multiplication procedure that handles input lengths divisible by, but
3562 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3563 and DH private keys are impossible. This is because the subroutine in
3564 question is not used in operations with the private key itself and an input
3565 of the attacker's direct choice. Otherwise the bug can manifest itself as
3566 transient authentication and key negotiation failures or reproducible
3567 erroneous outcome of public-key operations with specially crafted input.
3568 Among EC algorithms only Brainpool P-512 curves are affected and one
3569 presumably can attack ECDH key negotiation. Impact was not analyzed in
3570 detail, because pre-requisites for attack are considered unlikely. Namely
3571 multiple clients have to choose the curve in question and the server has to
3572 share the private key among them, neither of which is default behaviour.
3573 Even then only clients that chose the curve will be affected.
3575 This issue was publicly reported as transient failures and was not
3576 initially recognized as a security issue. Thanks to Richard Morgan for
3577 providing reproducible case.
3582 * Removed automatic addition of RPATH in shared libraries and executables,
3583 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3587 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3589 * Fix Use After Free for large message sizes
3591 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3592 message larger than approx 16k is received then the underlying buffer to
3593 store the incoming message is reallocated and moved. Unfortunately a
3594 dangling pointer to the old location is left which results in an attempt to
3595 write to the previously freed location. This is likely to result in a
3596 crash, however it could potentially lead to execution of arbitrary code.
3598 This issue only affects OpenSSL 1.1.0a.
3600 This issue was reported to OpenSSL by Robert Święcki.
3605 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3607 * OCSP Status Request extension unbounded memory growth
3609 A malicious client can send an excessively large OCSP Status Request
3610 extension. If that client continually requests renegotiation, sending a
3611 large OCSP Status Request extension each time, then there will be unbounded
3612 memory growth on the server. This will eventually lead to a Denial Of
3613 Service attack through memory exhaustion. Servers with a default
3614 configuration are vulnerable even if they do not support OCSP. Builds using
3615 the "no-ocsp" build time option are not affected.
3617 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3622 * SSL_peek() hang on empty record
3624 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3625 sends an empty record. This could be exploited by a malicious peer in a
3626 Denial Of Service attack.
3628 This issue was reported to OpenSSL by Alex Gaynor.
3633 * Excessive allocation of memory in tls_get_message_header() and
3634 dtls1_preprocess_fragment()
3636 A (D)TLS message includes 3 bytes for its length in the header for the
3637 message. This would allow for messages up to 16Mb in length. Messages of
3638 this length are excessive and OpenSSL includes a check to ensure that a
3639 peer is sending reasonably sized messages in order to avoid too much memory
3640 being consumed to service a connection. A flaw in the logic of version
3641 1.1.0 means that memory for the message is allocated too early, prior to
3642 the excessive message length check. Due to way memory is allocated in
3643 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3644 to service a connection. This could lead to a Denial of Service through
3645 memory exhaustion. However, the excessive message length check still takes
3646 place, and this would cause the connection to immediately fail. Assuming
3647 that the application calls SSL_free() on the failed connection in a timely
3648 manner then the 21Mb of allocated memory will then be immediately freed
3649 again. Therefore, the excessive memory allocation will be transitory in
3650 nature. This then means that there is only a security impact if:
3652 1) The application does not call SSL_free() in a timely manner in the event
3653 that the connection fails
3655 2) The application is working in a constrained environment where there is
3656 very little free memory
3658 3) The attacker initiates multiple connection attempts such that there are
3659 multiple connections in a state where memory has been allocated for the
3660 connection; SSL_free() has not yet been called; and there is insufficient
3661 memory to service the multiple requests.
3663 Except in the instance of (1) above any Denial Of Service is likely to be
3664 transitory because as soon as the connection fails the memory is
3665 subsequently freed again in the SSL_free() call. However there is an
3666 increased risk during this period of application crashes due to the lack of
3667 memory - which would then mean a more serious Denial of Service.
3669 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3670 (CVE-2016-6307 and CVE-2016-6308)
3674 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3675 had to be removed. Primary reason is that vendor assembler can't
3676 assemble our modules with -KPIC flag. As result it, assembly
3677 support, was not even available as option. But its lack means
3678 lack of side-channel resistant code, which is incompatible with
3679 security by todays standards. Fortunately gcc is readily available
3680 prepackaged option, which we firmly point at...
3684 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3686 * Windows command-line tool supports UTF-8 opt-in option for arguments
3687 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3688 (to any value) allows Windows user to access PKCS#12 file generated
3689 with Windows CryptoAPI and protected with non-ASCII password, as well
3690 as files generated under UTF-8 locale on Linux also protected with
3695 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3696 have been disabled by default and removed from DEFAULT, just like RC4.
3697 See the RC4 item below to re-enable both.
3701 * The method for finding the storage location for the Windows RAND seed file
3702 has changed. First we check %RANDFILE%. If that is not set then we check
3703 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3704 all else fails we fall back to C:\.
3708 * The EVP_EncryptUpdate() function has had its return type changed from void
3709 to int. A return of 0 indicates and error while a return of 1 indicates
3714 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3715 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3716 off the constant time implementation for RSA, DSA and DH have been made
3717 no-ops and deprecated.
3721 * Windows RAND implementation was simplified to only get entropy by
3722 calling CryptGenRandom(). Various other RAND-related tickets
3725 *Joseph Wylie Yandle, Rich Salz*
3727 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3728 and `OPENSSL_LH_`, respectively. The old names are available
3729 with API compatibility. They new names are now completely documented.
3733 * Unify TYPE_up_ref(obj) methods signature.
3734 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3735 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3736 int (instead of void) like all others TYPE_up_ref() methods.
3737 So now these methods also check the return value of CRYPTO_atomic_add(),
3738 and the validity of object reference counter.
3740 *fdasilvayy@gmail.com*
3742 * With Windows Visual Studio builds, the .pdb files are installed
3743 alongside the installed libraries and executables. For a static
3744 library installation, ossl_static.pdb is the associate compiler
3745 generated .pdb file to be used when linking programs.
3749 * Remove openssl.spec. Packaging files belong with the packagers.
3753 * Automatic Darwin/OSX configuration has had a refresh, it will now
3754 recognise x86_64 architectures automatically. You can still decide
3755 to build for a different bitness with the environment variable
3756 KERNEL_BITS (can be 32 or 64), for example:
3758 KERNEL_BITS=32 ./config
3762 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3763 256 bit AES and HMAC with SHA256.
3767 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3771 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3775 * To enable users to have their own config files and build file templates,
3776 Configure looks in the directory indicated by the environment variable
3777 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3778 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3779 name and is used as is.
3783 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3784 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3785 X509_CERT_FILE_CTX was removed.
3789 * "shared" builds are now the default. To create only static libraries use
3790 the "no-shared" Configure option.
3794 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3795 All of these option have not worked for some while and are fundamental
3800 * Make various cleanup routines no-ops and mark them as deprecated. Most
3801 global cleanup functions are no longer required because they are handled
3802 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3803 Explicitly de-initing can cause problems (e.g. where a library that uses
3804 OpenSSL de-inits, but an application is still using it). The affected
3805 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3806 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3807 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3808 COMP_zlib_cleanup().
3812 * --strict-warnings no longer enables runtime debugging options
3813 such as REF_DEBUG. Instead, debug options are automatically
3814 enabled with '--debug' builds.
3816 *Andy Polyakov, Emilia Käsper*
3818 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3819 have been moved out of the public header files. New functions for managing
3820 these have been added.
3824 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3825 objects have been moved out of the public header files. New
3826 functions for managing these have been added.
3830 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3831 have been moved out of the public header files. New functions for managing
3832 these have been added.
3836 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3837 moved out of the public header files. New functions for managing these
3842 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3846 * Removed the mk1mf build scripts.
3850 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3851 it is always safe to #include a header now.
3855 * Removed the aged BC-32 config and all its supporting scripts
3859 * Removed support for Ultrix, Netware, and OS/2.
3863 * Add support for HKDF.
3865 *Alessandro Ghedini*
3867 * Add support for blake2b and blake2s
3871 * Added support for "pipelining". Ciphers that have the
3872 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3873 encryptions/decryptions simultaneously. There are currently no built-in
3874 ciphers with this property but the expectation is that engines will be able
3875 to offer it to significantly improve throughput. Support has been extended
3876 into libssl so that multiple records for a single connection can be
3877 processed in one go (for >=TLS 1.1).
3881 * Added the AFALG engine. This is an async capable engine which is able to
3882 offload work to the Linux kernel. In this initial version it only supports
3883 AES128-CBC. The kernel must be version 4.1.0 or greater.
3887 * OpenSSL now uses a new threading API. It is no longer necessary to
3888 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3889 are two supported threading models: pthreads and windows threads. It is
3890 also possible to configure OpenSSL at compile time for "no-threads". The
3891 old threading API should no longer be used. The functions have been
3892 replaced with "no-op" compatibility macros.
3894 *Alessandro Ghedini, Matt Caswell*
3896 * Modify behavior of ALPN to invoke callback after SNI/servername
3897 callback, such that updates to the SSL_CTX affect ALPN.
3901 * Add SSL_CIPHER queries for authentication and key-exchange.
3905 * Changes to the DEFAULT cipherlist:
3906 - Prefer (EC)DHE handshakes over plain RSA.
3907 - Prefer AEAD ciphers over legacy ciphers.
3908 - Prefer ECDSA over RSA when both certificates are available.
3909 - Prefer TLSv1.2 ciphers/PRF.
3910 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3915 * Change the ECC default curve list to be this, in order: x25519,
3916 secp256r1, secp521r1, secp384r1.
3920 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3921 disabled by default. They can be re-enabled using the
3922 enable-weak-ssl-ciphers option to Configure.
3926 * If the server has ALPN configured, but supports no protocols that the
3927 client advertises, send a fatal "no_application_protocol" alert.
3928 This behaviour is SHALL in RFC 7301, though it isn't universally
3929 implemented by other servers.
3933 * Add X25519 support.
3934 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3935 for public and private key encoding using the format documented in
3936 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3937 key generation and key derivation.
3939 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3944 * Deprecate SRP_VBASE_get_by_user.
3945 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3946 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3947 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3948 seed, even if the seed is configured.
3950 Users should use SRP_VBASE_get1_by_user instead. Note that in
3951 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3952 also that even though configuring the SRP seed attempts to hide
3953 invalid usernames by continuing the handshake with fake
3954 credentials, this behaviour is not constant time and no strong
3955 guarantees are made that the handshake is indistinguishable from
3956 that of a valid user.
3960 * Configuration change; it's now possible to build dynamic engines
3961 without having to build shared libraries and vice versa. This
3962 only applies to the engines in `engines/`, those in `crypto/engine/`
3963 will always be built into libcrypto (i.e. "static").
3965 Building dynamic engines is enabled by default; to disable, use
3966 the configuration option "disable-dynamic-engine".
3968 The only requirements for building dynamic engines are the
3969 presence of the DSO module and building with position independent
3970 code, so they will also automatically be disabled if configuring
3971 with "disable-dso" or "disable-pic".
3973 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3974 are also taken away from openssl/opensslconf.h, as they are
3979 * Configuration change; if there is a known flag to compile
3980 position independent code, it will always be applied on the
3981 libcrypto and libssl object files, and never on the application
3982 object files. This means other libraries that use routines from
3983 libcrypto / libssl can be made into shared libraries regardless
3984 of how OpenSSL was configured.
3986 If this isn't desirable, the configuration options "disable-pic"
3987 or "no-pic" can be used to disable the use of PIC. This will
3988 also disable building shared libraries and dynamic engines.
3992 * Removed JPAKE code. It was experimental and has no wide use.
3996 * The INSTALL_PREFIX Makefile variable has been renamed to
3997 DESTDIR. That makes for less confusion on what this variable
3998 is for. Also, the configuration option --install_prefix is
4003 * Heartbeat for TLS has been removed and is disabled by default
4004 for DTLS; configure with enable-heartbeats. Code that uses the
4005 old #define's might need to be updated.
4007 *Emilia Käsper, Rich Salz*
4009 * Rename REF_CHECK to REF_DEBUG.
4013 * New "unified" build system
4015 The "unified" build system is aimed to be a common system for all
4016 platforms we support. With it comes new support for VMS.
4018 This system builds supports building in a different directory tree
4019 than the source tree. It produces one Makefile (for unix family
4020 or lookalikes), or one descrip.mms (for VMS).
4022 The source of information to make the Makefile / descrip.mms is
4023 small files called 'build.info', holding the necessary
4024 information for each directory with source to compile, and a
4025 template in Configurations, like unix-Makefile.tmpl or
4028 With this change, the library names were also renamed on Windows
4029 and on VMS. They now have names that are closer to the standard
4030 on Unix, and include the major version number, and in certain
4031 cases, the architecture they are built for. See "Notes on shared
4032 libraries" in INSTALL.
4034 We rely heavily on the perl module Text::Template.
4038 * Added support for auto-initialisation and de-initialisation of the library.
4039 OpenSSL no longer requires explicit init or deinit routines to be called,
4040 except in certain circumstances. See the OPENSSL_init_crypto() and
4041 OPENSSL_init_ssl() man pages for further information.
4045 * The arguments to the DTLSv1_listen function have changed. Specifically the
4046 "peer" argument is now expected to be a BIO_ADDR object.
4048 * Rewrite of BIO networking library. The BIO library lacked consistent
4049 support of IPv6, and adding it required some more extensive
4050 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4051 which hold all types of addresses and chains of address information.
4052 It also introduces a new API, with functions like BIO_socket,
4053 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4054 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4055 have been adapted accordingly.
4059 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4064 * CRIME protection: disable compression by default, even if OpenSSL is
4065 compiled with zlib enabled. Applications can still enable compression
4066 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4067 using the SSL_CONF library to configure compression.
4071 * The signature of the session callback configured with
4072 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4073 was explicitly marked as `const unsigned char*` instead of
4078 * Always DPURIFY. Remove the use of uninitialized memory in the
4079 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4083 * Removed many obsolete configuration items, including
4084 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4085 MD2_CHAR, MD2_INT, MD2_LONG
4087 IDEA_SHORT, IDEA_LONG
4088 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4090 *Rich Salz, with advice from Andy Polyakov*
4092 * Many BN internals have been moved to an internal header file.
4094 *Rich Salz with help from Andy Polyakov*
4096 * Configuration and writing out the results from it has changed.
4097 Files such as Makefile include/openssl/opensslconf.h and are now
4098 produced through general templates, such as Makefile.in and
4099 crypto/opensslconf.h.in and some help from the perl module
4102 Also, the center of configuration information is no longer
4103 Makefile. Instead, Configure produces a perl module in
4104 configdata.pm which holds most of the config data (in the hash
4105 table %config), the target data that comes from the target
4106 configuration in one of the `Configurations/*.conf` files (in
4111 * To clarify their intended purposes, the Configure options
4112 --prefix and --openssldir change their semantics, and become more
4113 straightforward and less interdependent.
4115 --prefix shall be used exclusively to give the location INSTALLTOP
4116 where programs, scripts, libraries, include files and manuals are
4117 going to be installed. The default is now /usr/local.
4119 --openssldir shall be used exclusively to give the default
4120 location OPENSSLDIR where certificates, private keys, CRLs are
4121 managed. This is also where the default openssl.cnf gets
4123 If the directory given with this option is a relative path, the
4124 values of both the --prefix value and the --openssldir value will
4125 be combined to become OPENSSLDIR.
4126 The default for --openssldir is INSTALLTOP/ssl.
4128 Anyone who uses --openssldir to specify where OpenSSL is to be
4129 installed MUST change to use --prefix instead.
4133 * The GOST engine was out of date and therefore it has been removed. An up
4134 to date GOST engine is now being maintained in an external repository.
4135 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4136 support for GOST ciphersuites (these are only activated if a GOST engine
4141 * EGD is no longer supported by default; use enable-egd when
4144 *Ben Kaduk and Rich Salz*
4146 * The distribution now has Makefile.in files, which are used to
4147 create Makefile's when Configure is run. *Configure must be run
4148 before trying to build now.*
4152 * The return value for SSL_CIPHER_description() for error conditions
4157 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4159 Obtaining and performing DNSSEC validation of TLSA records is
4160 the application's responsibility. The application provides
4161 the TLSA records of its choice to OpenSSL, and these are then
4162 used to authenticate the peer.
4164 The TLSA records need not even come from DNS. They can, for
4165 example, be used to implement local end-entity certificate or
4166 trust-anchor "pinning", where the "pin" data takes the form
4167 of TLSA records, which can augment or replace verification
4168 based on the usual WebPKI public certification authorities.
4172 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4173 continues to support deprecated interfaces in default builds.
4174 However, applications are strongly advised to compile their
4175 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4176 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4177 or the 1.1.0 releases.
4179 In environments in which all applications have been ported to
4180 not use any deprecated interfaces OpenSSL's Configure script
4181 should be used with the --api=1.1.0 option to entirely remove
4182 support for the deprecated features from the library and
4183 unconditionally disable them in the installed headers.
4184 Essentially the same effect can be achieved with the "no-deprecated"
4185 argument to Configure, except that this will always restrict
4186 the build to just the latest API, rather than a fixed API
4189 As applications are ported to future revisions of the API,
4190 they should update their compile-time OPENSSL_API_COMPAT define
4191 accordingly, but in most cases should be able to continue to
4192 compile with later releases.
4194 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4195 0x10000000L and 0x00908000L, respectively. However those
4196 versions did not support the OPENSSL_API_COMPAT feature, and
4197 so applications are not typically tested for explicit support
4198 of just the undeprecated features of either release.
4202 * Add support for setting the minimum and maximum supported protocol.
4203 It can bet set via the SSL_set_min_proto_version() and
4204 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4205 MaxProtocol. It's recommended to use the new APIs to disable
4206 protocols instead of disabling individual protocols using
4207 SSL_set_options() or SSL_CONF's Protocol. This change also
4208 removes support for disabling TLS 1.2 in the OpenSSL TLS
4209 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4213 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4217 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4218 and integrates ECDSA and ECDH functionality into EC. Implementations can
4219 now redirect key generation and no longer need to convert to or from
4222 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4223 include the ec.h header file instead.
4227 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4228 ciphers who are no longer supported and drops support the ephemeral RSA key
4229 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4233 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4234 opaque. For HMAC_CTX, the following constructors and destructors
4237 HMAC_CTX *HMAC_CTX_new(void);
4238 void HMAC_CTX_free(HMAC_CTX *ctx);
4240 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4241 destroy such methods has been added. See EVP_MD_meth_new(3) and
4242 EVP_CIPHER_meth_new(3) for documentation.
4245 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4246 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4247 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4248 an already created structure.
4249 2) For consistency with the majority of our object creators and
4250 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4251 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4252 for deprecated builds.
4256 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4257 cryptographic operations to be performed asynchronously as long as an
4258 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4259 further details. Libssl has also had this capability integrated with the
4260 introduction of the new mode SSL_MODE_ASYNC and associated error
4261 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4262 pages. This work was developed in partnership with Intel Corp.
4266 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4267 always enabled now. If you want to disable the support you should
4268 exclude it using the list of supported ciphers. This also means that the
4269 "-no_ecdhe" option has been removed from s_server.
4273 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4274 SSL_{CTX_}set1_curves() which can set a list.
4278 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4279 curve you want to support using SSL_{CTX_}set1_curves().
4283 * State machine rewrite. The state machine code has been significantly
4284 refactored in order to remove much duplication of code and solve issues
4285 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4286 further details). This change does have some associated API changes.
4287 Notably the SSL_state() function has been removed and replaced by
4288 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4289 SSL_set_state() has been removed altogether. The previous handshake states
4290 defined in ssl.h and ssl3.h have also been removed.
4294 * All instances of the string "ssleay" in the public API were replaced
4295 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4296 Some error codes related to internal RSA_eay API's were renamed.
4300 * The demo files in crypto/threads were moved to demo/threads.
4304 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4307 *Matt Caswell, Rich Salz*
4309 * New ASN.1 embed macro.
4311 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4312 structure is not allocated: it is part of the parent. That is instead of
4320 This reduces memory fragmentation and make it impossible to accidentally
4321 set a mandatory field to NULL.
4323 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4324 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4325 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4330 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4334 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4335 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4336 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4337 DES and RC4 ciphersuites.
4341 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4342 This changes the decoding behaviour for some invalid messages,
4343 though the change is mostly in the more lenient direction, and
4344 legacy behaviour is preserved as much as possible.
4348 * Fix no-stdio build.
4349 *David Woodhouse <David.Woodhouse@intel.com> and also*
4350 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4352 * New testing framework
4353 The testing framework has been largely rewritten and is now using
4354 perl and the perl modules Test::Harness and an extended variant of
4355 Test::More called OpenSSL::Test to do its work. All test scripts in
4356 test/ have been rewritten into test recipes, and all direct calls to
4357 executables in test/Makefile have become individual recipes using the
4358 simplified testing OpenSSL::Test::Simple.
4360 For documentation on our testing modules, do:
4362 perldoc test/testlib/OpenSSL/Test/Simple.pm
4363 perldoc test/testlib/OpenSSL/Test.pm
4367 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4368 are used; the latter aborts on memory leaks (usually checked on exit).
4369 Some undocumented "set malloc, etc., hooks" functions were removed
4370 and others were changed. All are now documented.
4374 * In DSA_generate_parameters_ex, if the provided seed is too short,
4377 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4379 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4380 from RFC4279, RFC4785, RFC5487, RFC5489.
4382 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4383 original RSA_PSK patch.
4387 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4388 era flag was never set throughout the codebase (only read). Also removed
4389 SSL3_FLAGS_POP_BUFFER which was only used if
4390 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4394 * Changed the default name options in the "ca", "crl", "req" and "x509"
4395 to be "oneline" instead of "compat".
4399 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4400 not aware of clients that still exhibit this bug, and the workaround
4401 hasn't been working properly for a while.
4405 * The return type of BIO_number_read() and BIO_number_written() as well as
4406 the corresponding num_read and num_write members in the BIO structure has
4407 changed from unsigned long to uint64_t. On platforms where an unsigned
4408 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4413 * Given the pervasive nature of TLS extensions it is inadvisable to run
4414 OpenSSL without support for them. It also means that maintaining
4415 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4416 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4420 * Removed support for the two export grade static DH ciphersuites
4421 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4422 were newly added (along with a number of other static DH ciphersuites) to
4423 1.0.2. However the two export ones have *never* worked since they were
4424 introduced. It seems strange in any case to be adding new export
4425 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4429 * Version negotiation has been rewritten. In particular SSLv23_method(),
4430 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4431 and turned into macros which simply call the new preferred function names
4432 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4433 should use the new names instead. Also as part of this change the ssl23.h
4434 header file has been removed.
4438 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4439 code and the associated standard is no longer considered fit-for-purpose.
4443 * RT2547 was closed. When generating a private key, try to make the
4444 output file readable only by the owner. This behavior change might
4445 be noticeable when interacting with other software.
4447 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4452 * Added HTTP GET support to the ocsp command.
4456 * Changed default digest for the dgst and enc commands from MD5 to
4461 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4465 * Added support for TLS extended master secret from
4466 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4467 initial patch which was a great help during development.
4471 * All libssl internal structures have been removed from the public header
4472 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4473 now redundant). Users should not attempt to access internal structures
4474 directly. Instead they should use the provided API functions.
4478 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4479 Access to deprecated functions can be re-enabled by running config with
4480 "enable-deprecated". In addition applications wishing to use deprecated
4481 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4482 will, by default, disable some transitive includes that previously existed
4483 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4487 * Added support for OCB mode. OpenSSL has been granted a patent license
4488 compatible with the OpenSSL license for use of OCB. Details are available
4489 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4490 for OCB can be removed by calling config with no-ocb.
4494 * SSLv2 support has been removed. It still supports receiving an SSLv2
4495 compatible client hello.
4499 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4500 done while fixing the error code for the key-too-small case.
4502 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4504 * CA.sh has been removed; use CA.pl instead.
4508 * Removed old DES API.
4512 * Remove various unsupported platforms:
4518 Sinix/ReliantUNIX RM400
4523 16-bit platforms such as WIN16
4527 * Clean up OPENSSL_NO_xxx #define's
4528 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4529 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4530 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4531 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4532 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4533 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4534 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4535 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4536 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4537 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4541 * Cleaned up dead code
4542 Remove all but one '#ifdef undef' which is to be looked at.
4546 * Clean up calling of xxx_free routines.
4547 Just like free(), fix most of the xxx_free routines to accept
4548 NULL. Remove the non-null checks from callers. Save much code.
4552 * Add secure heap for storage of private keys (when possible).
4553 Add BIO_s_secmem(), CBIGNUM, etc.
4554 Contributed by Akamai Technologies under our Corporate CLA.
4558 * Experimental support for a new, fast, unbiased prime candidate generator,
4559 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4561 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4563 * New output format NSS in the sess_id command line tool. This allows
4564 exporting the session id and the master key in NSS keylog format.
4566 *Martin Kaiser <martin@kaiser.cx>*
4568 * Harmonize version and its documentation. -f flag is used to display
4571 *mancha <mancha1@zoho.com>*
4573 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4574 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4576 *mancha <mancha1@zoho.com>*
4578 * Fix some double frees. These are not thought to be exploitable.
4580 *mancha <mancha1@zoho.com>*
4582 * A missing bounds check in the handling of the TLS heartbeat extension
4583 can be used to reveal up to 64k of memory to a connected client or
4586 Thanks for Neel Mehta of Google Security for discovering this bug and to
4587 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4588 preparing the fix ([CVE-2014-0160])
4590 *Adam Langley, Bodo Moeller*
4592 * Fix for the attack described in the paper "Recovering OpenSSL
4593 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4594 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4595 <http://eprint.iacr.org/2014/140>
4597 Thanks to Yuval Yarom and Naomi Benger for discovering this
4598 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4600 *Yuval Yarom and Naomi Benger*
4602 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4603 this fixes a limitation in previous versions of OpenSSL.
4607 * Experimental encrypt-then-mac support.
4609 Experimental support for encrypt then mac from
4610 draft-gutmann-tls-encrypt-then-mac-02.txt
4612 To enable it set the appropriate extension number (0x42 for the test
4613 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4615 For non-compliant peers (i.e. just about everything) this should have no
4618 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4622 * Add EVP support for key wrapping algorithms, to avoid problems with
4623 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4624 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4625 algorithms and include tests cases.
4629 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4634 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4635 MGF1 digest and OAEP label.
4639 * Make openssl verify return errors.
4641 *Chris Palmer <palmer@google.com> and Ben Laurie*
4643 * New function ASN1_TIME_diff to calculate the difference between two
4644 ASN1_TIME structures or one structure and the current time.
4648 * Update fips_test_suite to support multiple command line options. New
4649 test to induce all self test errors in sequence and check expected
4654 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4655 sign or verify all in one operation.
4659 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4660 test programs and fips_test_suite. Includes functionality to parse
4661 the minimal script output of fipsalgest.pl directly.
4665 * Add authorisation parameter to FIPS_module_mode_set().
4669 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4673 * Use separate DRBG fields for internal and external flags. New function
4674 FIPS_drbg_health_check() to perform on demand health checking. Add
4675 generation tests to fips_test_suite with reduced health check interval to
4676 demonstrate periodic health checking. Add "nodh" option to
4677 fips_test_suite to skip very slow DH test.
4681 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4686 * More extensive health check for DRBG checking many more failure modes.
4687 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4688 combination: call this in fips_test_suite.
4692 * Add support for canonical generation of DSA parameter 'g'. See
4695 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4696 POST to handle HMAC cases.
4700 * Add functions FIPS_module_version() and FIPS_module_version_text()
4701 to return numerical and string versions of the FIPS module number.
4705 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4706 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4707 outside the validated module in the FIPS capable OpenSSL.
4711 * Minor change to DRBG entropy callback semantics. In some cases
4712 there is no multiple of the block length between min_len and
4713 max_len. Allow the callback to return more than max_len bytes
4714 of entropy but discard any extra: it is the callback's responsibility
4715 to ensure that the extra data discarded does not impact the
4716 requested amount of entropy.
4720 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4721 information in FIPS186-3, SP800-57 and SP800-131A.
4725 * CCM support via EVP. Interface is very similar to GCM case except we
4726 must supply all data in one chunk (i.e. no update, final) and the
4727 message length must be supplied if AAD is used. Add algorithm test
4732 * Initial version of POST overhaul. Add POST callback to allow the status
4733 of POST to be monitored and/or failures induced. Modify fips_test_suite
4734 to use callback. Always run all selftests even if one fails.
4738 * XTS support including algorithm test driver in the fips_gcmtest program.
4739 Note: this does increase the maximum key length from 32 to 64 bytes but
4740 there should be no binary compatibility issues as existing applications
4741 will never use XTS mode.
4745 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4746 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4747 performs algorithm blocking for unapproved PRNG types. Also do not
4748 set PRNG type in FIPS_mode_set(): leave this to the application.
4749 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4750 the standard OpenSSL PRNG: set additional data to a date time vector.
4754 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4755 This shouldn't present any incompatibility problems because applications
4756 shouldn't be using these directly and any that are will need to rethink
4757 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4761 * Extensive self tests and health checking required by SP800-90 DRBG.
4762 Remove strength parameter from FIPS_drbg_instantiate and always
4763 instantiate at maximum supported strength.
4767 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4771 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4775 * New function DH_compute_key_padded() to compute a DH key and pad with
4776 leading zeroes if needed: this complies with SP800-56A et al.
4780 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4781 anything, incomplete, subject to change and largely untested at present.
4785 * Modify fipscanisteronly build option to only build the necessary object
4786 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4790 * Add experimental option FIPSSYMS to give all symbols in
4791 fipscanister.o and FIPS or fips prefix. This will avoid
4792 conflicts with future versions of OpenSSL. Add perl script
4793 util/fipsas.pl to preprocess assembly language source files
4794 and rename any affected symbols.
4798 * Add selftest checks and algorithm block of non-fips algorithms in
4799 FIPS mode. Remove DES2 from selftests.
4803 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4804 return internal method without any ENGINE dependencies. Add new
4805 tiny fips sign and verify functions.
4809 * New build option no-ec2m to disable characteristic 2 code.
4813 * New build option "fipscanisteronly". This only builds fipscanister.o
4814 and (currently) associated fips utilities. Uses the file Makefile.fips
4815 instead of Makefile.org as the prototype.
4819 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4820 Update fips_gcmtest to use IV generator.
4824 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4825 setting output buffer to NULL. The `*Final` function must be
4826 called although it will not retrieve any additional data. The tag
4827 can be set or retrieved with a ctrl. The IV length is by default 12
4828 bytes (96 bits) but can be set to an alternative value. If the IV
4829 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4834 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4835 underlying do_cipher function handles all cipher semantics itself
4836 including padding and finalisation. This is useful if (for example)
4837 an ENGINE cipher handles block padding itself. The behaviour of
4838 do_cipher is subtly changed if this flag is set: the return value
4839 is the number of characters written to the output buffer (zero is
4840 no longer an error code) or a negative error code. Also if the
4841 input buffer is NULL and length 0 finalisation should be performed.
4845 * If a candidate issuer certificate is already part of the constructed
4846 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4850 * Improve forward-security support: add functions
4852 void SSL_CTX_set_not_resumable_session_callback(
4853 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4854 void SSL_set_not_resumable_session_callback(
4855 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4857 for use by SSL/TLS servers; the callback function will be called whenever a
4858 new session is created, and gets to decide whether the session may be
4859 cached to make it resumable (return 0) or not (return 1). (As by the
4860 SSL/TLS protocol specifications, the session_id sent by the server will be
4861 empty to indicate that the session is not resumable; also, the server will
4862 not generate RFC 4507 (RFC 5077) session tickets.)
4864 A simple reasonable callback implementation is to return is_forward_secure.
4865 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4866 by the SSL/TLS server library, indicating whether it can provide forward
4869 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4871 * New -verify_name option in command line utilities to set verification
4876 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4877 Add CMAC pkey methods.
4881 * Experimental renegotiation in s_server -www mode. If the client
4882 browses /reneg connection is renegotiated. If /renegcert it is
4883 renegotiated requesting a certificate.
4887 * Add an "external" session cache for debugging purposes to s_server. This
4888 should help trace issues which normally are only apparent in deployed
4889 multi-process servers.
4893 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4894 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4895 BIO_set_cipher() and some obscure PEM functions were changed so they
4896 can now return an error. The RAND changes required a change to the
4897 RAND_METHOD structure.
4901 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4902 a gcc attribute to warn if the result of a function is ignored. This
4903 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4904 whose return value is often ignored.
4908 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4909 These allow SCTs (signed certificate timestamps) to be requested and
4910 validated when establishing a connection.
4912 *Rob Percival <robpercival@google.com>*
4917 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4919 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4920 used even when parsing explicit parameters, when loading a encoded key
4921 or calling `EC_GROUP_new_from_ecpkparameters()`/
4922 `EC_GROUP_new_from_ecparameters()`.
4923 This prevents bypass of security hardening and performance gains,
4924 especially for curves with specialized EC_METHODs.
4925 By default, if a key encoded with explicit parameters is loaded and later
4926 encoded, the output is still encoded with explicit parameters, even if
4927 internally a "named" EC_GROUP is used for computation.
4931 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4932 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4933 NULL. After this change, only the cofactor parameter can be NULL. It also
4934 does some minimal sanity checks on the passed order.
4939 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4940 An attack is simple, if the first CMS_recipientInfo is valid but the
4941 second CMS_recipientInfo is chosen ciphertext. If the second
4942 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4943 encryption key will be replaced by garbage, and the message cannot be
4944 decoded, but if the RSA decryption fails, the correct encryption key is
4945 used and the recipient will not notice the attack.
4946 As a work around for this potential attack the length of the decrypted
4947 key must be equal to the cipher default key length, in case the
4948 certificate is not given and all recipientInfo are tried out.
4949 The old behaviour can be re-enabled in the CMS code by setting the
4950 CMS_DEBUG_DECRYPT flag.
4955 * Document issue with installation paths in diverse Windows builds
4957 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4958 binaries and run-time config file.
4963 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4965 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4966 This changes the size when using the `genpkey` command when no size is given.
4967 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4968 generation commands to use 2048 bits by default.
4972 * Add FIPS support for Android Arm 64-bit
4974 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4975 Module in Version 2.0.10. For some reason, the corresponding target
4976 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4977 built with FIPS support on Android Arm 64-bit. This omission has been
4980 *Matthias St. Pierre*
4982 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4984 * 0-byte record padding oracle
4986 If an application encounters a fatal protocol error and then calls
4987 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4988 then OpenSSL can respond differently to the calling application if a 0 byte
4989 record is received with invalid padding compared to if a 0 byte record is
4990 received with an invalid MAC. If the application then behaves differently
4991 based on that in a way that is detectable to the remote peer, then this
4992 amounts to a padding oracle that could be used to decrypt data.
4994 In order for this to be exploitable "non-stitched" ciphersuites must be in
4995 use. Stitched ciphersuites are optimised implementations of certain
4996 commonly used ciphersuites. Also the application must call SSL_shutdown()
4997 twice even if a protocol error has occurred (applications should not do
4998 this but some do anyway).
5000 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5001 Aviram, with additional investigation by Steven Collison and Andrew
5002 Hourselt. It was reported to OpenSSL on 10th December 2018.
5007 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5011 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5013 * Microarchitecture timing vulnerability in ECC scalar multiplication
5015 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5016 shown to be vulnerable to a microarchitecture timing side channel attack.
5017 An attacker with sufficient access to mount local timing attacks during
5018 ECDSA signature generation could recover the private key.
5020 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5021 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5027 * Timing vulnerability in DSA signature generation
5029 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5030 timing side channel attack. An attacker could use variations in the signing
5031 algorithm to recover the private key.
5033 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5038 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5039 Module, accidentally introduced while backporting security fixes from the
5040 development branch and hindering the use of ECC in FIPS mode.
5044 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5046 * Client DoS due to large DH parameter
5048 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5049 malicious server can send a very large prime value to the client. This will
5050 cause the client to spend an unreasonably long period of time generating a
5051 key for this prime resulting in a hang until the client has finished. This
5052 could be exploited in a Denial Of Service attack.
5054 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5059 * Cache timing vulnerability in RSA Key Generation
5061 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5062 a cache timing side channel attack. An attacker with sufficient access to
5063 mount cache timing attacks during the RSA key generation process could
5064 recover the private key.
5066 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5067 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5072 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5073 parameter is no longer accepted, as it leads to a corrupt table. NULL
5074 pem_str is reserved for alias entries only.
5078 * Revert blinding in ECDSA sign and instead make problematic addition
5079 length-invariant. Switch even to fixed-length Montgomery multiplication.
5083 * Change generating and checking of primes so that the error rate of not
5084 being prime depends on the intended use based on the size of the input.
5085 For larger primes this will result in more rounds of Miller-Rabin.
5086 The maximal error rate for primes with more than 1080 bits is lowered
5089 *Kurt Roeckx, Annie Yousar*
5091 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5095 * Add blinding to ECDSA and DSA signatures to protect against side channel
5096 attacks discovered by Keegan Ryan (NCC Group).
5100 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5101 now allow empty (zero character) pass phrases.
5105 * Certificate time validation (X509_cmp_time) enforces stricter
5106 compliance with RFC 5280. Fractional seconds and timezone offsets
5107 are no longer allowed.
5111 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5113 * Constructed ASN.1 types with a recursive definition could exceed the stack
5115 Constructed ASN.1 types with a recursive definition (such as can be found
5116 in PKCS7) could eventually exceed the stack given malicious input with
5117 excessive recursion. This could result in a Denial Of Service attack. There
5118 are no such structures used within SSL/TLS that come from untrusted sources
5119 so this is considered safe.
5121 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5127 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5129 * Read/write after SSL object in error state
5131 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5132 mechanism. The intent was that if a fatal error occurred during a handshake
5133 then OpenSSL would move into the error state and would immediately fail if
5134 you attempted to continue the handshake. This works as designed for the
5135 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5136 SSL_connect()), however due to a bug it does not work correctly if
5137 SSL_read() or SSL_write() is called directly. In that scenario, if the
5138 handshake fails then a fatal error will be returned in the initial function
5139 call. If SSL_read()/SSL_write() is subsequently called by the application
5140 for the same SSL object then it will succeed and the data is passed without
5141 being decrypted/encrypted directly from the SSL/TLS record layer.
5143 In order to exploit this issue an application bug would have to be present
5144 that resulted in a call to SSL_read()/SSL_write() being issued after having
5145 already received a fatal error.
5147 This issue was reported to OpenSSL by David Benjamin (Google).
5152 * rsaz_1024_mul_avx2 overflow bug on x86_64
5154 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5155 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5156 Analysis suggests that attacks against RSA and DSA as a result of this
5157 defect would be very difficult to perform and are not believed likely.
5158 Attacks against DH1024 are considered just feasible, because most of the
5159 work necessary to deduce information about a private key may be performed
5160 offline. The amount of resources required for such an attack would be
5161 significant. However, for an attack on TLS to be meaningful, the server
5162 would have to share the DH1024 private key among multiple clients, which is
5163 no longer an option since CVE-2016-0701.
5165 This only affects processors that support the AVX2 but not ADX extensions
5166 like Intel Haswell (4th generation).
5168 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5169 was originally found via the OSS-Fuzz project.
5174 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5176 * bn_sqrx8x_internal carry bug on x86_64
5178 There is a carry propagating bug in the x86_64 Montgomery squaring
5179 procedure. No EC algorithms are affected. Analysis suggests that attacks
5180 against RSA and DSA as a result of this defect would be very difficult to
5181 perform and are not believed likely. Attacks against DH are considered just
5182 feasible (although very difficult) because most of the work necessary to
5183 deduce information about a private key may be performed offline. The amount
5184 of resources required for such an attack would be very significant and
5185 likely only accessible to a limited number of attackers. An attacker would
5186 additionally need online access to an unpatched system using the target
5187 private key in a scenario with persistent DH parameters and a private
5188 key that is shared between multiple clients.
5190 This only affects processors that support the BMI1, BMI2 and ADX extensions
5191 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5193 This issue was reported to OpenSSL by the OSS-Fuzz project.
5198 * Malformed X.509 IPAddressFamily could cause OOB read
5200 If an X.509 certificate has a malformed IPAddressFamily extension,
5201 OpenSSL could do a one-byte buffer overread. The most likely result
5202 would be an erroneous display of the certificate in text format.
5204 This issue was reported to OpenSSL by the OSS-Fuzz project.
5208 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5210 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5211 platform rather than 'mingw'.
5215 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5217 * Truncated packet could crash via OOB read
5219 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5220 cipher is being used, then a truncated packet can cause that host to
5221 perform an out-of-bounds read, usually resulting in a crash.
5223 This issue was reported to OpenSSL by Robert Święcki of Google.
5228 * BN_mod_exp may produce incorrect results on x86_64
5230 There is a carry propagating bug in the x86_64 Montgomery squaring
5231 procedure. No EC algorithms are affected. Analysis suggests that attacks
5232 against RSA and DSA as a result of this defect would be very difficult to
5233 perform and are not believed likely. Attacks against DH are considered just
5234 feasible (although very difficult) because most of the work necessary to
5235 deduce information about a private key may be performed offline. The amount
5236 of resources required for such an attack would be very significant and
5237 likely only accessible to a limited number of attackers. An attacker would
5238 additionally need online access to an unpatched system using the target
5239 private key in a scenario with persistent DH parameters and a private
5240 key that is shared between multiple clients. For example this can occur by
5241 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5242 similar to CVE-2015-3193 but must be treated as a separate problem.
5244 This issue was reported to OpenSSL by the OSS-Fuzz project.
5249 * Montgomery multiplication may produce incorrect results
5251 There is a carry propagating bug in the Broadwell-specific Montgomery
5252 multiplication procedure that handles input lengths divisible by, but
5253 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5254 and DH private keys are impossible. This is because the subroutine in
5255 question is not used in operations with the private key itself and an input
5256 of the attacker's direct choice. Otherwise the bug can manifest itself as
5257 transient authentication and key negotiation failures or reproducible
5258 erroneous outcome of public-key operations with specially crafted input.
5259 Among EC algorithms only Brainpool P-512 curves are affected and one
5260 presumably can attack ECDH key negotiation. Impact was not analyzed in
5261 detail, because pre-requisites for attack are considered unlikely. Namely
5262 multiple clients have to choose the curve in question and the server has to
5263 share the private key among them, neither of which is default behaviour.
5264 Even then only clients that chose the curve will be affected.
5266 This issue was publicly reported as transient failures and was not
5267 initially recognized as a security issue. Thanks to Richard Morgan for
5268 providing reproducible case.
5273 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5274 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5275 prevent issues where no progress is being made and the peer continually
5276 sends unrecognised record types, using up resources processing them.
5280 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5282 * Missing CRL sanity check
5284 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5285 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5286 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5288 This issue only affects the OpenSSL 1.0.2i
5293 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5295 * OCSP Status Request extension unbounded memory growth
5297 A malicious client can send an excessively large OCSP Status Request
5298 extension. If that client continually requests renegotiation, sending a
5299 large OCSP Status Request extension each time, then there will be unbounded
5300 memory growth on the server. This will eventually lead to a Denial Of
5301 Service attack through memory exhaustion. Servers with a default
5302 configuration are vulnerable even if they do not support OCSP. Builds using
5303 the "no-ocsp" build time option are not affected.
5305 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5310 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5313 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5319 * OOB write in MDC2_Update()
5321 An overflow can occur in MDC2_Update() either if called directly or
5322 through the EVP_DigestUpdate() function using MDC2. If an attacker
5323 is able to supply very large amounts of input data after a previous
5324 call to EVP_EncryptUpdate() with a partial block then a length check
5325 can overflow resulting in a heap corruption.
5327 The amount of data needed is comparable to SIZE_MAX which is impractical
5330 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5335 * Malformed SHA512 ticket DoS
5337 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5338 DoS attack where a malformed ticket will result in an OOB read which will
5341 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5342 a custom server callback and ticket lookup mechanism.
5344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5349 * OOB write in BN_bn2dec()
5351 The function BN_bn2dec() does not check the return value of BN_div_word().
5352 This can cause an OOB write if an application uses this function with an
5353 overly large BIGNUM. This could be a problem if an overly large certificate
5354 or CRL is printed out from an untrusted source. TLS is not affected because
5355 record limits will reject an oversized certificate before it is parsed.
5357 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5362 * OOB read in TS_OBJ_print_bio()
5364 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5365 the total length the OID text representation would use and not the amount
5366 of data written. This will result in OOB reads when large OIDs are
5369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5374 * Pointer arithmetic undefined behaviour
5376 Avoid some undefined pointer arithmetic
5378 A common idiom in the codebase is to check limits in the following manner:
5381 Where "p" points to some malloc'd data of SIZE bytes and
5384 "len" here could be from some externally supplied data (e.g. from a TLS
5387 The rules of C pointer arithmetic are such that "p + len" is only well
5388 defined where len <= SIZE. Therefore the above idiom is actually
5389 undefined behaviour.
5391 For example this could cause problems if some malloc implementation
5392 provides an address for "p" such that "p + len" actually overflows for
5393 values of len that are too big and therefore p + len < limit.
5395 This issue was reported to OpenSSL by Guido Vranken
5400 * Constant time flag not preserved in DSA signing
5402 Operations in the DSA signing algorithm should run in constant time in
5403 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5404 implementation means that a non-constant time codepath is followed for
5405 certain operations. This has been demonstrated through a cache-timing
5406 attack to be sufficient for an attacker to recover the private DSA key.
5408 This issue was reported by César Pereida (Aalto University), Billy Brumley
5409 (Tampere University of Technology), and Yuval Yarom (The University of
5410 Adelaide and NICTA).
5415 * DTLS buffered message DoS
5417 In a DTLS connection where handshake messages are delivered out-of-order
5418 those messages that OpenSSL is not yet ready to process will be buffered
5419 for later use. Under certain circumstances, a flaw in the logic means that
5420 those messages do not get removed from the buffer even though the handshake
5421 has been completed. An attacker could force up to approx. 15 messages to
5422 remain in the buffer when they are no longer required. These messages will
5423 be cleared when the DTLS connection is closed. The default maximum size for
5424 a message is 100k. Therefore, the attacker could force an additional 1500k
5425 to be consumed per connection. By opening many simultaneous connections an
5426 attacker could cause a DoS attack through memory exhaustion.
5428 This issue was reported to OpenSSL by Quan Luo.
5433 * DTLS replay protection DoS
5435 A flaw in the DTLS replay attack protection mechanism means that records
5436 that arrive for future epochs update the replay protection "window" before
5437 the MAC for the record has been validated. This could be exploited by an
5438 attacker by sending a record for the next epoch (which does not have to
5439 decrypt or have a valid MAC), with a very large sequence number. This means
5440 that all subsequent legitimate packets are dropped causing a denial of
5441 service for a specific DTLS connection.
5443 This issue was reported to OpenSSL by the OCAP audit team.
5448 * Certificate message OOB reads
5450 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5451 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5452 theoretical DoS risk but this has not been observed in practice on common
5455 The messages affected are client certificate, client certificate request
5456 and server certificate. As a result the attack can only be performed
5457 against a client or a server which enables client authentication.
5459 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5464 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5466 * Prevent padding oracle in AES-NI CBC MAC check
5468 A MITM attacker can use a padding oracle attack to decrypt traffic
5469 when the connection uses an AES CBC cipher and the server support
5472 This issue was introduced as part of the fix for Lucky 13 padding
5473 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5474 constant time by making sure that always the same bytes are read and
5475 compared against either the MAC or padding bytes. But it no longer
5476 checked that there was enough data to have both the MAC and padding
5479 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5483 * Fix EVP_EncodeUpdate overflow
5485 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5486 Base64 encoding of binary data. If an attacker is able to supply very large
5487 amounts of input data then a length check can overflow resulting in a heap
5490 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5491 the `PEM_write_bio*` family of functions. These are mainly used within the
5492 OpenSSL command line applications, so any application which processes data
5493 from an untrusted source and outputs it as a PEM file should be considered
5494 vulnerable to this issue. User applications that call these APIs directly
5495 with large amounts of untrusted data may also be vulnerable.
5497 This issue was reported by Guido Vranken.
5502 * Fix EVP_EncryptUpdate overflow
5504 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5505 is able to supply very large amounts of input data after a previous call to
5506 EVP_EncryptUpdate() with a partial block then a length check can overflow
5507 resulting in a heap corruption. Following an analysis of all OpenSSL
5508 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5509 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5510 the first called function after an EVP_EncryptInit(), and therefore that
5511 specific call must be safe. The second form is where the length passed to
5512 EVP_EncryptUpdate() can be seen from the code to be some small value and
5513 therefore there is no possibility of an overflow. Since all instances are
5514 one of these two forms, it is believed that there can be no overflows in
5515 internal code due to this problem. It should be noted that
5516 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5517 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5518 of these calls have also been analysed too and it is believed there are no
5519 instances in internal usage where an overflow could occur.
5521 This issue was reported by Guido Vranken.
5526 * Prevent ASN.1 BIO excessive memory allocation
5528 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5529 a short invalid encoding can cause allocation of large amounts of memory
5530 potentially consuming excessive resources or exhausting memory.
5532 Any application parsing untrusted data through d2i BIO functions is
5533 affected. The memory based functions such as d2i_X509() are *not* affected.
5534 Since the memory based functions are used by the TLS library, TLS
5535 applications are not affected.
5537 This issue was reported by Brian Carpenter.
5544 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5545 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5546 in arbitrary stack data being returned in the buffer.
5548 This issue was reported by Guido Vranken.
5553 * Modify behavior of ALPN to invoke callback after SNI/servername
5554 callback, such that updates to the SSL_CTX affect ALPN.
5558 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5563 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5564 methods are enabled and ssl2 is disabled the methods return NULL.
5568 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5570 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5571 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5572 provide any "EXPORT" or "LOW" strength ciphers.
5576 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5577 is by default disabled at build-time. Builds that are not configured with
5578 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5579 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5580 will need to explicitly call either of:
5582 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5584 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5586 as appropriate. Even if either of those is used, or the application
5587 explicitly uses the version-specific SSLv2_method() or its client and
5588 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5589 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5590 ciphers, and SSLv2 56-bit DES are no longer available.
5595 * Fix a double-free in DSA code
5597 A double free bug was discovered when OpenSSL parses malformed DSA private
5598 keys and could lead to a DoS attack or memory corruption for applications
5599 that receive DSA private keys from untrusted sources. This scenario is
5602 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5608 * Disable SRP fake user seed to address a server memory leak.
5610 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5612 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5613 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5614 was changed to ignore the "fake user" SRP seed, even if the seed
5617 Users should use SRP_VBASE_get1_by_user instead. Note that in
5618 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5619 also that even though configuring the SRP seed attempts to hide
5620 invalid usernames by continuing the handshake with fake
5621 credentials, this behaviour is not constant time and no strong
5622 guarantees are made that the handshake is indistinguishable from
5623 that of a valid user.
5628 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5630 In the BN_hex2bn function the number of hex digits is calculated using an
5631 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5632 large values of `i` this can result in `bn_expand` not allocating any
5633 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5634 field as NULL leading to a subsequent NULL ptr deref. For very large values
5635 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5636 In this case memory is allocated to the internal BIGNUM data field, but it
5637 is insufficiently sized leading to heap corruption. A similar issue exists
5638 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5639 is ever called by user applications with very large untrusted hex/dec data.
5640 This is anticipated to be a rare occurrence.
5642 All OpenSSL internal usage of these functions use data that is not expected
5643 to be untrusted, e.g. config file data or application command line
5644 arguments. If user developed applications generate config file data based
5645 on untrusted data then it is possible that this could also lead to security
5646 consequences. This is also anticipated to be rare.
5648 This issue was reported to OpenSSL by Guido Vranken.
5653 * Fix memory issues in `BIO_*printf` functions
5655 The internal `fmtstr` function used in processing a "%s" format string in
5656 the `BIO_*printf` functions could overflow while calculating the length of a
5657 string and cause an OOB read when printing very long strings.
5659 Additionally the internal `doapr_outch` function can attempt to write to an
5660 OOB memory location (at an offset from the NULL pointer) in the event of a
5661 memory allocation failure. In 1.0.2 and below this could be caused where
5662 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5663 could be in processing a very long "%s" format string. Memory leaks can
5666 The first issue may mask the second issue dependent on compiler behaviour.
5667 These problems could enable attacks where large amounts of untrusted data
5668 is passed to the `BIO_*printf` functions. If applications use these functions
5669 in this way then they could be vulnerable. OpenSSL itself uses these
5670 functions when printing out human-readable dumps of ASN.1 data. Therefore
5671 applications that print this data could be vulnerable if the data is from
5672 untrusted sources. OpenSSL command line applications could also be
5673 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5674 as command line arguments.
5676 Libssl is not considered directly vulnerable. Additionally certificates etc
5677 received via remote connections via libssl are also unlikely to be able to
5678 trigger these issues because of message size limits enforced within libssl.
5680 This issue was reported to OpenSSL Guido Vranken.
5685 * Side channel attack on modular exponentiation
5687 A side-channel attack was found which makes use of cache-bank conflicts on
5688 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5689 of RSA keys. The ability to exploit this issue is limited as it relies on
5690 an attacker who has control of code in a thread running on the same
5691 hyper-threaded core as the victim thread which is performing decryptions.
5693 This issue was reported to OpenSSL by Yuval Yarom, The University of
5694 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5695 Nadia Heninger, University of Pennsylvania with more information at
5696 <http://cachebleed.info>.
5701 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5702 if no keysize is specified with default_bits. This fixes an
5703 omission in an earlier change that changed all RSA/DSA key generation
5704 commands to use 2048 bits by default.
5708 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5710 * DH small subgroups
5712 Historically OpenSSL only ever generated DH parameters based on "safe"
5713 primes. More recently (in version 1.0.2) support was provided for
5714 generating X9.42 style parameter files such as those required for RFC 5114
5715 support. The primes used in such files may not be "safe". Where an
5716 application is using DH configured with parameters based on primes that are
5717 not "safe" then an attacker could use this fact to find a peer's private
5718 DH exponent. This attack requires that the attacker complete multiple
5719 handshakes in which the peer uses the same private DH exponent. For example
5720 this could be used to discover a TLS server's private DH exponent if it's
5721 reusing the private DH exponent or it's using a static DH ciphersuite.
5723 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5724 TLS. It is not on by default. If the option is not set then the server
5725 reuses the same private DH exponent for the life of the server process and
5726 would be vulnerable to this attack. It is believed that many popular
5727 applications do set this option and would therefore not be at risk.
5729 The fix for this issue adds an additional check where a "q" parameter is
5730 available (as is the case in X9.42 based parameters). This detects the
5731 only known attack, and is the only possible defense for static DH
5732 ciphersuites. This could have some performance impact.
5734 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5735 default and cannot be disabled. This could have some performance impact.
5737 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5742 * SSLv2 doesn't block disabled ciphers
5744 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5745 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5746 been disabled, provided that the SSLv2 protocol was not also disabled via
5749 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5750 and Sebastian Schinzel.
5755 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5757 * BN_mod_exp may produce incorrect results on x86_64
5759 There is a carry propagating bug in the x86_64 Montgomery squaring
5760 procedure. No EC algorithms are affected. Analysis suggests that attacks
5761 against RSA and DSA as a result of this defect would be very difficult to
5762 perform and are not believed likely. Attacks against DH are considered just
5763 feasible (although very difficult) because most of the work necessary to
5764 deduce information about a private key may be performed offline. The amount
5765 of resources required for such an attack would be very significant and
5766 likely only accessible to a limited number of attackers. An attacker would
5767 additionally need online access to an unpatched system using the target
5768 private key in a scenario with persistent DH parameters and a private
5769 key that is shared between multiple clients. For example this can occur by
5770 default in OpenSSL DHE based SSL/TLS ciphersuites.
5772 This issue was reported to OpenSSL by Hanno Böck.
5777 * Certificate verify crash with missing PSS parameter
5779 The signature verification routines will crash with a NULL pointer
5780 dereference if presented with an ASN.1 signature using the RSA PSS
5781 algorithm and absent mask generation function parameter. Since these
5782 routines are used to verify certificate signature algorithms this can be
5783 used to crash any certificate verification operation and exploited in a
5784 DoS attack. Any application which performs certificate verification is
5785 vulnerable including OpenSSL clients and servers which enable client
5788 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5793 * X509_ATTRIBUTE memory leak
5795 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5796 memory. This structure is used by the PKCS#7 and CMS routines so any
5797 application which reads PKCS#7 or CMS data from untrusted sources is
5798 affected. SSL/TLS is not affected.
5800 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5807 This changes the decoding behaviour for some invalid messages,
5808 though the change is mostly in the more lenient direction, and
5809 legacy behaviour is preserved as much as possible.
5813 * In DSA_generate_parameters_ex, if the provided seed is too short,
5816 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5818 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5820 * Alternate chains certificate forgery
5822 During certificate verification, OpenSSL will attempt to find an
5823 alternative certificate chain if the first attempt to build such a chain
5824 fails. An error in the implementation of this logic can mean that an
5825 attacker could cause certain checks on untrusted certificates to be
5826 bypassed, such as the CA flag, enabling them to use a valid leaf
5827 certificate to act as a CA and "issue" an invalid certificate.
5829 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5834 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5836 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5837 incompatibility in the handling of HMAC. The previous ABI has now been
5842 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5844 * Malformed ECParameters causes infinite loop
5846 When processing an ECParameters structure OpenSSL enters an infinite loop
5847 if the curve specified is over a specially malformed binary polynomial
5850 This can be used to perform denial of service against any
5851 system which processes public keys, certificate requests or
5852 certificates. This includes TLS clients and TLS servers with
5853 client authentication enabled.
5855 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5860 * Exploitable out-of-bounds read in X509_cmp_time
5862 X509_cmp_time does not properly check the length of the ASN1_TIME
5863 string and can read a few bytes out of bounds. In addition,
5864 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5867 An attacker can use this to craft malformed certificates and CRLs of
5868 various sizes and potentially cause a segmentation fault, resulting in
5869 a DoS on applications that verify certificates or CRLs. TLS clients
5870 that verify CRLs are affected. TLS clients and servers with client
5871 authentication enabled may be affected if they use custom verification
5874 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5875 independently by Hanno Böck.
5880 * PKCS7 crash with missing EnvelopedContent
5882 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5883 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5884 with missing content and trigger a NULL pointer dereference on parsing.
5886 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5887 structures from untrusted sources are affected. OpenSSL clients and
5888 servers are not affected.
5890 This issue was reported to OpenSSL by Michal Zalewski (Google).
5895 * CMS verify infinite loop with unknown hash function
5897 When verifying a signedData message the CMS code can enter an infinite loop
5898 if presented with an unknown hash function OID. This can be used to perform
5899 denial of service against any system which verifies signedData messages using
5901 This issue was reported to OpenSSL by Johannes Bauer.
5906 * Race condition handling NewSessionTicket
5908 If a NewSessionTicket is received by a multi-threaded client when attempting to
5909 reuse a previous ticket then a race condition can occur potentially leading to
5910 a double free of the ticket data.
5915 * Only support 256-bit or stronger elliptic curves with the
5916 'ecdh_auto' setting (server) or by default (client). Of supported
5917 curves, prefer P-256 (both).
5921 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5923 * ClientHello sigalgs DoS fix
5925 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5926 invalid signature algorithms extension a NULL pointer dereference will
5927 occur. This can be exploited in a DoS attack against the server.
5929 This issue was was reported to OpenSSL by David Ramos of Stanford
5933 *Stephen Henson and Matt Caswell*
5935 * Multiblock corrupted pointer fix
5937 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5938 feature only applies on 64 bit x86 architecture platforms that support AES
5939 NI instructions. A defect in the implementation of "multiblock" can cause
5940 OpenSSL's internal write buffer to become incorrectly set to NULL when
5941 using non-blocking IO. Typically, when the user application is using a
5942 socket BIO for writing, this will only result in a failed connection.
5943 However if some other BIO is used then it is likely that a segmentation
5944 fault will be triggered, thus enabling a potential DoS attack.
5946 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5951 * Segmentation fault in DTLSv1_listen fix
5953 The DTLSv1_listen function is intended to be stateless and processes the
5954 initial ClientHello from many peers. It is common for user code to loop
5955 over the call to DTLSv1_listen until a valid ClientHello is received with
5956 an associated cookie. A defect in the implementation of DTLSv1_listen means
5957 that state is preserved in the SSL object from one invocation to the next
5958 that can lead to a segmentation fault. Errors processing the initial
5959 ClientHello can trigger this scenario. An example of such an error could be
5960 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5963 This issue was reported to OpenSSL by Per Allansson.
5968 * Segmentation fault in ASN1_TYPE_cmp fix
5970 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5971 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5972 certificate signature algorithm consistency this can be used to crash any
5973 certificate verification operation and exploited in a DoS attack. Any
5974 application which performs certificate verification is vulnerable including
5975 OpenSSL clients and servers which enable client authentication.
5980 * Segmentation fault for invalid PSS parameters fix
5982 The signature verification routines will crash with a NULL pointer
5983 dereference if presented with an ASN.1 signature using the RSA PSS
5984 algorithm and invalid parameters. Since these routines are used to verify
5985 certificate signature algorithms this can be used to crash any
5986 certificate verification operation and exploited in a DoS attack. Any
5987 application which performs certificate verification is vulnerable including
5988 OpenSSL clients and servers which enable client authentication.
5990 This issue was was reported to OpenSSL by Brian Carpenter.
5995 * ASN.1 structure reuse memory corruption fix
5997 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5998 memory corruption via an invalid write. Such reuse is and has been
5999 strongly discouraged and is believed to be rare.
6001 Applications that parse structures containing CHOICE or ANY DEFINED BY
6002 components may be affected. Certificate parsing (d2i_X509 and related
6003 functions) are however not affected. OpenSSL clients and servers are
6009 * PKCS7 NULL pointer dereferences fix
6011 The PKCS#7 parsing code does not handle missing outer ContentInfo
6012 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6013 missing content and trigger a NULL pointer dereference on parsing.
6015 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6016 otherwise parse PKCS#7 structures from untrusted sources are
6017 affected. OpenSSL clients and servers are not affected.
6019 This issue was reported to OpenSSL by Michal Zalewski (Google).
6024 * DoS via reachable assert in SSLv2 servers fix
6026 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6027 servers that both support SSLv2 and enable export cipher suites by sending
6028 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6030 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6031 (OpenSSL development team).
6036 * Empty CKE with client auth and DHE fix
6038 If client auth is used then a server can seg fault in the event of a DHE
6039 ciphersuite being selected and a zero length ClientKeyExchange message
6040 being sent by the client. This could be exploited in a DoS attack.
6045 * Handshake with unseeded PRNG fix
6047 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6048 with an unseeded PRNG. The conditions are:
6049 - The client is on a platform where the PRNG has not been seeded
6050 automatically, and the user has not seeded manually
6051 - A protocol specific client method version has been used (i.e. not
6052 SSL_client_methodv23)
6053 - A ciphersuite is used that does not require additional random data from
6054 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6056 If the handshake succeeds then the client random that has been used will
6057 have been generated from a PRNG with insufficient entropy and therefore the
6058 output may be predictable.
6060 For example using the following command with an unseeded openssl will
6061 succeed on an unpatched platform:
6063 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6068 * Use After Free following d2i_ECPrivatekey error fix
6070 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6071 could cause a use after free condition. This, in turn, could cause a double
6072 free in several private key parsing functions (such as d2i_PrivateKey
6073 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6074 for applications that receive EC private keys from untrusted
6075 sources. This scenario is considered rare.
6077 This issue was discovered by the BoringSSL project and fixed in their
6083 * X509_to_X509_REQ NULL pointer deref fix
6085 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6086 the certificate key is invalid. This function is rarely used in practice.
6088 This issue was discovered by Brian Carpenter.
6093 * Removed the export ciphers from the DEFAULT ciphers
6097 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6099 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6100 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6101 So far those who have to target multiple platforms would compromise
6102 and argue that binary targeting say ARMv5 would still execute on
6103 ARMv8. "Universal" build resolves this compromise by providing
6104 near-optimal performance even on newer platforms.
6108 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6109 (other platforms pending).
6111 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6113 * Add support for the SignedCertificateTimestampList certificate and
6114 OCSP response extensions from RFC6962.
6118 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6119 for corner cases. (Certain input points at infinity could lead to
6120 bogus results, with non-infinity inputs mapped to infinity too.)
6124 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6125 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6126 common cases are optimized and there still is room for further
6127 improvements. Vector Permutation AES for Altivec is also added.
6131 * Add support for little-endian ppc64 Linux target.
6133 *Marcelo Cerri (IBM)*
6135 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6136 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6137 are optimized and there still is room for further improvements.
6138 Both 32- and 64-bit modes are supported.
6140 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6142 * Improved ARMv7 NEON support.
6146 * Support for SPARC Architecture 2011 crypto extensions, first
6147 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6148 SHA256/512, MD5, GHASH and modular exponentiation.
6150 *Andy Polyakov, David Miller*
6152 * Accelerated modular exponentiation for Intel processors, a.k.a.
6155 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6157 * Support for new and upcoming Intel processors, including AVX2,
6158 BMI and SHA ISA extensions. This includes additional "stitched"
6159 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6162 This work was sponsored by Intel Corp.
6166 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6167 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6168 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6172 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6173 this fixes a limitation in previous versions of OpenSSL.
6177 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6178 MGF1 digest and OAEP label.
6182 * Add EVP support for key wrapping algorithms, to avoid problems with
6183 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6184 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6185 algorithms and include tests cases.
6189 * Add functions to allocate and set the fields of an ECDSA_METHOD
6192 *Douglas E. Engert, Steve Henson*
6194 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6195 difference in days and seconds between two tm or ASN1_TIME structures.
6199 * Add -rev test option to s_server to just reverse order of characters
6200 received by client and send back to server. Also prints an abbreviated
6201 summary of the connection parameters.
6205 * New option -brief for s_client and s_server to print out a brief summary
6206 of connection parameters.
6210 * Add callbacks for arbitrary TLS extensions.
6212 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6214 * New option -crl_download in several openssl utilities to download CRLs
6215 from CRLDP extension in certificates.
6219 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6223 * New function X509_CRL_diff to generate a delta CRL from the difference
6224 of two full CRLs. Add support to "crl" utility.
6228 * New functions to set lookup_crls function and to retrieve
6229 X509_STORE from X509_STORE_CTX.
6233 * Print out deprecated issuer and subject unique ID fields in
6238 * Extend OCSP I/O functions so they can be used for simple general purpose
6239 HTTP as well as OCSP. New wrapper function which can be used to download
6240 CRLs using the OCSP API.
6244 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6248 * `SSL_CONF*` functions. These provide a common framework for application
6249 configuration using configuration files or command lines.
6253 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6254 message callback and prints the results. Needs compile time option
6255 "enable-ssl-trace". New options to s_client and s_server to enable
6260 * New ctrl and macro to retrieve supported points extensions.
6261 Print out extension in s_server and s_client.
6265 * New functions to retrieve certificate signature and signature
6270 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6275 * New Suite B modes for TLS code. These use and enforce the requirements
6276 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6277 only use Suite B curves. The Suite B modes can be set by using the
6278 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6282 * New chain verification flags for Suite B levels of security. Check
6283 algorithms are acceptable when flags are set in X509_verify_cert.
6287 * Make tls1_check_chain return a set of flags indicating checks passed
6288 by a certificate chain. Add additional tests to handle client
6289 certificates: checks for matching certificate type and issuer name
6294 * If an attempt is made to use a signature algorithm not in the peer
6295 preference list abort the handshake. If client has no suitable
6296 signature algorithms in response to a certificate request do not
6297 use the certificate.
6301 * If server EC tmp key is not in client preference list abort handshake.
6305 * Add support for certificate stores in CERT structure. This makes it
6306 possible to have different stores per SSL structure or one store in
6307 the parent SSL_CTX. Include distinct stores for certificate chain
6308 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6309 to build and store a certificate chain in CERT structure: returning
6310 an error if the chain cannot be built: this will allow applications
6311 to test if a chain is correctly configured.
6313 Note: if the CERT based stores are not set then the parent SSL_CTX
6314 store is used to retain compatibility with existing behaviour.
6318 * New function ssl_set_client_disabled to set a ciphersuite disabled
6319 mask based on the current session, check mask when sending client
6320 hello and checking the requested ciphersuite.
6324 * New ctrls to retrieve and set certificate types in a certificate
6325 request message. Print out received values in s_client. If certificate
6326 types is not set with custom values set sensible values based on
6327 supported signature algorithms.
6331 * Support for distinct client and server supported signature algorithms.
6335 * Add certificate callback. If set this is called whenever a certificate
6336 is required by client or server. An application can decide which
6337 certificate chain to present based on arbitrary criteria: for example
6338 supported signature algorithms. Add very simple example to s_server.
6339 This fixes many of the problems and restrictions of the existing client
6340 certificate callback: for example you can now clear an existing
6341 certificate and specify the whole chain.
6345 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6346 the certificate can be used for (if anything). Set valid_flags field
6347 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6348 to have similar checks in it.
6350 Add new "cert_flags" field to CERT structure and include a "strict mode".
6351 This enforces some TLS certificate requirements (such as only permitting
6352 certificate signature algorithms contained in the supported algorithms
6353 extension) which some implementations ignore: this option should be used
6354 with caution as it could cause interoperability issues.
6358 * Update and tidy signature algorithm extension processing. Work out
6359 shared signature algorithms based on preferences and peer algorithms
6360 and print them out in s_client and s_server. Abort handshake if no
6361 shared signature algorithms.
6365 * Add new functions to allow customised supported signature algorithms
6366 for SSL and SSL_CTX structures. Add options to s_client and s_server
6371 * New function SSL_certs_clear() to delete all references to certificates
6372 from an SSL structure. Before this once a certificate had been added
6373 it couldn't be removed.
6377 * Integrate hostname, email address and IP address checking with certificate
6378 verification. New verify options supporting checking in openssl utility.
6382 * Fixes and wildcard matching support to hostname and email checking
6383 functions. Add manual page.
6385 *Florian Weimer (Red Hat Product Security Team)*
6387 * New functions to check a hostname email or IP address against a
6388 certificate. Add options x509 utility to print results of checks against
6393 * Fix OCSP checking.
6395 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6397 * Initial experimental support for explicitly trusted non-root CAs.
6398 OpenSSL still tries to build a complete chain to a root but if an
6399 intermediate CA has a trust setting included that is used. The first
6400 setting is used: whether to trust (e.g., -addtrust option to the x509
6405 * Add -trusted_first option which attempts to find certificates in the
6406 trusted store even if an untrusted chain is also supplied.
6410 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6411 platform support for Linux and Android.
6415 * Support for linux-x32, ILP32 environment in x86_64 framework.
6419 * Experimental multi-implementation support for FIPS capable OpenSSL.
6420 When in FIPS mode the approved implementations are used as normal,
6421 when not in FIPS mode the internal unapproved versions are used instead.
6422 This means that the FIPS capable OpenSSL isn't forced to use the
6423 (often lower performance) FIPS implementations outside FIPS mode.
6427 * Transparently support X9.42 DH parameters when calling
6428 PEM_read_bio_DHparameters. This means existing applications can handle
6429 the new parameter format automatically.
6433 * Initial experimental support for X9.42 DH parameter format: mainly
6434 to support use of 'q' parameter for RFC5114 parameters.
6438 * Add DH parameters from RFC5114 including test data to dhtest.
6442 * Support for automatic EC temporary key parameter selection. If enabled
6443 the most preferred EC parameters are automatically used instead of
6444 hardcoded fixed parameters. Now a server just has to call:
6445 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6446 support ECDH and use the most appropriate parameters.
6450 * Enhance and tidy EC curve and point format TLS extension code. Use
6451 static structures instead of allocation if default values are used.
6452 New ctrls to set curves we wish to support and to retrieve shared curves.
6453 Print out shared curves in s_server. New options to s_server and s_client
6454 to set list of supported curves.
6458 * New ctrls to retrieve supported signature algorithms and
6459 supported curve values as an array of NIDs. Extend openssl utility
6460 to print out received values.
6464 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6465 between NIDs and the more common NIST names such as "P-256". Enhance
6466 ecparam utility and ECC method to recognise the NIST names for curves.
6470 * Enhance SSL/TLS certificate chain handling to support different
6471 chains for each certificate instead of one chain in the parent SSL_CTX.
6475 * Support for fixed DH ciphersuite client authentication: where both
6476 server and client use DH certificates with common parameters.
6480 * Support for fixed DH ciphersuites: those requiring DH server
6485 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6487 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6488 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6489 X509_CINF_get_signature were reverted post internal team review.
6494 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6496 * OCSP Status Request extension unbounded memory growth
6498 A malicious client can send an excessively large OCSP Status Request
6499 extension. If that client continually requests renegotiation, sending a
6500 large OCSP Status Request extension each time, then there will be unbounded
6501 memory growth on the server. This will eventually lead to a Denial Of
6502 Service attack through memory exhaustion. Servers with a default
6503 configuration are vulnerable even if they do not support OCSP. Builds using
6504 the "no-ocsp" build time option are not affected.
6506 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6511 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6514 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6520 * OOB write in MDC2_Update()
6522 An overflow can occur in MDC2_Update() either if called directly or
6523 through the EVP_DigestUpdate() function using MDC2. If an attacker
6524 is able to supply very large amounts of input data after a previous
6525 call to EVP_EncryptUpdate() with a partial block then a length check
6526 can overflow resulting in a heap corruption.
6528 The amount of data needed is comparable to SIZE_MAX which is impractical
6531 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6536 * Malformed SHA512 ticket DoS
6538 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6539 DoS attack where a malformed ticket will result in an OOB read which will
6542 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6543 a custom server callback and ticket lookup mechanism.
6545 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6550 * OOB write in BN_bn2dec()
6552 The function BN_bn2dec() does not check the return value of BN_div_word().
6553 This can cause an OOB write if an application uses this function with an
6554 overly large BIGNUM. This could be a problem if an overly large certificate
6555 or CRL is printed out from an untrusted source. TLS is not affected because
6556 record limits will reject an oversized certificate before it is parsed.
6558 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6563 * OOB read in TS_OBJ_print_bio()
6565 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6566 the total length the OID text representation would use and not the amount
6567 of data written. This will result in OOB reads when large OIDs are
6570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6575 * Pointer arithmetic undefined behaviour
6577 Avoid some undefined pointer arithmetic
6579 A common idiom in the codebase is to check limits in the following manner:
6582 Where "p" points to some malloc'd data of SIZE bytes and
6585 "len" here could be from some externally supplied data (e.g. from a TLS
6588 The rules of C pointer arithmetic are such that "p + len" is only well
6589 defined where len <= SIZE. Therefore, the above idiom is actually
6590 undefined behaviour.
6592 For example this could cause problems if some malloc implementation
6593 provides an address for "p" such that "p + len" actually overflows for
6594 values of len that are too big and therefore p + len < limit.
6596 This issue was reported to OpenSSL by Guido Vranken
6601 * Constant time flag not preserved in DSA signing
6603 Operations in the DSA signing algorithm should run in constant time in
6604 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6605 implementation means that a non-constant time codepath is followed for
6606 certain operations. This has been demonstrated through a cache-timing
6607 attack to be sufficient for an attacker to recover the private DSA key.
6609 This issue was reported by César Pereida (Aalto University), Billy Brumley
6610 (Tampere University of Technology), and Yuval Yarom (The University of
6611 Adelaide and NICTA).
6616 * DTLS buffered message DoS
6618 In a DTLS connection where handshake messages are delivered out-of-order
6619 those messages that OpenSSL is not yet ready to process will be buffered
6620 for later use. Under certain circumstances, a flaw in the logic means that
6621 those messages do not get removed from the buffer even though the handshake
6622 has been completed. An attacker could force up to approx. 15 messages to
6623 remain in the buffer when they are no longer required. These messages will
6624 be cleared when the DTLS connection is closed. The default maximum size for
6625 a message is 100k. Therefore, the attacker could force an additional 1500k
6626 to be consumed per connection. By opening many simultaneous connections an
6627 attacker could cause a DoS attack through memory exhaustion.
6629 This issue was reported to OpenSSL by Quan Luo.
6634 * DTLS replay protection DoS
6636 A flaw in the DTLS replay attack protection mechanism means that records
6637 that arrive for future epochs update the replay protection "window" before
6638 the MAC for the record has been validated. This could be exploited by an
6639 attacker by sending a record for the next epoch (which does not have to
6640 decrypt or have a valid MAC), with a very large sequence number. This means
6641 that all subsequent legitimate packets are dropped causing a denial of
6642 service for a specific DTLS connection.
6644 This issue was reported to OpenSSL by the OCAP audit team.
6649 * Certificate message OOB reads
6651 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6652 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6653 theoretical DoS risk but this has not been observed in practice on common
6656 The messages affected are client certificate, client certificate request
6657 and server certificate. As a result the attack can only be performed
6658 against a client or a server which enables client authentication.
6660 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6665 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6667 * Prevent padding oracle in AES-NI CBC MAC check
6669 A MITM attacker can use a padding oracle attack to decrypt traffic
6670 when the connection uses an AES CBC cipher and the server support
6673 This issue was introduced as part of the fix for Lucky 13 padding
6674 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6675 constant time by making sure that always the same bytes are read and
6676 compared against either the MAC or padding bytes. But it no longer
6677 checked that there was enough data to have both the MAC and padding
6680 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6685 * Fix EVP_EncodeUpdate overflow
6687 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6688 Base64 encoding of binary data. If an attacker is able to supply very large
6689 amounts of input data then a length check can overflow resulting in a heap
6692 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6693 the `PEM_write_bio*` family of functions. These are mainly used within the
6694 OpenSSL command line applications, so any application which processes data
6695 from an untrusted source and outputs it as a PEM file should be considered
6696 vulnerable to this issue. User applications that call these APIs directly
6697 with large amounts of untrusted data may also be vulnerable.
6699 This issue was reported by Guido Vranken.
6704 * Fix EVP_EncryptUpdate overflow
6706 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6707 is able to supply very large amounts of input data after a previous call to
6708 EVP_EncryptUpdate() with a partial block then a length check can overflow
6709 resulting in a heap corruption. Following an analysis of all OpenSSL
6710 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6711 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6712 the first called function after an EVP_EncryptInit(), and therefore that
6713 specific call must be safe. The second form is where the length passed to
6714 EVP_EncryptUpdate() can be seen from the code to be some small value and
6715 therefore there is no possibility of an overflow. Since all instances are
6716 one of these two forms, it is believed that there can be no overflows in
6717 internal code due to this problem. It should be noted that
6718 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6719 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6720 of these calls have also been analysed too and it is believed there are no
6721 instances in internal usage where an overflow could occur.
6723 This issue was reported by Guido Vranken.
6728 * Prevent ASN.1 BIO excessive memory allocation
6730 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6731 a short invalid encoding can casuse allocation of large amounts of memory
6732 potentially consuming excessive resources or exhausting memory.
6734 Any application parsing untrusted data through d2i BIO functions is
6735 affected. The memory based functions such as d2i_X509() are *not* affected.
6736 Since the memory based functions are used by the TLS library, TLS
6737 applications are not affected.
6739 This issue was reported by Brian Carpenter.
6746 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6747 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6748 in arbitrary stack data being returned in the buffer.
6750 This issue was reported by Guido Vranken.
6755 * Modify behavior of ALPN to invoke callback after SNI/servername
6756 callback, such that updates to the SSL_CTX affect ALPN.
6760 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6765 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6766 methods are enabled and ssl2 is disabled the methods return NULL.
6770 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6772 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6773 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6774 provide any "EXPORT" or "LOW" strength ciphers.
6778 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6779 is by default disabled at build-time. Builds that are not configured with
6780 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6781 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6782 will need to explicitly call either of:
6784 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6786 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6788 as appropriate. Even if either of those is used, or the application
6789 explicitly uses the version-specific SSLv2_method() or its client and
6790 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6791 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6792 ciphers, and SSLv2 56-bit DES are no longer available.
6797 * Fix a double-free in DSA code
6799 A double free bug was discovered when OpenSSL parses malformed DSA private
6800 keys and could lead to a DoS attack or memory corruption for applications
6801 that receive DSA private keys from untrusted sources. This scenario is
6804 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6810 * Disable SRP fake user seed to address a server memory leak.
6812 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6814 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6815 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6816 was changed to ignore the "fake user" SRP seed, even if the seed
6819 Users should use SRP_VBASE_get1_by_user instead. Note that in
6820 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6821 also that even though configuring the SRP seed attempts to hide
6822 invalid usernames by continuing the handshake with fake
6823 credentials, this behaviour is not constant time and no strong
6824 guarantees are made that the handshake is indistinguishable from
6825 that of a valid user.
6830 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6832 In the BN_hex2bn function the number of hex digits is calculated using an
6833 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6834 large values of `i` this can result in `bn_expand` not allocating any
6835 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6836 field as NULL leading to a subsequent NULL ptr deref. For very large values
6837 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6838 In this case memory is allocated to the internal BIGNUM data field, but it
6839 is insufficiently sized leading to heap corruption. A similar issue exists
6840 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6841 is ever called by user applications with very large untrusted hex/dec data.
6842 This is anticipated to be a rare occurrence.
6844 All OpenSSL internal usage of these functions use data that is not expected
6845 to be untrusted, e.g. config file data or application command line
6846 arguments. If user developed applications generate config file data based
6847 on untrusted data then it is possible that this could also lead to security
6848 consequences. This is also anticipated to be rare.
6850 This issue was reported to OpenSSL by Guido Vranken.
6855 * Fix memory issues in `BIO_*printf` functions
6857 The internal `fmtstr` function used in processing a "%s" format string in
6858 the `BIO_*printf` functions could overflow while calculating the length of a
6859 string and cause an OOB read when printing very long strings.
6861 Additionally the internal `doapr_outch` function can attempt to write to an
6862 OOB memory location (at an offset from the NULL pointer) in the event of a
6863 memory allocation failure. In 1.0.2 and below this could be caused where
6864 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6865 could be in processing a very long "%s" format string. Memory leaks can
6868 The first issue may mask the second issue dependent on compiler behaviour.
6869 These problems could enable attacks where large amounts of untrusted data
6870 is passed to the `BIO_*printf` functions. If applications use these functions
6871 in this way then they could be vulnerable. OpenSSL itself uses these
6872 functions when printing out human-readable dumps of ASN.1 data. Therefore
6873 applications that print this data could be vulnerable if the data is from
6874 untrusted sources. OpenSSL command line applications could also be
6875 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6876 as command line arguments.
6878 Libssl is not considered directly vulnerable. Additionally certificates etc
6879 received via remote connections via libssl are also unlikely to be able to
6880 trigger these issues because of message size limits enforced within libssl.
6882 This issue was reported to OpenSSL Guido Vranken.
6887 * Side channel attack on modular exponentiation
6889 A side-channel attack was found which makes use of cache-bank conflicts on
6890 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6891 of RSA keys. The ability to exploit this issue is limited as it relies on
6892 an attacker who has control of code in a thread running on the same
6893 hyper-threaded core as the victim thread which is performing decryptions.
6895 This issue was reported to OpenSSL by Yuval Yarom, The University of
6896 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6897 Nadia Heninger, University of Pennsylvania with more information at
6898 <http://cachebleed.info>.
6903 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6904 if no keysize is specified with default_bits. This fixes an
6905 omission in an earlier change that changed all RSA/DSA key generation
6906 commands to use 2048 bits by default.
6910 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6912 * Protection for DH small subgroup attacks
6914 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6915 switched on by default and cannot be disabled. This could have some
6920 * SSLv2 doesn't block disabled ciphers
6922 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6923 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6924 been disabled, provided that the SSLv2 protocol was not also disabled via
6927 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6928 and Sebastian Schinzel.
6933 * Reject DH handshakes with parameters shorter than 1024 bits.
6937 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6939 * Certificate verify crash with missing PSS parameter
6941 The signature verification routines will crash with a NULL pointer
6942 dereference if presented with an ASN.1 signature using the RSA PSS
6943 algorithm and absent mask generation function parameter. Since these
6944 routines are used to verify certificate signature algorithms this can be
6945 used to crash any certificate verification operation and exploited in a
6946 DoS attack. Any application which performs certificate verification is
6947 vulnerable including OpenSSL clients and servers which enable client
6950 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6955 * X509_ATTRIBUTE memory leak
6957 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6958 memory. This structure is used by the PKCS#7 and CMS routines so any
6959 application which reads PKCS#7 or CMS data from untrusted sources is
6960 affected. SSL/TLS is not affected.
6962 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6968 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6969 This changes the decoding behaviour for some invalid messages,
6970 though the change is mostly in the more lenient direction, and
6971 legacy behaviour is preserved as much as possible.
6975 * In DSA_generate_parameters_ex, if the provided seed is too short,
6976 use a random seed, as already documented.
6978 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6980 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6982 * Alternate chains certificate forgery
6984 During certificate verfification, OpenSSL will attempt to find an
6985 alternative certificate chain if the first attempt to build such a chain
6986 fails. An error in the implementation of this logic can mean that an
6987 attacker could cause certain checks on untrusted certificates to be
6988 bypassed, such as the CA flag, enabling them to use a valid leaf
6989 certificate to act as a CA and "issue" an invalid certificate.
6991 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6997 * Race condition handling PSK identify hint
6999 If PSK identity hints are received by a multi-threaded client then
7000 the values are wrongly updated in the parent SSL_CTX structure. This can
7001 result in a race condition potentially leading to a double free of the
7007 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7009 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7010 incompatibility in the handling of HMAC. The previous ABI has now been
7013 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7015 * Malformed ECParameters causes infinite loop
7017 When processing an ECParameters structure OpenSSL enters an infinite loop
7018 if the curve specified is over a specially malformed binary polynomial
7021 This can be used to perform denial of service against any
7022 system which processes public keys, certificate requests or
7023 certificates. This includes TLS clients and TLS servers with
7024 client authentication enabled.
7026 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7031 * Exploitable out-of-bounds read in X509_cmp_time
7033 X509_cmp_time does not properly check the length of the ASN1_TIME
7034 string and can read a few bytes out of bounds. In addition,
7035 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7038 An attacker can use this to craft malformed certificates and CRLs of
7039 various sizes and potentially cause a segmentation fault, resulting in
7040 a DoS on applications that verify certificates or CRLs. TLS clients
7041 that verify CRLs are affected. TLS clients and servers with client
7042 authentication enabled may be affected if they use custom verification
7045 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7046 independently by Hanno Böck.
7051 * PKCS7 crash with missing EnvelopedContent
7053 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7054 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7055 with missing content and trigger a NULL pointer dereference on parsing.
7057 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7058 structures from untrusted sources are affected. OpenSSL clients and
7059 servers are not affected.
7061 This issue was reported to OpenSSL by Michal Zalewski (Google).
7066 * CMS verify infinite loop with unknown hash function
7068 When verifying a signedData message the CMS code can enter an infinite loop
7069 if presented with an unknown hash function OID. This can be used to perform
7070 denial of service against any system which verifies signedData messages using
7072 This issue was reported to OpenSSL by Johannes Bauer.
7077 * Race condition handling NewSessionTicket
7079 If a NewSessionTicket is received by a multi-threaded client when attempting to
7080 reuse a previous ticket then a race condition can occur potentially leading to
7081 a double free of the ticket data.
7086 * Reject DH handshakes with parameters shorter than 768 bits.
7088 *Kurt Roeckx and Emilia Kasper*
7090 * dhparam: generate 2048-bit parameters by default.
7092 *Kurt Roeckx and Emilia Kasper*
7094 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7096 * Segmentation fault in ASN1_TYPE_cmp fix
7098 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7099 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7100 certificate signature algorithm consistency this can be used to crash any
7101 certificate verification operation and exploited in a DoS attack. Any
7102 application which performs certificate verification is vulnerable including
7103 OpenSSL clients and servers which enable client authentication.
7108 * ASN.1 structure reuse memory corruption fix
7110 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7111 memory corruption via an invalid write. Such reuse is and has been
7112 strongly discouraged and is believed to be rare.
7114 Applications that parse structures containing CHOICE or ANY DEFINED BY
7115 components may be affected. Certificate parsing (d2i_X509 and related
7116 functions) are however not affected. OpenSSL clients and servers are
7122 * PKCS7 NULL pointer dereferences fix
7124 The PKCS#7 parsing code does not handle missing outer ContentInfo
7125 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7126 missing content and trigger a NULL pointer dereference on parsing.
7128 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7129 otherwise parse PKCS#7 structures from untrusted sources are
7130 affected. OpenSSL clients and servers are not affected.
7132 This issue was reported to OpenSSL by Michal Zalewski (Google).
7137 * DoS via reachable assert in SSLv2 servers fix
7139 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7140 servers that both support SSLv2 and enable export cipher suites by sending
7141 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7143 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7144 (OpenSSL development team).
7149 * Use After Free following d2i_ECPrivatekey error fix
7151 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7152 could cause a use after free condition. This, in turn, could cause a double
7153 free in several private key parsing functions (such as d2i_PrivateKey
7154 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7155 for applications that receive EC private keys from untrusted
7156 sources. This scenario is considered rare.
7158 This issue was discovered by the BoringSSL project and fixed in their
7164 * X509_to_X509_REQ NULL pointer deref fix
7166 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7167 the certificate key is invalid. This function is rarely used in practice.
7169 This issue was discovered by Brian Carpenter.
7174 * Removed the export ciphers from the DEFAULT ciphers
7178 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7180 * Build fixes for the Windows and OpenVMS platforms
7182 *Matt Caswell and Richard Levitte*
7184 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7186 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7187 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7188 dereference. This could lead to a Denial Of Service attack. Thanks to
7189 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7194 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7195 dtls1_buffer_record function under certain conditions. In particular this
7196 could occur if an attacker sent repeated DTLS records with the same
7197 sequence number but for the next epoch. The memory leak could be exploited
7198 by an attacker in a Denial of Service attack through memory exhaustion.
7199 Thanks to Chris Mueller for reporting this issue.
7204 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7205 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7206 method would be set to NULL which could later result in a NULL pointer
7207 dereference. Thanks to Frank Schmirler for reporting this issue.
7212 * Abort handshake if server key exchange message is omitted for ephemeral
7215 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7216 reporting this issue.
7221 * Remove non-export ephemeral RSA code on client and server. This code
7222 violated the TLS standard by allowing the use of temporary RSA keys in
7223 non-export ciphersuites and could be used by a server to effectively
7224 downgrade the RSA key length used to a value smaller than the server
7225 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7226 INRIA or reporting this issue.
7231 * Fixed issue where DH client certificates are accepted without verification.
7232 An OpenSSL server will accept a DH certificate for client authentication
7233 without the certificate verify message. This effectively allows a client to
7234 authenticate without the use of a private key. This only affects servers
7235 which trust a client certificate authority which issues certificates
7236 containing DH keys: these are extremely rare and hardly ever encountered.
7237 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7243 * Ensure that the session ID context of an SSL is updated when its
7244 SSL_CTX is updated via SSL_set_SSL_CTX.
7246 The session ID context is typically set from the parent SSL_CTX,
7247 and can vary with the CTX.
7251 * Fix various certificate fingerprint issues.
7253 By using non-DER or invalid encodings outside the signed portion of a
7254 certificate the fingerprint can be changed without breaking the signature.
7255 Although no details of the signed portion of the certificate can be changed
7256 this can cause problems with some applications: e.g. those using the
7257 certificate fingerprint for blacklists.
7259 1. Reject signatures with non zero unused bits.
7261 If the BIT STRING containing the signature has non zero unused bits reject
7262 the signature. All current signature algorithms require zero unused bits.
7264 2. Check certificate algorithm consistency.
7266 Check the AlgorithmIdentifier inside TBS matches the one in the
7267 certificate signature. NB: this will result in signature failure
7268 errors for some broken certificates.
7270 Thanks to Konrad Kraszewski from Google for reporting this issue.
7272 3. Check DSA/ECDSA signatures use DER.
7274 Re-encode DSA/ECDSA signatures and compare with the original received
7275 signature. Return an error if there is a mismatch.
7277 This will reject various cases including garbage after signature
7278 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7279 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7280 (negative or with leading zeroes).
7282 Further analysis was conducted and fixes were developed by Stephen Henson
7283 of the OpenSSL core team.
7289 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7290 results on some platforms, including x86_64. This bug occurs at random
7291 with a very low probability, and is not known to be exploitable in any
7292 way, though its exact impact is difficult to determine. Thanks to Pieter
7293 Wuille (Blockstream) who reported this issue and also suggested an initial
7294 fix. Further analysis was conducted by the OpenSSL development team and
7295 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7296 the OpenSSL core team.
7301 * Do not resume sessions on the server if the negotiated protocol
7302 version does not match the session's version. Resuming with a different
7303 version, while not strictly forbidden by the RFC, is of questionable
7304 sanity and breaks all known clients.
7306 *David Benjamin, Emilia Käsper*
7308 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7309 early CCS messages during renegotiation. (Note that because
7310 renegotiation is encrypted, this early CCS was not exploitable.)
7314 * Tighten client-side session ticket handling during renegotiation:
7315 ensure that the client only accepts a session ticket if the server sends
7316 the extension anew in the ServerHello. Previously, a TLS client would
7317 reuse the old extension state and thus accept a session ticket if one was
7318 announced in the initial ServerHello.
7320 Similarly, ensure that the client requires a session ticket if one
7321 was advertised in the ServerHello. Previously, a TLS client would
7322 ignore a missing NewSessionTicket message.
7326 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7330 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7331 sends a carefully crafted handshake message, to cause OpenSSL to fail
7332 to free up to 64k of memory causing a memory leak. This could be
7333 exploited in a Denial Of Service attack. This issue affects OpenSSL
7334 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7335 whether SRTP is used or configured. Implementations of OpenSSL that
7336 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7338 The fix was developed by the OpenSSL team.
7343 * Session Ticket Memory Leak.
7345 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7346 integrity of that ticket is first verified. In the event of a session
7347 ticket integrity check failing, OpenSSL will fail to free memory
7348 causing a memory leak. By sending a large number of invalid session
7349 tickets an attacker could exploit this issue in a Denial Of Service
7355 * Build option no-ssl3 is incomplete.
7357 When OpenSSL is configured with "no-ssl3" as a build option, servers
7358 could accept and complete an SSL 3.0 handshake, and clients could be
7359 configured to send them.
7362 *Akamai and the OpenSSL team*
7364 * Add support for TLS_FALLBACK_SCSV.
7365 Client applications doing fallback retries should call
7366 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7369 *Adam Langley, Bodo Moeller*
7371 * Add additional DigestInfo checks.
7373 Re-encode DigestInto in DER and check against the original when
7374 verifying RSA signature: this will reject any improperly encoded
7375 DigestInfo structures.
7377 Note: this is a precautionary measure and no attacks are currently known.
7381 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7383 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7384 SRP code can be overrun an internal buffer. Add sanity check that
7385 g, A, B < N to SRP code.
7387 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7388 Group for discovering this issue.
7393 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7394 TLS 1.0 instead of higher protocol versions when the ClientHello message
7395 is badly fragmented. This allows a man-in-the-middle attacker to force a
7396 downgrade to TLS 1.0 even if both the server and the client support a
7397 higher protocol version, by modifying the client's TLS records.
7399 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7400 researching this issue.
7405 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7406 to a denial of service attack. A malicious server can crash the client
7407 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7408 ciphersuite and sending carefully crafted handshake messages.
7410 Thanks to Felix Gröbert (Google) for discovering and researching this
7416 * By sending carefully crafted DTLS packets an attacker could cause openssl
7417 to leak memory. This can be exploited through a Denial of Service attack.
7418 Thanks to Adam Langley for discovering and researching this issue.
7423 * An attacker can force openssl to consume large amounts of memory whilst
7424 processing DTLS handshake messages. This can be exploited through a
7425 Denial of Service attack.
7426 Thanks to Adam Langley for discovering and researching this issue.
7431 * An attacker can force an error condition which causes openssl to crash
7432 whilst processing DTLS packets due to memory being freed twice. This
7433 can be exploited through a Denial of Service attack.
7434 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7440 * If a multithreaded client connects to a malicious server using a resumed
7441 session and the server sends an ec point format extension it could write
7442 up to 255 bytes to freed memory.
7444 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7450 * A malicious server can crash an OpenSSL client with a null pointer
7451 dereference (read) by specifying an SRP ciphersuite even though it was not
7452 properly negotiated with the client. This can be exploited through a
7453 Denial of Service attack.
7455 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7456 discovering and researching this issue.
7461 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7462 X509_name_oneline, X509_name_print_ex et al. to leak some information
7463 from the stack. Applications may be affected if they echo pretty printing
7464 output to the attacker.
7466 Thanks to Ivan Fratric (Google) for discovering this issue.
7469 *Emilia Käsper, and Steve Henson*
7471 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7472 for corner cases. (Certain input points at infinity could lead to
7473 bogus results, with non-infinity inputs mapped to infinity too.)
7477 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7479 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7480 handshake can force the use of weak keying material in OpenSSL
7481 SSL/TLS clients and servers.
7483 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7484 researching this issue. ([CVE-2014-0224])
7486 *KIKUCHI Masashi, Steve Henson*
7488 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7489 OpenSSL DTLS client the code can be made to recurse eventually crashing
7492 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7495 *Imre Rad, Steve Henson*
7497 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7498 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7499 client or server. This is potentially exploitable to run arbitrary
7500 code on a vulnerable client or server.
7502 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7504 *Jüri Aedla, Steve Henson*
7506 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7507 are subject to a denial of service attack.
7509 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7510 this issue. ([CVE-2014-3470])
7512 *Felix Gröbert, Ivan Fratric, Steve Henson*
7514 * Harmonize version and its documentation. -f flag is used to display
7517 *mancha <mancha1@zoho.com>*
7519 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7520 in i2d_ECPrivateKey.
7522 *mancha <mancha1@zoho.com>*
7524 * Fix some double frees. These are not thought to be exploitable.
7526 *mancha <mancha1@zoho.com>*
7528 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7530 * A missing bounds check in the handling of the TLS heartbeat extension
7531 can be used to reveal up to 64k of memory to a connected client or
7534 Thanks for Neel Mehta of Google Security for discovering this bug and to
7535 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7536 preparing the fix ([CVE-2014-0160])
7538 *Adam Langley, Bodo Moeller*
7540 * Fix for the attack described in the paper "Recovering OpenSSL
7541 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7542 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7543 <http://eprint.iacr.org/2014/140>
7545 Thanks to Yuval Yarom and Naomi Benger for discovering this
7546 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7548 *Yuval Yarom and Naomi Benger*
7550 * TLS pad extension: draft-agl-tls-padding-03
7552 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7553 TLS client Hello record length value would otherwise be > 255 and
7554 less that 512 pad with a dummy extension containing zeroes so it
7555 is at least 512 bytes long.
7557 *Adam Langley, Steve Henson*
7559 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7561 * Fix for TLS record tampering bug. A carefully crafted invalid
7562 handshake could crash OpenSSL with a NULL pointer exception.
7563 Thanks to Anton Johansson for reporting this issues.
7566 * Keep original DTLS digest and encryption contexts in retransmission
7567 structures so we can use the previous session parameters if they need
7568 to be resent. ([CVE-2013-6450])
7572 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7573 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7574 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7575 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7576 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7577 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7579 *Rob Stradling, Adam Langley*
7581 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7583 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7584 supporting platforms or when small records were transferred.
7586 *Andy Polyakov, Steve Henson*
7588 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7590 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7592 This addresses the flaw in CBC record processing discovered by
7593 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7594 at: <http://www.isg.rhul.ac.uk/tls/>
7596 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7597 Security Group at Royal Holloway, University of London
7598 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7599 Emilia Käsper for the initial patch.
7602 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7604 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7605 ciphersuites which can be exploited in a denial of service attack.
7606 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7607 and detecting this bug and to Wolfgang Ettlinger
7608 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7613 * Return an error when checking OCSP signatures when key is NULL.
7614 This fixes a DoS attack. ([CVE-2013-0166])
7618 * Make openssl verify return errors.
7620 *Chris Palmer <palmer@google.com> and Ben Laurie*
7622 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7623 the right response is stapled. Also change SSL_get_certificate()
7624 so it returns the certificate actually sent.
7625 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7627 *Rob Stradling <rob.stradling@comodo.com>*
7629 * Fix possible deadlock when decoding public keys.
7633 * Don't use TLS 1.0 record version number in initial client hello
7638 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7640 * Sanity check record length before skipping explicit IV in TLS
7641 1.2, 1.1 and DTLS to fix DoS attack.
7643 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7644 fuzzing as a service testing platform.
7649 * Initialise tkeylen properly when encrypting CMS messages.
7650 Thanks to Solar Designer of Openwall for reporting this issue.
7654 * In FIPS mode don't try to use composite ciphers as they are not
7659 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7661 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7662 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7663 mean any application compiled against OpenSSL 1.0.0 headers setting
7664 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7665 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7666 0x10000000L Any application which was previously compiled against
7667 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7668 will need to be recompiled as a result. Letting be results in
7669 inability to disable specifically TLS 1.1 and in client context,
7670 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7674 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7675 disable just protocol X, but all protocols above X *if* there are
7676 protocols *below* X still enabled. In more practical terms it means
7677 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7678 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7679 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7684 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7686 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7687 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7688 in CRYPTO_realloc_clean.
7690 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7691 issue and to Adam Langley <agl@chromium.org> for fixing it.
7694 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7696 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7700 * Workarounds for some broken servers that "hang" if a client hello
7701 record length exceeds 255 bytes.
7703 1. Do not use record version number > TLS 1.0 in initial client
7704 hello: some (but not all) hanging servers will now work.
7705 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7706 the number of ciphers sent in the client hello. This should be
7707 set to an even number, such as 50, for example by passing:
7708 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7709 Most broken servers should now work.
7710 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7711 TLS 1.2 client support entirely.
7715 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7719 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7721 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7722 STRING form instead of a DigestInfo.
7726 * The format used for MDC2 RSA signatures is inconsistent between EVP
7727 and the RSA_sign/RSA_verify functions. This was made more apparent when
7728 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7729 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7730 the correct format in RSA_verify so both forms transparently work.
7734 * Some servers which support TLS 1.0 can choke if we initially indicate
7735 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7736 encrypted premaster secret. As a workaround use the maximum permitted
7737 client version in client hello, this should keep such servers happy
7738 and still work with previous versions of OpenSSL.
7742 * Add support for TLS/DTLS heartbeats.
7744 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7746 * Add support for SCTP.
7748 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7750 * Improved PRNG seeding for VOS.
7752 *Paul Green <Paul.Green@stratus.com>*
7754 * Extensive assembler packs updates, most notably:
7756 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7757 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7758 - x86_64: bit-sliced AES implementation;
7759 - ARM: NEON support, contemporary platforms optimizations;
7760 - s390x: z196 support;
7761 - `*`: GHASH and GF(2^m) multiplication implementations;
7765 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7766 (removal of unnecessary code)
7768 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7770 * Add TLS key material exporter from RFC 5705.
7774 * Add DTLS-SRTP negotiation from RFC 5764.
7778 * Add Next Protocol Negotiation,
7779 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7780 disabled with a no-npn flag to config or Configure. Code donated
7783 *Adam Langley <agl@google.com> and Ben Laurie*
7785 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7786 NIST-P256, NIST-P521, with constant-time single point multiplication on
7787 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7788 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7789 Code made available under Apache License version 2.0.
7791 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7792 line to include this in your build of OpenSSL, and run "make depend" (or
7793 "make update"). This enables the following EC_METHODs:
7795 EC_GFp_nistp224_method()
7796 EC_GFp_nistp256_method()
7797 EC_GFp_nistp521_method()
7799 EC_GROUP_new_by_curve_name() will automatically use these (while
7800 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7803 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7805 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7806 all platforms. Move ssize_t definition from e_os.h to the public
7807 header file e_os2.h as it now appears in public header file cms.h
7811 * New -sigopt option to the ca, req and x509 utilities. Additional
7812 signature parameters can be passed using this option and in
7817 * Add RSA PSS signing function. This will generate and set the
7818 appropriate AlgorithmIdentifiers for PSS based on those in the
7819 corresponding EVP_MD_CTX structure. No application support yet.
7823 * Support for companion algorithm specific ASN1 signing routines.
7824 New function ASN1_item_sign_ctx() signs a pre-initialised
7825 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7826 the appropriate parameters.
7830 * Add new algorithm specific ASN1 verification initialisation function
7831 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7832 handling will be the same no matter what EVP_PKEY_METHOD is used.
7833 Add a PSS handler to support verification of PSS signatures: checked
7834 against a number of sample certificates.
7838 * Add signature printing for PSS. Add PSS OIDs.
7840 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7842 * Add algorithm specific signature printing. An individual ASN1 method
7843 can now print out signatures instead of the standard hex dump.
7845 More complex signatures (e.g. PSS) can print out more meaningful
7846 information. Include DSA version that prints out the signature
7851 * Password based recipient info support for CMS library: implementing
7856 * Split password based encryption into PBES2 and PBKDF2 functions. This
7857 neatly separates the code into cipher and PBE sections and is required
7858 for some algorithms that split PBES2 into separate pieces (such as
7859 password based CMS).
7863 * Session-handling fixes:
7864 - Fix handling of connections that are resuming with a session ID,
7865 but also support Session Tickets.
7866 - Fix a bug that suppressed issuing of a new ticket if the client
7867 presented a ticket with an expired session.
7868 - Try to set the ticket lifetime hint to something reasonable.
7869 - Make tickets shorter by excluding irrelevant information.
7870 - On the client side, don't ignore renewed tickets.
7872 *Adam Langley, Bodo Moeller (Google)*
7874 * Fix PSK session representation.
7878 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7880 This work was sponsored by Intel.
7884 * Add GCM support to TLS library. Some custom code is needed to split
7885 the IV between the fixed (from PRF) and explicit (from TLS record)
7886 portions. This adds all GCM ciphersuites supported by RFC5288 and
7887 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7888 add a special AESGCM string for GCM only.
7892 * Expand range of ctrls for AES GCM. Permit setting invocation
7893 field on decrypt and retrieval of invocation field only on encrypt.
7897 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7898 As required by RFC5289 these ciphersuites cannot be used if for
7899 versions of TLS earlier than 1.2.
7903 * For FIPS capable OpenSSL interpret a NULL default public key method
7904 as unset and return the appropriate default but do *not* set the default.
7905 This means we can return the appropriate method in applications that
7906 switch between FIPS and non-FIPS modes.
7910 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7911 ENGINE is used then we cannot handle that in the FIPS module so we
7912 keep original code iff non-FIPS operations are allowed.
7916 * Add -attime option to openssl utilities.
7918 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7920 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7924 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7925 FIPS EC methods unconditionally for now.
7929 * New build option no-ec2m to disable characteristic 2 code.
7933 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7934 all cases can be covered as some introduce binary incompatibilities.
7938 * Redirect RSA operations to FIPS module including keygen,
7939 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7943 * Add similar low-level API blocking to ciphers.
7947 * low-level digest APIs are not approved in FIPS mode: any attempt
7948 to use these will cause a fatal error. Applications that *really* want
7949 to use them can use the `private_*` version instead.
7953 * Redirect cipher operations to FIPS module for FIPS builds.
7957 * Redirect digest operations to FIPS module for FIPS builds.
7961 * Update build system to add "fips" flag which will link in fipscanister.o
7962 for static and shared library builds embedding a signature if needed.
7966 * Output TLS supported curves in preference order instead of numerical
7967 order. This is currently hardcoded for the highest order curves first.
7968 This should be configurable so applications can judge speed vs strength.
7972 * Add TLS v1.2 server support for client authentication.
7976 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7981 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7982 FIPS modules versions.
7986 * Add TLS v1.2 client side support for client authentication. Keep cache
7987 of handshake records longer as we don't know the hash algorithm to use
7988 until after the certificate request message is received.
7992 * Initial TLS v1.2 client support. Add a default signature algorithms
7993 extension including all the algorithms we support. Parse new signature
7994 format in client key exchange. Relax some ECC signing restrictions for
7995 TLS v1.2 as indicated in RFC5246.
7999 * Add server support for TLS v1.2 signature algorithms extension. Switch
8000 to new signature format when needed using client digest preference.
8001 All server ciphersuites should now work correctly in TLS v1.2. No client
8002 support yet and no support for client certificates.
8006 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8007 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8008 ciphersuites. At present only RSA key exchange ciphersuites work with
8009 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8010 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8011 and version checking.
8015 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8016 with this defined it will not be affected by any changes to ssl internal
8017 structures. Add several utility functions to allow openssl application
8018 to work with OPENSSL_NO_SSL_INTERN defined.
8022 * A long standing patch to add support for SRP from EdelWeb (Peter
8023 Sylvester and Christophe Renou) was integrated.
8024 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8025 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8028 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8032 * Permit abbreviated handshakes when renegotiating using the function
8033 SSL_renegotiate_abbreviated().
8035 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8037 * Add call to ENGINE_register_all_complete() to
8038 ENGINE_load_builtin_engines(), so some implementations get used
8039 automatically instead of needing explicit application support.
8043 * Add support for TLS key exporter as described in RFC5705.
8045 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8047 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8048 a few changes are required:
8050 Add SSL_OP_NO_TLSv1_1 flag.
8051 Add TLSv1_1 methods.
8052 Update version checking logic to handle version 1.1.
8053 Add explicit IV handling (ported from DTLS code).
8054 Add command line options to s_client/s_server.
8061 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8063 * X509_ATTRIBUTE memory leak
8065 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8066 memory. This structure is used by the PKCS#7 and CMS routines so any
8067 application which reads PKCS#7 or CMS data from untrusted sources is
8068 affected. SSL/TLS is not affected.
8070 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8076 * Race condition handling PSK identify hint
8078 If PSK identity hints are received by a multi-threaded client then
8079 the values are wrongly updated in the parent SSL_CTX structure. This can
8080 result in a race condition potentially leading to a double free of the
8086 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8088 * Malformed ECParameters causes infinite loop
8090 When processing an ECParameters structure OpenSSL enters an infinite loop
8091 if the curve specified is over a specially malformed binary polynomial
8094 This can be used to perform denial of service against any
8095 system which processes public keys, certificate requests or
8096 certificates. This includes TLS clients and TLS servers with
8097 client authentication enabled.
8099 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8104 * Exploitable out-of-bounds read in X509_cmp_time
8106 X509_cmp_time does not properly check the length of the ASN1_TIME
8107 string and can read a few bytes out of bounds. In addition,
8108 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8111 An attacker can use this to craft malformed certificates and CRLs of
8112 various sizes and potentially cause a segmentation fault, resulting in
8113 a DoS on applications that verify certificates or CRLs. TLS clients
8114 that verify CRLs are affected. TLS clients and servers with client
8115 authentication enabled may be affected if they use custom verification
8118 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8119 independently by Hanno Böck.
8124 * PKCS7 crash with missing EnvelopedContent
8126 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8127 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8128 with missing content and trigger a NULL pointer dereference on parsing.
8130 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8131 structures from untrusted sources are affected. OpenSSL clients and
8132 servers are not affected.
8134 This issue was reported to OpenSSL by Michal Zalewski (Google).
8139 * CMS verify infinite loop with unknown hash function
8141 When verifying a signedData message the CMS code can enter an infinite loop
8142 if presented with an unknown hash function OID. This can be used to perform
8143 denial of service against any system which verifies signedData messages using
8145 This issue was reported to OpenSSL by Johannes Bauer.
8150 * Race condition handling NewSessionTicket
8152 If a NewSessionTicket is received by a multi-threaded client when attempting to
8153 reuse a previous ticket then a race condition can occur potentially leading to
8154 a double free of the ticket data.
8159 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8161 * Segmentation fault in ASN1_TYPE_cmp fix
8163 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8164 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8165 certificate signature algorithm consistency this can be used to crash any
8166 certificate verification operation and exploited in a DoS attack. Any
8167 application which performs certificate verification is vulnerable including
8168 OpenSSL clients and servers which enable client authentication.
8173 * ASN.1 structure reuse memory corruption fix
8175 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8176 memory corruption via an invalid write. Such reuse is and has been
8177 strongly discouraged and is believed to be rare.
8179 Applications that parse structures containing CHOICE or ANY DEFINED BY
8180 components may be affected. Certificate parsing (d2i_X509 and related
8181 functions) are however not affected. OpenSSL clients and servers are
8187 * PKCS7 NULL pointer dereferences fix
8189 The PKCS#7 parsing code does not handle missing outer ContentInfo
8190 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8191 missing content and trigger a NULL pointer dereference on parsing.
8193 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8194 otherwise parse PKCS#7 structures from untrusted sources are
8195 affected. OpenSSL clients and servers are not affected.
8197 This issue was reported to OpenSSL by Michal Zalewski (Google).
8202 * DoS via reachable assert in SSLv2 servers fix
8204 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8205 servers that both support SSLv2 and enable export cipher suites by sending
8206 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8208 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8209 (OpenSSL development team).
8214 * Use After Free following d2i_ECPrivatekey error fix
8216 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8217 could cause a use after free condition. This, in turn, could cause a double
8218 free in several private key parsing functions (such as d2i_PrivateKey
8219 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8220 for applications that receive EC private keys from untrusted
8221 sources. This scenario is considered rare.
8223 This issue was discovered by the BoringSSL project and fixed in their
8229 * X509_to_X509_REQ NULL pointer deref fix
8231 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8232 the certificate key is invalid. This function is rarely used in practice.
8234 This issue was discovered by Brian Carpenter.
8239 * Removed the export ciphers from the DEFAULT ciphers
8243 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8245 * Build fixes for the Windows and OpenVMS platforms
8247 *Matt Caswell and Richard Levitte*
8249 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8251 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8252 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8253 dereference. This could lead to a Denial Of Service attack. Thanks to
8254 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8259 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8260 dtls1_buffer_record function under certain conditions. In particular this
8261 could occur if an attacker sent repeated DTLS records with the same
8262 sequence number but for the next epoch. The memory leak could be exploited
8263 by an attacker in a Denial of Service attack through memory exhaustion.
8264 Thanks to Chris Mueller for reporting this issue.
8269 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8270 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8271 method would be set to NULL which could later result in a NULL pointer
8272 dereference. Thanks to Frank Schmirler for reporting this issue.
8277 * Abort handshake if server key exchange message is omitted for ephemeral
8280 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8281 reporting this issue.
8286 * Remove non-export ephemeral RSA code on client and server. This code
8287 violated the TLS standard by allowing the use of temporary RSA keys in
8288 non-export ciphersuites and could be used by a server to effectively
8289 downgrade the RSA key length used to a value smaller than the server
8290 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8291 INRIA or reporting this issue.
8296 * Fixed issue where DH client certificates are accepted without verification.
8297 An OpenSSL server will accept a DH certificate for client authentication
8298 without the certificate verify message. This effectively allows a client to
8299 authenticate without the use of a private key. This only affects servers
8300 which trust a client certificate authority which issues certificates
8301 containing DH keys: these are extremely rare and hardly ever encountered.
8302 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8308 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8309 results on some platforms, including x86_64. This bug occurs at random
8310 with a very low probability, and is not known to be exploitable in any
8311 way, though its exact impact is difficult to determine. Thanks to Pieter
8312 Wuille (Blockstream) who reported this issue and also suggested an initial
8313 fix. Further analysis was conducted by the OpenSSL development team and
8314 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8315 the OpenSSL core team.
8320 * Fix various certificate fingerprint issues.
8322 By using non-DER or invalid encodings outside the signed portion of a
8323 certificate the fingerprint can be changed without breaking the signature.
8324 Although no details of the signed portion of the certificate can be changed
8325 this can cause problems with some applications: e.g. those using the
8326 certificate fingerprint for blacklists.
8328 1. Reject signatures with non zero unused bits.
8330 If the BIT STRING containing the signature has non zero unused bits reject
8331 the signature. All current signature algorithms require zero unused bits.
8333 2. Check certificate algorithm consistency.
8335 Check the AlgorithmIdentifier inside TBS matches the one in the
8336 certificate signature. NB: this will result in signature failure
8337 errors for some broken certificates.
8339 Thanks to Konrad Kraszewski from Google for reporting this issue.
8341 3. Check DSA/ECDSA signatures use DER.
8343 Reencode DSA/ECDSA signatures and compare with the original received
8344 signature. Return an error if there is a mismatch.
8346 This will reject various cases including garbage after signature
8347 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8348 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8349 (negative or with leading zeroes).
8351 Further analysis was conducted and fixes were developed by Stephen Henson
8352 of the OpenSSL core team.
8358 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8360 * Session Ticket Memory Leak.
8362 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8363 integrity of that ticket is first verified. In the event of a session
8364 ticket integrity check failing, OpenSSL will fail to free memory
8365 causing a memory leak. By sending a large number of invalid session
8366 tickets an attacker could exploit this issue in a Denial Of Service
8372 * Build option no-ssl3 is incomplete.
8374 When OpenSSL is configured with "no-ssl3" as a build option, servers
8375 could accept and complete an SSL 3.0 handshake, and clients could be
8376 configured to send them.
8379 *Akamai and the OpenSSL team*
8381 * Add support for TLS_FALLBACK_SCSV.
8382 Client applications doing fallback retries should call
8383 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8386 *Adam Langley, Bodo Moeller*
8388 * Add additional DigestInfo checks.
8390 Reencode DigestInto in DER and check against the original when
8391 verifying RSA signature: this will reject any improperly encoded
8392 DigestInfo structures.
8394 Note: this is a precautionary measure and no attacks are currently known.
8398 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8400 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8401 to a denial of service attack. A malicious server can crash the client
8402 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8403 ciphersuite and sending carefully crafted handshake messages.
8405 Thanks to Felix Gröbert (Google) for discovering and researching this
8411 * By sending carefully crafted DTLS packets an attacker could cause openssl
8412 to leak memory. This can be exploited through a Denial of Service attack.
8413 Thanks to Adam Langley for discovering and researching this issue.
8418 * An attacker can force openssl to consume large amounts of memory whilst
8419 processing DTLS handshake messages. This can be exploited through a
8420 Denial of Service attack.
8421 Thanks to Adam Langley for discovering and researching this issue.
8426 * An attacker can force an error condition which causes openssl to crash
8427 whilst processing DTLS packets due to memory being freed twice. This
8428 can be exploited through a Denial of Service attack.
8429 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8435 * If a multithreaded client connects to a malicious server using a resumed
8436 session and the server sends an ec point format extension it could write
8437 up to 255 bytes to freed memory.
8439 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8445 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8446 X509_name_oneline, X509_name_print_ex et al. to leak some information
8447 from the stack. Applications may be affected if they echo pretty printing
8448 output to the attacker.
8450 Thanks to Ivan Fratric (Google) for discovering this issue.
8453 *Emilia Käsper, and Steve Henson*
8455 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8456 for corner cases. (Certain input points at infinity could lead to
8457 bogus results, with non-infinity inputs mapped to infinity too.)
8461 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8463 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8464 handshake can force the use of weak keying material in OpenSSL
8465 SSL/TLS clients and servers.
8467 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8468 researching this issue. ([CVE-2014-0224])
8470 *KIKUCHI Masashi, Steve Henson*
8472 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8473 OpenSSL DTLS client the code can be made to recurse eventually crashing
8476 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8479 *Imre Rad, Steve Henson*
8481 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8482 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8483 client or server. This is potentially exploitable to run arbitrary
8484 code on a vulnerable client or server.
8486 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8488 *Jüri Aedla, Steve Henson*
8490 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8491 are subject to a denial of service attack.
8493 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8494 this issue. ([CVE-2014-3470])
8496 *Felix Gröbert, Ivan Fratric, Steve Henson*
8498 * Harmonize version and its documentation. -f flag is used to display
8501 *mancha <mancha1@zoho.com>*
8503 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8504 in i2d_ECPrivateKey.
8506 *mancha <mancha1@zoho.com>*
8508 * Fix some double frees. These are not thought to be exploitable.
8510 *mancha <mancha1@zoho.com>*
8512 * Fix for the attack described in the paper "Recovering OpenSSL
8513 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8514 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8515 <http://eprint.iacr.org/2014/140>
8517 Thanks to Yuval Yarom and Naomi Benger for discovering this
8518 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8520 *Yuval Yarom and Naomi Benger*
8522 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8524 * Keep original DTLS digest and encryption contexts in retransmission
8525 structures so we can use the previous session parameters if they need
8526 to be resent. ([CVE-2013-6450])
8530 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8531 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8532 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8533 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8534 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8535 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8537 *Rob Stradling, Adam Langley*
8539 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8541 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8543 This addresses the flaw in CBC record processing discovered by
8544 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8545 at: <http://www.isg.rhul.ac.uk/tls/>
8547 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8548 Security Group at Royal Holloway, University of London
8549 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8550 Emilia Käsper for the initial patch.
8553 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8555 * Return an error when checking OCSP signatures when key is NULL.
8556 This fixes a DoS attack. ([CVE-2013-0166])
8560 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8561 the right response is stapled. Also change SSL_get_certificate()
8562 so it returns the certificate actually sent.
8563 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8564 (This is a backport)
8566 *Rob Stradling <rob.stradling@comodo.com>*
8568 * Fix possible deadlock when decoding public keys.
8572 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8574 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8577 * Sanity check record length before skipping explicit IV in DTLS
8580 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8581 fuzzing as a service testing platform.
8586 * Initialise tkeylen properly when encrypting CMS messages.
8587 Thanks to Solar Designer of Openwall for reporting this issue.
8591 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8593 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8594 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8595 in CRYPTO_realloc_clean.
8597 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8598 issue and to Adam Langley <agl@chromium.org> for fixing it.
8601 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8603 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8605 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8606 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8607 content decryption and always return the same error. Note: this attack
8608 needs on average 2^20 messages so it only affects automated senders. The
8609 old behaviour can be re-enabled in the CMS code by setting the
8610 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8611 an MMA defence is not necessary.
8612 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8613 this issue. ([CVE-2012-0884])
8617 * Fix CVE-2011-4619: make sure we really are receiving a
8618 client hello before rejecting multiple SGC restarts. Thanks to
8619 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8623 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8625 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8626 Thanks to Antonio Martin, Enterprise Secure Access Research and
8627 Development, Cisco Systems, Inc. for discovering this bug and
8628 preparing a fix. ([CVE-2012-0050])
8632 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8634 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8635 of the Vaudenay padding oracle attack on CBC mode encryption
8636 which enables an efficient plaintext recovery attack against
8637 the OpenSSL implementation of DTLS. Their attack exploits timing
8638 differences arising during decryption processing. A research
8639 paper describing this attack can be found at:
8640 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8641 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8642 Security Group at Royal Holloway, University of London
8643 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8644 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8645 for preparing the fix. ([CVE-2011-4108])
8647 *Robin Seggelmann, Michael Tuexen*
8649 * Clear bytes used for block padding of SSL 3.0 records.
8652 *Adam Langley (Google)*
8654 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8655 Kadianakis <desnacked@gmail.com> for discovering this issue and
8656 Adam Langley for preparing the fix. ([CVE-2011-4619])
8658 *Adam Langley (Google)*
8660 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8662 *Andrey Kulikov <amdeich@gmail.com>*
8664 * Prevent malformed RFC3779 data triggering an assertion failure.
8665 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8666 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8668 *Rob Austein <sra@hactrn.net>*
8670 * Improved PRNG seeding for VOS.
8672 *Paul Green <Paul.Green@stratus.com>*
8674 * Fix ssl_ciph.c set-up race.
8676 *Adam Langley (Google)*
8678 * Fix spurious failures in ecdsatest.c.
8680 *Emilia Käsper (Google)*
8682 * Fix the BIO_f_buffer() implementation (which was mixing different
8683 interpretations of the `..._len` fields).
8685 *Adam Langley (Google)*
8687 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8688 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8689 threads won't reuse the same blinding coefficients.
8691 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8692 lock to call BN_BLINDING_invert_ex, and avoids one use of
8693 BN_BLINDING_update for each BN_BLINDING structure (previously,
8694 the last update always remained unused).
8696 *Emilia Käsper (Google)*
8698 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8700 *Bob Buckholz (Google)*
8702 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8704 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8705 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8707 *Kaspar Brand <ossl@velox.ch>*
8709 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8710 for multi-threaded use of ECDH. ([CVE-2011-3210])
8712 *Adam Langley (Google)*
8714 * Fix x509_name_ex_d2i memory leak on bad inputs.
8718 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8719 signature public key algorithm by using OID xref utilities instead.
8720 Before this you could only use some ECC ciphersuites with SHA1 only.
8724 * Add protection against ECDSA timing attacks as mentioned in the paper
8725 by Billy Bob Brumley and Nicola Tuveri, see:
8726 <http://eprint.iacr.org/2011/232.pdf>
8728 *Billy Bob Brumley and Nicola Tuveri*
8730 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8732 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8734 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8736 * Fix bug in string printing code: if *any* escaping is enabled we must
8737 escape the escape character (backslash) or the resulting string is
8742 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8744 * Disable code workaround for ancient and obsolete Netscape browsers
8745 and servers: an attacker can use it in a ciphersuite downgrade attack.
8746 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8750 * Fixed J-PAKE implementation error, originally discovered by
8751 Sebastien Martini, further info and confirmation from Stefan
8752 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8756 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8758 * Fix extension code to avoid race conditions which can result in a buffer
8759 overrun vulnerability: resumed sessions must not be modified as they can
8760 be shared by multiple threads. CVE-2010-3864
8764 * Fix WIN32 build system to correctly link an ENGINE directory into
8769 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8771 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8774 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8776 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8778 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8779 context. The operation can be customised via the ctrl mechanism in
8780 case ENGINEs want to include additional functionality.
8784 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8788 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8789 output hashes compatible with older versions of OpenSSL.
8791 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8793 * Fix compression algorithm handling: if resuming a session use the
8794 compression algorithm of the resumed session instead of determining
8795 it from client hello again. Don't allow server to change algorithm.
8799 * Add load_crls() function to commands tidying load_certs() too. Add option
8800 to verify utility to allow additional CRLs to be included.
8804 * Update OCSP request code to permit adding custom headers to the request:
8805 some responders need this.
8809 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8812 *Julia Lawall <julia@diku.dk>*
8814 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8815 needlessly dereferenced structures, used obsolete functions and
8816 didn't handle all updated verify codes correctly.
8820 * Disable MD2 in the default configuration.
8824 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8825 indicate the initial BIO being pushed or popped. This makes it possible
8826 to determine whether the BIO is the one explicitly called or as a result
8827 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8828 it handles reference counts correctly and doesn't zero out the I/O bio
8829 when it is not being explicitly popped. WARNING: applications which
8830 included workarounds for the old buggy behaviour will need to be modified
8831 or they could free up already freed BIOs.
8835 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8836 renaming to all platforms (within the 0.9.8 branch, this was
8837 done conditionally on Netware platforms to avoid a name clash).
8839 *Guenter <lists@gknw.net>*
8841 * Add ECDHE and PSK support to DTLS.
8843 *Michael Tuexen <tuexen@fh-muenster.de>*
8845 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8850 * Add "missing" function EVP_MD_flags() (without this the only way to
8851 retrieve a digest flags is by accessing the structure directly. Update
8852 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8853 or cipher is registered as in the "from" argument. Print out all
8854 registered digests in the dgst usage message instead of manually
8855 attempting to work them out.
8859 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8860 this allows the use of compression and extensions. Change default cipher
8861 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8862 by default unless an application cipher string requests it.
8866 * Alter match criteria in PKCS12_parse(). It used to try to use local
8867 key ids to find matching certificates and keys but some PKCS#12 files
8868 don't follow the (somewhat unwritten) rules and this strategy fails.
8869 Now just gather all certificates together and the first private key
8870 then look for the first certificate that matches the key.
8874 * Support use of registered digest and cipher names for dgst and cipher
8875 commands instead of having to add each one as a special case. So now
8882 openssl dgst -sha256 foo
8884 and this works for ENGINE based algorithms too.
8888 * Update Gost ENGINE to support parameter files.
8890 *Victor B. Wagner <vitus@cryptocom.ru>*
8892 * Support GeneralizedTime in ca utility.
8894 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8896 * Enhance the hash format used for certificate directory links. The new
8897 form uses the canonical encoding (meaning equivalent names will work
8898 even if they aren't identical) and uses SHA1 instead of MD5. This form
8899 is incompatible with the older format and as a result c_rehash should
8900 be used to rebuild symbolic links.
8904 * Make PKCS#8 the default write format for private keys, replacing the
8905 traditional format. This form is standardised, more secure and doesn't
8906 include an implicit MD5 dependency.
8910 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8911 committed to OpenSSL should pass this lot as a minimum.
8915 * Add session ticket override functionality for use by EAP-FAST.
8917 *Jouni Malinen <j@w1.fi>*
8919 * Modify HMAC functions to return a value. Since these can be implemented
8920 in an ENGINE errors can occur.
8924 * Type-checked OBJ_bsearch_ex.
8928 * Type-checked OBJ_bsearch. Also some constification necessitated
8929 by type-checking. Still to come: TXT_DB, bsearch(?),
8930 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8935 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8936 seconds to a tm structure directly, instead of going through OS
8937 specific date routines. This avoids any issues with OS routines such
8938 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8939 and X509_time_adj_ex() to cover the extended range. The existing
8940 X509_time_adj() is still usable and will no longer have any date issues.
8944 * Delta CRL support. New use deltas option which will attempt to locate
8945 and search any appropriate delta CRLs available.
8947 This work was sponsored by Google.
8951 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8952 code and add additional score elements. Validate alternate CRL paths
8953 as part of the CRL checking and indicate a new error "CRL path validation
8954 error" in this case. Applications wanting additional details can use
8955 the verify callback and check the new "parent" field. If this is not
8956 NULL CRL path validation is taking place. Existing applications won't
8957 see this because it requires extended CRL support which is off by
8960 This work was sponsored by Google.
8964 * Support for freshest CRL extension.
8966 This work was sponsored by Google.
8970 * Initial indirect CRL support. Currently only supported in the CRLs
8971 passed directly and not via lookup. Process certificate issuer
8972 CRL entry extension and lookup CRL entries by bother issuer name
8973 and serial number. Check and process CRL issuer entry in IDP extension.
8975 This work was sponsored by Google.
8979 * Add support for distinct certificate and CRL paths. The CRL issuer
8980 certificate is validated separately in this case. Only enabled if
8981 an extended CRL support flag is set: this flag will enable additional
8982 CRL functionality in future.
8984 This work was sponsored by Google.
8988 * Add support for policy mappings extension.
8990 This work was sponsored by Google.
8994 * Fixes to pathlength constraint, self issued certificate handling,
8995 policy processing to align with RFC3280 and PKITS tests.
8997 This work was sponsored by Google.
9001 * Support for name constraints certificate extension. DN, email, DNS
9002 and URI types are currently supported.
9004 This work was sponsored by Google.
9008 * To cater for systems that provide a pointer-based thread ID rather
9009 than numeric, deprecate the current numeric thread ID mechanism and
9010 replace it with a structure and associated callback type. This
9011 mechanism allows a numeric "hash" to be extracted from a thread ID in
9012 either case, and on platforms where pointers are larger than 'long',
9013 mixing is done to help ensure the numeric 'hash' is usable even if it
9014 can't be guaranteed unique. The default mechanism is to use "&errno"
9015 as a pointer-based thread ID to distinguish between threads.
9017 Applications that want to provide their own thread IDs should now use
9018 CRYPTO_THREADID_set_callback() to register a callback that will call
9019 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9021 Note that ERR_remove_state() is now deprecated, because it is tied
9022 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9023 to free the current thread's error state should be replaced by
9024 ERR_remove_thread_state(NULL).
9026 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9027 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9028 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9029 application was previously providing a numeric thread callback that
9030 was inappropriate for distinguishing threads, then uniqueness might
9031 have been obtained with &errno that happened immediately in the
9032 intermediate development versions of OpenSSL; this is no longer the
9033 case, the numeric thread callback will now override the automatic use
9036 *Geoff Thorpe, with help from Bodo Moeller*
9038 * Initial support for different CRL issuing certificates. This covers a
9039 simple case where the self issued certificates in the chain exist and
9040 the real CRL issuer is higher in the existing chain.
9042 This work was sponsored by Google.
9046 * Removed effectively defunct crypto/store from the build.
9050 * Revamp of STACK to provide stronger type-checking. Still to come:
9051 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9052 ASN1_STRING, CONF_VALUE.
9056 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9057 RAM on SSL connections. This option can save about 34k per idle SSL.
9061 * Revamp of LHASH to provide stronger type-checking. Still to come:
9062 STACK, TXT_DB, bsearch, qsort.
9066 * Initial support for Cryptographic Message Syntax (aka CMS) based
9067 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9068 support for data, signedData, compressedData, digestedData and
9069 encryptedData, envelopedData types included. Scripts to check against
9070 RFC4134 examples draft and interop and consistency checks of many
9071 content types and variants.
9075 * Add options to enc utility to support use of zlib compression BIO.
9079 * Extend mk1mf to support importing of options and assembly language
9080 files from Configure script, currently only included in VC-WIN32.
9081 The assembly language rules can now optionally generate the source
9082 files from the associated perl scripts.
9086 * Implement remaining functionality needed to support GOST ciphersuites.
9087 Interop testing has been performed using CryptoPro implementations.
9089 *Victor B. Wagner <vitus@cryptocom.ru>*
9091 * s390x assembler pack.
9095 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9100 * Implement Opaque PRF Input TLS extension as specified in
9101 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9102 official specification yet and no extension type assignment by
9103 IANA exists, this extension (for now) will have to be explicitly
9104 enabled when building OpenSSL by providing the extension number
9105 to use. For example, specify an option
9107 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9109 to the "config" or "Configure" script to enable the extension,
9110 assuming extension number 0x9527 (which is a completely arbitrary
9111 and unofficial assignment based on the MD5 hash of the Internet
9112 Draft). Note that by doing so, you potentially lose
9113 interoperability with other TLS implementations since these might
9114 be using the same extension number for other purposes.
9116 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9117 opaque PRF input value to use in the handshake. This will create
9118 an internal copy of the length-'len' string at 'src', and will
9119 return non-zero for success.
9121 To get more control and flexibility, provide a callback function
9124 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9125 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9129 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9132 Callback function 'cb' will be called in handshakes, and is
9133 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9134 Argument 'arg' is for application purposes (the value as given to
9135 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9136 be provided to the callback function). The callback function
9137 has to return non-zero to report success: usually 1 to use opaque
9138 PRF input just if possible, or 2 to enforce use of the opaque PRF
9139 input. In the latter case, the library will abort the handshake
9140 if opaque PRF input is not successfully negotiated.
9142 Arguments 'peerinput' and 'len' given to the callback function
9143 will always be NULL and 0 in the case of a client. A server will
9144 see the client's opaque PRF input through these variables if
9145 available (NULL and 0 otherwise). Note that if the server
9146 provides an opaque PRF input, the length must be the same as the
9147 length of the client's opaque PRF input.
9149 Note that the callback function will only be called when creating
9150 a new session (session resumption can resume whatever was
9151 previously negotiated), and will not be called in SSL 2.0
9152 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9153 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9154 for applications that need to enforce opaque PRF input.
9158 * Update ssl code to support digests other than SHA1+MD5 for handshake
9161 *Victor B. Wagner <vitus@cryptocom.ru>*
9163 * Add RFC4507 support to OpenSSL. This includes the corrections in
9164 RFC4507bis. The encrypted ticket format is an encrypted encoded
9165 SSL_SESSION structure, that way new session features are automatically
9168 If a client application caches session in an SSL_SESSION structure
9169 support is transparent because tickets are now stored in the encoded
9172 The SSL_CTX structure automatically generates keys for ticket
9173 protection in servers so again support should be possible
9174 with no application modification.
9176 If a client or server wishes to disable RFC4507 support then the option
9177 SSL_OP_NO_TICKET can be set.
9179 Add a TLS extension debugging callback to allow the contents of any client
9180 or server extensions to be examined.
9182 This work was sponsored by Google.
9186 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9187 OpenSSL should now compile cleanly on gcc 4.2
9189 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9191 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9192 support including streaming MAC support: this is required for GOST
9193 ciphersuite support.
9195 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9197 * Add option -stream to use PKCS#7 streaming in smime utility. New
9198 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9199 to output in BER and PEM format.
9203 * Experimental support for use of HMAC via EVP_PKEY interface. This
9204 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9205 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9206 ENGINE support for HMAC keys which are unextractable. New -mac and
9207 -macopt options to dgst utility.
9211 * New option -sigopt to dgst utility. Update dgst to use
9212 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9213 alternative signing parameters such as X9.31 or PSS in the dgst
9218 * Change ssl_cipher_apply_rule(), the internal function that does
9219 the work each time a ciphersuite string requests enabling
9220 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9221 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9222 the order of disabled ciphersuites such that those ciphersuites
9223 that most recently went from enabled to disabled not only stay
9224 in order with respect to each other, but also have higher priority
9225 than other disabled ciphersuites the next time ciphersuites are
9228 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9229 the same ciphersuites as with "HIGH" alone, but in a specific
9230 order where the PSK ciphersuites come first (since they are the
9231 most recently disabled ciphersuites when "HIGH" is parsed).
9233 Also, change ssl_create_cipher_list() (using this new
9234 functionality) such that between otherwise identical
9235 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9240 * Change ssl_create_cipher_list() so that it automatically
9241 arranges the ciphersuites in reasonable order before starting
9242 to process the rule string. Thus, the definition for "DEFAULT"
9243 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9244 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9245 This makes it much easier to arrive at a reasonable default order
9246 in applications for which anonymous ciphers are OK (meaning
9247 that you can't actually use DEFAULT).
9249 *Bodo Moeller; suggested by Victor Duchovni*
9251 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9252 processing) into multiple integers instead of setting
9253 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9254 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9255 (These masks as well as the individual bit definitions are hidden
9256 away into the non-exported interface ssl/ssl_locl.h, so this
9257 change to the definition of the SSL_CIPHER structure shouldn't
9258 affect applications.) This give us more bits for each of these
9259 categories, so there is no longer a need to coagulate AES128 and
9260 AES256 into a single algorithm bit, and to coagulate Camellia128
9261 and Camellia256 into a single algorithm bit, which has led to all
9264 Thus, among other things, the kludge introduced in 0.9.7m and
9265 0.9.8e for masking out AES256 independently of AES128 or masking
9266 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9268 With the change, we also introduce new ciphersuite aliases that
9269 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9274 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9275 Use the leftmost N bytes of the signature input if the input is
9276 larger than the prime q (with N being the size in bytes of q).
9280 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9281 it yet and it is largely untested.
9285 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9289 * Initial incomplete changes to avoid need for function casts in OpenSSL
9290 some compilers (gcc 4.2 and later) reject their use. Safestack is
9291 reimplemented. Update ASN1 to avoid use of legacy functions.
9295 * Win32/64 targets are linked with Winsock2.
9299 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9300 to external functions. This can be used to increase CRL handling
9301 efficiency especially when CRLs are very large by (for example) storing
9302 the CRL revoked certificates in a database.
9306 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9307 new CRLs added to a directory can be used. New command line option
9308 -verify_return_error to s_client and s_server. This causes real errors
9309 to be returned by the verify callback instead of carrying on no matter
9310 what. This reflects the way a "real world" verify callback would behave.
9314 * GOST engine, supporting several GOST algorithms and public key formats.
9315 Kindly donated by Cryptocom.
9319 * Partial support for Issuing Distribution Point CRL extension. CRLs
9320 partitioned by DP are handled but no indirect CRL or reason partitioning
9321 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9322 selected via a scoring technique which handles IDP and AKID in CRLs.
9326 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9327 will ultimately be used for all verify operations: this will remove the
9328 X509_STORE dependency on certificate verification and allow alternative
9329 lookup methods. X509_STORE based implementations of these two callbacks.
9333 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9334 Modify get_crl() to find a valid (unexpired) CRL if possible.
9338 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9339 this would be called X509_CRL_cmp() but that name is already used by
9340 a function that just compares CRL issuer names. Cache several CRL
9341 extensions in X509_CRL structure and cache CRLDP in X509.
9345 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9346 this maps equivalent X509_NAME structures into a consistent structure.
9347 Name comparison can then be performed rapidly using memcmp().
9351 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9356 * Allow digests to supply their own micalg string for S/MIME type using
9357 the ctrl EVP_MD_CTRL_MICALG.
9361 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9362 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9363 ctrl. It can then customise the structure before and/or after signing
9368 * New function OBJ_add_sigid() to allow application defined signature OIDs
9369 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9370 to free up any added signature OIDs.
9374 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9375 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9376 digest and cipher tables. New options added to openssl utility:
9377 list-message-digest-algorithms and list-cipher-algorithms.
9381 * Change the array representation of binary polynomials: the list
9382 of degrees of non-zero coefficients is now terminated with -1.
9383 Previously it was terminated with 0, which was also part of the
9384 value; thus, the array representation was not applicable to
9385 polynomials where t^0 has coefficient zero. This change makes
9386 the array representation useful in a more general context.
9390 * Various modifications and fixes to SSL/TLS cipher string
9391 handling. For ECC, the code now distinguishes between fixed ECDH
9392 with RSA certificates on the one hand and with ECDSA certificates
9393 on the other hand, since these are separate ciphersuites. The
9394 unused code for Fortezza ciphersuites has been removed.
9396 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9397 (not "ECDHE"). For consistency with the code for DH
9398 certificates, use of ECDH certificates is now considered ECDH
9399 authentication, not RSA or ECDSA authentication (the latter is
9400 merely the CA's signing algorithm and not actively used in the
9403 The temporary ciphersuite alias "ECCdraft" is no longer
9404 available, and ECC ciphersuites are no longer excluded from "ALL"
9405 and "DEFAULT". The following aliases now exist for RFC 4492
9406 ciphersuites, most of these by analogy with the DH case:
9408 kECDHr - ECDH cert, signed with RSA
9409 kECDHe - ECDH cert, signed with ECDSA
9410 kECDH - ECDH cert (signed with either RSA or ECDSA)
9411 kEECDH - ephemeral ECDH
9412 ECDH - ECDH cert or ephemeral ECDH
9418 AECDH - anonymous ECDH
9419 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9423 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9424 Use correct micalg parameters depending on digest(s) in signed message.
9428 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9429 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9433 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9434 an engine to register a method. Add ENGINE lookups for methods and
9435 functional reference processing.
9439 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9440 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9445 * New -resign option to smime utility. This adds one or more signers
9446 to an existing PKCS#7 signedData structure. Also -md option to use an
9447 alternative message digest algorithm for signing.
9451 * Tidy up PKCS#7 routines and add new functions to make it easier to
9452 create PKCS7 structures containing multiple signers. Update smime
9453 application to support multiple signers.
9457 * New -macalg option to pkcs12 utility to allow setting of an alternative
9462 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9463 Reorganize PBE internals to lookup from a static table using NIDs,
9464 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9465 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9466 PRF which will be automatically used with PBES2.
9470 * Replace the algorithm specific calls to generate keys in "req" with the
9475 * Update PKCS#7 enveloped data routines to use new API. This is now
9476 supported by any public key method supporting the encrypt operation. A
9477 ctrl is added to allow the public key algorithm to examine or modify
9478 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9483 * Add a ctrl to asn1 method to allow a public key algorithm to express
9484 a default digest type to use. In most cases this will be SHA1 but some
9485 algorithms (such as GOST) need to specify an alternative digest. The
9486 return value indicates how strong the preference is 1 means optional and
9487 2 is mandatory (that is it is the only supported type). Modify
9488 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9489 use the default md. Update openssl utilities to use the default digest
9490 type for signing if it is not explicitly indicated.
9494 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9495 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9496 signing method from the key type. This effectively removes the link
9497 between digests and public key types.
9501 * Add an OID cross reference table and utility functions. Its purpose is to
9502 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9503 rsaEncryption. This will allow some of the algorithm specific hackery
9504 needed to use the correct OID to be removed.
9508 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9509 structures for PKCS7_sign(). They are now set up by the relevant public
9514 * Add provisional EC pkey method with support for ECDSA and ECDH.
9518 * Add support for key derivation (agreement) in the API, DH method and
9523 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9524 public and private key formats. As a side effect these add additional
9525 command line functionality not previously available: DSA signatures can be
9526 generated and verified using pkeyutl and DH key support and generation in
9533 *Oliver Tappe <zooey@hirschkaefer.de>*
9535 * New make target "install_html_docs" installs HTML renditions of the
9538 *Oliver Tappe <zooey@hirschkaefer.de>*
9540 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9541 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9542 support key and parameter generation and add initial key generation
9543 functionality for RSA.
9547 * Add functions for main EVP_PKEY_method operations. The undocumented
9548 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9549 `EVP_PKEY_{encrypt,decrypt}_old`.
9553 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9554 key API, doesn't do much yet.
9558 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9559 public key algorithms. New option to openssl utility:
9560 "list-public-key-algorithms" to print out info.
9564 * Implement the Supported Elliptic Curves Extension for
9565 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9569 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9570 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9574 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9575 utilities such as rsa, dsa, dsaparam etc except they process any key
9580 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9581 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9582 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9587 * Initial support for pluggable public key ASN1.
9588 De-spaghettify the public key ASN1 handling. Move public and private
9589 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9590 algorithm specific handling to a single module within the relevant
9591 algorithm directory. Add functions to allow (near) opaque processing
9592 of public and private key structures.
9596 * Implement the Supported Point Formats Extension for
9597 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9601 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9602 for the psk identity [hint] and the psk callback functions to the
9603 SSL_SESSION, SSL and SSL_CTX structure.
9606 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9610 SSL_CTX_use_psk_identity_hint
9611 SSL_get_psk_identity_hint
9612 SSL_get_psk_identity
9613 SSL_use_psk_identity_hint
9615 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9617 * Add RFC 3161 compliant time stamp request creation, response generation
9618 and response verification functionality.
9620 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9622 * Add initial support for TLS extensions, specifically for the server_name
9623 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9624 have new members for a hostname. The SSL data structure has an
9625 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9626 stored in that context to allow for session resumption, even after the
9627 SSL has been switched to a new SSL_CTX in reaction to a client's
9628 server_name extension.
9630 New functions (subject to change):
9632 SSL_get_servername()
9633 SSL_get_servername_type()
9636 New CTRL codes and macros (subject to change):
9638 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9639 - SSL_CTX_set_tlsext_servername_callback()
9640 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9641 - SSL_CTX_set_tlsext_servername_arg()
9642 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9644 openssl s_client has a new '-servername ...' option.
9646 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9647 '-key2 ...', '-servername_fatal' (subject to change). This allows
9648 testing the HostName extension for a specific single hostname ('-cert'
9649 and '-key' remain fallbacks for handshakes without HostName
9650 negotiation). If the unrecognized_name alert has to be sent, this by
9651 default is a warning; it becomes fatal with the '-servername_fatal'
9654 *Peter Sylvester, Remy Allais, Christophe Renou*
9656 * Whirlpool hash implementation is added.
9660 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9661 bn(64,32). Because of instruction set limitations it doesn't have
9662 any negative impact on performance. This was done mostly in order
9663 to make it possible to share assembler modules, such as bn_mul_mont
9664 implementations, between 32- and 64-bit builds without hassle.
9668 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9669 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9674 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9675 dedicated Montgomery multiplication procedure, is introduced.
9676 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9677 "64-bit" performance on certain 32-bit targets.
9681 * New option SSL_OP_NO_COMP to disable use of compression selectively
9682 in SSL structures. New SSL ctrl to set maximum send fragment size.
9683 Save memory by setting the I/O buffer sizes dynamically instead of
9684 using the maximum available value.
9688 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9689 in addition to the text details.
9693 * Very, very preliminary EXPERIMENTAL support for printing of general
9694 ASN1 structures. This currently produces rather ugly output and doesn't
9695 handle several customised structures at all.
9699 * Integrated support for PVK file format and some related formats such
9700 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9701 these in the 'rsa' and 'dsa' utilities.
9705 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9709 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9710 place for the (very old) "NETSCAPE" format certificates which are now
9711 handled using new ASN1 code equivalents.
9715 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9716 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9717 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9721 * Modify CRL distribution points extension code to print out previously
9722 unsupported fields. Enhance extension setting code to allow setting of
9727 * Add print and set support for Issuing Distribution Point CRL extension.
9731 * Change 'Configure' script to enable Camellia by default.
9738 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9740 * When rejecting SSL/TLS records due to an incorrect version number, never
9741 update s->server with a new major version number. As of
9742 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9743 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9744 the previous behavior could result in a read attempt at NULL when
9745 receiving specific incorrect SSL/TLS records once record payload
9746 protection is active. ([CVE-2010-0740])
9748 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9750 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9751 could be crashed if the relevant tables were not present (e.g. chrooted).
9753 *Tomas Hoger <thoger@redhat.com>*
9755 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9757 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9759 *Martin Olsson, Neel Mehta*
9761 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9762 accommodate for stack sorting, always a write lock!).
9766 * On some versions of WIN32 Heap32Next is very slow. This can cause
9767 excessive delays in the RAND_poll(): over a minute. As a workaround
9768 include a time check in the inner Heap32Next loop too.
9772 * The code that handled flushing of data in SSL/TLS originally used the
9773 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9774 the problem outlined in PR#1949. The fix suggested there however can
9775 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9776 of Apache). So instead simplify the code to flush unconditionally.
9777 This should be fine since flushing with no data to flush is a no op.
9781 * Handle TLS versions 2.0 and later properly and correctly use the
9782 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9783 off ancient servers have a habit of sticking around for a while...
9787 * Modify compression code so it frees up structures without using the
9788 ex_data callbacks. This works around a problem where some applications
9789 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9790 restarting) then use compression (e.g. SSL with compression) later.
9791 This results in significant per-connection memory leaks and
9792 has caused some security issues including CVE-2008-1678 and
9797 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9798 change when encrypting or decrypting.
9802 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9803 connect and renegotiate with servers which do not support RI.
9804 Until RI is more widely deployed this option is enabled by default.
9808 * Add "missing" ssl ctrls to clear options and mode.
9812 * If client attempts to renegotiate and doesn't support RI respond with
9813 a no_renegotiation alert as required by RFC5746. Some renegotiating
9814 TLS clients will continue a connection gracefully when they receive
9815 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9816 waiting for a server hello which it will never receive. Now we treat a
9817 received no_renegotiation alert as a fatal error. This is because
9818 applications requesting a renegotiation might well expect it to succeed
9819 and would have no code in place to handle the server denying it so the
9820 only safe thing to do is to terminate the connection.
9824 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9825 peer supports secure renegotiation and 0 otherwise. Print out peer
9826 renegotiation support in s_client/s_server.
9830 * Replace the highly broken and deprecated SPKAC certification method with
9831 the updated NID creation version. This should correctly handle UTF8.
9835 * Implement RFC5746. Re-enable renegotiation but require the extension
9836 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9837 turns out to be a bad idea. It has been replaced by
9838 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9839 SSL_CTX_set_options(). This is really not recommended unless you
9840 know what you are doing.
9842 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9844 * Fixes to stateless session resumption handling. Use initial_ctx when
9845 issuing and attempting to decrypt tickets in case it has changed during
9846 servername handling. Use a non-zero length session ID when attempting
9847 stateless session resumption: this makes it possible to determine if
9848 a resumption has occurred immediately after receiving server hello
9849 (several places in OpenSSL subtly assume this) instead of later in
9854 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9855 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9856 fixes for a few places where the return code is not checked
9859 *Julia Lawall <julia@diku.dk>*
9861 * Add --strict-warnings option to Configure script to include devteam
9862 warnings in other configurations.
9866 * Add support for --libdir option and LIBDIR variable in makefiles. This
9867 makes it possible to install openssl libraries in locations which
9868 have names other than "lib", for example "/usr/lib64" which some
9871 *Steve Henson, based on patch from Jeremy Utley*
9873 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9874 X690 8.9.12 and can produce some misleading textual output of OIDs.
9876 *Steve Henson, reported by Dan Kaminsky*
9878 * Delete MD2 from algorithm tables. This follows the recommendation in
9879 several standards that it is not used in new applications due to
9880 several cryptographic weaknesses. For binary compatibility reasons
9881 the MD2 API is still compiled in by default.
9885 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9890 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9891 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9894 *Guenter <lists@gknw.net>*
9896 * Fix the server certificate chain building code to use X509_verify_cert(),
9897 it used to have an ad-hoc builder which was unable to cope with anything
9898 other than a simple chain.
9900 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9902 * Don't check self signed certificate signatures in X509_verify_cert()
9903 by default (a flag can override this): it just wastes time without
9904 adding any security. As a useful side effect self signed root CAs
9905 with non-FIPS digests are now usable in FIPS mode.
9909 * In dtls1_process_out_of_seq_message() the check if the current message
9910 is already buffered was missing. For every new message was memory
9911 allocated, allowing an attacker to perform an denial of service attack
9912 with sending out of seq handshake messages until there is no memory
9913 left. Additionally every future message was buffered, even if the
9914 sequence number made no sense and would be part of another handshake.
9915 So only messages with sequence numbers less than 10 in advance will be
9916 buffered. ([CVE-2009-1378])
9918 *Robin Seggelmann, discovered by Daniel Mentz*
9920 * Records are buffered if they arrive with a future epoch to be
9921 processed after finishing the corresponding handshake. There is
9922 currently no limitation to this buffer allowing an attacker to perform
9923 a DOS attack with sending records with future epochs until there is no
9924 memory left. This patch adds the pqueue_size() function to determine
9925 the size of a buffer and limits the record buffer to 100 entries.
9928 *Robin Seggelmann, discovered by Daniel Mentz*
9930 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9931 parent structure is freed. ([CVE-2009-1379])
9935 * Handle non-blocking I/O properly in SSL_shutdown() call.
9937 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9939 * Add `2.5.4.*` OIDs
9941 *Ilya O. <vrghost@gmail.com>*
9943 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9945 * Disable renegotiation completely - this fixes a severe security
9946 problem ([CVE-2009-3555]) at the cost of breaking all
9947 renegotiation. Renegotiation can be re-enabled by setting
9948 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9949 run-time. This is really not recommended unless you know what
9954 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9956 * Don't set val to NULL when freeing up structures, it is freed up by
9957 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9958 zeroing past the valid field. ([CVE-2009-0789])
9960 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9962 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9963 checked correctly. This would allow some invalid signed attributes to
9964 appear to verify correctly. ([CVE-2009-0591])
9966 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9968 * Reject UniversalString and BMPString types with invalid lengths. This
9969 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9970 a legal length. ([CVE-2009-0590])
9974 * Set S/MIME signing as the default purpose rather than setting it
9975 unconditionally. This allows applications to override it at the store
9980 * Permit restricted recursion of ASN1 strings. This is needed in practice
9981 to handle some structures.
9985 * Improve efficiency of mem_gets: don't search whole buffer each time
9988 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9990 * New -hex option for openssl rand.
9994 * Print out UTF8String and NumericString when parsing ASN1.
9998 * Support NumericString type for name components.
10002 * Allow CC in the environment to override the automatically chosen
10003 compiler. Note that nothing is done to ensure flags work with the
10008 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10010 * Properly check EVP_VerifyFinal() and similar return values
10013 *Ben Laurie, Bodo Moeller, Google Security Team*
10015 * Enable TLS extensions by default.
10019 * Allow the CHIL engine to be loaded, whether the application is
10020 multithreaded or not. (This does not release the developer from the
10021 obligation to set up the dynamic locking callbacks.)
10023 *Sander Temme <sander@temme.net>*
10025 * Use correct exit code if there is an error in dgst command.
10027 *Steve Henson; problem pointed out by Roland Dirlewanger*
10029 * Tweak Configure so that you need to say "experimental-jpake" to enable
10030 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10034 * Add experimental JPAKE support, including demo authentication in
10035 s_client and s_server.
10039 * Set the comparison function in v3_addr_canonize().
10041 *Rob Austein <sra@hactrn.net>*
10043 * Add support for XMPP STARTTLS in s_client.
10045 *Philip Paeps <philip@freebsd.org>*
10047 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10048 to ensure that even with this option, only ciphersuites in the
10049 server's preference list will be accepted. (Note that the option
10050 applies only when resuming a session, so the earlier behavior was
10051 just about the algorithm choice for symmetric cryptography.)
10055 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10057 * Fix NULL pointer dereference if a DTLS server received
10058 ChangeCipherSpec as first record ([CVE-2009-1386]).
10062 * Fix a state transition in s3_srvr.c and d1_srvr.c
10063 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10065 *Nagendra Modadugu*
10067 * The fix in 0.9.8c that supposedly got rid of unsafe
10068 double-checked locking was incomplete for RSA blinding,
10069 addressing just one layer of what turns out to have been
10070 doubly unsafe triple-checked locking.
10072 So now fix this for real by retiring the MONT_HELPER macro
10073 in crypto/rsa/rsa_eay.c.
10075 *Bodo Moeller; problem pointed out by Marius Schilder*
10077 * Various precautionary measures:
10079 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10081 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10082 (NB: This would require knowledge of the secret session ticket key
10083 to exploit, in which case you'd be SOL either way.)
10085 - Change bn_nist.c so that it will properly handle input BIGNUMs
10086 outside the expected range.
10088 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10091 *Neel Mehta, Bodo Moeller*
10093 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10094 the load fails. Useful for distros.
10096 *Ben Laurie and the FreeBSD team*
10098 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10102 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10106 * Expand ENGINE to support engine supplied SSL client certificate functions.
10108 This work was sponsored by Logica.
10112 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10113 keystores. Support for SSL/TLS client authentication too.
10114 Not compiled unless enable-capieng specified to Configure.
10116 This work was sponsored by Logica.
10120 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10121 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10122 attribute creation routines such as certificate requests and PKCS#12
10127 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10129 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10130 handshake which could lead to a client crash as found using the
10131 Codenomicon TLS test suite ([CVE-2008-1672])
10133 *Steve Henson, Mark Cox*
10135 * Fix double free in TLS server name extensions which could lead to
10136 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10140 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10142 Clear the error queue to ensure that error entries left from
10143 older function calls do not interfere with the correct operation.
10145 *Lutz Jaenicke, Erik de Castro Lopo*
10147 * Remove root CA certificates of commercial CAs:
10149 The OpenSSL project does not recommend any specific CA and does not
10150 have any policy with respect to including or excluding any CA.
10151 Therefore, it does not make any sense to ship an arbitrary selection
10152 of root CA certificates with the OpenSSL software.
10156 * RSA OAEP patches to fix two separate invalid memory reads.
10157 The first one involves inputs when 'lzero' is greater than
10158 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10159 before the beginning of from). The second one involves inputs where
10160 the 'db' section contains nothing but zeroes (there is a one-byte
10161 invalid read after the end of 'db').
10163 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10165 * Partial backport from 0.9.9-dev:
10167 Introduce bn_mul_mont (dedicated Montgomery multiplication
10168 procedure) as a candidate for BIGNUM assembler implementation.
10169 While 0.9.9-dev uses assembler for various architectures, only
10170 x86_64 is available by default here in the 0.9.8 branch, and
10171 32-bit x86 is available through a compile-time setting.
10173 To try the 32-bit x86 assembler implementation, use Configure
10174 option "enable-montasm" (which exists only for this backport).
10176 As "enable-montasm" for 32-bit x86 disclaims code stability
10177 anyway, in this constellation we activate additional code
10178 backported from 0.9.9-dev for further performance improvements,
10179 namely BN_from_montgomery_word. (To enable this otherwise,
10180 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10182 *Andy Polyakov (backport partially by Bodo Moeller)*
10184 * Add TLS session ticket callback. This allows an application to set
10185 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10186 values. This is useful for key rollover for example where several key
10187 sets may exist with different names.
10191 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10192 This was broken until now in 0.9.8 releases, such that the only way
10193 a registered ENGINE could be used (assuming it initialises
10194 successfully on the host) was to explicitly set it as the default
10195 for the relevant algorithms. This is in contradiction with 0.9.7
10196 behaviour and the documentation. With this fix, when an ENGINE is
10197 registered into a given algorithm's table of implementations, the
10198 'uptodate' flag is reset so that auto-discovery will be used next
10199 time a new context for that algorithm attempts to select an
10202 *Ian Lister (tweaked by Geoff Thorpe)*
10204 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10205 implementation in the following ways:
10207 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10210 Lack of BER streaming support means one pass streaming processing is
10211 only supported if data is detached: setting the streaming flag is
10212 ignored for embedded content.
10214 CMS support is disabled by default and must be explicitly enabled
10215 with the enable-cms configuration option.
10219 * Update the GMP engine glue to do direct copies between BIGNUM and
10220 mpz_t when openssl and GMP use the same limb size. Otherwise the
10221 existing "conversion via a text string export" trick is still used.
10223 *Paul Sheer <paulsheer@gmail.com>*
10225 * Zlib compression BIO. This is a filter BIO which compressed and
10226 uncompresses any data passed through it.
10230 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10231 RFC3394 compatible AES key wrapping.
10235 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10236 sets string data without copying. X509_ALGOR_set0() and
10237 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10238 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10239 from an X509_ATTRIBUTE structure optionally checking it occurs only
10240 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10245 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10246 to get the expected BN_FLG_CONSTTIME behavior.
10248 *Bodo Moeller (Google)*
10252 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10253 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10254 - added some more tests to do_tests.pl
10255 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10256 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10257 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10258 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10259 - various changes to netware.pl to enable gcc-cross builds on Win32
10261 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10262 - various changes to fix missing prototype warnings
10263 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10264 - added AES, WHIRLPOOL and CPUID assembler code to build files
10265 - added missing AES assembler make rules to mk1mf.pl
10266 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10268 *Guenter Knauf <eflash@gmx.net>*
10270 * Implement certificate status request TLS extension defined in RFC3546.
10271 A client can set the appropriate parameters and receive the encoded
10272 OCSP response via a callback. A server can query the supplied parameters
10273 and set the encoded OCSP response in the callback. Add simplified examples
10274 to s_client and s_server.
10278 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10280 * Fix various bugs:
10281 + Binary incompatibility of ssl_ctx_st structure
10282 + DTLS interoperation with non-compliant servers
10283 + Don't call get_session_cb() without proposed session
10284 + Fix ia64 assembler code
10286 *Andy Polyakov, Steve Henson*
10288 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10290 * DTLS Handshake overhaul. There were longstanding issues with
10291 OpenSSL DTLS implementation, which were making it impossible for
10292 RFC 4347 compliant client to communicate with OpenSSL server.
10293 Unfortunately just fixing these incompatibilities would "cut off"
10294 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10295 server keeps tolerating non RFC compliant syntax. The opposite is
10296 not true, 0.9.8f client can not communicate with earlier server.
10297 This update even addresses CVE-2007-4995.
10301 * Changes to avoid need for function casts in OpenSSL: some compilers
10302 (gcc 4.2 and later) reject their use.
10303 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10306 * Add RFC4507 support to OpenSSL. This includes the corrections in
10307 RFC4507bis. The encrypted ticket format is an encrypted encoded
10308 SSL_SESSION structure, that way new session features are automatically
10311 If a client application caches session in an SSL_SESSION structure
10312 support is transparent because tickets are now stored in the encoded
10315 The SSL_CTX structure automatically generates keys for ticket
10316 protection in servers so again support should be possible
10317 with no application modification.
10319 If a client or server wishes to disable RFC4507 support then the option
10320 SSL_OP_NO_TICKET can be set.
10322 Add a TLS extension debugging callback to allow the contents of any client
10323 or server extensions to be examined.
10325 This work was sponsored by Google.
10329 * Add initial support for TLS extensions, specifically for the server_name
10330 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10331 have new members for a hostname. The SSL data structure has an
10332 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10333 stored in that context to allow for session resumption, even after the
10334 SSL has been switched to a new SSL_CTX in reaction to a client's
10335 server_name extension.
10337 New functions (subject to change):
10339 SSL_get_servername()
10340 SSL_get_servername_type()
10343 New CTRL codes and macros (subject to change):
10345 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10346 - SSL_CTX_set_tlsext_servername_callback()
10347 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10348 - SSL_CTX_set_tlsext_servername_arg()
10349 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10351 openssl s_client has a new '-servername ...' option.
10353 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10354 '-key2 ...', '-servername_fatal' (subject to change). This allows
10355 testing the HostName extension for a specific single hostname ('-cert'
10356 and '-key' remain fallbacks for handshakes without HostName
10357 negotiation). If the unrecognized_name alert has to be sent, this by
10358 default is a warning; it becomes fatal with the '-servername_fatal'
10361 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10363 * Add AES and SSE2 assembly language support to VC++ build.
10367 * Mitigate attack on final subtraction in Montgomery reduction.
10371 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10372 (which previously caused an internal error).
10376 * Squeeze another 10% out of IGE mode when in != out.
10380 * AES IGE mode speedup.
10382 *Dean Gaudet (Google)*
10384 * Add the Korean symmetric 128-bit cipher SEED (see
10385 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10386 add SEED ciphersuites from RFC 4162:
10388 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10389 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10390 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10391 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10393 To minimize changes between patchlevels in the OpenSSL 0.9.8
10394 series, SEED remains excluded from compilation unless OpenSSL
10395 is configured with 'enable-seed'.
10397 *KISA, Bodo Moeller*
10399 * Mitigate branch prediction attacks, which can be practical if a
10400 single processor is shared, allowing a spy process to extract
10401 information. For detailed background information, see
10402 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10403 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10404 and Necessary Software Countermeasures"). The core of the change
10405 are new versions BN_div_no_branch() and
10406 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10407 respectively, which are slower, but avoid the security-relevant
10408 conditional branches. These are automatically called by BN_div()
10409 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10410 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10411 remove a conditional branch.
10413 BN_FLG_CONSTTIME is the new name for the previous
10414 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10415 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10416 in the exponent causes BN_mod_exp_mont() to use the alternative
10417 implementation in BN_mod_exp_mont_consttime().) The old name
10418 remains as a deprecated alias.
10420 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10421 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10422 constant-time implementations for more than just exponentiation.
10423 Here too the old name is kept as a deprecated alias.
10425 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10426 the BN_BLINDING structure gets an independent copy of the
10427 modulus. This means that the previous `BIGNUM *m` argument to
10428 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10429 essentially becomes `const BIGNUM *m`, although we can't actually
10430 change this in the header file before 0.9.9. It allows
10431 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10432 enable BN_FLG_CONSTTIME.
10434 *Matthew D Wood (Intel Corp)*
10436 * In the SSL/TLS server implementation, be strict about session ID
10437 context matching (which matters if an application uses a single
10438 external cache for different purposes). Previously,
10439 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10440 set. This did ensure strict client verification, but meant that,
10441 with applications using a single external cache for quite
10442 different requirements, clients could circumvent ciphersuite
10443 restrictions for a given session ID context by starting a session
10444 in a different context.
10448 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10449 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10450 authentication-only ciphersuites.
10454 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10455 not complete and could lead to a possible single byte overflow
10456 ([CVE-2007-5135]) [Ben Laurie]
10458 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10460 * Since AES128 and AES256 (and similarly Camellia128 and
10461 Camellia256) share a single mask bit in the logic of
10462 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10463 kludge to work properly if AES128 is available and AES256 isn't
10464 (or if Camellia128 is available and Camellia256 isn't).
10468 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10469 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10470 When a point or a seed is encoded in a BIT STRING, we need to
10471 prevent the removal of trailing zero bits to get the proper DER
10472 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10473 of a NamedBitList, for which trailing 0 bits need to be removed.)
10477 * Have SSL/TLS server implementation tolerate "mismatched" record
10478 protocol version while receiving ClientHello even if the
10479 ClientHello is fragmented. (The server can't insist on the
10480 particular protocol version it has chosen before the ServerHello
10481 message has informed the client about his choice.)
10485 * Add RFC 3779 support.
10487 *Rob Austein for ARIN, Ben Laurie*
10489 * Load error codes if they are not already present instead of using a
10490 static variable. This allows them to be cleanly unloaded and reloaded.
10491 Improve header file function name parsing.
10495 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10496 or CAPABILITY handshake as required by RFCs.
10498 *Goetz Babin-Ebell*
10500 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10502 * Introduce limits to prevent malicious keys being able to
10503 cause a denial of service. ([CVE-2006-2940])
10505 *Steve Henson, Bodo Moeller*
10507 * Fix ASN.1 parsing of certain invalid structures that can result
10508 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10510 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10511 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10513 * Fix SSL client code which could crash if connecting to a
10514 malicious SSLv2 server. ([CVE-2006-4343])
10516 *Tavis Ormandy and Will Drewry, Google Security Team*
10518 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10519 match only those. Before that, "AES256-SHA" would be interpreted
10520 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10521 the same strength classification in 0.9.7h) as we currently only
10522 have a single AES bit in the ciphersuite description bitmap.
10523 That change, however, also applied to ciphersuite strings such as
10524 "RC4-MD5" that intentionally matched multiple ciphersuites --
10525 namely, SSL 2.0 ciphersuites in addition to the more common ones
10526 from SSL 3.0/TLS 1.0.
10528 So we change the selection algorithm again: Naming an explicit
10529 ciphersuite selects this one ciphersuite, and any other similar
10530 ciphersuite (same bitmap) from *other* protocol versions.
10531 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10532 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10534 Since SSL 2.0 does not have any ciphersuites for which the
10535 128/256 bit distinction would be relevant, this works for now.
10536 The proper fix will be to use different bits for AES128 and
10537 AES256, which would have avoided the problems from the beginning;
10538 however, bits are scarce, so we can only do this in a new release
10539 (not just a patchlevel) when we can change the SSL_CIPHER
10540 definition to split the single 'unsigned long mask' bitmap into
10541 multiple values to extend the available space.
10545 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10547 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10548 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10550 * Add AES IGE and biIGE modes.
10554 * Change the Unix randomness entropy gathering to use poll() when
10555 possible instead of select(), since the latter has some
10556 undesirable limitations.
10558 *Darryl Miles via Richard Levitte and Bodo Moeller*
10560 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10561 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10562 cannot be implicitly activated as part of, e.g., the "AES" alias.
10563 However, please upgrade to OpenSSL 0.9.9[-dev] for
10564 non-experimental use of the ECC ciphersuites to get TLS extension
10565 support, which is required for curve and point format negotiation
10566 to avoid potential handshake problems.
10570 * Disable rogue ciphersuites:
10572 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10573 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10574 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10576 The latter two were purportedly from
10577 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10580 Also deactivate the remaining ciphersuites from
10581 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10582 unofficial, and the ID has long expired.
10586 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10587 dual-core machines) and other potential thread-safety issues.
10591 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10592 versions), which is now available for royalty-free use
10593 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10594 Also, add Camellia TLS ciphersuites from RFC 4132.
10596 To minimize changes between patchlevels in the OpenSSL 0.9.8
10597 series, Camellia remains excluded from compilation unless OpenSSL
10598 is configured with 'enable-camellia'.
10602 * Disable the padding bug check when compression is in use. The padding
10603 bug check assumes the first packet is of even length, this is not
10604 necessarily true if compression is enabled and can result in false
10605 positives causing handshake failure. The actual bug test is ancient
10606 code so it is hoped that implementations will either have fixed it by
10607 now or any which still have the bug do not support compression.
10611 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10613 * When applying a cipher rule check to see if string match is an explicit
10614 cipher suite and only match that one cipher suite if it is.
10618 * Link in manifests for VC++ if needed.
10620 *Austin Ziegler <halostatue@gmail.com>*
10622 * Update support for ECC-based TLS ciphersuites according to
10623 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10624 TLS extensions, which are supported starting with the 0.9.9
10625 branch, not in the OpenSSL 0.9.8 branch).
10629 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10630 opaque EVP_CIPHER_CTX handling.
10634 * Fixes and enhancements to zlib compression code. We now only use
10635 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10636 to conform with the standards mentioned here:
10637 <http://www.zlib.net/DLL_FAQ.txt>
10638 Static zlib linking now works on Windows and the new --with-zlib-include
10639 --with-zlib-lib options to Configure can be used to supply the location
10640 of the headers and library. Gracefully handle case where zlib library
10645 * Several fixes and enhancements to the OID generation code. The old code
10646 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10647 handle numbers larger than ULONG_MAX, truncated printing and had a
10648 non standard OBJ_obj2txt() behaviour.
10652 * Add support for building of engines under engine/ as shared libraries
10653 under VC++ build system.
10657 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10658 Hopefully, we will not see any false combination of paths any more.
10662 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10664 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10665 (part of SSL_OP_ALL). This option used to disable the
10666 countermeasure against man-in-the-middle protocol-version
10667 rollback in the SSL 2.0 server implementation, which is a bad
10668 idea. ([CVE-2005-2969])
10670 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10671 for Information Security, National Institute of Advanced Industrial
10672 Science and Technology [AIST], Japan)*
10674 * Add two function to clear and return the verify parameter flags.
10678 * Keep cipherlists sorted in the source instead of sorting them at
10679 runtime, thus removing the need for a lock.
10683 * Avoid some small subgroup attacks in Diffie-Hellman.
10685 *Nick Mathewson and Ben Laurie*
10687 * Add functions for well-known primes.
10691 * Extended Windows CE support.
10693 *Satoshi Nakamura and Andy Polyakov*
10695 * Initialize SSL_METHOD structures at compile time instead of during
10696 runtime, thus removing the need for a lock.
10700 * Make PKCS7_decrypt() work even if no certificate is supplied by
10701 attempting to decrypt each encrypted key in turn. Add support to
10706 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10708 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10711 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10715 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10716 key into the same file any more.
10720 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10724 * Add -utf8 command line and config file option to 'ca'.
10726 *Stefan <stf@udoma.org*
10728 * Removed the macro des_crypt(), as it seems to conflict with some
10729 libraries. Use DES_crypt().
10733 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10734 involves renaming the source and generated shared-libs for
10735 both. The engines will accept the corrected or legacy ids
10736 ('ncipher' and '4758_cca' respectively) when binding. NB,
10737 this only applies when building 'shared'.
10739 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10741 * Add attribute functions to EVP_PKEY structure. Modify
10742 PKCS12_create() to recognize a CSP name attribute and
10743 use it. Make -CSP option work again in pkcs12 utility.
10747 * Add new functionality to the bn blinding code:
10748 - automatic re-creation of the BN_BLINDING parameters after
10749 a fixed number of uses (currently 32)
10750 - add new function for parameter creation
10751 - introduce flags to control the update behaviour of the
10752 BN_BLINDING parameters
10753 - hide BN_BLINDING structure
10754 Add a second BN_BLINDING slot to the RSA structure to improve
10755 performance when a single RSA object is shared among several
10760 * Add support for DTLS.
10762 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10764 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10765 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10769 * Remove buggy and incomplete DH cert support from
10770 ssl/ssl_rsa.c and ssl/s3_both.c
10774 * Use SHA-1 instead of MD5 as the default digest algorithm for
10775 the `apps/openssl` commands.
10779 * Compile clean with "-Wall -Wmissing-prototypes
10780 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10781 DEBUG_SAFESTACK must also be set.
10785 * Change ./Configure so that certain algorithms can be disabled by default.
10786 The new counterpiece to "no-xxx" is "enable-xxx".
10788 The patented RC5 and MDC2 algorithms will now be disabled unless
10789 "enable-rc5" and "enable-mdc2", respectively, are specified.
10791 (IDEA remains enabled despite being patented. This is because IDEA
10792 is frequently required for interoperability, and there is no license
10793 fee for non-commercial use. As before, "no-idea" can be used to
10794 avoid this algorithm.)
10798 * Add processing of proxy certificates (see RFC 3820). This work was
10799 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10800 EGEE (Enabling Grids for E-science in Europe).
10804 * RC4 performance overhaul on modern architectures/implementations, such
10805 as Intel P4, IA-64 and AMD64.
10809 * New utility extract-section.pl. This can be used specify an alternative
10810 section number in a pod file instead of having to treat each file as
10811 a separate case in Makefile. This can be done by adding two lines to the
10814 =for comment openssl_section:XXX
10816 The blank line is mandatory.
10820 * New arguments -certform, -keyform and -pass for s_client and s_server
10821 to allow alternative format key and certificate files and passphrase
10826 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10827 update associated structures and add various utility functions.
10829 Add new policy related verify parameters, include policy checking in
10830 standard verify code. Enhance 'smime' application with extra parameters
10831 to support policy checking and print out.
10835 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10836 Nehemiah processors. These extensions support AES encryption in hardware
10837 as well as RNG (though RNG support is currently disabled).
10839 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10841 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10845 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10847 *Andy Polyakov and a number of other people*
10849 * Improved PowerPC platform support. Most notably BIGNUM assembler
10850 implementation contributed by IBM.
10852 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10854 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10855 exponent rather than 'unsigned long'. There is a corresponding change to
10856 the new 'rsa_keygen' element of the RSA_METHOD structure.
10858 *Jelte Jansen, Geoff Thorpe*
10860 * Functionality for creating the initial serial number file is now
10861 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10863 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10864 number file to 1, which is bound to cause problems. To avoid
10865 the problems while respecting compatibility between different 0.9.7
10866 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10867 CA.pl for serial number initialization. With the new release 0.9.8,
10868 we can fix the problem directly in the 'ca' utility.)
10872 * Reduced header interdependencies by declaring more opaque objects in
10873 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10874 give fewer recursive includes, which could break lazy source code - so
10875 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10876 developers should define this symbol when building and using openssl to
10877 ensure they track the recommended behaviour, interfaces, [etc], but
10878 backwards-compatible behaviour prevails when this isn't defined.
10882 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10886 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10887 This will generate a random key of the appropriate length based on the
10888 cipher context. The EVP_CIPHER can provide its own random key generation
10889 routine to support keys of a specific form. This is used in the des and
10890 3des routines to generate a key of the correct parity. Update S/MIME
10891 code to use new functions and hence generate correct parity DES keys.
10892 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10893 valid (weak or incorrect parity).
10897 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10898 as looking them up. This is useful when the verified structure may contain
10899 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10900 present unless the new PKCS7_NO_CRL flag is asserted.
10904 * Extend ASN1 oid configuration module. It now additionally accepts the
10907 shortName = some long name, 1.2.3.4
10911 * Reimplemented the BN_CTX implementation. There is now no more static
10912 limitation on the number of variables it can handle nor the depth of the
10913 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10914 information can now expand as required, and rather than having a single
10915 static array of bignums, BN_CTX now uses a linked-list of such arrays
10916 allowing it to expand on demand whilst maintaining the usefulness of
10917 BN_CTX's "bundling".
10921 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10922 to allow all RSA operations to function using a single BN_CTX.
10926 * Preliminary support for certificate policy evaluation and checking. This
10927 is initially intended to pass the tests outlined in "Conformance Testing
10928 of Relying Party Client Certificate Path Processing Logic" v1.07.
10932 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10933 remained unused and not that useful. A variety of other little bignum
10934 tweaks and fixes have also been made continuing on from the audit (see
10939 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10940 associated ASN1, EVP and SSL functions and old ASN1 macros.
10944 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10945 and this should never fail. So the return value from the use of
10946 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10947 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10951 * BN_CTX_get() should return zero-valued bignums, providing the same
10952 initialised value as BN_new().
10954 *Geoff Thorpe, suggested by Ulf Möller*
10956 * Support for inhibitAnyPolicy certificate extension.
10960 * An audit of the BIGNUM code is underway, for which debugging code is
10961 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10962 is considered valid when processing BIGNUMs, and causes execution to
10963 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10964 further steps are taken to deliberately pollute unused data in BIGNUM
10965 structures to try and expose faulty code further on. For now, openssl will
10966 (in its default mode of operation) continue to tolerate the inconsistent
10967 forms that it has tolerated in the past, but authors and packagers should
10968 consider trying openssl and their own applications when compiled with
10969 these debugging symbols defined. It will help highlight potential bugs in
10970 their own code, and will improve the test coverage for OpenSSL itself. At
10971 some point, these tighter rules will become openssl's default to improve
10972 maintainability, though the assert()s and other overheads will remain only
10973 in debugging configurations. See bn.h for more details.
10975 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10977 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10978 that can only be obtained through BN_CTX_new() (which implicitly
10979 initialises it). The presence of this function only made it possible
10980 to overwrite an existing structure (and cause memory leaks).
10984 * Because of the callback-based approach for implementing LHASH as a
10985 template type, lh_insert() adds opaque objects to hash-tables and
10986 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10987 to clean up those corresponding objects before destroying the hash table
10988 (and losing the object pointers). So some over-zealous constifications in
10989 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10990 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10991 prototyped to have "const" restrictions on the object pointers they are
10992 given (and so aren't required to cast them away any more).
10996 * The tmdiff.h API was so ugly and minimal that our own timing utility
10997 (speed) prefers to use its own implementation. The two implementations
10998 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10999 its object type properly exposed (MS_TM) instead of casting to/from
11000 `char *`. This may still change yet if someone realises MS_TM and
11002 aren't necessarily the greatest nomenclatures - but this is what was used
11003 internally to the implementation so I've used that for now.
11007 * Ensure that deprecated functions do not get compiled when
11008 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11009 the self-tests were still using deprecated key-generation functions so
11010 these have been updated also.
11014 * Reorganise PKCS#7 code to separate the digest location functionality
11015 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11016 New function PKCS7_set_digest() to set the digest type for PKCS#7
11017 digestedData type. Add additional code to correctly generate the
11018 digestedData type and add support for this type in PKCS7 initialization
11023 * New function PKCS7_set0_type_other() this initializes a PKCS7
11024 structure of type "other".
11028 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11029 sure the loop does correctly stop and breaking ("division by zero")
11030 modulus operations are not performed. The (pre-generated) prime
11031 table crypto/bn/bn_prime.h was already correct, but it could not be
11032 re-generated on some platforms because of the "division by zero"
11033 situation in the script.
11035 *Ralf S. Engelschall*
11037 * Update support for ECC-based TLS ciphersuites according to
11038 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11039 SHA-1 now is only used for "small" curves (where the
11040 representation of a field element takes up to 24 bytes); for
11041 larger curves, the field element resulting from ECDH is directly
11042 used as premaster secret.
11044 *Douglas Stebila (Sun Microsystems Laboratories)*
11046 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11047 curve secp160r1 to the tests.
11049 *Douglas Stebila (Sun Microsystems Laboratories)*
11051 * Add the possibility to load symbols globally with DSO.
11053 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11055 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11056 control of the error stack.
11060 * Add support for STORE in ENGINE.
11064 * Add the STORE type. The intention is to provide a common interface
11065 to certificate and key stores, be they simple file-based stores, or
11066 HSM-type store, or LDAP stores, or...
11067 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11071 * Add a generic structure called OPENSSL_ITEM. This can be used to
11072 pass a list of arguments to any function as well as provide a way
11073 for a function to pass data back to the caller.
11077 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11078 works like BUF_strdup() but can be used to duplicate a portion of
11079 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11084 * Add the function sk_find_ex() which works like sk_find(), but will
11085 return an index to an element even if an exact match couldn't be
11086 found. The index is guaranteed to point at the element where the
11087 searched-for key would be inserted to preserve sorting order.
11091 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11092 takes an extra flags argument for optional functionality. Currently,
11093 the following flags are defined:
11095 OBJ_BSEARCH_VALUE_ON_NOMATCH
11096 This one gets OBJ_bsearch_ex() to return a pointer to the first
11097 element where the comparing function returns a negative or zero
11100 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11101 This one gets OBJ_bsearch_ex() to return a pointer to the first
11102 element where the comparing function returns zero. This is useful
11103 if there are more than one element where the comparing function
11108 * Make it possible to create self-signed certificates with 'openssl ca'
11109 in such a way that the self-signed certificate becomes part of the
11110 CA database and uses the same mechanisms for serial number generation
11111 as all other certificate signing. The new flag '-selfsign' enables
11112 this functionality. Adapt CA.sh and CA.pl.in.
11116 * Add functionality to check the public key of a certificate request
11117 against a given private. This is useful to check that a certificate
11118 request can be signed by that key (self-signing).
11122 * Make it possible to have multiple active certificates with the same
11123 subject in the CA index file. This is done only if the keyword
11124 'unique_subject' is set to 'no' in the main CA section (default
11125 if 'CA_default') of the configuration file. The value is saved
11126 with the database itself in a separate index attribute file,
11127 named like the index file with '.attr' appended to the name.
11131 * Generate multi-valued AVAs using '+' notation in config files for
11136 * Support for nameConstraints certificate extension.
11140 * Support for policyConstraints certificate extension.
11144 * Support for policyMappings certificate extension.
11148 * Make sure the default DSA_METHOD implementation only uses its
11149 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11150 and change its own handlers to be NULL so as to remove unnecessary
11151 indirection. This lets alternative implementations fallback to the
11152 default implementation more easily.
11156 * Support for directoryName in GeneralName related extensions
11161 * Make it possible to link applications using Makefile.shared.
11162 Make that possible even when linking against static libraries!
11166 * Support for single pass processing for S/MIME signing. This now
11167 means that S/MIME signing can be done from a pipe, in addition
11168 cleartext signing (multipart/signed type) is effectively streaming
11169 and the signed data does not need to be all held in memory.
11171 This is done with a new flag PKCS7_STREAM. When this flag is set
11172 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11173 is done after the data is output (and digests calculated) in
11174 SMIME_write_PKCS7().
11178 * Add full support for -rpath/-R, both in shared libraries and
11179 applications, at least on the platforms where it's known how
11184 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11185 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11186 will now compute a table of multiples of the generator that
11187 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11188 faster (notably in the case of a single point multiplication,
11189 scalar * generator).
11191 *Nils Larsch, Bodo Moeller*
11193 * IPv6 support for certificate extensions. The various extensions
11194 which use the IP:a.b.c.d can now take IPv6 addresses using the
11195 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11200 * Added an ENGINE that implements RSA by performing private key
11201 exponentiations with the GMP library. The conversions to and from
11202 GMP's mpz_t format aren't optimised nor are any montgomery forms
11203 cached, and on x86 it appears OpenSSL's own performance has caught up.
11204 However there are likely to be other architectures where GMP could
11205 provide a boost. This ENGINE is not built in by default, but it can be
11206 specified at Configure time and should be accompanied by the necessary
11207 linker additions, eg;
11208 ./config -DOPENSSL_USE_GMP -lgmp
11212 * "openssl engine" will not display ENGINE/DSO load failure errors when
11213 testing availability of engines with "-t" - the old behaviour is
11214 produced by increasing the feature's verbosity with "-tt".
11218 * ECDSA routines: under certain error conditions uninitialized BN objects
11219 could be freed. Solution: make sure initialization is performed early
11220 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11225 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11226 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11227 software implementations. For DSA and DH, parameter generation can
11228 also be overridden by providing the appropriate method callbacks.
11232 * Change the "progress" mechanism used in key-generation and
11233 primality testing to functions that take a new BN_GENCB pointer in
11234 place of callback/argument pairs. The new API functions have `_ex`
11235 postfixes and the older functions are reimplemented as wrappers for
11236 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11237 declarations of the old functions to help (graceful) attempts to
11238 migrate to the new functions. Also, the new key-generation API
11239 functions operate on a caller-supplied key-structure and return
11240 success/failure rather than returning a key or NULL - this is to
11241 help make "keygen" another member function of RSA_METHOD etc.
11243 Example for using the new callback interface:
11245 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11246 void *my_arg = ...;
11249 BN_GENCB_set(&my_cb, my_callback, my_arg);
11251 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11252 /* For the meaning of a, b in calls to my_callback(), see the
11253 * documentation of the function that calls the callback.
11254 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11255 * my_callback should return 1 if it wants BN_is_prime_ex()
11256 * to continue, or 0 to stop.
11261 * Change the ZLIB compression method to be stateful, and make it
11262 available to TLS with the number defined in
11263 draft-ietf-tls-compression-04.txt.
11267 * Add the ASN.1 structures and functions for CertificatePair, which
11268 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11270 CertificatePair ::= SEQUENCE {
11271 forward [0] Certificate OPTIONAL,
11272 reverse [1] Certificate OPTIONAL,
11273 -- at least one of the pair shall be present -- }
11275 Also implement the PEM functions to read and write certificate
11276 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11278 This needed to be defined, mostly for the sake of the LDAP
11279 attribute crossCertificatePair, but may prove useful elsewhere as
11284 * Make it possible to inhibit symlinking of shared libraries in
11285 Makefile.shared, for Cygwin's sake.
11289 * Extend the BIGNUM API by creating a function
11290 void BN_set_negative(BIGNUM *a, int neg);
11291 and a macro that behave like
11292 int BN_is_negative(const BIGNUM *a);
11294 to avoid the need to access 'a->neg' directly in applications.
11298 * Implement fast modular reduction for pseudo-Mersenne primes
11299 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11300 EC_GROUP_new_curve_GFp() will now automatically use this
11303 *Nils Larsch <nla@trustcenter.de>*
11305 * Add new lock type (CRYPTO_LOCK_BN).
11309 * Change the ENGINE framework to automatically load engines
11310 dynamically from specific directories unless they could be
11311 found to already be built in or loaded. Move all the
11312 current engines except for the cryptodev one to a new
11313 directory engines/.
11314 The engines in engines/ are built as shared libraries if
11315 the "shared" options was given to ./Configure or ./config.
11316 Otherwise, they are inserted in libcrypto.a.
11317 /usr/local/ssl/engines is the default directory for dynamic
11318 engines, but that can be overridden at configure time through
11319 the usual use of --prefix and/or --openssldir, and at run
11320 time with the environment variable OPENSSL_ENGINES.
11322 *Geoff Thorpe and Richard Levitte*
11324 * Add Makefile.shared, a helper makefile to build shared
11325 libraries. Adapt Makefile.org.
11329 * Add version info to Win32 DLLs.
11331 *Peter 'Luna' Runestig" <peter@runestig.com>*
11333 * Add new 'medium level' PKCS#12 API. Certificates and keys
11334 can be added using this API to created arbitrary PKCS#12
11335 files while avoiding the low-level API.
11337 New options to PKCS12_create(), key or cert can be NULL and
11338 will then be omitted from the output file. The encryption
11339 algorithm NIDs can be set to -1 for no encryption, the mac
11340 iteration count can be set to 0 to omit the mac.
11342 Enhance pkcs12 utility by making the -nokeys and -nocerts
11343 options work when creating a PKCS#12 file. New option -nomac
11344 to omit the mac, NONE can be set for an encryption algorithm.
11345 New code is modified to use the enhanced PKCS12_create()
11346 instead of the low-level API.
11350 * Extend ASN1 encoder to support indefinite length constructed
11351 encoding. This can output sequences tags and octet strings in
11352 this form. Modify pk7_asn1.c to support indefinite length
11353 encoding. This is experimental and needs additional code to
11354 be useful, such as an ASN1 bio and some enhanced streaming
11357 Extend template encode functionality so that tagging is passed
11358 down to the template encoder.
11362 * Let 'openssl req' fail if an argument to '-newkey' is not
11363 recognized instead of using RSA as a default.
11367 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11368 As these are not official, they are not included in "ALL";
11369 the "ECCdraft" ciphersuite group alias can be used to select them.
11371 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11373 * Add ECDH engine support.
11375 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11377 * Add ECDH in new directory crypto/ecdh/.
11379 *Douglas Stebila (Sun Microsystems Laboratories)*
11381 * Let BN_rand_range() abort with an error after 100 iterations
11382 without success (which indicates a broken PRNG).
11386 * Change BN_mod_sqrt() so that it verifies that the input value
11387 is really the square of the return value. (Previously,
11388 BN_mod_sqrt would show GIGO behaviour.)
11392 * Add named elliptic curves over binary fields from X9.62, SECG,
11393 and WAP/WTLS; add OIDs that were still missing.
11395 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11397 * Extend the EC library for elliptic curves over binary fields
11398 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11401 EC_GF2m_simple_method
11405 EC_GROUP_new_curve_GF2m
11406 EC_GROUP_set_curve_GF2m
11407 EC_GROUP_get_curve_GF2m
11408 EC_POINT_set_affine_coordinates_GF2m
11409 EC_POINT_get_affine_coordinates_GF2m
11410 EC_POINT_set_compressed_coordinates_GF2m
11412 Point compression for binary fields is disabled by default for
11413 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11416 As binary polynomials are represented as BIGNUMs, various members
11417 of the EC_GROUP and EC_POINT data structures can be shared
11418 between the implementations for prime fields and binary fields;
11419 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11420 are essentially identical to their `..._GFp` counterparts.
11421 (For simplicity, the `..._GFp` prefix has been dropped from
11422 various internal method names.)
11424 An internal 'field_div' method (similar to 'field_mul' and
11425 'field_sqr') has been added; this is used only for binary fields.
11427 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11429 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11430 through methods ('mul', 'precompute_mult').
11432 The generic implementations (now internally called 'ec_wNAF_mul'
11433 and 'ec_wNAF_precomputed_mult') remain the default if these
11434 methods are undefined.
11436 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11438 * New function EC_GROUP_get_degree, which is defined through
11439 EC_METHOD. For curves over prime fields, this returns the bit
11440 length of the modulus.
11442 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11444 * New functions EC_GROUP_dup, EC_POINT_dup.
11445 (These simply call ..._new and ..._copy).
11447 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11449 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11450 Polynomials are represented as BIGNUMs (where the sign bit is not
11451 used) in the following functions [macros]:
11454 BN_GF2m_sub [= BN_GF2m_add]
11455 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11456 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11457 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11459 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11460 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11461 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11462 BN_GF2m_cmp [= BN_ucmp]
11464 (Note that only the 'mod' functions are actually for fields GF(2^m).
11465 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11467 For some functions, an the irreducible polynomial defining a
11468 field can be given as an 'unsigned int[]' with strictly
11469 decreasing elements giving the indices of those bits that are set;
11470 i.e., p[] represents the polynomial
11471 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11473 p[0] > p[1] > ... > p[k] = 0.
11474 This applies to the following functions:
11477 BN_GF2m_mod_mul_arr
11478 BN_GF2m_mod_sqr_arr
11479 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11480 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11481 BN_GF2m_mod_exp_arr
11482 BN_GF2m_mod_sqrt_arr
11483 BN_GF2m_mod_solve_quad_arr
11487 Conversion can be performed by the following functions:
11492 bntest.c has additional tests for binary polynomial arithmetic.
11494 Two implementations for BN_GF2m_mod_div() are available.
11495 The default algorithm simply uses BN_GF2m_mod_inv() and
11496 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11497 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11498 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11500 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11502 * Add new error code 'ERR_R_DISABLED' that can be used when some
11503 functionality is disabled at compile-time.
11505 *Douglas Stebila <douglas.stebila@sun.com>*
11507 * Change default behaviour of 'openssl asn1parse' so that more
11508 information is visible when viewing, e.g., a certificate:
11510 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11511 mode the content of non-printable OCTET STRINGs is output in a
11512 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11513 avoid the appearance of a printable string.
11515 *Nils Larsch <nla@trustcenter.de>*
11517 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11519 EC_GROUP_set_asn1_flag()
11520 EC_GROUP_get_asn1_flag()
11521 EC_GROUP_set_point_conversion_form()
11522 EC_GROUP_get_point_conversion_form()
11523 These control ASN1 encoding details:
11524 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11525 has been set to OPENSSL_EC_NAMED_CURVE.
11526 - Points are encoded in uncompressed form by default; options for
11527 asn1_for are as for point2oct, namely
11528 POINT_CONVERSION_COMPRESSED
11529 POINT_CONVERSION_UNCOMPRESSED
11530 POINT_CONVERSION_HYBRID
11532 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11534 EC_GROUP_set_seed()
11535 EC_GROUP_get0_seed()
11536 EC_GROUP_get_seed_len()
11537 This is used only for ASN1 purposes (so far).
11539 *Nils Larsch <nla@trustcenter.de>*
11541 * Add 'field_type' member to EC_METHOD, which holds the NID
11542 of the appropriate field type OID. The new function
11543 EC_METHOD_get_field_type() returns this value.
11545 *Nils Larsch <nla@trustcenter.de>*
11548 EC_POINT_point2bn()
11549 EC_POINT_bn2point()
11550 EC_POINT_point2hex()
11551 EC_POINT_hex2point()
11552 providing useful interfaces to EC_POINT_point2oct() and
11553 EC_POINT_oct2point().
11555 *Nils Larsch <nla@trustcenter.de>*
11557 * Change internals of the EC library so that the functions
11558 EC_GROUP_set_generator()
11559 EC_GROUP_get_generator()
11560 EC_GROUP_get_order()
11561 EC_GROUP_get_cofactor()
11562 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11563 to methods, which would lead to unnecessary code duplication when
11564 adding different types of curves.
11566 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11568 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11569 arithmetic, and such that modified wNAFs are generated
11570 (which avoid length expansion in many cases).
11574 * Add a function EC_GROUP_check_discriminant() (defined via
11575 EC_METHOD) that verifies that the curve discriminant is non-zero.
11577 Add a function EC_GROUP_check() that makes some sanity tests
11578 on a EC_GROUP, its generator and order. This includes
11579 EC_GROUP_check_discriminant().
11581 *Nils Larsch <nla@trustcenter.de>*
11583 * Add ECDSA in new directory crypto/ecdsa/.
11585 Add applications 'openssl ecparam' and 'openssl ecdsa'
11586 (these are based on 'openssl dsaparam' and 'openssl dsa').
11588 ECDSA support is also included in various other files across the
11589 library. Most notably,
11590 - 'openssl req' now has a '-newkey ecdsa:file' option;
11591 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11592 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11593 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11594 them suitable for ECDSA where domain parameters must be
11595 extracted before the specific public key;
11596 - ECDSA engine support has been added.
11598 *Nils Larsch <nla@trustcenter.de>*
11600 * Include some named elliptic curves, and add OIDs from X9.62,
11601 SECG, and WAP/WTLS. Each curve can be obtained from the new
11603 EC_GROUP_new_by_curve_name(),
11604 and the list of available named curves can be obtained with
11605 EC_get_builtin_curves().
11606 Also add a 'curve_name' member to EC_GROUP objects, which can be
11608 EC_GROUP_set_curve_name()
11609 EC_GROUP_get_curve_name()
11611 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11613 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11614 was actually never needed) and in BN_mul(). The removal in BN_mul()
11615 required a small change in bn_mul_part_recursive() and the addition
11616 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11617 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11618 bn_sub_words() and bn_add_words() except they take arrays with
11623 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11625 * Cleanse PEM buffers before freeing them since they may contain
11628 *Benjamin Bennett <ben@psc.edu>*
11630 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11631 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11632 authentication-only ciphersuites.
11636 * Since AES128 and AES256 share a single mask bit in the logic of
11637 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11638 kludge to work properly if AES128 is available and AES256 isn't.
11642 * Expand security boundary to match 1.1.1 module.
11646 * Remove redundant features: hash file source, editing of test vectors
11647 modify fipsld to use external fips_premain.c signature.
11651 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11652 run algorithm test programs.
11656 * Make algorithm test programs more tolerant of whitespace.
11660 * Have SSL/TLS server implementation tolerate "mismatched" record
11661 protocol version while receiving ClientHello even if the
11662 ClientHello is fragmented. (The server can't insist on the
11663 particular protocol version it has chosen before the ServerHello
11664 message has informed the client about his choice.)
11668 * Load error codes if they are not already present instead of using a
11669 static variable. This allows them to be cleanly unloaded and reloaded.
11673 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11675 * Introduce limits to prevent malicious keys being able to
11676 cause a denial of service. ([CVE-2006-2940])
11678 *Steve Henson, Bodo Moeller*
11680 * Fix ASN.1 parsing of certain invalid structures that can result
11681 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11683 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11684 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11686 * Fix SSL client code which could crash if connecting to a
11687 malicious SSLv2 server. ([CVE-2006-4343])
11689 *Tavis Ormandy and Will Drewry, Google Security Team*
11691 * Change ciphersuite string processing so that an explicit
11692 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11693 will no longer include "AES128-SHA"), and any other similar
11694 ciphersuite (same bitmap) from *other* protocol versions (so that
11695 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11696 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11697 changes from 0.9.8b and 0.9.8d.
11701 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11703 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11704 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11706 * Change the Unix randomness entropy gathering to use poll() when
11707 possible instead of select(), since the latter has some
11708 undesirable limitations.
11710 *Darryl Miles via Richard Levitte and Bodo Moeller*
11712 * Disable rogue ciphersuites:
11714 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11715 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11716 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11718 The latter two were purportedly from
11719 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11722 Also deactivate the remaining ciphersuites from
11723 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11724 unofficial, and the ID has long expired.
11728 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11729 dual-core machines) and other potential thread-safety issues.
11733 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11735 * Adapt fipsld and the build system to link against the validated FIPS
11736 module in FIPS mode.
11740 * Fixes for VC++ 2005 build under Windows.
11744 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11745 from a Windows bash shell such as MSYS. It is autodetected from the
11746 "config" script when run from a VC++ environment. Modify standard VC++
11747 build to use fipscanister.o from the GNU make build.
11751 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11753 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11754 The value now differs depending on if you build for FIPS or not.
11755 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11756 safely run with a non-FIPSed libcrypto, as it may crash because of
11757 the difference induced by this change.
11761 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11763 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11764 (part of SSL_OP_ALL). This option used to disable the
11765 countermeasure against man-in-the-middle protocol-version
11766 rollback in the SSL 2.0 server implementation, which is a bad
11767 idea. ([CVE-2005-2969])
11769 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11770 for Information Security, National Institute of Advanced Industrial
11771 Science and Technology [AIST, Japan)]*
11773 * Minimal support for X9.31 signatures and PSS padding modes. This is
11774 mainly for FIPS compliance and not fully integrated at this stage.
11778 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11779 the exponentiation using a fixed-length exponent. (Otherwise,
11780 the information leaked through timing could expose the secret key
11781 after many signatures; cf. Bleichenbacher's attack on DSA with
11786 * Make a new fixed-window mod_exp implementation the default for
11787 RSA, DSA, and DH private-key operations so that the sequence of
11788 squares and multiplies and the memory access pattern are
11789 independent of the particular secret key. This will mitigate
11790 cache-timing and potential related attacks.
11792 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11793 and this is automatically used by BN_mod_exp_mont() if the new flag
11794 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11795 will use this BN flag for private exponents unless the flag
11796 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11797 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11799 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11801 * Change the client implementation for SSLv23_method() and
11802 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11803 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11804 (Previously, the SSL 2.0 backwards compatible Client Hello
11805 message format would be used even with SSL_OP_NO_SSLv2.)
11809 * Add support for smime-type MIME parameter in S/MIME messages which some
11814 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11815 a threadsafe manner. Modify rsa code to use new function and add calls
11816 to dsa and dh code (which had race conditions before).
11820 * Include the fixed error library code in the C error file definitions
11821 instead of fixing them up at runtime. This keeps the error code
11822 structures constant.
11826 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11828 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11831 * Fixes for newer kerberos headers. NB: the casts are needed because
11832 the 'length' field is signed on one version and unsigned on another
11833 with no (?) obvious way to tell the difference, without these VC++
11834 complains. Also the "definition" of FAR (blank) is no longer included
11835 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11836 some needed definitions.
11840 * Undo Cygwin change.
11844 * Added support for proxy certificates according to RFC 3820.
11845 Because they may be a security thread to unaware applications,
11846 they must be explicitly allowed in run-time. See
11847 docs/HOWTO/proxy_certificates.txt for further information.
11851 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11853 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11854 server and client random values. Previously
11855 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11856 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11858 This change has negligible security impact because:
11860 1. Server and client random values still have 24 bytes of pseudo random
11863 2. Server and client random values are sent in the clear in the initial
11866 3. The master secret is derived using the premaster secret (48 bytes in
11867 size for static RSA ciphersuites) as well as client server and random
11870 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11873 *Stephen Henson, reported by UK NISCC*
11875 * Use Windows randomness collection on Cygwin.
11879 * Fix hang in EGD/PRNGD query when communication socket is closed
11880 prematurely by EGD/PRNGD.
11882 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11884 * Prompt for pass phrases when appropriate for PKCS12 input format.
11888 * Back-port of selected performance improvements from development
11889 branch, as well as improved support for PowerPC platforms.
11893 * Add lots of checks for memory allocation failure, error codes to indicate
11894 failure and freeing up memory if a failure occurs.
11896 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11898 * Add new -passin argument to dgst.
11902 * Perform some character comparisons of different types in X509_NAME_cmp:
11903 this is needed for some certificates that re-encode DNs into UTF8Strings
11904 (in violation of RFC3280) and can't or won't issue name rollover
11909 * Make an explicit check during certificate validation to see that
11910 the CA setting in each certificate on the chain is correct. As a
11911 side effect always do the following basic checks on extensions,
11912 not just when there's an associated purpose to the check:
11914 - if there is an unhandled critical extension (unless the user
11915 has chosen to ignore this fault)
11916 - if the path length has been exceeded (if one is set at all)
11917 - that certain extensions fit the associated purpose (if one has
11922 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11924 * Avoid a race condition when CRLs are checked in a multi threaded
11925 environment. This would happen due to the reordering of the revoked
11926 entries during signature checking and serial number lookup. Now the
11927 encoding is cached and the serial number sort performed under a lock.
11928 Add new STACK function sk_is_sorted().
11932 * Add Delta CRL to the extension code.
11936 * Various fixes to s3_pkt.c so alerts are sent properly.
11938 *David Holmes <d.holmes@f5.com>*
11940 * Reduce the chances of duplicate issuer name and serial numbers (in
11941 violation of RFC3280) using the OpenSSL certificate creation utilities.
11942 This is done by creating a random 64 bit value for the initial serial
11943 number when a serial number file is created or when a self signed
11944 certificate is created using 'openssl req -x509'. The initial serial
11945 number file is created using 'openssl x509 -next_serial' in CA.pl
11946 rather than being initialized to 1.
11950 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11952 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11953 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11955 *Joe Orton, Steve Henson*
11957 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11960 *Joe Orton, Steve Henson*
11962 * Make it possible to have multiple active certificates with the same
11963 subject in the CA index file. This is done only if the keyword
11964 'unique_subject' is set to 'no' in the main CA section (default
11965 if 'CA_default') of the configuration file. The value is saved
11966 with the database itself in a separate index attribute file,
11967 named like the index file with '.attr' appended to the name.
11971 * X509 verify fixes. Disable broken certificate workarounds when
11972 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11973 keyUsage extension present. Don't accept CRLs with unhandled critical
11974 extensions: since verify currently doesn't process CRL extensions this
11975 rejects a CRL with *any* critical extensions. Add new verify error codes
11980 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11981 A clarification of RFC2560 will require the use of OCTET STRINGs and
11982 some implementations cannot handle the current raw format. Since OpenSSL
11983 copies and compares OCSP nonces as opaque blobs without any attempt at
11984 parsing them this should not create any compatibility issues.
11988 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11989 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11990 this HMAC (and other) operations are several times slower than OpenSSL
11995 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11997 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11999 * Use the correct content when signing type "other".
12003 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12005 * Fix various bugs revealed by running the NISCC test suite:
12007 Stop out of bounds reads in the ASN1 code when presented with
12008 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12010 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12012 If verify callback ignores invalid public key errors don't try to check
12013 certificate signature with the NULL public key.
12017 * New -ignore_err option in ocsp application to stop the server
12018 exiting on the first error in a request.
12022 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12023 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12028 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12029 extra data after the compression methods not only for TLS 1.0
12030 but also for SSL 3.0 (as required by the specification).
12032 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12034 * Change X509_certificate_type() to mark the key as exported/exportable
12035 when it's 512 *bits* long, not 512 bytes.
12039 * Change AES_cbc_encrypt() so it outputs exact multiple of
12040 blocks during encryption.
12044 * Various fixes to base64 BIO and non blocking I/O. On write
12045 flushes were not handled properly if the BIO retried. On read
12046 data was not being buffered properly and had various logic bugs.
12047 This also affects blocking I/O when the data being decoded is a
12052 * Various S/MIME bugfixes and compatibility changes:
12053 output correct application/pkcs7 MIME type if
12054 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12055 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12056 of files as .eml work). Correctly handle very long lines in MIME
12061 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12063 * Countermeasure against the Klima-Pokorny-Rosa extension of
12064 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12065 a protocol version number mismatch like a decryption error
12066 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12070 * Turn on RSA blinding by default in the default implementation
12071 to avoid a timing attack. Applications that don't want it can call
12072 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12073 They would be ill-advised to do so in most cases.
12075 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12077 * Change RSA blinding code so that it works when the PRNG is not
12078 seeded (in this case, the secret RSA exponent is abused as
12079 an unpredictable seed -- if it is not unpredictable, there
12080 is no point in blinding anyway). Make RSA blinding thread-safe
12081 by remembering the creator's thread ID in rsa->blinding and
12082 having all other threads use local one-time blinding factors
12083 (this requires more computation than sharing rsa->blinding, but
12084 avoids excessive locking; and if an RSA object is not shared
12085 between threads, blinding will still be very fast).
12089 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12090 ENGINE as defaults for all supported algorithms irrespective of
12091 the 'flags' parameter. 'flags' is now honoured, so applications
12092 should make sure they are passing it correctly.
12096 * Target "mingw" now allows native Windows code to be generated in
12097 the Cygwin environment as well as with the MinGW compiler.
12101 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12103 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12104 via timing by performing a MAC computation even if incorrect
12105 block cipher padding has been found. This is a countermeasure
12106 against active attacks where the attacker has to distinguish
12107 between bad padding and a MAC verification error. ([CVE-2003-0078])
12109 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12110 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12111 Martin Vuagnoux (EPFL, Ilion)*
12113 * Make the no-err option work as intended. The intention with no-err
12114 is not to have the whole error stack handling routines removed from
12115 libcrypto, it's only intended to remove all the function name and
12116 reason texts, thereby removing some of the footprint that may not
12117 be interesting if those errors aren't displayed anyway.
12119 NOTE: it's still possible for any application or module to have its
12120 own set of error texts inserted. The routines are there, just not
12121 used by default when no-err is given.
12125 * Add support for FreeBSD on IA64.
12127 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12129 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12130 Kerberos function mit_des_cbc_cksum(). Before this change,
12131 the value returned by DES_cbc_cksum() was like the one from
12132 mit_des_cbc_cksum(), except the bytes were swapped.
12134 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12136 * Allow an application to disable the automatic SSL chain building.
12137 Before this a rather primitive chain build was always performed in
12138 ssl3_output_cert_chain(): an application had no way to send the
12139 correct chain if the automatic operation produced an incorrect result.
12141 Now the chain builder is disabled if either:
12143 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12145 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12147 The reasoning behind this is that an application would not want the
12148 auto chain building to take place if extra chain certificates are
12149 present and it might also want a means of sending no additional
12150 certificates (for example the chain has two certificates and the
12155 * Add the possibility to build without the ENGINE framework.
12157 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12159 * Under Win32 gmtime() can return NULL: check return value in
12160 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12164 * DSA routines: under certain error conditions uninitialized BN objects
12165 could be freed. Solution: make sure initialization is performed early
12166 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12167 Nils Larsch <nla@trustcenter.de> via PR#459)
12171 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12172 checked on reconnect on the client side, therefore session resumption
12173 could still fail with a "ssl session id is different" error. This
12174 behaviour is masked when SSL_OP_ALL is used due to
12175 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12176 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12177 followup to PR #377.
12181 * IA-32 assembler support enhancements: unified ELF targets, support
12182 for SCO/Caldera platforms, fix for Cygwin shared build.
12186 * Add support for FreeBSD on sparc64. As a consequence, support for
12187 FreeBSD on non-x86 processors is separate from x86 processors on
12188 the config script, much like the NetBSD support.
12190 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12192 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12194 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12197 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12198 code (06) was taken as the first octet of the session ID and the last
12199 octet was ignored consequently. As a result SSLv2 client side session
12200 caching could not have worked due to the session ID mismatch between
12202 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12207 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12208 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12213 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12214 seems that in spite of existing for more than a year, many application
12215 author have done nothing to provide the necessary callbacks, which
12216 means that this particular engine will not work properly anywhere.
12217 This is a very unfortunate situation which forces us, in the name
12218 of usability, to give the hw_ncipher.c a static lock, which is part
12220 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12221 appear in 0.9.8 or later. We EXPECT application authors to have
12222 dealt properly with this when 0.9.8 is released (unless we actually
12223 make such changes in the libcrypto locking code that changes will
12224 have to be made anyway).
12228 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12229 octets have been read, EOF or an error occurs. Without this change
12230 some truncated ASN1 structures will not produce an error.
12234 * Disable Heimdal support, since it hasn't been fully implemented.
12235 Still give the possibility to force the use of Heimdal, but with
12236 warnings and a request that patches get sent to openssl-dev.
12240 * Add the VC-CE target, introduce the WINCE sysname, and add
12241 INSTALL.WCE and appropriate conditionals to make it build.
12243 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12245 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12246 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12247 edit numbers of the version.
12249 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12251 * Introduce safe string copy and catenation functions
12252 (BUF_strlcpy() and BUF_strlcat()).
12254 *Ben Laurie (CHATS) and Richard Levitte*
12256 * Avoid using fixed-size buffers for one-line DNs.
12258 *Ben Laurie (CHATS)*
12260 * Add BUF_MEM_grow_clean() to avoid information leakage when
12261 resizing buffers containing secrets, and use where appropriate.
12263 *Ben Laurie (CHATS)*
12265 * Avoid using fixed size buffers for configuration file location.
12267 *Ben Laurie (CHATS)*
12269 * Avoid filename truncation for various CA files.
12271 *Ben Laurie (CHATS)*
12273 * Use sizeof in preference to magic numbers.
12275 *Ben Laurie (CHATS)*
12277 * Avoid filename truncation in cert requests.
12279 *Ben Laurie (CHATS)*
12281 * Add assertions to check for (supposedly impossible) buffer
12284 *Ben Laurie (CHATS)*
12286 * Don't cache truncated DNS entries in the local cache (this could
12287 potentially lead to a spoofing attack).
12289 *Ben Laurie (CHATS)*
12291 * Fix various buffers to be large enough for hex/decimal
12292 representations in a platform independent manner.
12294 *Ben Laurie (CHATS)*
12296 * Add CRYPTO_realloc_clean() to avoid information leakage when
12297 resizing buffers containing secrets, and use where appropriate.
12299 *Ben Laurie (CHATS)*
12301 * Add BIO_indent() to avoid much slightly worrying code to do
12304 *Ben Laurie (CHATS)*
12306 * Convert sprintf()/BIO_puts() to BIO_printf().
12308 *Ben Laurie (CHATS)*
12310 * buffer_gets() could terminate with the buffer only half
12313 *Ben Laurie (CHATS)*
12315 * Add assertions to prevent user-supplied crypto functions from
12316 overflowing internal buffers by having large block sizes, etc.
12318 *Ben Laurie (CHATS)*
12320 * New OPENSSL_assert() macro (similar to assert(), but enabled
12323 *Ben Laurie (CHATS)*
12325 * Eliminate unused copy of key in RC4.
12327 *Ben Laurie (CHATS)*
12329 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12331 *Ben Laurie (CHATS)*
12333 * Fix off-by-one error in EGD path.
12335 *Ben Laurie (CHATS)*
12337 * If RANDFILE path is too long, ignore instead of truncating.
12339 *Ben Laurie (CHATS)*
12341 * Eliminate unused and incorrectly sized X.509 structure
12344 *Ben Laurie (CHATS)*
12346 * Eliminate unused and dangerous function knumber().
12348 *Ben Laurie (CHATS)*
12350 * Eliminate unused and dangerous structure, KSSL_ERR.
12352 *Ben Laurie (CHATS)*
12354 * Protect against overlong session ID context length in an encoded
12355 session object. Since these are local, this does not appear to be
12358 *Ben Laurie (CHATS)*
12360 * Change from security patch (see 0.9.6e below) that did not affect
12361 the 0.9.6 release series:
12363 Remote buffer overflow in SSL3 protocol - an attacker could
12364 supply an oversized master key in Kerberos-enabled versions.
12367 *Ben Laurie (CHATS)*
12369 * Change the SSL kerb5 codes to match RFC 2712.
12373 * Make -nameopt work fully for req and add -reqopt switch.
12375 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12377 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12379 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12381 * Make sure tests can be performed even if the corresponding algorithms
12382 have been removed entirely. This was also the last step to make
12383 OpenSSL compilable with DJGPP under all reasonable conditions.
12385 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12387 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12388 to allow version independent disabling of normally unselected ciphers,
12389 which may be activated as a side-effect of selecting a single cipher.
12391 (E.g., cipher list string "RSA" enables ciphersuites that are left
12392 out of "ALL" because they do not provide symmetric encryption.
12393 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12395 *Lutz Jaenicke, Bodo Moeller*
12397 * Add appropriate support for separate platform-dependent build
12398 directories. The recommended way to make a platform-dependent
12399 build directory is the following (tested on Linux), maybe with
12402 # Place yourself outside of the OpenSSL source tree. In
12403 # this example, the environment variable OPENSSL_SOURCE
12404 # is assumed to contain the absolute OpenSSL source directory.
12405 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12406 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12407 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12408 mkdir -p `dirname $F`
12409 ln -s $OPENSSL_SOURCE/$F $F
12412 To be absolutely sure not to disturb the source tree, a "make clean"
12413 is a good thing. If it isn't successful, don't worry about it,
12414 it probably means the source directory is very clean.
12418 * Make sure any ENGINE control commands make local copies of string
12419 pointers passed to them whenever necessary. Otherwise it is possible
12420 the caller may have overwritten (or deallocated) the original string
12421 data when a later ENGINE operation tries to use the stored values.
12423 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12425 * Improve diagnostics in file reading and command-line digests.
12427 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12429 * Add AES modes CFB and OFB to the object database. Correct an
12430 error in AES-CFB decryption.
12434 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12435 allows existing EVP_CIPHER_CTX structures to be reused after
12436 calling `EVP_*Final()`. This behaviour is used by encryption
12437 BIOs and some applications. This has the side effect that
12438 applications must explicitly clean up cipher contexts with
12439 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12443 * Check the values of dna and dnb in bn_mul_recursive before calling
12444 bn_mul_comba (a non zero value means the a or b arrays do not contain
12445 n2 elements) and fallback to bn_mul_normal if either is not zero.
12449 * Fix escaping of non-ASCII characters when using the -subj option
12450 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12454 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12455 form for "surname", serialNumber has no short form.
12456 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12457 therefore remove "mail" short name for "internet 7".
12458 The OID for unique identifiers in X509 certificates is
12459 x500UniqueIdentifier, not uniqueIdentifier.
12460 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12464 * Add an "init" command to the ENGINE config module and auto initialize
12465 ENGINEs. Without any "init" command the ENGINE will be initialized
12466 after all ctrl commands have been executed on it. If init=1 the
12467 ENGINE is initialized at that point (ctrls before that point are run
12468 on the uninitialized ENGINE and after on the initialized one). If
12469 init=0 then the ENGINE will not be initialized at all.
12473 * Fix the 'app_verify_callback' interface so that the user-defined
12474 argument is actually passed to the callback: In the
12475 SSL_CTX_set_cert_verify_callback() prototype, the callback
12476 declaration has been changed from
12479 int (*cb)(X509_STORE_CTX *,void *);
12480 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12481 i=s->ctx->app_verify_callback(&ctx)
12482 has been changed into
12483 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12485 To update applications using SSL_CTX_set_cert_verify_callback(),
12486 a dummy argument can be added to their callback functions.
12488 *D. K. Smetters <smetters@parc.xerox.com>*
12490 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12492 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12494 * Add and OPENSSL_LOAD_CONF define which will cause
12495 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12496 This allows older applications to transparently support certain
12497 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12498 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12499 load the config file and OPENSSL_add_all_algorithms_conf() which will
12500 always load it have also been added.
12504 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12505 Adjust NIDs and EVP layer.
12507 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12509 * Config modules support in openssl utility.
12511 Most commands now load modules from the config file,
12512 though in a few (such as version) this isn't done
12513 because it couldn't be used for anything.
12515 In the case of ca and req the config file used is
12516 the same as the utility itself: that is the -config
12517 command line option can be used to specify an
12522 * Move default behaviour from OPENSSL_config(). If appname is NULL
12523 use "openssl_conf" if filename is NULL use default openssl config file.
12527 * Add an argument to OPENSSL_config() to allow the use of an alternative
12528 config section name. Add a new flag to tolerate a missing config file
12529 and move code to CONF_modules_load_file().
12533 * Support for crypto accelerator cards from Accelerated Encryption
12534 Processing, www.aep.ie. (Use engine 'aep')
12535 The support was copied from 0.9.6c [engine] and adapted/corrected
12536 to work with the new engine framework.
12538 *AEP Inc. and Richard Levitte*
12540 * Support for SureWare crypto accelerator cards from Baltimore
12541 Technologies. (Use engine 'sureware')
12542 The support was copied from 0.9.6c [engine] and adapted
12543 to work with the new engine framework.
12547 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12548 make the newer ENGINE framework commands for the CHIL engine work.
12550 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12552 * Make it possible to produce shared libraries on ReliantUNIX.
12554 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12556 * Add the configuration target debug-linux-ppro.
12557 Make 'openssl rsa' use the general key loading routines
12558 implemented in `apps.c`, and make those routines able to
12559 handle the key format FORMAT_NETSCAPE and the variant
12562 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12564 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12566 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12568 * Add -keyform to rsautl, and document -engine.
12570 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12572 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12573 BIO_R_NO_SUCH_FILE error code rather than the generic
12574 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12578 * Add new functions
12579 ERR_peek_last_error
12580 ERR_peek_last_error_line
12581 ERR_peek_last_error_line_data.
12582 These are similar to
12584 ERR_peek_error_line
12585 ERR_peek_error_line_data,
12586 but report on the latest error recorded rather than the first one
12587 still in the error queue.
12589 *Ben Laurie, Bodo Moeller*
12591 * default_algorithms option in ENGINE config module. This allows things
12593 default_algorithms = ALL
12594 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12598 * Preliminary ENGINE config module.
12602 * New experimental application configuration code.
12606 * Change the AES code to follow the same name structure as all other
12607 symmetric ciphers, and behave the same way. Move everything to
12608 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12610 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12612 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12614 *Ben Laurie and Theo de Raadt*
12616 * Add option to output public keys in req command.
12618 *Massimiliano Pala madwolf@openca.org*
12620 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12621 (up to about 10% better than before for P-192 and P-224).
12625 * New functions/macros
12627 SSL_CTX_set_msg_callback(ctx, cb)
12628 SSL_CTX_set_msg_callback_arg(ctx, arg)
12629 SSL_set_msg_callback(ssl, cb)
12630 SSL_set_msg_callback_arg(ssl, arg)
12632 to request calling a callback function
12634 void cb(int write_p, int version, int content_type,
12635 const void *buf, size_t len, SSL *ssl, void *arg)
12637 whenever a protocol message has been completely received
12638 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12639 protocol version according to which the SSL library interprets
12640 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12641 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12642 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12643 specification (change_cipher_spec(20), alert(21), handshake(22)).
12644 'buf' and 'len' point to the actual message, 'ssl' to the
12645 SSL object, and 'arg' is the application-defined value set by
12646 SSL[_CTX]_set_msg_callback_arg().
12648 'openssl s_client' and 'openssl s_server' have new '-msg' options
12649 to enable a callback that displays all protocol messages.
12653 * Change the shared library support so shared libraries are built as
12654 soon as the corresponding static library is finished, and thereby get
12655 openssl and the test programs linked against the shared library.
12656 This still only happens when the keyword "shard" has been given to
12657 the configuration scripts.
12659 NOTE: shared library support is still an experimental thing, and
12660 backward binary compatibility is still not guaranteed.
12662 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12664 * Add support for Subject Information Access extension.
12666 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12668 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12669 additional bytes when new memory had to be allocated, not just
12670 when reusing an existing buffer.
12674 * New command line and configuration option 'utf8' for the req command.
12675 This allows field values to be specified as UTF8 strings.
12679 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12680 runs for the former and machine-readable output for the latter.
12684 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12685 of the e-mail address in the DN (i.e., it will go into a certificate
12686 extension only). The new configuration file option 'email_in_dn = no'
12687 has the same effect.
12689 *Massimiliano Pala madwolf@openca.org*
12691 * Change all functions with names starting with `des_` to be starting
12692 with `DES_` instead. Add wrappers that are compatible with libdes,
12693 but are named `_ossl_old_des_*`. Finally, add macros that map the
12694 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12695 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12696 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12699 Since we provide two compatibility mappings, the user needs to
12700 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12701 compatibility is desired. The default (i.e., when that macro
12702 isn't defined) is OpenSSL 0.9.6c compatibility.
12704 There are also macros that enable and disable the support of old
12705 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12706 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12707 are defined, the default will apply: to support the old des routines.
12709 In either case, one must include openssl/des.h to get the correct
12710 definitions. Do not try to just include openssl/des_old.h, that
12713 NOTE: This is a major break of an old API into a new one. Software
12714 authors are encouraged to switch to the `DES_` style functions. Some
12715 time in the future, des_old.h and the libdes compatibility functions
12716 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12717 default), and then completely removed.
12721 * Test for certificates which contain unsupported critical extensions.
12722 If such a certificate is found during a verify operation it is
12723 rejected by default: this behaviour can be overridden by either
12724 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12725 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12726 X509_supported_extension() has also been added which returns 1 if a
12727 particular extension is supported.
12731 * Modify the behaviour of EVP cipher functions in similar way to digests
12732 to retain compatibility with existing code.
12736 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12737 compatibility with existing code. In particular the 'ctx' parameter does
12738 not have to be to be initialized before the call to EVP_DigestInit() and
12739 it is tidied up after a call to EVP_DigestFinal(). New function
12740 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12741 EVP_MD_CTX_copy() changed to not require the destination to be
12742 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12743 requires the destination to be valid.
12745 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12746 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12750 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12751 so that complete 'Handshake' protocol structures are kept in memory
12752 instead of overwriting 'msg_type' and 'length' with 'body' data.
12756 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12758 *Massimo Santin via Richard Levitte*
12760 * Major restructuring to the underlying ENGINE code. This includes
12761 reduction of linker bloat, separation of pure "ENGINE" manipulation
12762 (initialisation, etc) from functionality dealing with implementations
12763 of specific crypto interfaces. This change also introduces integrated
12764 support for symmetric ciphers and digest implementations - so ENGINEs
12765 can now accelerate these by providing EVP_CIPHER and EVP_MD
12766 implementations of their own. This is detailed in
12767 [crypto/engine/README.md](crypto/engine/README.md)
12768 as it couldn't be adequately described here. However, there are a few
12769 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12770 were changed in the original introduction of ENGINE code have now
12771 reverted back - the hooking from this code to ENGINE is now a good
12772 deal more passive and at run-time, operations deal directly with
12773 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12774 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12775 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12776 they were not being used by the framework as there is no concept of a
12777 BIGNUM_METHOD and they could not be generalised to the new
12778 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12779 ENGINE_cpy() has been removed as it cannot be consistently defined in
12784 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12788 * Change mkdef.pl to sort symbols that get the same entry number,
12789 and make sure the automatically generated functions `ERR_load_*`
12790 become part of libeay.num as well.
12794 * New function SSL_renegotiate_pending(). This returns true once
12795 renegotiation has been requested (either SSL_renegotiate() call
12796 or HelloRequest/ClientHello received from the peer) and becomes
12797 false once a handshake has been completed.
12798 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12799 sends a HelloRequest, but does not ensure that a handshake takes
12800 place. SSL_renegotiate_pending() is useful for checking if the
12801 client has followed the request.)
12805 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12806 By default, clients may request session resumption even during
12807 renegotiation (if session ID contexts permit); with this option,
12808 session resumption is possible only in the first handshake.
12810 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12811 more bits available for options that should not be part of
12812 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12816 * Add some demos for certificate and certificate request creation.
12820 * Make maximum certificate chain size accepted from the peer application
12821 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12822 "Douglas E. Engert" <deengert@anl.gov>.
12826 * Add support for shared libraries for Unixware-7
12827 (Boyd Lynn Gerber <gerberb@zenez.com>).
12831 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12832 be done prior to destruction. Use this to unload error strings from
12833 ENGINEs that load their own error strings. NB: This adds two new API
12834 functions to "get" and "set" this destroy handler in an ENGINE.
12838 * Alter all existing ENGINE implementations (except "openssl" and
12839 "openbsd") to dynamically instantiate their own error strings. This
12840 makes them more flexible to be built both as statically-linked ENGINEs
12841 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12842 Also, add stub code to each that makes building them as self-contained
12843 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12847 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12848 implementations into applications that are completely implemented in
12849 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12850 commands that can be used to configure what shared-library to load and
12851 to control aspects of the way it is handled. Also, made an update to
12852 the [README-Engine.md](README-Engine.md) file
12853 that brings its information up-to-date and
12854 provides some information and instructions on the "dynamic" ENGINE
12855 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12859 * Make it possible to unload ranges of ERR strings with a new
12860 "ERR_unload_strings" function.
12864 * Add a copy() function to EVP_MD.
12868 * Make EVP_MD routines take a context pointer instead of just the
12869 md_data void pointer.
12873 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12874 that the digest can only process a single chunk of data
12875 (typically because it is provided by a piece of
12876 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12877 is only going to provide a single chunk of data, and hence the
12878 framework needn't accumulate the data for oneshot drivers.
12882 * As with "ERR", make it possible to replace the underlying "ex_data"
12883 functions. This change also alters the storage and management of global
12884 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12885 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12886 index counters. The API functions that use this state have been changed
12887 to take a "class_index" rather than pointers to the class's local STACK
12888 and counter, and there is now an API function to dynamically create new
12889 classes. This centralisation allows us to (a) plug a lot of the
12890 thread-safety problems that existed, and (b) makes it possible to clean
12891 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12892 such data would previously have always leaked in application code and
12893 workarounds were in place to make the memory debugging turn a blind eye
12894 to it. Application code that doesn't use this new function will still
12895 leak as before, but their memory debugging output will announce it now
12896 rather than letting it slide.
12898 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12899 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12900 has a return value to indicate success or failure.
12904 * Make it possible to replace the underlying "ERR" functions such that the
12905 global state (2 LHASH tables and 2 locks) is only used by the "default"
12906 implementation. This change also adds two functions to "get" and "set"
12907 the implementation prior to it being automatically set the first time
12908 any other ERR function takes place. Ie. an application can call "get",
12909 pass the return value to a module it has just loaded, and that module
12910 can call its own "set" function using that value. This means the
12911 module's "ERR" operations will use (and modify) the error state in the
12912 application and not in its own statically linked copy of OpenSSL code.
12916 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12917 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12918 the operation, and provides a more encapsulated way for external code
12919 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12920 to use these functions rather than manually incrementing the counts.
12922 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12926 * Add EVP test program.
12930 * Add symmetric cipher support to ENGINE. Expect the API to change!
12934 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12935 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12936 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12937 These allow a CRL to be built without having to access X509_CRL fields
12938 directly. Modify 'ca' application to use new functions.
12942 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12943 bug workarounds. Rollback attack detection is a security feature.
12944 The problem will only arise on OpenSSL servers when TLSv1 is not
12945 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12946 Software authors not wanting to support TLSv1 will have special reasons
12947 for their choice and can explicitly enable this option.
12949 *Bodo Moeller, Lutz Jaenicke*
12951 * Rationalise EVP so it can be extended: don't include a union of
12952 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12953 (similar to those existing for EVP_CIPHER_CTX).
12958 EVP_MD_CTX_init(&md); /* new function call */
12959 EVP_DigestInit(&md, EVP_sha1());
12960 EVP_DigestUpdate(&md, in, len);
12961 EVP_DigestFinal(&md, out, NULL);
12962 EVP_MD_CTX_cleanup(&md); /* new function call */
12966 * Make DES key schedule conform to the usual scheme, as well as
12967 correcting its structure. This means that calls to DES functions
12968 now have to pass a pointer to a des_key_schedule instead of a
12969 plain des_key_schedule (which was actually always a pointer
12972 des_key_schedule ks;
12974 des_set_key_checked(..., &ks);
12975 des_ncbc_encrypt(..., &ks, ...);
12977 (Note that a later change renames 'des_...' into 'DES_...'.)
12981 * Initial reduction of linker bloat: the use of some functions, such as
12982 PEM causes large amounts of unused functions to be linked in due to
12983 poor organisation. For example pem_all.c contains every PEM function
12984 which has a knock on effect of linking in large amounts of (unused)
12985 ASN1 code. Grouping together similar functions and splitting unrelated
12986 functions prevents this.
12990 * Cleanup of EVP macros.
12994 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12995 correct `_ecb suffix`.
12999 * Add initial OCSP responder support to ocsp application. The
13000 revocation information is handled using the text based index
13001 use by the ca application. The responder can either handle
13002 requests generated internally, supplied in files (for example
13003 via a CGI script) or using an internal minimal server.
13007 * Add configuration choices to get zlib compression for TLS.
13011 * Changes to Kerberos SSL for RFC 2712 compliance:
13012 1. Implemented real KerberosWrapper, instead of just using
13013 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13014 2. Implemented optional authenticator field of KerberosWrapper.
13016 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13017 and authenticator structs; see crypto/krb5/.
13019 Generalized Kerberos calls to support multiple Kerberos libraries.
13020 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13021 via Richard Levitte*
13023 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13024 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13025 values for each of the key sizes rather than having just
13026 parameters (and 'speed' generating keys each time).
13030 * Speed up EVP routines.
13033 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13034 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13035 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13036 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13038 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13039 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13040 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13043 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13045 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13049 * Added the OS2-EMX target.
13051 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13053 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13054 New functions to support `NCONF` routines in extension code.
13055 New function `CONF_set_nconf()`
13056 to allow functions which take an `NCONF` to also handle the old `LHASH`
13057 structure: this means that the old `CONF` compatible routines can be
13058 retained (in particular w.rt. extensions) without having to duplicate the
13059 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13063 * Enhance the general user interface with mechanisms for inner control
13064 and with possibilities to have yes/no kind of prompts.
13068 * Change all calls to low-level digest routines in the library and
13069 applications to use EVP. Add missing calls to HMAC_cleanup() and
13070 don't assume HMAC_CTX can be copied using memcpy().
13072 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13074 * Add the possibility to control engines through control names but with
13075 arbitrary arguments instead of just a string.
13076 Change the key loaders to take a UI_METHOD instead of a callback
13077 function pointer. NOTE: this breaks binary compatibility with earlier
13078 versions of OpenSSL [engine].
13079 Adapt the nCipher code for these new conditions and add a card insertion
13084 * Enhance the general user interface with mechanisms to better support
13085 dialog box interfaces, application-defined prompts, the possibility
13086 to use defaults (for example default passwords from somewhere else)
13087 and interrupts/cancellations.
13091 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13092 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13096 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13097 tidy up some unnecessarily weird code in 'sk_new()').
13099 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13101 * Change the key loading routines for ENGINEs to use the same kind
13102 callback (pem_password_cb) as all other routines that need this
13107 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13108 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13109 than this minimum value is recommended.
13113 * New random seeder for OpenVMS, using the system process statistics
13114 that are easily reachable.
13118 * Windows apparently can't transparently handle global
13119 variables defined in DLLs. Initialisations such as:
13121 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13123 won't compile. This is used by the any applications that need to
13124 declare their own ASN1 modules. This was fixed by adding the option
13125 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13126 needed for static libraries under Win32.
13130 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13131 setting of purpose and trust fields. New X509_STORE trust and
13132 purpose functions and tidy up setting in other SSL functions.
13136 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13137 structure. These are inherited by X509_STORE_CTX when it is
13138 initialised. This allows various defaults to be set in the
13139 X509_STORE structure (such as flags for CRL checking and custom
13140 purpose or trust settings) for functions which only use X509_STORE_CTX
13141 internally such as S/MIME.
13143 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13144 trust settings if they are not set in X509_STORE. This allows X509_STORE
13145 purposes and trust (in S/MIME for example) to override any set by default.
13147 Add command line options for CRL checking to smime, s_client and s_server
13152 * Initial CRL based revocation checking. If the CRL checking flag(s)
13153 are set then the CRL is looked up in the X509_STORE structure and
13154 its validity and signature checked, then if the certificate is found
13155 in the CRL the verify fails with a revoked error.
13157 Various new CRL related callbacks added to X509_STORE_CTX structure.
13159 Command line options added to 'verify' application to support this.
13161 This needs some additional work, such as being able to handle multiple
13162 CRLs with different times, extension based lookup (rather than just
13163 by subject name) and ultimately more complete V2 CRL extension
13168 * Add a general user interface API (crypto/ui/). This is designed
13169 to replace things like des_read_password and friends (backward
13170 compatibility functions using this new API are provided).
13171 The purpose is to remove prompting functions from the DES code
13172 section as well as provide for prompting through dialog boxes in
13173 a window system and the like.
13177 * Add "ex_data" support to ENGINE so implementations can add state at a
13178 per-structure level rather than having to store it globally.
13182 * Make it possible for ENGINE structures to be copied when retrieved by
13183 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13184 This causes the "original" ENGINE structure to act like a template,
13185 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13186 operational state can be localised to each ENGINE structure, despite the
13187 fact they all share the same "methods". New ENGINE structures returned in
13188 this case have no functional references and the return value is the single
13189 structural reference. This matches the single structural reference returned
13190 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13195 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13196 needs to match any other type at all we need to manually clear the
13201 * Changes to the "openssl engine" utility to include;
13202 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13203 about an ENGINE's available control commands.
13204 - executing control commands from command line arguments using the
13205 '-pre' and '-post' switches. '-post' is only used if '-t' is
13206 specified and the ENGINE is successfully initialised. The syntax for
13207 the individual commands are colon-separated, for example;
13208 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13212 * New dynamic control command support for ENGINEs. ENGINEs can now
13213 declare their own commands (numbers), names (strings), descriptions,
13214 and input types for run-time discovery by calling applications. A
13215 subset of these commands are implicitly classed as "executable"
13216 depending on their input type, and only these can be invoked through
13217 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13218 can be based on user input, config files, etc). The distinction is
13219 that "executable" commands cannot return anything other than a boolean
13220 result and can only support numeric or string input, whereas some
13221 discoverable commands may only be for direct use through
13222 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13223 pointers, or other custom uses. The "executable" commands are to
13224 support parameterisations of ENGINE behaviour that can be
13225 unambiguously defined by ENGINEs and used consistently across any
13226 OpenSSL-based application. Commands have been added to all the
13227 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13228 control over shared-library paths without source code alterations.
13232 * Changed all ENGINE implementations to dynamically allocate their
13233 ENGINEs rather than declaring them statically. Apart from this being
13234 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13235 this also allows the implementations to compile without using the
13236 internal engine_int.h header.
13240 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13241 'const' value. Any code that should be able to modify a RAND_METHOD
13242 should already have non-const pointers to it (ie. they should only
13243 modify their own ones).
13247 * Made a variety of little tweaks to the ENGINE code.
13248 - "atalla" and "ubsec" string definitions were moved from header files
13249 to C code. "nuron" string definitions were placed in variables
13250 rather than hard-coded - allowing parameterisation of these values
13251 later on via ctrl() commands.
13252 - Removed unused "#if 0"'d code.
13253 - Fixed engine list iteration code so it uses ENGINE_free() to release
13254 structural references.
13255 - Constified the RAND_METHOD element of ENGINE structures.
13256 - Constified various get/set functions as appropriate and added
13257 missing functions (including a catch-all ENGINE_cpy that duplicates
13258 all ENGINE values onto a new ENGINE except reference counts/state).
13259 - Removed NULL parameter checks in get/set functions. Setting a method
13260 or function to NULL is a way of cancelling out a previously set
13261 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13262 and doesn't justify the extra error symbols and code.
13263 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13264 flags from engine_int.h to engine.h.
13265 - Changed prototypes for ENGINE handler functions (init(), finish(),
13266 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13270 * Implement binary inversion algorithm for BN_mod_inverse in addition
13271 to the algorithm using long division. The binary algorithm can be
13272 used only if the modulus is odd. On 32-bit systems, it is faster
13273 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13274 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13275 up to 450 bits. In 64-bit environments, the binary algorithm
13276 appears to be advantageous for much longer moduli; here we use it
13277 for moduli up to 2048 bits.
13281 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13282 could not support the combine flag in choice fields.
13286 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13287 extensions from a certificate request to the certificate.
13291 * Allow multiple 'certopt' and 'nameopt' options to be separated
13292 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13293 file: this allows the display of the certificate about to be
13294 signed to be customised, to allow certain fields to be included
13295 or excluded and extension details. The old system didn't display
13296 multicharacter strings properly, omitted fields not in the policy
13297 and couldn't display additional details such as extensions.
13301 * Function EC_POINTs_mul for multiple scalar multiplication
13302 of an arbitrary number of elliptic curve points
13303 \sum scalars[i]*points[i],
13304 optionally including the generator defined for the EC_GROUP:
13305 scalar*generator + \sum scalars[i]*points[i].
13307 EC_POINT_mul is a simple wrapper function for the typical case
13308 that the point list has just one item (besides the optional
13313 * First EC_METHODs for curves over GF(p):
13315 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13316 operations and provides various method functions that can also
13317 operate with faster implementations of modular arithmetic.
13319 EC_GFp_mont_method() reuses most functions that are part of
13320 EC_GFp_simple_method, but uses Montgomery arithmetic.
13322 *Bodo Moeller; point addition and point doubling
13323 implementation directly derived from source code provided by
13324 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13326 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13327 crypto/ec/ec_lib.c):
13329 Curves are EC_GROUP objects (with an optional group generator)
13330 based on EC_METHODs that are built into the library.
13332 Points are EC_POINT objects based on EC_GROUP objects.
13334 Most of the framework would be able to handle curves over arbitrary
13335 finite fields, but as there are no obvious types for fields other
13336 than GF(p), some functions are limited to that for now.
13340 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13341 that the file contains a complete HTTP response.
13345 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13346 change the def and num file printf format specifier from "%-40sXXX"
13347 to "%-39s XXX". The latter will always guarantee a space after the
13348 field while the former will cause them to run together if the field
13349 is 40 of more characters long.
13353 * Constify the cipher and digest 'method' functions and structures
13354 and modify related functions to take constant EVP_MD and EVP_CIPHER
13359 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13360 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13364 * Modify `EVP_Digest*()` routines so they now return values. Although the
13365 internal software routines can never fail additional hardware versions
13370 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13372 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13373 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13376 ERR_R_NESTED_ASN1_ERROR
13378 ERR_R_MISSING_ASN1_EOS
13379 were 4 .. 9, conflicting with
13380 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13382 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13383 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13385 Add new error code 'ERR_R_INTERNAL_ERROR'.
13389 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13394 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13395 sets the subject name for a new request or supersedes the
13396 subject name in a given request. Formats that can be parsed are
13397 'CN=Some Name, OU=myOU, C=IT'
13399 'CN=Some Name/OU=myOU/C=IT'.
13401 Add options '-batch' and '-verbose' to 'openssl req'.
13403 *Massimiliano Pala <madwolf@hackmasters.net>*
13405 * Introduce the possibility to access global variables through
13406 functions on platform were that's the best way to handle exporting
13407 global variables in shared libraries. To enable this functionality,
13408 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13409 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13410 is normally done by Configure or something similar).
13412 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13413 in the source file (foo.c) like this:
13415 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13416 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13418 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13419 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13421 OPENSSL_DECLARE_GLOBAL(int,foo);
13422 #define foo OPENSSL_GLOBAL_REF(foo)
13423 OPENSSL_DECLARE_GLOBAL(double,bar);
13424 #define bar OPENSSL_GLOBAL_REF(bar)
13426 The #defines are very important, and therefore so is including the
13427 header file everywhere where the defined globals are used.
13429 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13430 of ASN.1 items, but that structure is a bit different.
13432 The largest change is in util/mkdef.pl which has been enhanced with
13433 better and easier to understand logic to choose which symbols should
13434 go into the Windows .def files as well as a number of fixes and code
13435 cleanup (among others, algorithm keywords are now sorted
13436 lexicographically to avoid constant rewrites).
13440 * In BN_div() keep a copy of the sign of 'num' before writing the
13441 result to 'rm' because if rm==num the value will be overwritten
13442 and produce the wrong result if 'num' is negative: this caused
13443 problems with BN_mod() and BN_nnmod().
13447 * Function OCSP_request_verify(). This checks the signature on an
13448 OCSP request and verifies the signer certificate. The signer
13449 certificate is just checked for a generic purpose and OCSP request
13454 * Add OCSP_check_validity() function to check the validity of OCSP
13455 responses. OCSP responses are prepared in real time and may only
13456 be a few seconds old. Simply checking that the current time lies
13457 between thisUpdate and nextUpdate max reject otherwise valid responses
13458 caused by either OCSP responder or client clock inaccuracy. Instead
13459 we allow thisUpdate and nextUpdate to fall within a certain period of
13460 the current time. The age of the response can also optionally be
13461 checked. Two new options -validity_period and -status_age added to
13466 * If signature or public key algorithm is unrecognized print out its
13467 OID rather that just UNKNOWN.
13471 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13472 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13473 ID to be generated from the issuer certificate alone which can then be
13474 passed to OCSP_id_issuer_cmp().
13478 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13479 ASN1 modules to export functions returning ASN1_ITEM pointers
13480 instead of the ASN1_ITEM structures themselves. This adds several
13481 new macros which allow the underlying ASN1 function/structure to
13482 be accessed transparently. As a result code should not use ASN1_ITEM
13483 references directly (such as &X509_it) but instead use the relevant
13484 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13485 use of the new ASN1 code on platforms where exporting structures
13486 is problematical (for example in shared libraries) but exporting
13487 functions returning pointers to structures is not.
13491 * Add support for overriding the generation of SSL/TLS session IDs.
13492 These callbacks can be registered either in an SSL_CTX or per SSL.
13493 The purpose of this is to allow applications to control, if they wish,
13494 the arbitrary values chosen for use as session IDs, particularly as it
13495 can be useful for session caching in multiple-server environments. A
13496 command-line switch for testing this (and any client code that wishes
13497 to use such a feature) has been added to "s_server".
13499 *Geoff Thorpe, Lutz Jaenicke*
13501 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13502 of the form `#if defined(...) || defined(...) || ...` and
13503 `#if !defined(...) && !defined(...) && ...`. This also avoids
13504 the growing number of special cases it was previously handling.
13508 * Make all configuration macros available for application by making
13509 sure they are available in opensslconf.h, by giving them names starting
13510 with `OPENSSL_` to avoid conflicts with other packages and by making
13511 sure e_os2.h will cover all platform-specific cases together with
13513 Additionally, it is now possible to define configuration/platform-
13514 specific names (called "system identities"). In the C code, these
13515 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13516 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13517 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13522 * New option -set_serial to 'req' and 'x509' this allows the serial
13523 number to use to be specified on the command line. Previously self
13524 signed certificates were hard coded with serial number 0 and the
13525 CA options of 'x509' had to use a serial number in a file which was
13530 * New options to 'ca' utility to support V2 CRL entry extensions.
13531 Currently CRL reason, invalidity date and hold instruction are
13532 supported. Add new CRL extensions to V3 code and some new objects.
13536 * New function EVP_CIPHER_CTX_set_padding() this is used to
13537 disable standard block padding (aka PKCS#5 padding) in the EVP
13538 API, which was previously mandatory. This means that the data is
13539 not padded in any way and so the total length much be a multiple
13540 of the block size, otherwise an error occurs.
13544 * Initial (incomplete) OCSP SSL support.
13548 * New function OCSP_parse_url(). This splits up a URL into its host,
13549 port and path components: primarily to parse OCSP URLs. New -url
13550 option to ocsp utility.
13554 * New nonce behavior. The return value of OCSP_check_nonce() now
13555 reflects the various checks performed. Applications can decide
13556 whether to tolerate certain situations such as an absent nonce
13557 in a response when one was present in a request: the ocsp application
13558 just prints out a warning. New function OCSP_add1_basic_nonce()
13559 this is to allow responders to include a nonce in a response even if
13560 the request is nonce-less.
13564 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13565 skipped when using openssl x509 multiple times on a single input file,
13566 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13570 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13571 set string type: to handle setting ASN1_TIME structures. Fix ca
13572 utility to correctly initialize revocation date of CRLs.
13576 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13577 the clients preferred ciphersuites and rather use its own preferences.
13578 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13579 Internet Explorer by ensuring unchanged hash method during stepup.
13580 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13584 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13585 to aes and add a new 'exist' option to print out symbols that don't
13590 * Additional options to ocsp utility to allow flags to be set and
13591 additional certificates supplied.
13595 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13596 OCSP client a number of certificate to only verify the response
13601 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13602 handle the new API. Currently only ECB, CBC modes supported. Add new
13605 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13606 Encryption Standard (AES) Ciphersuites for Transport Layer
13607 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13608 not enabled by default and were not part of the "ALL" ciphersuite
13609 alias because they were not yet official; they could be
13610 explicitly requested by specifying the "AESdraft" ciphersuite
13611 group alias. In the final release of OpenSSL 0.9.7, the group
13612 alias is called "AES" and is part of "ALL".)
13614 *Ben Laurie, Steve Henson, Bodo Moeller*
13616 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13617 request to response.
13621 * Functions for OCSP responders. OCSP_request_onereq_count(),
13622 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13623 extract information from a certificate request. OCSP_response_create()
13624 creates a response and optionally adds a basic response structure.
13625 OCSP_basic_add1_status() adds a complete single response to a basic
13626 response and returns the OCSP_SINGLERESP structure just added (to allow
13627 extensions to be included for example). OCSP_basic_add1_cert() adds a
13628 certificate to a basic response and OCSP_basic_sign() signs a basic
13629 response with various flags. New helper functions ASN1_TIME_check()
13630 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13631 (converts ASN1_TIME to GeneralizedTime).
13635 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13636 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13637 structure from a certificate. X509_pubkey_digest() digests the public_key
13638 contents: this is used in various key identifiers.
13642 * Make sk_sort() tolerate a NULL argument.
13644 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13646 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13647 passed by the function are trusted implicitly. If any of them signed the
13648 response then it is assumed to be valid and is not verified.
13652 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13653 to data. This was previously part of the PKCS7 ASN1 code. This
13654 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13655 *Steve Henson, reported by Kenneth R. Robinette
13656 <support@securenetterm.com>*
13658 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13659 routines: without these tracing memory leaks is very painful.
13660 Fix leaks in PKCS12 and PKCS7 routines.
13664 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13665 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13666 effectively meant GeneralizedTime would never be used. Now it
13667 is initialised to -1 but X509_time_adj() now has to check the value
13668 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13669 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13670 *Steve Henson, reported by Kenneth R. Robinette
13671 <support@securenetterm.com>*
13673 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13674 result in a zero length in the ASN1_INTEGER structure which was
13675 not consistent with the structure when d2i_ASN1_INTEGER() was used
13676 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13677 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13678 where it did not print out a minus for negative ASN1_INTEGER.
13682 * Add summary printout to ocsp utility. The various functions which
13683 convert status values to strings have been renamed to:
13684 OCSP_response_status_str(), OCSP_cert_status_str() and
13685 OCSP_crl_reason_str() and are no longer static. New options
13686 to verify nonce values and to disable verification. OCSP response
13687 printout format cleaned up.
13691 * Add additional OCSP certificate checks. These are those specified
13692 in RFC2560. This consists of two separate checks: the CA of the
13693 certificate being checked must either be the OCSP signer certificate
13694 or the issuer of the OCSP signer certificate. In the latter case the
13695 OCSP signer certificate must contain the OCSP signing extended key
13696 usage. This check is performed by attempting to match the OCSP
13697 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13698 in the OCSP_CERTID structures of the response.
13702 * Initial OCSP certificate verification added to OCSP_basic_verify()
13703 and related routines. This uses the standard OpenSSL certificate
13704 verify routines to perform initial checks (just CA validity) and
13705 to obtain the certificate chain. Then additional checks will be
13706 performed on the chain. Currently the root CA is checked to see
13707 if it is explicitly trusted for OCSP signing. This is used to set
13708 a root CA as a global signing root: that is any certificate that
13709 chains to that CA is an acceptable OCSP signing certificate.
13713 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13714 extensions from a separate configuration file.
13715 As when reading extensions from the main configuration file,
13716 the '-extensions ...' option may be used for specifying the
13719 *Massimiliano Pala <madwolf@comune.modena.it>*
13721 * New OCSP utility. Allows OCSP requests to be generated or
13722 read. The request can be sent to a responder and the output
13723 parsed, outputted or printed in text form. Not complete yet:
13724 still needs to check the OCSP response validity.
13728 * New subcommands for 'openssl ca':
13729 `openssl ca -status <serial>` prints the status of the cert with
13730 the given serial number (according to the index file).
13731 `openssl ca -updatedb` updates the expiry status of certificates
13734 *Massimiliano Pala <madwolf@comune.modena.it>*
13736 * New '-newreq-nodes' command option to CA.pl. This is like
13737 '-newreq', but calls 'openssl req' with the '-nodes' option
13738 so that the resulting key is not encrypted.
13740 *Damien Miller <djm@mindrot.org>*
13742 * New configuration for the GNU Hurd.
13744 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13746 * Initial code to implement OCSP basic response verify. This
13747 is currently incomplete. Currently just finds the signer's
13748 certificate and verifies the signature on the response.
13752 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13753 value of OPENSSLDIR. This is available via the new '-d' option
13754 to 'openssl version', and is also included in 'openssl version -a'.
13758 * Allowing defining memory allocation callbacks that will be given
13759 file name and line number information in additional arguments
13760 (a `const char*` and an int). The basic functionality remains, as
13761 well as the original possibility to just replace malloc(),
13762 realloc() and free() by functions that do not know about these
13763 additional arguments. To register and find out the current
13764 settings for extended allocation functions, the following
13765 functions are provided:
13767 CRYPTO_set_mem_ex_functions
13768 CRYPTO_set_locked_mem_ex_functions
13769 CRYPTO_get_mem_ex_functions
13770 CRYPTO_get_locked_mem_ex_functions
13772 These work the same way as CRYPTO_set_mem_functions and friends.
13773 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13774 extended allocation function is enabled.
13775 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13776 a conventional allocation function is enabled.
13778 *Richard Levitte, Bodo Moeller*
13780 * Finish off removing the remaining LHASH function pointer casts.
13781 There should no longer be any prototype-casting required when using
13782 the LHASH abstraction, and any casts that remain are "bugs". See
13783 the callback types and macros at the head of lhash.h for details
13784 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13788 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13789 If /dev/[u]random devices are not available or do not return enough
13790 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13792 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13793 /etc/entropy will be queried once each in this sequence, querying stops
13794 when enough entropy was collected without querying more sockets.
13798 * Change the Unix RAND_poll() variant to be able to poll several
13799 random devices, as specified by DEVRANDOM, until a sufficient amount
13800 of data has been collected. We spend at most 10 ms on each file
13801 (select timeout) and read in non-blocking mode. DEVRANDOM now
13802 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13803 (previously it was just the string "/dev/urandom"), so on typical
13804 platforms the 10 ms delay will never occur.
13805 Also separate out the Unix variant to its own file, rand_unix.c.
13806 For VMS, there's a currently-empty rand_vms.c.
13810 * Move OCSP client related routines to ocsp_cl.c. These
13811 provide utility functions which an application needing
13812 to issue a request to an OCSP responder and analyse the
13813 response will typically need: as opposed to those which an
13814 OCSP responder itself would need which will be added later.
13816 OCSP_request_sign() signs an OCSP request with an API similar
13817 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13818 response. OCSP_response_get1_basic() extracts basic response
13819 from response. OCSP_resp_find_status(): finds and extracts status
13820 information from an OCSP_CERTID structure (which will be created
13821 when the request structure is built). These are built from lower
13822 level functions which work on OCSP_SINGLERESP structures but
13823 won't normally be used unless the application wishes to examine
13824 extensions in the OCSP response for example.
13826 Replace nonce routines with a pair of functions.
13827 OCSP_request_add1_nonce() adds a nonce value and optionally
13828 generates a random value. OCSP_check_nonce() checks the
13829 validity of the nonce in an OCSP response.
13833 * Change function OCSP_request_add() to OCSP_request_add0_id().
13834 This doesn't copy the supplied OCSP_CERTID and avoids the
13835 need to free up the newly created id. Change return type
13836 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13837 This can then be used to add extensions to the request.
13838 Deleted OCSP_request_new(), since most of its functionality
13839 is now in OCSP_REQUEST_new() (and the case insensitive name
13840 clash) apart from the ability to set the request name which
13841 will be added elsewhere.
13845 * Update OCSP API. Remove obsolete extensions argument from
13846 various functions. Extensions are now handled using the new
13847 OCSP extension code. New simple OCSP HTTP function which
13848 can be used to send requests and parse the response.
13852 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13853 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13854 uses the special reorder version of SET OF to sort the attributes
13855 and reorder them to match the encoded order. This resolves a long
13856 standing problem: a verify on a PKCS7 structure just after signing
13857 it used to fail because the attribute order did not match the
13858 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13859 it uses the received order. This is necessary to tolerate some broken
13860 software that does not order SET OF. This is handled by encoding
13861 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13862 to produce the required SET OF.
13866 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13867 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13868 files to get correct declarations of the ASN.1 item variables.
13872 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13873 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13874 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13875 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13876 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13877 ASN1_ITEM and no wrapper functions.
13881 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13882 replace the old function pointer based I/O routines. Change most of
13883 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13887 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13888 lines, recognize more "algorithms" that can be deselected, and make
13889 it complain about algorithm deselection that isn't recognised.
13893 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13894 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13895 to use new functions. Add NO_ASN1_OLD which can be set to remove
13896 some old style ASN1 functions: this can be used to determine if old
13897 code will still work when these eventually go away.
13901 * New extension functions for OCSP structures, these follow the
13902 same conventions as certificates and CRLs.
13906 * New function X509V3_add1_i2d(). This automatically encodes and
13907 adds an extension. Its behaviour can be customised with various
13908 flags to append, replace or delete. Various wrappers added for
13909 certificates and CRLs.
13913 * Fix to avoid calling the underlying ASN1 print routine when
13914 an extension cannot be parsed. Correct a typo in the
13915 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13919 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13920 entries for variables.
13924 * Add functionality to `apps/openssl.c` for detecting locking
13925 problems: As the program is single-threaded, all we have
13926 to do is register a locking callback using an array for
13927 storing which locks are currently held by the program.
13931 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13932 SSL_get_ex_data_X509_STORE_idx(), which is used in
13933 ssl_verify_cert_chain() and thus can be called at any time
13934 during TLS/SSL handshakes so that thread-safety is essential.
13935 Unfortunately, the ex_data design is not at all suited
13936 for multi-threaded use, so it probably should be abolished.
13940 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13942 *Broadcom, tweaked and integrated by Geoff Thorpe*
13944 * Move common extension printing code to new function
13945 X509V3_print_extensions(). Reorganise OCSP print routines and
13946 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13950 * New function X509_signature_print() to remove duplication in some
13955 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13956 set (this was treated exactly the same as SET OF previously). This
13957 is used to reorder the STACK representing the structure to match the
13958 encoding. This will be used to get round a problem where a PKCS7
13959 structure which was signed could not be verified because the STACK
13960 order did not reflect the encoded order.
13964 * Reimplement the OCSP ASN1 module using the new code.
13968 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13969 for its ASN1 operations. The old style function pointers still exist
13970 for now but they will eventually go away.
13974 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13975 completely replaces the old ASN1 functionality with a table driven
13976 encoder and decoder which interprets an ASN1_ITEM structure describing
13977 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13978 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13979 has also been converted to the new form.
13983 * Change BN_mod_exp_recp so that negative moduli are tolerated
13984 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13985 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13986 for negative moduli.
13990 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13991 of not touching the result's sign bit.
13995 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14000 * Changed the LHASH code to use prototypes for callbacks, and created
14001 macros to declare and implement thin (optionally static) functions
14002 that provide type-safety and avoid function pointer casting for the
14003 type-specific callbacks.
14007 * Added Kerberos Cipher Suites to be used with TLS, as written in
14009 *Veers Staats <staatsvr@asc.hpc.mil>,
14010 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14012 * Reformat the FAQ so the different questions and answers can be divided
14013 in sections depending on the subject.
14017 * Have the zlib compression code load ZLIB.DLL dynamically under
14022 * New function BN_mod_sqrt for computing square roots modulo a prime
14023 (using the probabilistic Tonelli-Shanks algorithm unless
14024 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14025 be handled deterministically).
14027 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14029 * Make BN_mod_inverse faster by explicitly handling small quotients
14030 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14031 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14035 * New function BN_kronecker.
14039 * Fix BN_gcd so that it works on negative inputs; the result is
14040 positive unless both parameters are zero.
14041 Previously something reasonably close to an infinite loop was
14042 possible because numbers could be growing instead of shrinking
14043 in the implementation of Euclid's algorithm.
14047 * Fix BN_is_word() and BN_is_one() macros to take into account the
14048 sign of the number in question.
14050 Fix BN_is_word(a,w) to work correctly for w == 0.
14052 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14053 because its test if the absolute value of 'a' equals 'w'.
14054 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14055 it exists mostly for use in the implementations of BN_is_zero(),
14056 BN_is_one(), and BN_is_word().
14060 * New function BN_swap.
14064 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14065 the exponentiation functions are more likely to produce reasonable
14066 results on negative inputs.
14070 * Change BN_mod_mul so that the result is always non-negative.
14071 Previously, it could be negative if one of the factors was negative;
14072 I don't think anyone really wanted that behaviour.
14076 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14077 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14078 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14079 and add new functions:
14088 BN_mod_lshift1_quick
14090 BN_mod_lshift_quick
14092 These functions always generate non-negative results.
14094 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14095 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14097 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14098 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14099 be reduced modulo `m`.
14101 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14104 The following entry accidentally appeared in the CHANGES file
14105 distributed with OpenSSL 0.9.7. The modifications described in
14106 it do *not* apply to OpenSSL 0.9.7.
14108 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14109 was actually never needed) and in BN_mul(). The removal in BN_mul()
14110 required a small change in bn_mul_part_recursive() and the addition
14111 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14112 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14113 bn_sub_words() and bn_add_words() except they take arrays with
14119 * In 'openssl passwd', verify passwords read from the terminal
14120 unless the '-salt' option is used (which usually means that
14121 verification would just waste user's time since the resulting
14122 hash is going to be compared with some given password hash)
14123 or the new '-noverify' option is used.
14125 This is an incompatible change, but it does not affect
14126 non-interactive use of 'openssl passwd' (passwords on the command
14127 line, '-stdin' option, '-in ...' option) and thus should not
14128 cause any problems.
14132 * Remove all references to RSAref, since there's no more need for it.
14136 * Make DSO load along a path given through an environment variable
14137 (SHLIB_PATH) with shl_load().
14141 * Constify the ENGINE code as a result of BIGNUM constification.
14142 Also constify the RSA code and most things related to it. In a
14143 few places, most notable in the depth of the ASN.1 code, ugly
14144 casts back to non-const were required (to be solved at a later
14149 * Make it so the openssl application has all engines loaded by default.
14153 * Constify the BIGNUM routines a little more.
14157 * Add the following functions:
14159 ENGINE_load_cswift()
14161 ENGINE_load_atalla()
14162 ENGINE_load_nuron()
14163 ENGINE_load_builtin_engines()
14165 That way, an application can itself choose if external engines that
14166 are built-in in OpenSSL shall ever be used or not. The benefit is
14167 that applications won't have to be linked with libdl or other dso
14168 libraries unless it's really needed.
14170 Changed 'openssl engine' to load all engines on demand.
14171 Changed the engine header files to avoid the duplication of some
14172 declarations (they differed!).
14176 * 'openssl engine' can now list capabilities.
14180 * Better error reporting in 'openssl engine'.
14184 * Never call load_dh_param(NULL) in s_server.
14188 * Add engine application. It can currently list engines by name and
14189 identity, and test if they are actually available.
14193 * Improve RPM specification file by forcing symbolic linking and making
14194 sure the installed documentation is also owned by root.root.
14196 *Damien Miller <djm@mindrot.org>*
14198 * Give the OpenSSL applications more possibilities to make use of
14199 keys (public as well as private) handled by engines.
14203 * Add OCSP code that comes from CertCo.
14207 * Add VMS support for the Rijndael code.
14211 * Added untested support for Nuron crypto accelerator.
14215 * Add support for external cryptographic devices. This code was
14216 previously distributed separately as the "engine" branch.
14218 *Geoff Thorpe, Richard Levitte*
14220 * Rework the filename-translation in the DSO code. It is now possible to
14221 have far greater control over how a "name" is turned into a filename
14222 depending on the operating environment and any oddities about the
14223 different shared library filenames on each system.
14227 * Support threads on FreeBSD-elf in Configure.
14231 * Fix for SHA1 assembly problem with MASM: it produces
14232 warnings about corrupt line number information when assembling
14233 with debugging information. This is caused by the overlapping
14236 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14239 NCONF_get_number() has no error checking at all. As a replacement,
14240 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14241 promoted strongly. The old NCONF_get_number is kept around for
14242 binary backward compatibility.
14243 Make it possible for methods to load from something other than a BIO,
14244 by providing a function pointer that is given a name instead of a BIO.
14245 For example, this could be used to load configuration data from an
14250 * Fix for non blocking accept BIOs. Added new I/O special reason
14251 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14252 with non blocking I/O was not possible because no retry code was
14253 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14258 * Added the beginnings of Rijndael support.
14262 * Fix for bug in DirectoryString mask setting. Add support for
14263 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14264 to allow certificate printing to more controllable, additional
14265 'certopt' option to 'x509' to allow new printing options to be
14270 * Clean old EAY MD5 hack from e_os.h.
14274 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14276 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14277 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14279 *Joe Orton, Steve Henson*
14281 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14283 * Fix additional bug revealed by the NISCC test suite:
14285 Stop bug triggering large recursion when presented with
14286 certain ASN.1 tags ([CVE-2003-0851])
14290 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14292 * Fix various bugs revealed by running the NISCC test suite:
14294 Stop out of bounds reads in the ASN1 code when presented with
14295 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14297 If verify callback ignores invalid public key errors don't try to check
14298 certificate signature with the NULL public key.
14302 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14303 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14308 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14309 extra data after the compression methods not only for TLS 1.0
14310 but also for SSL 3.0 (as required by the specification).
14312 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14314 * Change X509_certificate_type() to mark the key as exported/exportable
14315 when it's 512 *bits* long, not 512 bytes.
14319 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14321 * Countermeasure against the Klima-Pokorny-Rosa extension of
14322 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14323 a protocol version number mismatch like a decryption error
14324 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14328 * Turn on RSA blinding by default in the default implementation
14329 to avoid a timing attack. Applications that don't want it can call
14330 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14331 They would be ill-advised to do so in most cases.
14333 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14335 * Change RSA blinding code so that it works when the PRNG is not
14336 seeded (in this case, the secret RSA exponent is abused as
14337 an unpredictable seed -- if it is not unpredictable, there
14338 is no point in blinding anyway). Make RSA blinding thread-safe
14339 by remembering the creator's thread ID in rsa->blinding and
14340 having all other threads use local one-time blinding factors
14341 (this requires more computation than sharing rsa->blinding, but
14342 avoids excessive locking; and if an RSA object is not shared
14343 between threads, blinding will still be very fast).
14347 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14349 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14350 via timing by performing a MAC computation even if incorrect
14351 block cipher padding has been found. This is a countermeasure
14352 against active attacks where the attacker has to distinguish
14353 between bad padding and a MAC verification error. ([CVE-2003-0078])
14355 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14356 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14357 Martin Vuagnoux (EPFL, Ilion)*
14359 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14361 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14362 memory from its contents. This is done with a counter that will
14363 place alternating values in each byte. This can be used to solve
14364 two issues: 1) the removal of calls to memset() by highly optimizing
14365 compilers, and 2) cleansing with other values than 0, since those can
14366 be read through on certain media, for example a swap space on disk.
14370 * Bugfix: client side session caching did not work with external caching,
14371 because the session->cipher setting was not restored when reloading
14372 from the external cache. This problem was masked, when
14373 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14374 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14378 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14379 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14381 *Zeev Lieber <zeev-l@yahoo.com>*
14383 * Undo an undocumented change introduced in 0.9.6e which caused
14384 repeated calls to OpenSSL_add_all_ciphers() and
14385 OpenSSL_add_all_digests() to be ignored, even after calling
14390 * Change the default configuration reader to deal with last line not
14391 being properly terminated.
14395 * Change X509_NAME_cmp() so it applies the special rules on handling
14396 DN values that are of type PrintableString, as well as RDNs of type
14397 emailAddress where the value has the type ia5String.
14399 *stefank@valicert.com via Richard Levitte*
14401 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14402 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14403 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14404 the bitwise-OR of the two for use by the majority of applications
14405 wanting this behaviour, and update the docs. The documented
14406 behaviour and actual behaviour were inconsistent and had been
14407 changing anyway, so this is more a bug-fix than a behavioural
14410 *Geoff Thorpe, diagnosed by Nadav Har'El*
14412 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14413 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14417 * Fix initialization code race conditions in
14418 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14419 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14420 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14421 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14422 ssl2_get_cipher_by_char(),
14423 ssl3_get_cipher_by_char().
14425 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14427 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14428 the cached sessions are flushed, as the remove_cb() might use ex_data
14429 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14430 (see [openssl.org #212]).
14432 *Geoff Thorpe, Lutz Jaenicke*
14434 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14435 length, instead of the encoding length to d2i_ASN1_OBJECT.
14439 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14441 * [In 0.9.6g-engine release:]
14442 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14444 *Lynn Gazis <lgazis@rainbow.com>*
14446 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14448 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14449 and get fix the header length calculation.
14450 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14451 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14453 * Use proper error handling instead of 'assertions' in buffer
14454 overflow checks added in 0.9.6e. This prevents DoS (the
14455 assertions could call abort()).
14457 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14459 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14461 * Add various sanity checks to asn1_get_length() to reject
14462 the ASN1 length bytes if they exceed sizeof(long), will appear
14463 negative or the content length exceeds the length of the
14466 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14468 * Fix cipher selection routines: ciphers without encryption had no flags
14469 for the cipher strength set and where therefore not handled correctly
14470 by the selection routines (PR #130).
14474 * Fix EVP_dsa_sha macro.
14479 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14480 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14481 that was added in OpenSSL 0.9.6d.
14483 As the countermeasure turned out to be incompatible with some
14484 broken SSL implementations, the new option is part of SSL_OP_ALL.
14485 SSL_OP_ALL is usually employed when compatibility with weird SSL
14486 implementations is desired (e.g. '-bugs' option to 's_client' and
14487 's_server'), so the new option is automatically set in many
14492 * Changes in security patch:
14494 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14495 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14496 Air Force Materiel Command, USAF, under agreement number
14499 * Add various sanity checks to asn1_get_length() to reject
14500 the ASN1 length bytes if they exceed sizeof(long), will appear
14501 negative or the content length exceeds the length of the
14502 supplied buffer. ([CVE-2002-0659])
14504 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14506 * Assertions for various potential buffer overflows, not known to
14507 happen in practice.
14509 *Ben Laurie (CHATS)*
14511 * Various temporary buffers to hold ASCII versions of integers were
14512 too small for 64 bit platforms. ([CVE-2002-0655])
14513 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14515 * Remote buffer overflow in SSL3 protocol - an attacker could
14516 supply an oversized session ID to a client. ([CVE-2002-0656])
14518 *Ben Laurie (CHATS)*
14520 * Remote buffer overflow in SSL2 protocol - an attacker could
14521 supply an oversized client master key. ([CVE-2002-0656])
14523 *Ben Laurie (CHATS)*
14525 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14527 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14528 encoded as NULL) with id-dsa-with-sha1.
14530 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14532 * Check various `X509_...()` return values in `apps/req.c`.
14534 *Nils Larsch <nla@trustcenter.de>*
14536 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14537 an end-of-file condition would erroneously be flagged, when the CRLF
14538 was just at the end of a processed block. The bug was discovered when
14539 processing data through a buffering memory BIO handing the data to a
14540 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14541 <ptsekov@syntrex.com> and Nedelcho Stanev.
14545 * Implement a countermeasure against a vulnerability recently found
14546 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14547 before application data chunks to avoid the use of known IVs
14548 with data potentially chosen by the attacker.
14552 * Fix length checks in ssl3_get_client_hello().
14556 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14557 to prevent ssl3_read_internal() from incorrectly assuming that
14558 ssl3_read_bytes() found application data while handshake
14559 processing was enabled when in fact s->s3->in_read_app_data was
14560 merely automatically cleared during the initial handshake.
14562 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14564 * Fix object definitions for Private and Enterprise: they were not
14565 recognized in their shortname (=lowercase) representation. Extend
14566 obj_dat.pl to issue an error when using undefined keywords instead
14567 of silently ignoring the problem (Svenning Sorensen
14568 <sss@sss.dnsalias.net>).
14572 * Fix DH_generate_parameters() so that it works for 'non-standard'
14573 generators, i.e. generators other than 2 and 5. (Previously, the
14574 code did not properly initialise the 'add' and 'rem' values to
14575 BN_generate_prime().)
14577 In the new general case, we do not insist that 'generator' is
14578 actually a primitive root: This requirement is rather pointless;
14579 a generator of the order-q subgroup is just as good, if not
14584 * Map new X509 verification errors to alerts. Discovered and submitted by
14585 Tom Wu <tom@arcot.com>.
14589 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14590 returning non-zero before the data has been completely received
14591 when using non-blocking I/O.
14593 *Bodo Moeller; problem pointed out by John Hughes*
14595 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14597 *Ben Laurie, Lutz Jaenicke*
14599 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14600 Yoram Zahavi <YoramZ@gilian.com>).
14604 * Add information about CygWin 1.3 and on, and preserve proper
14605 configuration for the versions before that.
14607 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14609 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14610 check whether we deal with a copy of a session and do not delete from
14611 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14612 <izhar@checkpoint.com>.
14616 * Do not store session data into the internal session cache, if it
14617 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14618 flag is set). Proposed by Aslam <aslam@funk.com>.
14622 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14627 * [In 0.9.6d-engine release:]
14628 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14630 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14632 * Add the configuration target linux-s390x.
14634 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14636 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14637 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14638 variable as an indication that a ClientHello message has been
14639 received. As the flag value will be lost between multiple
14640 invocations of ssl3_accept when using non-blocking I/O, the
14641 function may not be aware that a handshake has actually taken
14642 place, thus preventing a new session from being added to the
14645 To avoid this problem, we now set s->new_session to 2 instead of
14646 using a local variable.
14648 *Lutz Jaenicke, Bodo Moeller*
14650 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14651 if the SSL_R_LENGTH_MISMATCH error is detected.
14653 *Geoff Thorpe, Bodo Moeller*
14655 * New 'shared_ldflag' column in Configure platform table.
14659 * Fix EVP_CIPHER_mode macro.
14661 *"Dan S. Camper" <dan@bti.net>*
14663 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14664 type, we must throw them away by setting rr->length to 0.
14666 *D P Chang <dpc@qualys.com>*
14668 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14670 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14671 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14672 worked incorrectly for those cases where range = `10..._2` and
14673 `3*range` is two bits longer than range.)
14677 * Only add signing time to PKCS7 structures if it is not already
14682 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14683 OBJ_ld_ce should be OBJ_id_ce.
14684 Also some ip-pda OIDs in crypto/objects/objects.txt were
14685 incorrect (cf. RFC 3039).
14687 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14689 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14690 returns early because it has nothing to do.
14692 *Andy Schneider <andy.schneider@bjss.co.uk>*
14694 * [In 0.9.6c-engine release:]
14695 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14697 *Andy Schneider <andy.schneider@bjss.co.uk>*
14699 * [In 0.9.6c-engine release:]
14700 Add support for Cryptographic Appliance's keyserver technology.
14701 (Use engine 'keyclient')
14703 *Cryptographic Appliances and Geoff Thorpe*
14705 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14706 is called via tools/c89.sh because arguments have to be
14707 rearranged (all '-L' options must appear before the first object
14710 *Richard Shapiro <rshapiro@abinitio.com>*
14712 * [In 0.9.6c-engine release:]
14713 Add support for Broadcom crypto accelerator cards, backported
14716 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14718 * [In 0.9.6c-engine release:]
14719 Add support for SureWare crypto accelerator cards from
14720 Baltimore Technologies. (Use engine 'sureware')
14722 *Baltimore Technologies and Mark Cox*
14724 * [In 0.9.6c-engine release:]
14725 Add support for crypto accelerator cards from Accelerated
14726 Encryption Processing, www.aep.ie. (Use engine 'aep')
14728 *AEP Inc. and Mark Cox*
14730 * Add a configuration entry for gcc on UnixWare.
14732 *Gary Benson <gbenson@redhat.com>*
14734 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14735 messages are stored in a single piece (fixed-length part and
14736 variable-length part combined) and fix various bugs found on the way.
14740 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14741 instead. BIO_gethostbyname() does not know what timeouts are
14742 appropriate, so entries would stay in cache even when they have
14744 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14746 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14747 faced with a pathologically small ClientHello fragment that does
14748 not contain client_version: Instead of aborting with an error,
14749 simply choose the highest available protocol version (i.e.,
14750 TLS 1.0 unless it is disabled). In practice, ClientHello
14751 messages are never sent like this, but this change gives us
14752 strictly correct behaviour at least for TLS.
14756 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14757 never resets s->method to s->ctx->method when called from within
14758 one of the SSL handshake functions.
14760 *Bodo Moeller; problem pointed out by Niko Baric*
14762 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14763 (sent using the client's version number) if client_version is
14764 smaller than the protocol version in use. Also change
14765 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14766 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14767 the client will at least see that alert.
14771 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14776 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14777 client receives HelloRequest while in a handshake.
14779 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14781 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14782 should end in 'break', not 'goto end' which circumvents various
14783 cleanups done in state SSL_ST_OK. But session related stuff
14784 must be disabled for SSL_ST_OK in the case that we just sent a
14787 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14788 before just sending a HelloRequest.
14790 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14792 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14793 reveal whether illegal block cipher padding was found or a MAC
14794 verification error occurred. (Neither SSLerr() codes nor alerts
14795 are directly visible to potential attackers, but the information
14796 may leak via logfiles.)
14798 Similar changes are not required for the SSL 2.0 implementation
14799 because the number of padding bytes is sent in clear for SSL 2.0,
14800 and the extra bytes are just ignored. However ssl/s2_pkt.c
14801 failed to verify that the purported number of padding bytes is in
14806 * Add OpenUNIX-8 support including shared libraries
14807 (Boyd Lynn Gerber <gerberb@zenez.com>).
14811 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14812 'wristwatch attack' using huge encoding parameters (cf.
14813 James H. Manger's CRYPTO 2001 paper). Note that the
14814 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14815 encoding parameters and hence was not vulnerable.
14819 * BN_sqr() bug fix.
14821 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14823 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14824 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14825 followed by modular reduction.
14827 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14829 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14830 equivalent based on BN_pseudo_rand() instead of BN_rand().
14834 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14835 This function was broken, as the check for a new client hello message
14836 to handle SGC did not allow these large messages.
14837 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14841 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14845 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14846 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14850 * Rework the configuration and shared library support for Tru64 Unix.
14851 The configuration part makes use of modern compiler features and
14852 still retains old compiler behavior for those that run older versions
14853 of the OS. The shared library support part includes a variant that
14854 uses the RPATH feature, and is available through the special
14855 configuration target "alpha-cc-rpath", which will never be selected
14858 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14860 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14861 with the same message size as in ssl3_get_certificate_request().
14862 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14863 messages might inadvertently be reject as too long.
14865 *Petr Lampa <lampa@fee.vutbr.cz>*
14867 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14871 * Modified SSL library such that the verify_callback that has been set
14872 specifically for an SSL object with SSL_set_verify() is actually being
14873 used. Before the change, a verify_callback set with this function was
14874 ignored and the verify_callback() set in the SSL_CTX at the time of
14875 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14876 to allow the necessary settings.
14880 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14881 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14882 done automatically (in contradiction to the requirements of the C
14883 standard). This made problems when used from OpenSSH.
14887 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14888 dh->length and always used
14890 BN_rand_range(priv_key, dh->p).
14892 BN_rand_range() is not necessary for Diffie-Hellman, and this
14893 specific range makes Diffie-Hellman unnecessarily inefficient if
14894 dh->length (recommended exponent length) is much smaller than the
14895 length of dh->p. We could use BN_rand_range() if the order of
14896 the subgroup was stored in the DH structure, but we only have
14901 BN_rand(priv_key, l, ...)
14903 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14910 RSA_eay_public_encrypt
14911 RSA_eay_private_decrypt
14912 RSA_eay_private_encrypt (signing)
14913 RSA_eay_public_decrypt (signature verification)
14915 (default implementations for RSA_public_encrypt,
14916 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14917 always reject numbers >= n.
14921 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14922 to synchronize access to 'locking_thread'. This is necessary on
14923 systems where access to 'locking_thread' (an 'unsigned long'
14924 variable) is not atomic.
14928 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14929 *before* setting the 'crypto_lock_rand' flag. The previous code had
14930 a race condition if 0 is a valid thread ID.
14932 *Travis Vitek <vitek@roguewave.com>*
14934 * Add support for shared libraries under Irix.
14936 *Albert Chin-A-Young <china@thewrittenword.com>*
14938 * Add configuration option to build on Linux on both big-endian and
14939 little-endian MIPS.
14941 *Ralf Baechle <ralf@uni-koblenz.de>*
14943 * Add the possibility to create shared libraries on HP-UX.
14947 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14949 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14950 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14951 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14952 PRNG state recovery was possible based on the output of
14953 one PRNG request appropriately sized to gain knowledge on
14954 'md' followed by enough consecutive 1-byte PRNG requests
14955 to traverse all of 'state'.
14957 1. When updating 'md_local' (the current thread's copy of 'md')
14958 during PRNG output generation, hash all of the previous
14959 'md_local' value, not just the half used for PRNG output.
14961 2. Make the number of bytes from 'state' included into the hash
14962 independent from the number of PRNG bytes requested.
14964 The first measure alone would be sufficient to avoid
14965 Markku-Juhani's attack. (Actually it had never occurred
14966 to me that the half of 'md_local' used for chaining was the
14967 half from which PRNG output bytes were taken -- I had always
14968 assumed that the secret half would be used.) The second
14969 measure makes sure that additional data from 'state' is never
14970 mixed into 'md_local' in small portions; this heuristically
14971 further strengthens the PRNG.
14975 * Fix crypto/bn/asm/mips3.s.
14979 * When only the key is given to "enc", the IV is undefined. Print out
14980 an error message in this case.
14984 * Handle special case when X509_NAME is empty in X509 printing routines.
14988 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14989 positive and less than q.
14993 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14994 used: it isn't thread safe and the add_lock_callback should handle
14997 *Paul Rose <Paul.Rose@bridge.com>*
14999 * Verify that incoming data obeys the block size in
15000 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15006 *Ulf Möller, Bodo Möller*
15008 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15009 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15010 when fixing the server behaviour for backwards-compatible 'client
15011 hello' messages. (Note that the attack is impractical against
15012 SSL 3.0 and TLS 1.0 anyway because length and version checking
15013 means that the probability of guessing a valid ciphertext is
15014 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15017 Before 0.9.5, the countermeasure (hide the error by generating a
15018 random 'decryption result') did not work properly because
15019 ERR_clear_error() was missing, meaning that SSL_get_error() would
15020 detect the supposedly ignored error.
15022 Both problems are now fixed.
15026 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15027 (previously it was 1024).
15031 * Fix for compatibility mode trust settings: ignore trust settings
15032 unless some valid trust or reject settings are present.
15036 * Fix for blowfish EVP: its a variable length cipher.
15040 * Fix various bugs related to DSA S/MIME verification. Handle missing
15041 parameters in DSA public key structures and return an error in the
15042 DSA routines if parameters are absent.
15046 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15047 in the current directory if neither $RANDFILE nor $HOME was set.
15048 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15049 caused some confusion to Windows users who haven't defined $HOME.
15050 Thus RAND_file_name() is changed again: e_os.h can define a
15051 DEFAULT_HOME, which will be used if $HOME is not set.
15052 For Windows, we use "C:"; on other platforms, we still require
15053 environment variables.
15055 * Move 'if (!initialized) RAND_poll()' into regions protected by
15056 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15057 having multiple threads call RAND_poll() concurrently.
15061 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15062 combination of a flag and a thread ID variable.
15063 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15064 flag), *other* threads can enter ssleay_add_bytes without obeying
15065 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15066 that they do not hold after the first thread unsets add_do_not_lock).
15070 * Change bctest again: '-x' expressions are not available in all
15071 versions of 'test'.
15075 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15077 * Fix a couple of memory leaks in PKCS7_dataDecode()
15079 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15081 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15082 the default extension for executables, if any. Also, make the perl
15083 scripts that use symlink() to test if it really exists and use "cp"
15084 if it doesn't. All this made OpenSSL compilable and installable in
15089 * Fix for asn1_GetSequence() for indefinite length constructed data.
15090 If SEQUENCE is length is indefinite just set c->slen to the total
15091 amount of data available.
15093 *Steve Henson, reported by shige@FreeBSD.org*
15095 *This change does not apply to 0.9.7.*
15097 * Change bctest to avoid here-documents inside command substitution
15098 (workaround for FreeBSD /bin/sh bug).
15099 For compatibility with Ultrix, avoid shell functions (introduced
15100 in the bctest version that searches along $PATH).
15104 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15105 with des_encrypt() defined on some operating systems, like Solaris
15110 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15111 On the Importance of Eliminating Errors in Cryptographic
15112 Computations, J. Cryptology 14 (2001) 2, 101-119,
15113 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15117 * MIPS assembler BIGNUM division bug fix.
15121 * Disabled incorrect Alpha assembler code.
15125 * Fix PKCS#7 decode routines so they correctly update the length
15126 after reading an EOC for the EXPLICIT tag.
15130 *This change does not apply to 0.9.7.*
15132 * Fix bug in PKCS#12 key generation routines. This was triggered
15133 if a 3DES key was generated with a 0 initial byte. Include
15134 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15135 (but broken) behaviour.
15139 * Enhance bctest to search for a working bc along $PATH and print
15142 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15144 * Fix memory leaks in err.c: free err_data string if necessary;
15145 don't write to the wrong index in ERR_set_error_data.
15149 * Implement ssl23_peek (analogous to ssl23_read), which previously
15154 * Replace rdtsc with `_emit` statements for VC++ version 5.
15156 *Jeremy Cooper <jeremy@baymoo.org>*
15158 * Make it possible to reuse SSLv2 sessions.
15162 * In copy_email() check for >= 0 as a return value for
15163 X509_NAME_get_index_by_NID() since 0 is a valid index.
15165 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15167 * Avoid coredump with unsupported or invalid public keys by checking if
15168 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15169 PKCS7_verify() fails with non detached data.
15173 * Don't use getenv in library functions when run as setuid/setgid.
15174 New function OPENSSL_issetugid().
15178 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15179 due to incorrect handling of multi-threading:
15181 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15183 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15185 3. Count how many times MemCheck_off() has been called so that
15186 nested use can be treated correctly. This also avoids
15187 inband-signalling in the previous code (which relied on the
15188 assumption that thread ID 0 is impossible).
15192 * Add "-rand" option also to s_client and s_server.
15196 * Fix CPU detection on Irix 6.x.
15197 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15198 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15200 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15205 *This change does not apply to 0.9.7.*
15207 * Use the cached encoding of an X509_NAME structure rather than
15208 copying it. This is apparently the reason for the libsafe "errors"
15209 but the code is actually correct.
15213 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15214 Bleichenbacher's DSA attack.
15215 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15216 to be set and top=0 forces the highest bit to be set; top=-1 is new
15217 and leaves the highest bit random.
15219 *Ulf Moeller, Bodo Moeller*
15221 * In the `NCONF_...`-based implementations for `CONF_...` queries
15222 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15223 a temporary CONF structure with the data component set to NULL
15224 (which gives segmentation faults in lh_retrieve).
15225 Instead, use NULL for the CONF pointer in CONF_get_string and
15226 CONF_get_number (which may use environment variables) and directly
15227 return NULL from CONF_get_section.
15231 * Fix potential buffer overrun for EBCDIC.
15235 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15236 keyUsage if basicConstraints absent for a CA.
15240 * Make SMIME_write_PKCS7() write mail header values with a format that
15241 is more generally accepted (no spaces before the semicolon), since
15242 some programs can't parse those values properly otherwise. Also make
15243 sure BIO's that break lines after each write do not create invalid
15248 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15249 macros previously used would not encode an empty SEQUENCE OF
15250 and break the signature.
15254 *This change does not apply to 0.9.7.*
15256 * Zero the premaster secret after deriving the master secret in
15261 * Add some EVP_add_digest_alias registrations (as found in
15262 OpenSSL_add_all_digests()) to SSL_library_init()
15263 aka OpenSSL_add_ssl_algorithms(). This provides improved
15264 compatibility with peers using X.509 certificates
15265 with unconventional AlgorithmIdentifier OIDs.
15269 * Fix for Irix with NO_ASM.
15271 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15273 * ./config script fixes.
15275 *Ulf Moeller, Richard Levitte*
15277 * Fix 'openssl passwd -1'.
15281 * Change PKCS12_key_gen_asc() so it can cope with non null
15282 terminated strings whose length is passed in the passlen
15283 parameter, for example from PEM callbacks. This was done
15284 by adding an extra length parameter to asc2uni().
15286 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15288 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15289 call failed, free the DSA structure.
15293 * Fix to uni2asc() to cope with zero length Unicode strings.
15294 These are present in some PKCS#12 files.
15298 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15299 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15300 when writing a 32767 byte record.
15302 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15304 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15305 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15307 (RSA objects have a reference count access to which is protected
15308 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15309 so they are meant to be shared between threads.)
15310 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15311 "Reddie, Steven" <Steven.Reddie@ca.com>*
15313 * Fix a deadlock in CRYPTO_mem_leaks().
15317 * Use better test patterns in bntest.
15321 * rand_win.c fix for Borland C.
15325 * BN_rshift bugfix for n == 0.
15329 * Add a 'bctest' script that checks for some known 'bc' bugs
15330 so that 'make test' does not abort just because 'bc' is broken.
15334 * Store verify_result within SSL_SESSION also for client side to
15335 avoid potential security hole. (Re-used sessions on the client side
15336 always resulted in verify_result==X509_V_OK, not using the original
15337 result of the server certificate verification.)
15341 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15342 SSL3_RT_APPLICATION_DATA, return 0.
15343 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15348 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15349 releases, have been re-implemented by renaming the previous
15350 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15351 and ssl3_read_internal, respectively, and adding 'peek' parameters
15352 to them. The new ssl[23]_{read,peek} functions are calls to
15353 ssl[23]_read_internal with the 'peek' flag set appropriately.
15354 A 'peek' parameter has also been added to ssl3_read_bytes, which
15355 does the actual work for ssl3_read_internal.
15359 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15360 the method-specific "init()" handler. Also clean up ex_data after
15361 calling the method-specific "finish()" handler. Previously, this was
15362 happening the other way round.
15366 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15367 The previous value, 12, was not always sufficient for BN_mod_exp().
15371 * Make sure that shared libraries get the internal name engine with
15372 the full version number and not just 0. This should mark the
15373 shared libraries as not backward compatible. Of course, this should
15374 be changed again when we can guarantee backward binary compatibility.
15378 * Fix typo in get_cert_by_subject() in by_dir.c
15380 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15382 * Rework the system to generate shared libraries:
15384 - Make note of the expected extension for the shared libraries and
15385 if there is a need for symbolic links from for example libcrypto.so.0
15386 to libcrypto.so.0.9.7. There is extended info in Configure for
15389 - Make as few rebuilds of the shared libraries as possible.
15391 - Still avoid linking the OpenSSL programs with the shared libraries.
15393 - When installing, install the shared libraries separately from the
15398 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15400 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15401 and not in SSL_clear because the latter is also used by the
15402 accept/connect functions; previously, the settings made by
15403 SSL_set_read_ahead would be lost during the handshake.
15405 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15407 * Correct util/mkdef.pl to be selective about disabled algorithms.
15408 Previously, it would create entries for disabled algorithms no
15413 * Added several new manual pages for SSL_* function.
15417 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15419 * In ssl23_get_client_hello, generate an error message when faced
15420 with an initial SSL 3.0/TLS record that is too small to contain the
15421 first two bytes of the ClientHello message, i.e. client_version.
15422 (Note that this is a pathologic case that probably has never happened
15423 in real life.) The previous approach was to use the version number
15424 from the record header as a substitute; but our protocol choice
15425 should not depend on that one because it is not authenticated
15426 by the Finished messages.
15430 * More robust randomness gathering functions for Windows.
15432 *Jeffrey Altman <jaltman@columbia.edu>*
15434 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15435 not set then we don't setup the error code for issuer check errors
15436 to avoid possibly overwriting other errors which the callback does
15437 handle. If an application does set the flag then we assume it knows
15438 what it is doing and can handle the new informational codes
15443 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15444 a general "ANY" type, as such it should be able to decode anything
15445 including tagged types. However it didn't check the class so it would
15446 wrongly interpret tagged types in the same way as their universal
15447 counterpart and unknown types were just rejected. Changed so that the
15448 tagged and unknown types are handled in the same way as a SEQUENCE:
15449 that is the encoding is stored intact. There is also a new type
15450 "V_ASN1_OTHER" which is used when the class is not universal, in this
15451 case we have no idea what the actual type is so we just lump them all
15456 * On VMS, stdout may very well lead to a file that is written to
15457 in a record-oriented fashion. That means that every write() will
15458 write a separate record, which will be read separately by the
15459 programs trying to read from it. This can be very confusing.
15461 The solution is to put a BIO filter in the way that will buffer
15462 text until a linefeed is reached, and then write everything a
15463 line at a time, so every record written will be an actual line,
15464 not chunks of lines and not (usually doesn't happen, but I've
15465 seen it once) several lines in one record. BIO_f_linebuffer() is
15468 Currently, it's a VMS-only method, because that's where it has
15469 been tested well enough.
15473 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15474 it can return incorrect results.
15475 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15476 but it was in 0.9.6-beta[12].)
15480 * Disable the check for content being present when verifying detached
15481 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15482 include zero length content when signing messages.
15486 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15487 BIO_ctrl (for BIO pairs).
15491 * Add DSO method for VMS.
15495 * Bug fix: Montgomery multiplication could produce results with the
15500 * Add RPM specification openssl.spec and modify it to build three
15501 packages. The default package contains applications, application
15502 documentation and run-time libraries. The devel package contains
15503 include files, static libraries and function documentation. The
15504 doc package contains the contents of the doc directory. The original
15505 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15509 * Add a large number of documentation files for many SSL routines.
15511 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15513 * Add a configuration entry for Sony News 4.
15515 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15517 * Don't set the two most significant bits to one when generating a
15518 random number < q in the DSA library.
15522 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15523 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15524 the underlying transport is blocking) if a handshake took place.
15525 (The default behaviour is needed by applications such as s_client
15526 and s_server that use select() to determine when to use SSL_read;
15527 but for applications that know in advance when to expect data, it
15528 just makes things more complicated.)
15532 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15537 * Add a few more EBCDIC conditionals that make `req` and `x509`
15538 work better on such systems.
15540 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15542 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15543 Update PKCS12_parse() so it copies the friendlyName and the
15544 keyid to the certificates aux info.
15548 * Fix bug in PKCS7_verify() which caused an infinite loop
15549 if there was more than one signature.
15551 *Sven Uszpelkat <su@celocom.de>*
15553 * Major change in util/mkdef.pl to include extra information
15554 about each symbol, as well as presenting variables as well
15555 as functions. This change means that there's n more need
15556 to rebuild the .num files when some algorithms are excluded.
15560 * Allow the verify time to be set by an application,
15561 rather than always using the current time.
15565 * Phase 2 verify code reorganisation. The certificate
15566 verify code now looks up an issuer certificate by a
15567 number of criteria: subject name, authority key id
15568 and key usage. It also verifies self signed certificates
15569 by the same criteria. The main comparison function is
15570 X509_check_issued() which performs these checks.
15572 Lot of changes were necessary in order to support this
15573 without completely rewriting the lookup code.
15575 Authority and subject key identifier are now cached.
15577 The LHASH 'certs' is X509_STORE has now been replaced
15578 by a STACK_OF(X509_OBJECT). This is mainly because an
15579 LHASH can't store or retrieve multiple objects with
15580 the same hash value.
15582 As a result various functions (which were all internal
15583 use only) have changed to handle the new X509_STORE
15584 structure. This will break anything that messed round
15585 with X509_STORE internally.
15587 The functions X509_STORE_add_cert() now checks for an
15588 exact match, rather than just subject name.
15590 The X509_STORE API doesn't directly support the retrieval
15591 of multiple certificates matching a given criteria, however
15592 this can be worked round by performing a lookup first
15593 (which will fill the cache with candidate certificates)
15594 and then examining the cache for matches. This is probably
15595 the best we can do without throwing out X509_LOOKUP
15596 entirely (maybe later...).
15598 The X509_VERIFY_CTX structure has been enhanced considerably.
15600 All certificate lookup operations now go via a get_issuer()
15601 callback. Although this currently uses an X509_STORE it
15602 can be replaced by custom lookups. This is a simple way
15603 to bypass the X509_STORE hackery necessary to make this
15604 work and makes it possible to use more efficient techniques
15605 in future. A very simple version which uses a simple
15606 STACK for its trusted certificate store is also provided
15607 using X509_STORE_CTX_trusted_stack().
15609 The verify_cb() and verify() callbacks now have equivalents
15610 in the X509_STORE_CTX structure.
15612 X509_STORE_CTX also has a 'flags' field which can be used
15613 to customise the verify behaviour.
15617 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15618 excludes S/MIME capabilities.
15622 * When a certificate request is read in keep a copy of the
15623 original encoding of the signed data and use it when outputting
15624 again. Signatures then use the original encoding rather than
15625 a decoded, encoded version which may cause problems if the
15626 request is improperly encoded.
15630 * For consistency with other BIO_puts implementations, call
15631 buffer_write(b, ...) directly in buffer_puts instead of calling
15634 In BIO_puts, increment b->num_write as in BIO_write.
15636 *Peter.Sylvester@EdelWeb.fr*
15638 * Fix BN_mul_word for the case where the word is 0. (We have to use
15639 BN_zero, we may not return a BIGNUM with an array consisting of
15640 words set to zero.)
15644 * Avoid calling abort() from within the library when problems are
15645 detected, except if preprocessor symbols have been defined
15646 (such as REF_CHECK, BN_DEBUG etc.).
15650 * New openssl application 'rsautl'. This utility can be
15651 used for low-level RSA operations. DER public key
15652 BIO/fp routines also added.
15656 * New Configure entry and patches for compiling on QNX 4.
15658 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15660 * A demo state-machine implementation was sponsored by
15661 Nuron (<http://www.nuron.com/>) and is now available in
15662 demos/state_machine.
15666 * New options added to the 'dgst' utility for signature
15667 generation and verification.
15671 * Unrecognized PKCS#7 content types are now handled via a
15672 catch all ASN1_TYPE structure. This allows unsupported
15673 types to be stored as a "blob" and an application can
15674 encode and decode it manually.
15678 * Fix various signed/unsigned issues to make a_strex.c
15679 compile under VC++.
15681 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15683 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15684 length if passed a buffer. ASN1_INTEGER_to_BN failed
15685 if passed a NULL BN and its argument was negative.
15687 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15689 * Modification to PKCS#7 encoding routines to output definite
15690 length encoding. Since currently the whole structures are in
15691 memory there's not real point in using indefinite length
15692 constructed encoding. However if OpenSSL is compiled with
15693 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15697 * Added BIO_vprintf() and BIO_vsnprintf().
15701 * Added more prefixes to parse for in the strings written
15702 through a logging bio, to cover all the levels that are available
15703 through syslog. The prefixes are now:
15705 PANIC, EMERG, EMR => LOG_EMERG
15706 ALERT, ALR => LOG_ALERT
15707 CRIT, CRI => LOG_CRIT
15708 ERROR, ERR => LOG_ERR
15709 WARNING, WARN, WAR => LOG_WARNING
15710 NOTICE, NOTE, NOT => LOG_NOTICE
15711 INFO, INF => LOG_INFO
15712 DEBUG, DBG => LOG_DEBUG
15714 and as before, if none of those prefixes are present at the
15715 beginning of the string, LOG_ERR is chosen.
15717 On Win32, the `LOG_*` levels are mapped according to this:
15719 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15720 LOG_WARNING => EVENTLOG_WARNING_TYPE
15721 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15725 * Made it possible to reconfigure with just the configuration
15726 argument "reconf" or "reconfigure". The command line arguments
15727 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15728 and are retrieved from there when reconfiguring.
15734 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15736 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15740 * The obj_dat.pl script was messing up the sorting of object
15741 names. The reason was that it compared the quoted version
15742 of strings as a result "OCSP" > "OCSP Signing" because
15743 " > SPACE. Changed script to store unquoted versions of
15744 names and add quotes on output. It was also omitting some
15745 names from the lookup table if they were given a default
15746 value (that is if SN is missing it is given the same
15747 value as LN and vice versa), these are now added on the
15748 grounds that if an object has a name we should be able to
15749 look it up. Finally added warning output when duplicate
15750 short or long names are found.
15754 * Changes needed for Tandem NSK.
15756 *Scott Uroff <scott@xypro.com>*
15758 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15759 RSA_padding_check_SSLv23(), special padding was never detected
15760 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15761 version rollback attacks was not effective.
15763 In s23_clnt.c, don't use special rollback-attack detection padding
15764 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15765 client; similarly, in s23_srvr.c, don't do the rollback check if
15766 SSL 2.0 is the only protocol enabled in the server.
15770 * Make it possible to get hexdumps of unprintable data with 'openssl
15771 asn1parse'. By implication, the functions ASN1_parse_dump() and
15772 BIO_dump_indent() are added.
15776 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15777 these print out strings and name structures based on various
15778 flags including RFC2253 support and proper handling of
15779 multibyte characters. Added options to the 'x509' utility
15780 to allow the various flags to be set.
15784 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15785 Also change the functions X509_cmp_current_time() and
15786 X509_gmtime_adj() work with an ASN1_TIME structure,
15787 this will enable certificates using GeneralizedTime in validity
15788 dates to be checked.
15792 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15793 negative public key encodings) on by default,
15794 NO_NEG_PUBKEY_BUG can be set to disable it.
15798 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15799 content octets. An i2c_ASN1_OBJECT is unnecessary because
15800 the encoding can be trivially obtained from the structure.
15804 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15805 not read locks (`CRYPTO_r_[un]lock`).
15809 * A first attempt at creating official support for shared
15810 libraries through configuration. I've kept it so the
15811 default is static libraries only, and the OpenSSL programs
15812 are always statically linked for now, but there are
15813 preparations for dynamic linking in place.
15814 This has been tested on Linux and Tru64.
15818 * Randomness polling function for Win9x, as described in:
15819 Peter Gutmann, Software Generation of Practically Strong
15824 * Fix so PRNG is seeded in req if using an already existing
15829 * New options to smime application. -inform and -outform
15830 allow alternative formats for the S/MIME message including
15831 PEM and DER. The -content option allows the content to be
15832 specified separately. This should allow things like Netscape
15833 form signing output easier to verify.
15837 * Fix the ASN1 encoding of tags using the 'long form'.
15841 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15842 STRING types. These convert content octets to and from the
15843 underlying type. The actual tag and length octets are
15844 already assumed to have been read in and checked. These
15845 are needed because all other string types have virtually
15846 identical handling apart from the tag. By having versions
15847 of the ASN1 functions that just operate on content octets
15848 IMPLICIT tagging can be handled properly. It also allows
15849 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15850 and ASN1_INTEGER are identical apart from the tag.
15854 * Change the handling of OID objects as follows:
15856 - New object identifiers are inserted in objects.txt, following
15857 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15858 - objects.pl is used to process obj_mac.num and create a new
15860 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15863 This is currently kind of a hack, and the perl code in objects.pl
15864 isn't very elegant, but it works as I intended. The simplest way
15865 to check that it worked correctly is to look in obj_dat.h and
15866 check the array nid_objs and make sure the objects haven't moved
15867 around (this is important!). Additions are OK, as well as
15868 consistent name changes.
15872 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15876 * Addition of the command line parameter '-rand file' to 'openssl req'.
15877 The given file adds to whatever has already been seeded into the
15878 random pool through the RANDFILE configuration file option or
15879 environment variable, or the default random state file.
15883 * mkstack.pl now sorts each macro group into lexical order.
15884 Previously the output order depended on the order the files
15885 appeared in the directory, resulting in needless rewriting
15890 * Patches to make OpenSSL compile under Win32 again. Mostly
15891 work arounds for the VC++ problem that it treats func() as
15892 func(void). Also stripped out the parts of mkdef.pl that
15893 added extra typesafe functions: these no longer exist.
15897 * Reorganisation of the stack code. The macros are now all
15898 collected in safestack.h . Each macro is defined in terms of
15899 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15900 DEBUG_SAFESTACK is now handled in terms of function casts,
15901 this has the advantage of retaining type safety without the
15902 use of additional functions. If DEBUG_SAFESTACK is not defined
15903 then the non typesafe macros are used instead. Also modified the
15904 mkstack.pl script to handle the new form. Needs testing to see
15905 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15906 the default if no major problems. Similar behaviour for ASN1_SET_OF
15907 and PKCS12_STACK_OF.
15911 * When some versions of IIS use the 'NET' form of private key the
15912 key derivation algorithm is different. Normally MD5(password) is
15913 used as a 128 bit RC4 key. In the modified case
15914 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15915 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15916 as the old Netscape_RSA functions except they have an additional
15917 'sgckey' parameter which uses the modified algorithm. Also added
15918 an -sgckey command line option to the rsa utility. Thanks to
15919 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15920 algorithm to openssl-dev.
15924 * The evp_local.h macros were using 'c.##kname' which resulted in
15925 invalid expansion on some systems (SCO 5.0.5 for example).
15926 Corrected to 'c.kname'.
15928 *Phillip Porch <root@theporch.com>*
15930 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15931 a STACK of email addresses from a certificate or request, these look
15932 in the subject name and the subject alternative name extensions and
15933 omit any duplicate addresses.
15937 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15938 This makes DSA verification about 2 % faster.
15942 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15943 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15944 plus overhead for 1024 bit moduli).
15945 This makes exponentiations about 0.5 % faster for 1024 bit
15946 exponents (as measured by "openssl speed rsa2048").
15950 * Rename memory handling macros to avoid conflicts with other
15952 Malloc => OPENSSL_malloc
15953 Malloc_locked => OPENSSL_malloc_locked
15954 Realloc => OPENSSL_realloc
15955 Free => OPENSSL_free
15959 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15960 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15964 * CygWin32 support.
15966 *John Jarvie <jjarvie@newsguy.com>*
15968 * The type-safe stack code has been rejigged. It is now only compiled
15969 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15970 by default all type-specific stack functions are "#define"d back to
15971 standard stack functions. This results in more streamlined output
15972 but retains the type-safety checking possibilities of the original
15977 * The STACK code has been cleaned up, and certain type declarations
15978 that didn't make a lot of sense have been brought in line. This has
15979 also involved a cleanup of sorts in safestack.h to more correctly
15980 map type-safe stack functions onto their plain stack counterparts.
15981 This work has also resulted in a variety of "const"ifications of
15982 lots of the code, especially `_cmp` operations which should normally
15983 be prototyped with "const" parameters anyway.
15987 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15988 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15989 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15990 where all of 'md' is used each time the PRNG is used, but 'state'
15991 is used only indexed by a cyclic counter. As entropy may not be
15992 well distributed from the beginning, 'md' is important as a
15993 chaining variable. However, the output function chains only half
15994 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15995 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15996 in all of 'state' being rewritten, with the new values depending
15997 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16001 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16002 the handshake is continued after ssl_verify_cert_chain();
16003 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16004 can lead to 'unexplainable' connection aborts later.
16006 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16008 * Major EVP API cipher revision.
16009 Add hooks for extra EVP features. This allows various cipher
16010 parameters to be set in the EVP interface. Support added for variable
16011 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16012 setting of RC2 and RC5 parameters.
16014 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16017 Remove lots of duplicated code from the EVP library. For example *every*
16018 cipher init() function handles the 'iv' in the same way according to the
16019 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16020 for CFB and OFB modes they zero ctx->num.
16022 New functionality allows removal of S/MIME code RC2 hack.
16024 Most of the routines have the same form and so can be declared in terms
16027 By shifting this to the top level EVP_CipherInit() it can be removed from
16028 all individual ciphers. If the cipher wants to handle IVs or keys
16029 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16032 Change lots of functions like EVP_EncryptUpdate() to now return a
16033 value: although software versions of the algorithms cannot fail
16034 any installed hardware versions can.
16038 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16039 this option is set, tolerate broken clients that send the negotiated
16040 protocol version number instead of the requested protocol version
16045 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16046 i.e. non-zero for export ciphersuites, zero otherwise.
16047 Previous versions had this flag inverted, inconsistent with
16048 rsa_tmp_cb (..._TMP_RSA_CB).
16050 *Bodo Moeller; problem reported by Amit Chopra*
16052 * Add missing DSA library text string. Work around for some IIS
16053 key files with invalid SEQUENCE encoding.
16057 * Add a document (doc/standards.txt) that list all kinds of standards
16058 and so on that are implemented in OpenSSL.
16062 * Enhance c_rehash script. Old version would mishandle certificates
16063 with the same subject name hash and wouldn't handle CRLs at all.
16064 Added -fingerprint option to crl utility, to support new c_rehash
16069 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16073 * Fix for SSL server purpose checking. Server checking was
16074 rejecting certificates which had extended key usage present
16075 but no ssl client purpose.
16077 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16079 * Make PKCS#12 code work with no password. The PKCS#12 spec
16080 is a little unclear about how a blank password is handled.
16081 Since the password in encoded as a BMPString with terminating
16082 double NULL a zero length password would end up as just the
16083 double NULL. However no password at all is different and is
16084 handled differently in the PKCS#12 key generation code. NS
16085 treats a blank password as zero length. MSIE treats it as no
16086 password on export: but it will try both on import. We now do
16087 the same: PKCS12_parse() tries zero length and no password if
16088 the password is set to "" or NULL (NULL is now a valid password:
16089 it wasn't before) as does the pkcs12 application.
16093 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16094 perror when PEM_read_bio_X509_REQ fails, the error message must
16095 be obtained from the error queue.
16099 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16100 it in ERR_remove_state if appropriate, and change ERR_get_state
16101 accordingly to avoid race conditions (this is necessary because
16102 thread_hash is no longer constant once set).
16106 * Bugfix for linux-elf makefile.one.
16110 * RSA_get_default_method() will now cause a default
16111 RSA_METHOD to be chosen if one doesn't exist already.
16112 Previously this was only set during a call to RSA_new()
16113 or RSA_new_method(NULL) meaning it was possible for
16114 RSA_get_default_method() to return NULL.
16118 * Added native name translation to the existing DSO code
16119 that will convert (if the flag to do so is set) filenames
16120 that are sufficiently small and have no path information
16121 into a canonical native form. Eg. "blah" converted to
16122 "libblah.so" or "blah.dll" etc.
16126 * New function ERR_error_string_n(e, buf, len) which is like
16127 ERR_error_string(e, buf), but writes at most 'len' bytes
16128 including the 0 terminator. For ERR_error_string_n, 'buf'
16131 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16133 * CONF library reworked to become more general. A new CONF
16134 configuration file reader "class" is implemented as well as a
16135 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16136 old `CONF_*` functions are still there, but are reimplemented to
16137 work in terms of the new functions. Also, a set of functions
16138 to handle the internal storage of the configuration data is
16139 provided to make it easier to write new configuration file
16140 reader "classes" (I can definitely see something reading a
16141 configuration file in XML format, for example), called `_CONF_*`,
16142 or "the configuration storage API"...
16144 The new configuration file reading functions are:
16146 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16147 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16149 NCONF_default, NCONF_WIN32
16151 NCONF_dump_fp, NCONF_dump_bio
16153 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16154 NCONF_new creates a new CONF object. This works in the same way
16155 as other interfaces in OpenSSL, like the BIO interface.
16156 `NCONF_dump_*` dump the internal storage of the configuration file,
16157 which is useful for debugging. All other functions take the same
16158 arguments as the old `CONF_*` functions with the exception of the
16159 first that must be a `CONF *` instead of a `LHASH *`.
16161 To make it easier to use the new classes with the old `CONF_*` functions,
16162 the function CONF_set_default_method is provided.
16166 * Add '-tls1' option to 'openssl ciphers', which was already
16167 mentioned in the documentation but had not been implemented.
16168 (This option is not yet really useful because even the additional
16169 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16173 * Initial DSO code added into libcrypto for letting OpenSSL (and
16174 OpenSSL-based applications) load shared libraries and bind to
16175 them in a portable way.
16177 *Geoff Thorpe, with contributions from Richard Levitte*
16179 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16181 * Make sure _lrotl and _lrotr are only used with MSVC.
16183 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16184 (the default implementation of RAND_status).
16186 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16187 to '-clrext' (= clear extensions), as intended and documented.
16188 *Bodo Moeller; inconsistency pointed out by Michael Attili
16189 <attili@amaxo.com>*
16191 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16192 was larger than the MD block size.
16194 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16196 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16197 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16198 using the passed key: if the passed key was a private key the result
16199 of X509_print(), for example, would be to print out all the private key
16204 * des_quad_cksum() byte order bug fix.
16205 *Ulf Möller, using the problem description in krb4-0.9.7, where
16206 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16208 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16211 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16213 * For easily testing in shell scripts whether some command
16214 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16215 returns with exit code 0 iff no command of the given name is available.
16216 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16217 the output goes to stdout and nothing is printed to stderr.
16218 Additional arguments are always ignored.
16220 Since for each cipher there is a command of the same name,
16221 the 'no-cipher' compilation switches can be tested this way.
16223 ('openssl no-XXX' is not able to detect pseudo-commands such
16224 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16228 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16232 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16233 is set; it will be thrown away anyway because each handshake creates
16235 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16236 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16237 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16238 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16242 * New s_client option -ign_eof: EOF at stdin is ignored, and
16243 'Q' and 'R' lose their special meanings (quit/renegotiate).
16244 This is part of what -quiet does; unlike -quiet, -ign_eof
16245 does not suppress any output.
16249 * Add compatibility options to the purpose and trust code. The
16250 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16251 accepts a certificate or CA, this was the previous behaviour,
16252 with all the associated security issues.
16254 X509_TRUST_COMPAT is the old trust behaviour: only and
16255 automatically trust self signed roots in certificate store. A
16256 new trust setting X509_TRUST_DEFAULT is used to specify that
16257 a purpose has no associated trust setting and it should instead
16258 use the value in the default purpose.
16262 * Fix the PKCS#8 DSA private key code so it decodes keys again
16263 and fix a memory leak.
16267 * In util/mkerr.pl (which implements 'make errors'), preserve
16268 reason strings from the previous version of the .c file, as
16269 the default to have only downcase letters (and digits) in
16270 automatically generated reasons codes is not always appropriate.
16274 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16275 using strerror. Previously, ERR_reason_error_string() returned
16276 library names as reason strings for SYSerr; but SYSerr is a special
16277 case where small numbers are errno values, not library numbers.
16281 * Add '-dsaparam' option to 'openssl dhparam' application. This
16282 converts DSA parameters into DH parameters. (When creating parameters,
16283 DSA_generate_parameters is used.)
16287 * Include 'length' (recommended exponent length) in C code generated
16288 by 'openssl dhparam -C'.
16292 * The second argument to set_label in perlasm was already being used
16293 so couldn't be used as a "file scope" flag. Moved to third argument
16298 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16299 instead of RAND_bytes for encryption IVs and salts.
16303 * Include RAND_status() into RAND_METHOD instead of implementing
16304 it only for md_rand.c Otherwise replacing the PRNG by calling
16305 RAND_set_rand_method would be impossible.
16309 * Don't let DSA_generate_key() enter an infinite loop if the random
16310 number generation fails.
16314 * New 'rand' application for creating pseudo-random output.
16318 * Added configuration support for Linux/IA64
16320 *Rolf Haberrecker <rolf@suse.de>*
16322 * Assembler module support for Mingw32.
16326 * Shared library support for HPUX (in shlib/).
16328 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16330 * Shared library support for Solaris gcc.
16332 *Lutz Behnke <behnke@trustcenter.de>*
16334 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16336 * PKCS7_encrypt() was adding text MIME headers twice because they
16337 were added manually and by SMIME_crlf_copy().
16341 * In bntest.c don't call BN_rand with zero bits argument.
16343 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16345 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16346 case was implemented. This caused BN_div_recp() to fail occasionally.
16350 * Add an optional second argument to the set_label() in the perl
16351 assembly language builder. If this argument exists and is set
16352 to 1 it signals that the assembler should use a symbol whose
16353 scope is the entire file, not just the current function. This
16354 is needed with MASM which uses the format label:: for this scope.
16356 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16358 * Change the ASN1 types so they are typedefs by default. Before
16359 almost all types were #define'd to ASN1_STRING which was causing
16360 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16365 * Change names of new functions to the new get1/get0 naming
16366 convention: After 'get1', the caller owns a reference count
16367 and has to call `..._free`; 'get0' returns a pointer to some
16368 data structure without incrementing reference counters.
16369 (Some of the existing 'get' functions increment a reference
16370 counter, some don't.)
16371 Similarly, 'set1' and 'add1' functions increase reference
16372 counters or duplicate objects.
16376 * Allow for the possibility of temp RSA key generation failure:
16377 the code used to assume it always worked and crashed on failure.
16381 * Fix potential buffer overrun problem in BIO_printf().
16382 *Ulf Möller, using public domain code by Patrick Powell; problem
16383 pointed out by David Sacerdote <das33@cornell.edu>*
16385 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16386 RAND_egd() and RAND_status(). In the command line application,
16387 the EGD socket can be specified like a seed file using RANDFILE
16392 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16393 Some CAs (e.g. Verisign) distribute certificates in this form.
16397 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16398 list to exclude them. This means that no special compilation option
16399 is needed to use anonymous DH: it just needs to be included in the
16404 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16405 EVP_MD_type. The old functionality is available in a new macro called
16406 EVP_MD_md(). Change code that uses it and update docs.
16410 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16411 where the `void *` argument is replaced by a function pointer argument.
16412 Previously `void *` was abused to point to functions, which works on
16413 many platforms, but is not correct. As these functions are usually
16414 called by macros defined in OpenSSL header files, most source code
16415 should work without changes.
16419 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16420 sections with information on -D... compiler switches used for
16421 compiling the library so that applications can see them. To enable
16422 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16423 must be defined. E.g.,
16424 #define OPENSSL_ALGORITHM_DEFINES
16425 #include <openssl/opensslconf.h>
16426 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16428 *Richard Levitte, Ulf and Bodo Möller*
16430 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16435 * Change the 'other' type in certificate aux info to a STACK_OF
16436 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16437 the required ASN1 format: arbitrary types determined by an OID.
16441 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16442 argument to 'req'. This is not because the function is newer or
16443 better than others it just uses the work 'NEW' in the certificate
16444 request header lines. Some software needs this.
16448 * Reorganise password command line arguments: now passwords can be
16449 obtained from various sources. Delete the PEM_cb function and make
16450 it the default behaviour: i.e. if the callback is NULL and the
16451 usrdata argument is not NULL interpret it as a null terminated pass
16452 phrase. If usrdata and the callback are NULL then the pass phrase
16453 is prompted for as usual.
16457 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16458 the support is automatically enabled. The resulting binaries will
16459 autodetect the card and use it if present.
16461 *Ben Laurie and Compaq Inc.*
16463 * Work around for Netscape hang bug. This sends certificate request
16464 and server done in one record. Since this is perfectly legal in the
16465 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16466 the bugs/SSLv3 entry for more info.
16470 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16474 * Add -rand argument to smime and pkcs12 applications and read/write
16479 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16483 * Add command line password options to the remaining applications.
16487 * Bug fix for BN_div_recp() for numerators with an even number of
16492 * More tests in bntest.c, and changed test_bn output.
16496 * ./config recognizes MacOS X now.
16500 * Bug fix for BN_div() when the first words of num and divisor are
16501 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16505 * Add support for various broken PKCS#8 formats, and command line
16506 options to produce them.
16510 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16511 get temporary BIGNUMs from a BN_CTX.
16515 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16520 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16521 include a #define from the old name to the new. The original intent
16522 was that statically linked binaries could for example just call
16523 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16524 link with digests. This never worked because SSLeay_add_all_digests()
16525 and SSLeay_add_all_ciphers() were in the same source file so calling
16526 one would link with the other. They are now in separate source files.
16530 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16534 * Use a less unusual form of the Miller-Rabin primality test (it used
16535 a binary algorithm for exponentiation integrated into the Miller-Rabin
16536 loop, our standard modexp algorithms are faster).
16540 * Support for the EBCDIC character set completed.
16542 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16544 * Source code cleanups: use const where appropriate, eliminate casts,
16545 use `void *` instead of `char *` in lhash.
16549 * Bugfix: ssl3_send_server_key_exchange was not restartable
16550 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16551 this the server could overwrite ephemeral keys that the client
16556 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16557 using 50 iterations of the Rabin-Miller test.
16559 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16560 iterations of the Rabin-Miller test as required by the appendix
16561 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16562 As BN_is_prime_fasttest includes trial division, DSA parameter
16563 generation becomes much faster.
16565 This implies a change for the callback functions in DSA_is_prime
16566 and DSA_generate_parameters: The callback function is called once
16567 for each positive witness in the Rabin-Miller test, not just
16568 occasionally in the inner loop; and the parameters to the
16569 callback function now provide an iteration count for the outer
16570 loop rather than for the current invocation of the inner loop.
16571 DSA_generate_parameters additionally can call the callback
16572 function with an 'iteration count' of -1, meaning that a
16573 candidate has passed the trial division test (when q is generated
16574 from an application-provided seed, trial division is skipped).
16578 * New function BN_is_prime_fasttest that optionally does trial
16579 division before starting the Rabin-Miller test and has
16580 an additional BN_CTX * argument (whereas BN_is_prime always
16581 has to allocate at least one BN_CTX).
16582 'callback(1, -1, cb_arg)' is called when a number has passed the
16583 trial division stage.
16587 * Fix for bug in CRL encoding. The validity dates weren't being handled
16592 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16596 * New function BN_pseudo_rand().
16600 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16601 bignum version of BN_from_montgomery() with the working code from
16602 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16607 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16608 made it impossible to use the same SSL_SESSION data structure in
16609 SSL2 clients in multiple threads.
16613 * The return value of RAND_load_file() no longer counts bytes obtained
16614 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16615 to seed the PRNG (previously an explicit byte count was required).
16617 *Ulf Möller, Bodo Möller*
16619 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16620 used `char *` instead of `void *` and had casts all over the place.
16624 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16628 * Retain source code compatibility for BN_prime_checks macro:
16629 BN_is_prime(..., BN_prime_checks, ...) now uses
16630 BN_prime_checks_for_size to determine the appropriate number of
16631 Rabin-Miller iterations.
16635 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16636 DH_CHECK_P_NOT_SAFE_PRIME.
16637 (Check if this is true? OpenPGP calls them "strong".)
16641 * Merge the functionality of "dh" and "gendh" programs into a new program
16642 "dhparam". The old programs are retained for now but will handle DH keys
16643 (instead of parameters) in future.
16647 * Make the ciphers, s_server and s_client programs check the return values
16648 when a new cipher list is set.
16652 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16653 ciphers. Before when the 56bit ciphers were enabled the sorting was
16656 The syntax for the cipher sorting has been extended to support sorting by
16657 cipher-strength (using the strength_bits hard coded in the tables).
16658 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16660 Fix a bug in the cipher-command parser: when supplying a cipher command
16661 string with an "undefined" symbol (neither command nor alphanumeric
16662 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16663 an error is flagged.
16665 Due to the strength-sorting extension, the code of the
16666 ssl_create_cipher_list() function was completely rearranged. I hope that
16667 the readability was also increased :-)
16669 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16671 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16672 for the first serial number and places 2 in the serial number file. This
16673 avoids problems when the root CA is created with serial number zero and
16674 the first user certificate has the same issuer name and serial number
16679 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16680 the new code. Add documentation for this stuff.
16684 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16685 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16686 structures and behave in an analogous way to the X509v3 functions:
16687 they shouldn't be called directly but wrapper functions should be used
16690 So we also now have some wrapper functions that call the X509at functions
16691 when passed certificate requests. (TO DO: similar things can be done with
16692 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16693 things. Some of these need some d2i or i2d and print functionality
16694 because they handle more complex structures.)
16698 * Add missing #ifndefs that caused missing symbols when building libssl
16699 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16700 NO_RSA in `ssl/s2*.c`.
16702 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16704 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16705 has a return value which indicates the quality of the random data
16706 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16707 error queue. New function RAND_pseudo_bytes() generates output that is
16708 guaranteed to be unique but not unpredictable. RAND_add is like
16709 RAND_seed, but takes an extra argument for an entropy estimate
16710 (RAND_seed always assumes full entropy).
16714 * Do more iterations of Rabin-Miller probable prime test (specifically,
16715 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16716 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16717 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16718 false-positive rate of at most 2^-80 for random input.
16722 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16726 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16727 in the 0.9.5 release), this returns the chain
16728 from an X509_CTX structure with a dup of the stack and all
16729 the X509 reference counts upped: so the stack will exist
16730 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16733 Also make SSL_SESSION_print() print out the verify return
16738 * Add manpage for the pkcs12 command. Also change the default
16739 behaviour so MAC iteration counts are used unless the new
16740 -nomaciter option is used. This improves file security and
16741 only older versions of MSIE (4.0 for example) need it.
16745 * Honor the no-xxx Configure options when creating .DEF files.
16749 * Add PKCS#10 attributes to field table: challengePassword,
16750 unstructuredName and unstructuredAddress. These are taken from
16751 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16752 international characters are used.
16754 More changes to X509_ATTRIBUTE code: allow the setting of types
16755 based on strings. Remove the 'loc' parameter when adding
16756 attributes because these will be a SET OF encoding which is sorted
16761 * Initial changes to the 'req' utility to allow request generation
16762 automation. This will allow an application to just generate a template
16763 file containing all the field values and have req construct the
16766 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16767 used all over the place including certificate requests and PKCS#7
16768 structures. They are currently handled manually where necessary with
16769 some primitive wrappers for PKCS#7. The new functions behave in a
16770 manner analogous to the X509 extension functions: they allow
16771 attributes to be looked up by NID and added.
16773 Later something similar to the X509V3 code would be desirable to
16774 automatically handle the encoding, decoding and printing of the
16775 more complex types. The string types like challengePassword can
16776 be handled by the string table functions.
16778 Also modified the multi byte string table handling. Now there is
16779 a 'global mask' which masks out certain types. The table itself
16780 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16781 is useful when for example there is only one permissible type
16782 (as in countryName) and using the mask might result in no valid
16787 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16788 SSL_get_peer_finished to allow applications to obtain the latest
16789 Finished messages sent to the peer or expected from the peer,
16790 respectively. (SSL_get_peer_finished is usually the Finished message
16791 actually received from the peer, otherwise the protocol will be aborted.)
16793 As the Finished message are message digests of the complete handshake
16794 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16795 be used for external authentication procedures when the authentication
16796 provided by SSL/TLS is not desired or is not enough.
16800 * Enhanced support for Alpha Linux is added. Now ./config checks if
16801 the host supports BWX extension and if Compaq C is present on the
16802 $PATH. Just exploiting of the BWX extension results in 20-30%
16803 performance kick for some algorithms, e.g. DES and RC4 to mention
16804 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16809 * Add support for MS "fast SGC". This is arguably a violation of the
16810 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16811 weak crypto and after checking the certificate is SGC a second one
16812 with strong crypto. MS SGC stops the first handshake after receiving
16813 the server certificate message and sends a second client hello. Since
16814 a server will typically do all the time consuming operations before
16815 expecting any further messages from the client (server key exchange
16816 is the most expensive) there is little difference between the two.
16818 To get OpenSSL to support MS SGC we have to permit a second client
16819 hello message after we have sent server done. In addition we have to
16820 reset the MAC if we do get this second client hello.
16824 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16825 if a DER encoded private key is RSA or DSA traditional format. Changed
16826 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16827 format DER encoded private key. Newer code should use PKCS#8 format which
16828 has the key type encoded in the ASN1 structure. Added DER private key
16829 support to pkcs8 application.
16833 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16834 ciphersuites has been selected (as required by the SSL 3/TLS 1
16835 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16836 is set, we interpret this as a request to violate the specification
16837 (the worst that can happen is a handshake failure, and 'correct'
16838 behaviour would result in a handshake failure anyway).
16842 * In SSL_CTX_add_session, take into account that there might be multiple
16843 SSL_SESSION structures with the same session ID (e.g. when two threads
16844 concurrently obtain them from an external cache).
16845 The internal cache can handle only one SSL_SESSION with a given ID,
16846 so if there's a conflict, we now throw out the old one to achieve
16851 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16852 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16853 some routines that use cipher OIDs: some ciphers do not have OIDs
16854 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16859 * Simplify the trust setting structure and code. Now we just have
16860 two sequences of OIDs for trusted and rejected settings. These will
16861 typically have values the same as the extended key usage extension
16862 and any application specific purposes.
16864 The trust checking code now has a default behaviour: it will just
16865 check for an object with the same NID as the passed id. Functions can
16866 be provided to override either the default behaviour or the behaviour
16867 for a given id. SSL client, server and email already have functions
16868 in place for compatibility: they check the NID and also return "trusted"
16869 if the certificate is self signed.
16873 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16874 traditional format into an EVP_PKEY structure.
16878 * Add a password callback function PEM_cb() which either prompts for
16879 a password if usr_data is NULL or otherwise assumes it is a null
16880 terminated password. Allow passwords to be passed on command line
16881 environment or config files in a few more utilities.
16885 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16886 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16887 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16888 Update documentation.
16892 * Support for ASN1 "NULL" type. This could be handled before by using
16893 ASN1_TYPE but there wasn't any function that would try to read a NULL
16894 and produce an error if it couldn't. For compatibility we also have
16895 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16896 don't allocate anything because they don't need to.
16900 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16903 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16905 * Rebuild of the memory allocation routines used by OpenSSL code and
16906 possibly others as well. The purpose is to make an interface that
16907 provide hooks so anyone can build a separate set of allocation and
16908 deallocation routines to be used by OpenSSL, for example memory
16909 pool implementations, or something else, which was previously hard
16910 since Malloc(), Realloc() and Free() were defined as macros having
16911 the values malloc, realloc and free, respectively (except for Win32
16912 compilations). The same is provided for memory debugging code.
16913 OpenSSL already comes with functionality to find memory leaks, but
16914 this gives people a chance to debug other memory problems.
16916 With these changes, a new set of functions and macros have appeared:
16918 CRYPTO_set_mem_debug_functions() [F]
16919 CRYPTO_get_mem_debug_functions() [F]
16920 CRYPTO_dbg_set_options() [F]
16921 CRYPTO_dbg_get_options() [F]
16922 CRYPTO_malloc_debug_init() [M]
16924 The memory debug functions are NULL by default, unless the library
16925 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16926 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16927 gives the standard debugging functions that come with OpenSSL) or
16928 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16929 provided by the library user) must be used. When the standard
16930 debugging functions are used, CRYPTO_dbg_set_options can be used to
16931 request additional information:
16932 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16933 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16935 Also, things like CRYPTO_set_mem_functions will always give the
16936 expected result (the new set of functions is used for allocation
16937 and deallocation) at all times, regardless of platform and compiler
16940 To finish it up, some functions that were never use in any other
16941 way than through macros have a new API and new semantic:
16943 CRYPTO_dbg_malloc()
16944 CRYPTO_dbg_realloc()
16947 All macros of value have retained their old syntax.
16949 *Richard Levitte and Bodo Moeller*
16951 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16952 ordering of SMIMECapabilities wasn't in "strength order" and there
16953 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16958 * Some ASN1 types with illegal zero length encoding (INTEGER,
16959 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16961 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16963 * Merge in my S/MIME library for OpenSSL. This provides a simple
16964 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16965 functionality to handle multipart/signed properly) and a utility
16966 called 'smime' to call all this stuff. This is based on code I
16967 originally wrote for Celo who have kindly allowed it to be
16968 included in OpenSSL.
16972 * Add variants des_set_key_checked and des_set_key_unchecked of
16973 des_set_key (aka des_key_sched). Global variable des_check_key
16974 decides which of these is called by des_set_key; this way
16975 des_check_key behaves as it always did, but applications and
16976 the library itself, which was buggy for des_check_key == 1,
16977 have a cleaner way to pick the version they need.
16981 * New function PKCS12_newpass() which changes the password of a
16986 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16987 dynamic mix. In both cases the ids can be used as an index into the
16988 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16989 functions so they accept a list of the field values and the
16990 application doesn't need to directly manipulate the X509_TRUST
16995 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17000 * Modify the way the V3 extension code looks up extensions. This now
17001 works in a similar way to the object code: we have some "standard"
17002 extensions in a static table which is searched with OBJ_bsearch()
17003 and the application can add dynamic ones if needed. The file
17004 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17005 updated whenever a new extension is added to the core code and kept
17006 in ext_nid order. There is a simple program 'tabtest.c' which checks
17007 this. New extensions are not added too often so this file can readily
17008 be maintained manually.
17010 There are two big advantages in doing things this way. The extensions
17011 can be looked up immediately and no longer need to be "added" using
17012 X509V3_add_standard_extensions(): this function now does nothing.
17013 Side note: I get *lots* of email saying the extension code doesn't
17014 work because people forget to call this function.
17015 Also no dynamic allocation is done unless new extensions are added:
17016 so if we don't add custom extensions there is no need to call
17017 X509V3_EXT_cleanup().
17021 * Modify enc utility's salting as follows: make salting the default. Add a
17022 magic header, so unsalted files fail gracefully instead of just decrypting
17023 to garbage. This is because not salting is a big security hole, so people
17024 should be discouraged from doing it.
17028 * Fixes and enhancements to the 'x509' utility. It allowed a message
17029 digest to be passed on the command line but it only used this
17030 parameter when signing a certificate. Modified so all relevant
17031 operations are affected by the digest parameter including the
17032 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17033 DSA key was used because it didn't fix the digest.
17037 * Initial certificate chain verify code. Currently tests the untrusted
17038 certificates for consistency with the verify purpose (which is set
17039 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17041 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17042 this is because it will reject chains with invalid extensions whereas
17043 every previous version of OpenSSL and SSLeay made no checks at all.
17045 Trust code: checks the root CA for the relevant trust settings. Trust
17046 settings have an initial value consistent with the verify purpose: e.g.
17047 if the verify purpose is for SSL client use it expects the CA to be
17048 trusted for SSL client use. However the default value can be changed to
17049 permit custom trust settings: one example of this would be to only trust
17050 certificates from a specific "secure" set of CAs.
17052 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17053 which should be used for version portability: especially since the
17054 verify structure is likely to change more often now.
17056 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17057 to set them. If not set then assume SSL clients will verify SSL servers
17060 Two new options to the verify program: -untrusted allows a set of
17061 untrusted certificates to be passed in and -purpose which sets the
17062 intended purpose of the certificate. If a purpose is set then the
17063 new chain verify code is used to check extension consistency.
17067 * Support for the authority information access extension.
17071 * Modify RSA and DSA PEM read routines to transparently handle
17072 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17073 public keys in a format compatible with certificate
17074 SubjectPublicKeyInfo structures. Unfortunately there were already
17075 functions called *_PublicKey_* which used various odd formats so
17076 these are retained for compatibility: however the DSA variants were
17077 never in a public release so they have been deleted. Changed dsa/rsa
17078 utilities to handle the new format: note no releases ever handled public
17079 keys so we should be OK.
17081 The primary motivation for this change is to avoid the same fiasco
17082 that dogs private keys: there are several incompatible private key
17083 formats some of which are standard and some OpenSSL specific and
17084 require various evil hacks to allow partial transparent handling and
17085 even then it doesn't work with DER formats. Given the option anything
17086 other than PKCS#8 should be dumped: but the other formats have to
17087 stay in the name of compatibility.
17089 With public keys and the benefit of hindsight one standard format
17090 is used which works with EVP_PKEY, RSA or DSA structures: though
17091 it clearly returns an error if you try to read the wrong kind of key.
17093 Added a -pubkey option to the 'x509' utility to output the public key.
17094 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17095 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17096 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17097 that do the same as the `EVP_PKEY_assign_*()` except they up the
17098 reference count of the added key (they don't "swallow" the
17103 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17104 CRLs would fail if the file contained no certificates or no CRLs:
17105 added a new function to read in both types and return the number
17106 read: this means that if none are read it will be an error. The
17107 DER versions of the certificate and CRL reader would always fail
17108 because it isn't possible to mix certificates and CRLs in DER format
17109 without choking one or the other routine. Changed this to just read
17110 a certificate: this is the best we can do. Also modified the code
17111 in `apps/verify.c` to take notice of return codes: it was previously
17112 attempting to read in certificates from NULL pointers and ignoring
17113 any errors: this is one reason why the cert and CRL reader seemed
17114 to work. It doesn't check return codes from the default certificate
17115 routines: these may well fail if the certificates aren't installed.
17119 * Code to support otherName option in GeneralName.
17123 * First update to verify code. Change the verify utility
17124 so it warns if it is passed a self signed certificate:
17125 for consistency with the normal behaviour. X509_verify
17126 has been modified to it will now verify a self signed
17127 certificate if *exactly* the same certificate appears
17128 in the store: it was previously impossible to trust a
17129 single self signed certificate. This means that:
17130 openssl verify ss.pem
17131 now gives a warning about a self signed certificate but
17132 openssl verify -CAfile ss.pem ss.pem
17137 * For servers, store verify_result in SSL_SESSION data structure
17138 (and add it to external session representation).
17139 This is needed when client certificate verifications fails,
17140 but an application-provided verification callback (set by
17141 SSL_CTX_set_cert_verify_callback) allows accepting the session
17142 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17143 but returns 1): When the session is reused, we have to set
17144 ssl->verify_result to the appropriate error code to avoid
17147 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17149 * Fix a bug in the new PKCS#7 code: it didn't consider the
17150 case in PKCS7_dataInit() where the signed PKCS7 structure
17151 didn't contain any existing data because it was being created.
17153 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17155 * Add a salt to the key derivation routines in enc.c. This
17156 forms the first 8 bytes of the encrypted file. Also add a
17157 -S option to allow a salt to be input on the command line.
17161 * New function X509_cmp(). Oddly enough there wasn't a function
17162 to compare two certificates. We do this by working out the SHA1
17163 hash and comparing that. X509_cmp() will be needed by the trust
17168 * SSL_get1_session() is like SSL_get_session(), but increments
17169 the reference count in the SSL_SESSION returned.
17171 *Geoff Thorpe <geoff@eu.c2.net>*
17173 * Fix for 'req': it was adding a null to request attributes.
17174 Also change the X509_LOOKUP and X509_INFO code to handle
17175 certificate auxiliary information.
17179 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17184 * Add the possibility to add extra information to the memory leak
17185 detecting output, to form tracebacks, showing from where each
17186 allocation was originated: CRYPTO_push_info("constant string") adds
17187 the string plus current file name and line number to a per-thread
17188 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17189 is like calling CYRPTO_pop_info() until the stack is empty.
17190 Also updated memory leak detection code to be multi-thread-safe.
17194 * Add options -text and -noout to pkcs7 utility and delete the
17195 encryption options which never did anything. Update docs.
17199 * Add options to some of the utilities to allow the pass phrase
17200 to be included on either the command line (not recommended on
17201 OSes like Unix) or read from the environment. Update the
17202 manpages and fix a few bugs.
17206 * Add a few manpages for some of the openssl commands.
17210 * Fix the -revoke option in ca. It was freeing up memory twice,
17211 leaking and not finding already revoked certificates.
17215 * Extensive changes to support certificate auxiliary information.
17216 This involves the use of X509_CERT_AUX structure and X509_AUX
17217 functions. An X509_AUX function such as PEM_read_X509_AUX()
17218 can still read in a certificate file in the usual way but it
17219 will also read in any additional "auxiliary information". By
17220 doing things this way a fair degree of compatibility can be
17221 retained: existing certificates can have this information added
17222 using the new 'x509' options.
17224 Current auxiliary information includes an "alias" and some trust
17225 settings. The trust settings will ultimately be used in enhanced
17226 certificate chain verification routines: currently a certificate
17227 can only be trusted if it is self signed and then it is trusted
17232 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17233 The problem was that one of the replacement routines had not been working
17234 since SSLeay releases. For now the offending routine has been replaced
17235 with non-optimised assembler. Even so, this now gives around 95%
17236 performance improvement for 1024 bit RSA signs.
17240 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17241 handling. Most clients have the effective key size in bits equal to
17242 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17243 A few however don't do this and instead use the size of the decrypted key
17244 to determine the RC2 key length and the AlgorithmIdentifier to determine
17245 the effective key length. In this case the effective key length can still
17246 be 40 bits but the key length can be 168 bits for example. This is fixed
17247 by manually forcing an RC2 key into the EVP_PKEY structure because the
17248 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17249 the key length and effective key length are equal.
17253 * Add a bunch of functions that should simplify the creation of
17254 X509_NAME structures. Now you should be able to do:
17255 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17256 and have it automatically work out the correct field type and fill in
17257 the structures. The more adventurous can try:
17258 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17259 and it will (hopefully) work out the correct multibyte encoding.
17263 * Change the 'req' utility to use the new field handling and multibyte
17264 copy routines. Before the DN field creation was handled in an ad hoc
17265 way in req, ca, and x509 which was rather broken and didn't support
17266 BMPStrings or UTF8Strings. Since some software doesn't implement
17267 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17268 using the dirstring_type option. See the new comment in the default
17269 openssl.cnf for more info.
17273 * Make crypto/rand/md_rand.c more robust:
17274 - Assure unique random numbers after fork().
17275 - Make sure that concurrent threads access the global counter and
17276 md serializably so that we never lose entropy in them
17277 or use exactly the same state in multiple threads.
17278 Access to the large state is not always serializable because
17279 the additional locking could be a performance killer, and
17280 md should be large enough anyway.
17284 * New file `apps/app_rand.c` with commonly needed functionality
17285 for handling the random seed file.
17287 Use the random seed file in some applications that previously did not:
17289 dsaparam -genkey (which also ignored its '-rand' option),
17292 x509 (when signing).
17293 Except on systems with /dev/urandom, it is crucial to have a random
17294 seed file at least for key creation, DSA signing, and for DH exchanges;
17295 for RSA signatures we could do without one.
17297 gendh and gendsa (unlike genrsa) used to read only the first byte
17298 of each file listed in the '-rand' option. The function as previously
17299 found in genrsa is now in app_rand.c and is used by all programs
17300 that support '-rand'.
17304 * In RAND_write_file, use mode 0600 for creating files;
17305 don't just chmod when it may be too late.
17309 * Report an error from X509_STORE_load_locations
17310 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17314 * New function ASN1_mbstring_copy() this copies a string in either
17315 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17316 into an ASN1_STRING type. A mask of permissible types is passed
17317 and it chooses the "minimal" type to use or an error if not type
17322 * Add function equivalents to the various macros in asn1.h. The old
17323 macros are retained with an `M_` prefix. Code inside the library can
17324 use the `M_` macros. External code (including the openssl utility)
17325 should *NOT* in order to be "shared library friendly".
17329 * Add various functions that can check a certificate's extensions
17330 to see if it usable for various purposes such as SSL client,
17331 server or S/MIME and CAs of these types. This is currently
17332 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17333 verification. Also added a -purpose flag to x509 utility to
17334 print out all the purposes.
17338 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17343 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17344 for, obtain and decode and extension and obtain its critical flag.
17345 This allows all the necessary extension code to be handled in a
17346 single function call.
17350 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17351 platforms. See crypto/rc4/rc4_enc.c for further details.
17355 * New -noout option to asn1parse. This causes no output to be produced
17356 its main use is when combined with -strparse and -out to extract data
17357 from a file (which may not be in ASN.1 format).
17361 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17362 when producing the local key id.
17364 *Richard Levitte <levitte@stacken.kth.se>*
17366 * New option -dhparam in s_server. This allows a DH parameter file to be
17367 stated explicitly. If it is not stated then it tries the first server
17368 certificate file. The previous behaviour hard coded the filename
17373 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17374 a public key to be input or output. For example:
17375 openssl rsa -in key.pem -pubout -out pubkey.pem
17376 Also added necessary DSA public key functions to handle this.
17380 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17381 in the message. This was handled by allowing
17382 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17384 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17386 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17387 to the end of the strings whereas this didn't. This would cause problems
17388 if strings read with d2i_ASN1_bytes() were later modified.
17390 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17392 * Fix for base64 decode bug. When a base64 bio reads only one line of
17393 data and it contains EOF it will end up returning an error. This is
17394 caused by input 46 bytes long. The cause is due to the way base64
17395 BIOs find the start of base64 encoded data. They do this by trying a
17396 trial decode on each line until they find one that works. When they
17397 do a flag is set and it starts again knowing it can pass all the
17398 data directly through the decoder. Unfortunately it doesn't reset
17399 the context it uses. This means that if EOF is reached an attempt
17400 is made to pass two EOFs through the context and this causes the
17401 resulting error. This can also cause other problems as well. As is
17402 usual with these problems it takes *ages* to find and the fix is
17403 trivial: move one line.
17405 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17407 * Ugly workaround to get s_client and s_server working under Windows. The
17408 old code wouldn't work because it needed to select() on sockets and the
17409 tty (for keypresses and to see if data could be written). Win32 only
17410 supports select() on sockets so we select() with a 1s timeout on the
17411 sockets and then see if any characters are waiting to be read, if none
17412 are present then we retry, we also assume we can always write data to
17413 the tty. This isn't nice because the code then blocks until we've
17414 received a complete line of data and it is effectively polling the
17415 keyboard at 1s intervals: however it's quite a bit better than not
17416 working at all :-) A dedicated Windows application might handle this
17417 with an event loop for example.
17421 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17422 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17423 will be called when RSA_sign() and RSA_verify() are used. This is useful
17424 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17425 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17426 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17427 This necessitated the support of an extra signature type NID_md5_sha1
17428 for SSL signatures and modifications to the SSL library to use it instead
17429 of calling RSA_public_decrypt() and RSA_private_encrypt().
17433 * Add new -verify -CAfile and -CApath options to the crl program, these
17434 will lookup a CRL issuers certificate and verify the signature in a
17435 similar way to the verify program. Tidy up the crl program so it
17436 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17437 less strict. It will now permit CRL extensions even if it is not
17438 a V2 CRL: this will allow it to tolerate some broken CRLs.
17442 * Initialize all non-automatic variables each time one of the openssl
17443 sub-programs is started (this is necessary as they may be started
17444 multiple times from the "OpenSSL>" prompt).
17446 *Lennart Bang, Bodo Moeller*
17448 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17449 removing all other RSA functionality (this is what NO_RSA does). This
17450 is so (for example) those in the US can disable those operations covered
17451 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17456 * Non-copying interface to BIO pairs.
17457 (still largely untested)
17461 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17462 ASCII string. This was handled independently in various places before.
17466 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17467 UTF8 strings a character at a time.
17471 * Use client_version from client hello to select the protocol
17472 (s23_srvr.c) and for RSA client key exchange verification
17473 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17477 * Add various utility functions to handle SPKACs, these were previously
17478 handled by poking round in the structure internals. Added new function
17479 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17480 print, verify and generate SPKACs. Based on an original idea from
17481 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17485 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17489 * Allow the config file extension section to be overwritten on the
17490 command line. Based on an original idea from Massimiliano Pala
17491 <madwolf@comune.modena.it>. The new option is called -extensions
17492 and can be applied to ca, req and x509. Also -reqexts to override
17493 the request extensions in req and -crlexts to override the crl extensions
17498 * Add new feature to the SPKAC handling in ca. Now you can include
17499 the same field multiple times by preceding it by "XXXX." for example:
17502 this is the same syntax as used in the req config file.
17506 * Allow certificate extensions to be added to certificate requests. These
17507 are specified in a 'req_extensions' option of the req section of the
17508 config file. They can be printed out with the -text option to req but
17509 are otherwise ignored at present.
17513 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17514 data read consists of only the final block it would not decrypted because
17515 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17516 A misplaced 'break' also meant the decrypted final block might not be
17517 copied until the next read.
17521 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17522 a few extra parameters to the DH structure: these will be useful if
17523 for example we want the value of 'q' or implement X9.42 DH.
17527 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17528 provides hooks that allow the default DSA functions or functions on a
17529 "per key" basis to be replaced. This allows hardware acceleration and
17530 hardware key storage to be handled without major modification to the
17531 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17532 associated functions.
17536 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17537 as "read only": it can't be written to and the buffer it points to will
17538 not be freed. Reading from a read only BIO is much more efficient than
17539 a normal memory BIO. This was added because there are several times when
17540 an area of memory needs to be read from a BIO. The previous method was
17541 to create a memory BIO and write the data to it, this results in two
17542 copies of the data and an O(n^2) reading algorithm. There is a new
17543 function BIO_new_mem_buf() which creates a read only memory BIO from
17544 an area of memory. Also modified the PKCS#7 routines to use read only
17549 * Bugfix: ssl23_get_client_hello did not work properly when called in
17550 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17551 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17552 but a retry condition occurred while trying to read the rest.
17556 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17557 NID_pkcs7_encrypted by default: this was wrong since this should almost
17558 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17559 the encrypted data type: this is a more sensible place to put it and it
17560 allows the PKCS#12 code to be tidied up that duplicated this
17565 * Changed obj_dat.pl script so it takes its input and output files on
17566 the command line. This should avoid shell escape redirection problems
17571 * Initial support for certificate extension requests, these are included
17572 in things like Xenroll certificate requests. Included functions to allow
17573 extensions to be obtained and added.
17577 * -crlf option to s_client and s_server for sending newlines as
17578 CRLF (as required by many protocols).
17582 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17584 * Install libRSAglue.a when OpenSSL is built with RSAref.
17586 *Ralf S. Engelschall*
17588 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17590 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17592 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17597 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17598 DH parameters/keys (q is lost during that conversion, but the resulting
17599 DH parameters contain its length).
17601 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17602 much faster than DH_generate_parameters (which creates parameters
17603 where `p = 2*q + 1`), and also the smaller q makes DH computations
17604 much more efficient (160-bit exponentiation instead of 1024-bit
17605 exponentiation); so this provides a convenient way to support DHE
17606 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17607 utter importance to use
17608 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17610 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17611 when such DH parameters are used, because otherwise small subgroup
17612 attacks may become possible!
17616 * Avoid memory leak in i2d_DHparams.
17620 * Allow the -k option to be used more than once in the enc program:
17621 this allows the same encrypted message to be read by multiple recipients.
17625 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17626 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17627 it will always use the numerical form of the OID, even if it has a short
17632 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17633 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17634 otherwise bn_mod_exp was called. In the case of hardware keys for example
17635 no private key components need be present and it might store extra data
17636 in the RSA structure, which cannot be accessed from bn_mod_exp.
17637 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17638 private key operations.
17642 * Added support for SPARC Linux.
17646 * pem_password_cb function type incompatibly changed from
17647 typedef int pem_password_cb(char *buf, int size, int rwflag);
17649 ....(char *buf, int size, int rwflag, void *userdata);
17650 so that applications can pass data to their callbacks:
17651 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17652 additional void * argument, which is just handed through whenever
17653 the password callback is called.
17655 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17657 New function SSL_CTX_set_default_passwd_cb_userdata.
17659 Compatibility note: As many C implementations push function arguments
17660 onto the stack in reverse order, the new library version is likely to
17661 interoperate with programs that have been compiled with the old
17662 pem_password_cb definition (PEM_whatever takes some data that
17663 happens to be on the stack as its last argument, and the callback
17664 just ignores this garbage); but there is no guarantee whatsoever that
17667 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17668 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17669 problems not only on Windows, but also on some Unix platforms.
17670 To avoid problematic command lines, these definitions are now in an
17671 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17672 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17676 * MIPS III/IV assembler module is reimplemented.
17680 * More DES library cleanups: remove references to srand/rand and
17681 delete an unused file.
17685 * Add support for the free Netwide assembler (NASM) under Win32,
17686 since not many people have MASM (ml) and it can be hard to obtain.
17687 This is currently experimental but it seems to work OK and pass all
17688 the tests. Check out INSTALL.W32 for info.
17692 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17693 without temporary keys kept an extra copy of the server key,
17694 and connections with temporary keys did not free everything in case
17699 * New function RSA_check_key and new openssl rsa option -check
17700 for verifying the consistency of RSA keys.
17702 *Ulf Moeller, Bodo Moeller*
17704 * Various changes to make Win32 compile work:
17705 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17706 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17707 comparison" warnings.
17708 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17712 * Add a debugging option to PKCS#5 v2 key generation function: when
17713 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17714 derived keys are printed to stderr.
17718 * Copy the flags in ASN1_STRING_dup().
17720 *Roman E. Pavlov <pre@mo.msk.ru>*
17722 * The x509 application mishandled signing requests containing DSA
17723 keys when the signing key was also DSA and the parameters didn't match.
17725 It was supposed to omit the parameters when they matched the signing key:
17726 the verifying software was then supposed to automatically use the CA's
17727 parameters if they were absent from the end user certificate.
17729 Omitting parameters is no longer recommended. The test was also
17730 the wrong way round! This was probably due to unusual behaviour in
17731 EVP_cmp_parameters() which returns 1 if the parameters match.
17732 This meant that parameters were omitted when they *didn't* match and
17733 the certificate was useless. Certificates signed with 'ca' didn't have
17736 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17738 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17739 The interface is as follows:
17740 Applications can use
17741 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17742 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17743 "off" is now the default.
17744 The library internally uses
17745 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17746 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17747 to disable memory-checking temporarily.
17749 Some inconsistent states that previously were possible (and were
17750 even the default) are now avoided.
17752 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17753 with each memory chunk allocated; this is occasionally more helpful
17754 than just having a counter.
17756 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17758 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17763 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17764 which largely parallels "options", but is for changing API behaviour,
17765 whereas "options" are about protocol behaviour.
17766 Initial "mode" flags are:
17768 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17769 a single record has been written.
17770 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17771 retries use the same buffer location.
17772 (But all of the contents must be
17777 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17780 * Fix problems with no-hmac etc.
17782 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17784 * New functions RSA_get_default_method(), RSA_set_method() and
17785 RSA_get_method(). These allows replacement of RSA_METHODs without having
17786 to mess around with the internals of an RSA structure.
17790 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17791 Also really enable memory leak checks in openssl.c and in some
17794 *Chad C. Mulligan, Bodo Moeller*
17796 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17797 up the length of negative integers. This has now been simplified to just
17798 store the length when it is first determined and use it later, rather
17799 than trying to keep track of where data is copied and updating it to
17801 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17803 * Add a new function PKCS7_signatureVerify. This allows the verification
17804 of a PKCS#7 signature but with the signing certificate passed to the
17805 function itself. This contrasts with PKCS7_dataVerify which assumes the
17806 certificate is present in the PKCS#7 structure. This isn't always the
17807 case: certificates can be omitted from a PKCS#7 structure and be
17808 distributed by "out of band" means (such as a certificate database).
17812 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17813 function prototypes in pem.h, also change util/mkdef.pl to add the
17814 necessary function names.
17818 * mk1mf.pl (used by Windows builds) did not properly read the
17819 options set by Configure in the top level Makefile, and Configure
17820 was not even able to write more than one option correctly.
17821 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17825 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17826 file to be loaded from a BIO or FILE pointer. The BIO version will
17827 for example allow memory BIOs to contain config info.
17831 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17832 Whoever hopes to achieve shared-library compatibility across versions
17833 must use this, not the compile-time macro.
17834 (Exercise 0.9.4: Which is the minimum library version required by
17836 Note: All this applies only to multi-threaded programs, others don't
17841 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17842 through a BIO pair triggered the default case, i.e.
17843 SSLerr(...,SSL_R_UNKNOWN_STATE).
17847 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17848 can use the SSL library even if none of the specific BIOs is
17853 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17854 for the encoded length.
17856 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17858 * Add initial documentation of the X509V3 functions.
17862 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17863 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17864 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17865 secure PKCS#8 private key format with a high iteration count.
17869 * Fix determination of Perl interpreter: A perl or perl5
17870 *directory* in $PATH was also accepted as the interpreter.
17872 *Ralf S. Engelschall*
17874 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17875 wrong with it but it was very old and did things like calling
17876 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17877 unusual formatting.
17881 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17882 to use the new extension code.
17886 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17887 with macros. This should make it easier to change their form, add extra
17888 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17893 * Add to configuration table a new entry that can specify an alternative
17894 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17895 according to Mark Crispin <MRC@Panda.COM>.
17899 * DES CBC did not update the IV. Weird.
17903 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17904 Changing the behaviour of the former might break existing programs --
17905 where IV updating is needed, des_ncbc_encrypt can be used.
17908 * When bntest is run from "make test" it drives bc to check its
17909 calculations, as well as internally checking them. If an internal check
17910 fails, it needs to cause bc to give a non-zero result or make test carries
17911 on without noticing the failure. Fixed.
17915 * DES library cleanups.
17919 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17920 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17921 ciphers. NOTE: although the key derivation function has been verified
17922 against some published test vectors it has not been extensively tested
17923 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17928 * Instead of "mkdir -p", which is not fully portable, use new
17929 Perl script "util/mkdir-p.pl".
17933 * Rewrite the way password based encryption (PBE) is handled. It used to
17934 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17935 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17936 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17937 the 'parameter' field of the AlgorithmIdentifier is passed to the
17938 underlying key generation function so it must do its own ASN1 parsing.
17939 This has also changed the EVP_PBE_CipherInit() function which now has a
17940 'parameter' argument instead of literal salt and iteration count values
17941 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17945 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17946 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17947 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17948 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17949 value was just used as a "magic string" and not used directly its
17950 value doesn't matter.
17954 * Introduce some semblance of const correctness to BN. Shame C doesn't
17959 * "linux-sparc64" configuration (ultrapenguin).
17961 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17962 "linux-sparc" configuration.
17964 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17966 * config now generates no-xxx options for missing ciphers.
17970 * Support the EBCDIC character set (work in progress).
17971 File ebcdic.c not yet included because it has a different license.
17973 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17975 * Support BS2000/OSD-POSIX.
17977 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17979 * Make callbacks for key generation use `void *` instead of `char *`.
17983 * Make S/MIME samples compile (not yet tested).
17987 * Additional typesafe stacks.
17991 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17995 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17997 * New configuration variant "sco5-gcc".
17999 * Updated some demos.
18001 *Sean O Riordain, Wade Scholine*
18003 * Add missing BIO_free at exit of pkcs12 application.
18007 * Fix memory leak in conf.c.
18011 * Updates for Win32 to assembler version of MD5.
18015 * Set #! path to perl in `apps/der_chop` to where we found it
18016 instead of using a fixed path.
18020 * SHA library changes for irix64-mips4-cc.
18024 * Improvements for VMS support.
18028 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18030 * Bignum library bug fix. IRIX 6 passes "make test" now!
18031 This also avoids the problems with SC4.2 and unpatched SC5.
18033 *Andy Polyakov <appro@fy.chalmers.se>*
18035 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18036 These are required because of the typesafe stack would otherwise break
18037 existing code. If old code used a structure member which used to be STACK
18038 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18039 sk_num or sk_value it would produce an error because the num, data members
18040 are not present in STACK_OF. Now it just produces a warning. sk_set
18041 replaces the old method of assigning a value to sk_value
18042 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18043 that does this will no longer work (and should use sk_set instead) but
18044 this could be regarded as a "questionable" behaviour anyway.
18048 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18049 correctly handle encrypted S/MIME data.
18053 * Change type of various DES function arguments from des_cblock
18054 (which means, in function argument declarations, pointer to char)
18055 to des_cblock * (meaning pointer to array with 8 char elements),
18056 which allows the compiler to do more typechecking; it was like
18057 that back in SSLeay, but with lots of ugly casts.
18059 Introduce new type const_des_cblock.
18063 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18064 problems: find RecipientInfo structure that matches recipient certificate
18065 and initialise the ASN1 structures properly based on passed cipher.
18069 * Belatedly make the BN tests actually check the results.
18073 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18074 to and from BNs: it was completely broken. New compilation option
18075 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18076 key elements as negative integers.
18080 * Reorganize and speed up MD5.
18082 *Andy Polyakov <appro@fy.chalmers.se>*
18086 *Richard Levitte <richard@levitte.org>*
18088 * New option -out to asn1parse to allow the parsed structure to be
18089 output to a file. This is most useful when combined with the -strparse
18090 option to examine the output of things like OCTET STRINGS.
18094 * Make SSL library a little more fool-proof by not requiring any longer
18095 that `SSL_set_{accept,connect}_state` be called before
18096 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18097 in many applications because usually everything *appeared* to work as
18098 intended anyway -- now it really works as intended).
18102 * Move openssl.cnf out of lib/.
18106 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18107 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18108 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18110 *Ralf S. Engelschall*
18112 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18113 handle PKCS#7 enveloped data properly.
18115 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18117 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18118 copying pointers. The cert_st handling is changed by this in
18119 various ways (and thus what used to be known as ctx->default_cert
18120 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18121 any longer when s->cert does not give us what we need).
18122 ssl_cert_instantiate becomes obsolete by this change.
18123 As soon as we've got the new code right (possibly it already is?),
18124 we have solved a couple of bugs of the earlier code where s->cert
18125 was used as if it could not have been shared with other SSL structures.
18127 Note that using the SSL API in certain dirty ways now will result
18128 in different behaviour than observed with earlier library versions:
18129 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18130 does not influence s as it used to.
18132 In order to clean up things more thoroughly, inside SSL_SESSION
18133 we don't use CERT any longer, but a new structure SESS_CERT
18134 that holds per-session data (if available); currently, this is
18135 the peer's certificate chain and, for clients, the server's certificate
18136 and temporary key. CERT holds only those values that can have
18137 meaningful defaults in an SSL_CTX.
18141 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18142 from the internal representation. Various PKCS#7 fixes: remove some
18143 evil casts and set the enc_dig_alg field properly based on the signing
18148 * Allow PKCS#12 password to be set from the command line or the
18149 environment. Let 'ca' get its config file name from the environment
18150 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18155 * Allow certificate policies extension to use an IA5STRING for the
18156 organization field. This is contrary to the PKIX definition but
18157 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18162 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18163 without disallowing inline assembler and the like for non-pedantic builds.
18167 * Support Borland C++ builder.
18169 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18175 * SHA-1 cleanups and performance enhancements.
18177 *Andy Polyakov <appro@fy.chalmers.se>*
18179 * Sparc v8plus assembler for the bignum library.
18181 *Andy Polyakov <appro@fy.chalmers.se>*
18183 * Accept any -xxx and +xxx compiler options in Configure.
18187 * Update HPUX configuration.
18191 * Add missing `sk_<type>_unshift()` function to safestack.h
18193 *Ralf S. Engelschall*
18195 * New function SSL_CTX_use_certificate_chain_file that sets the
18196 "extra_cert"s in addition to the certificate. (This makes sense
18197 only for "PEM" format files, as chains as a whole are not
18202 * Support verify_depth from the SSL API.
18203 x509_vfy.c had what can be considered an off-by-one-error:
18204 Its depth (which was not part of the external interface)
18205 was actually counting the number of certificates in a chain;
18206 now it really counts the depth.
18210 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18211 instead of X509err, which often resulted in confusing error
18212 messages since the error codes are not globally unique
18213 (e.g. an alleged error in ssl3_accept when a certificate
18214 didn't match the private key).
18216 * New function SSL_CTX_set_session_id_context that allows to set a default
18217 value (so that you don't need SSL_set_session_id_context for each
18218 connection using the SSL_CTX).
18222 * OAEP decoding bug fix.
18226 * Support INSTALL_PREFIX for package builders, as proposed by
18231 * New Configure options "threads" and "no-threads". For systems
18232 where the proper compiler options are known (currently Solaris
18233 and Linux), "threads" is the default.
18237 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18241 * Install various scripts to $(OPENSSLDIR)/misc, not to
18242 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18243 such as /usr/local/bin.
18247 * "make linux-shared" to build shared libraries.
18249 *Niels Poppe <niels@netbox.org>*
18251 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18255 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18256 extension adding in x509 utility.
18260 * Remove NOPROTO sections and error code comments.
18264 * Partial rewrite of the DEF file generator to now parse the ANSI
18269 * New Configure options --prefix=DIR and --openssldir=DIR.
18273 * Complete rewrite of the error code script(s). It is all now handled
18274 by one script at the top level which handles error code gathering,
18275 header rewriting and C source file generation. It should be much better
18276 than the old method: it now uses a modified version of Ulf's parser to
18277 read the ANSI prototypes in all header files (thus the old K&R definitions
18278 aren't needed for error creation any more) and do a better job of
18279 translating function codes into names. The old 'ASN1 error code embedded
18280 in a comment' is no longer necessary and it doesn't use .err files which
18281 have now been deleted. Also the error code call doesn't have to appear all
18282 on one line (which resulted in some large lines...).
18286 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18290 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18291 0 (which usually indicates a closed connection), but continue reading.
18295 * Fix some race conditions.
18299 * Add support for CRL distribution points extension. Add Certificate
18300 Policies and CRL distribution points documentation.
18304 * Move the autogenerated header file parts to crypto/opensslconf.h.
18308 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18309 8 of keying material. Merlin has also confirmed interop with this fix
18310 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18312 *Merlin Hughes <merlin@baltimore.ie>*
18314 * Fix lots of warnings.
18316 *Richard Levitte <levitte@stacken.kth.se>*
18318 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18319 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18321 *Richard Levitte <levitte@stacken.kth.se>*
18323 * Fix problems with sizeof(long) == 8.
18325 *Andy Polyakov <appro@fy.chalmers.se>*
18327 * Change functions to ANSI C.
18331 * Fix typos in error codes.
18333 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18335 * Remove defunct assembler files from Configure.
18339 * SPARC v8 assembler BIGNUM implementation.
18341 *Andy Polyakov <appro@fy.chalmers.se>*
18343 * Support for Certificate Policies extension: both print and set.
18344 Various additions to support the r2i method this uses.
18348 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18349 return a const string when you are expecting an allocated buffer.
18353 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18354 types DirectoryString and DisplayText.
18358 * Add code to allow r2i extensions to access the configuration database,
18359 add an LHASH database driver and add several ctx helper functions.
18363 * Fix an evil bug in bn_expand2() which caused various BN functions to
18364 fail when they extended the size of a BIGNUM.
18368 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18369 support typesafe stack.
18373 * Fix typo in SSL_[gs]et_options().
18375 *Nils Frostberg <nils@medcom.se>*
18377 * Delete various functions and files that belonged to the (now obsolete)
18378 old X509V3 handling code.
18382 * New Configure option "rsaref".
18386 * Don't auto-generate pem.h.
18390 * Introduce type-safe ASN.1 SETs.
18394 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18396 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18398 * Introduce type-safe STACKs. This will almost certainly break lots of code
18399 that links with OpenSSL (well at least cause lots of warnings), but fear
18400 not: the conversion is trivial, and it eliminates loads of evil casts. A
18401 few STACKed things have been converted already. Feel free to convert more.
18402 In the fullness of time, I'll do away with the STACK type altogether.
18406 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18407 specified in `<certfile>` by updating the entry in the index.txt file.
18408 This way one no longer has to edit the index.txt file manually for
18409 revoking a certificate. The -revoke option does the gory details now.
18411 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18413 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18414 `-text` option at all and this way the `-noout -text` combination was
18415 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18417 *Ralf S. Engelschall*
18419 * Make sure a corresponding plain text error message exists for the
18420 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18421 verify callback function determined that a certificate was revoked.
18423 *Ralf S. Engelschall*
18425 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18426 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18427 all available ciphers including rc5, which was forgotten until now.
18428 In order to let the testing shell script know which algorithms
18429 are available, a new (up to now undocumented) command
18430 `openssl list-cipher-commands` is used.
18434 * Bugfix: s_client occasionally would sleep in select() when
18435 it should have checked SSL_pending() first.
18439 * New functions DSA_do_sign and DSA_do_verify to provide access to
18440 the raw DSA values prior to ASN.1 encoding.
18444 * Tweaks to Configure
18446 *Niels Poppe <niels@netbox.org>*
18448 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18453 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18457 * New config option to avoid instructions that are illegal on the 80386.
18458 The default code is faster, but requires at least a 486.
18462 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18463 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18464 same as SSL2_VERSION anyway.
18468 * New "-showcerts" option for s_client.
18472 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18473 application. Various cleanups and fixes.
18477 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18478 modify error routines to work internally. Add error codes and PBE init
18479 to library startup routines.
18483 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18484 packing functions to asn1 and evp. Changed function names and error
18485 codes along the way.
18489 * PKCS12 integration: and so it begins... First of several patches to
18490 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18491 objects to objects.h
18495 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18496 and display support for Thawte strong extranet extension.
18500 * Add LinuxPPC support.
18502 *Jeff Dubrule <igor@pobox.org>*
18504 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18505 bn_div_words in alpha.s.
18507 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18509 * Make sure the RSA OAEP test is skipped under -DRSAref because
18510 OAEP isn't supported when OpenSSL is built with RSAref.
18512 *Ulf Moeller <ulf@fitug.de>*
18514 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18515 so they no longer are missing under -DNOPROTO.
18517 *Soren S. Jorvang <soren@t.dk>*
18519 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18521 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18522 doesn't work when the session is reused. Coming soon!
18526 * Fix a security hole, that allows sessions to be reused in the wrong
18527 context thus bypassing client cert protection! All software that uses
18528 client certs and session caches in multiple contexts NEEDS PATCHING to
18529 allow session reuse! A fuller solution is in the works.
18531 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18533 * Some more source tree cleanups (removed obsolete files
18534 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18535 permission on "config" script to be executable) and a fix for the INSTALL
18538 *Ulf Moeller <ulf@fitug.de>*
18540 * Remove some legacy and erroneous uses of malloc, free instead of
18543 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18545 * Make rsa_oaep_test return non-zero on error.
18547 *Ulf Moeller <ulf@fitug.de>*
18549 * Add support for native Solaris shared libraries. Configure
18550 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18551 if someone would make that last step automatic.
18553 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18555 * ctx_size was not built with the right compiler during "make links". Fixed.
18559 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18560 except NULL ciphers". This means the default cipher list will no longer
18561 enable NULL ciphers. They need to be specifically enabled e.g. with
18562 the string "DEFAULT:eNULL".
18566 * Fix to RSA private encryption routines: if p < q then it would
18567 occasionally produce an invalid result. This will only happen with
18568 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18572 * Be less restrictive and allow also `perl util/perlpath.pl
18573 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18574 because this way one can also use an interpreter named `perl5` (which is
18575 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18576 installed as `perl`).
18578 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18580 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18582 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18584 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18585 advapi32.lib to Win32 build and change the pem test comparison
18586 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18587 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18588 and crypto/des/ede_cbcm_enc.c.
18592 * DES quad checksum was broken on big-endian architectures. Fixed.
18596 * Comment out two functions in bio.h that aren't implemented. Fix up the
18597 Win32 test batch file so it (might) work again. The Win32 test batch file
18598 is horrible: I feel ill....
18602 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18603 in e_os.h. Audit of header files to check ANSI and non ANSI
18604 sections: 10 functions were absent from non ANSI section and not exported
18605 from Windows DLLs. Fixed up libeay.num for new functions.
18609 * Make `openssl version` output lines consistent.
18611 *Ralf S. Engelschall*
18613 * Fix Win32 symbol export lists for BIO functions: Added
18614 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18615 to ms/libeay{16,32}.def.
18617 *Ralf S. Engelschall*
18619 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18620 fine under Unix and passes some trivial tests I've now added. But the
18621 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18622 added to make sure no one expects that this stuff really works in the
18623 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18624 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18627 *Ralf S. Engelschall*
18629 * Fix the generation of two part addresses in perl.
18631 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18633 * Add config entry for Linux on MIPS.
18635 *John Tobey <jtobey@channel1.com>*
18637 * Make links whenever Configure is run, unless we are on Windoze.
18641 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18642 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18647 * Add a useful kludge to allow package maintainers to specify compiler and
18648 other platforms details on the command line without having to patch the
18649 Configure script every time: One now can use
18650 `perl Configure <id>:<details>`,
18651 i.e. platform ids are allowed to have details appended
18652 to them (separated by colons). This is treated as there would be a static
18653 pre-configured entry in Configure's %table under key `<id>` with value
18654 `<details>` and `perl Configure <id>` is called. So, when you want to
18655 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18656 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18657 now, which overrides the FreeBSD-elf entry on-the-fly.
18659 *Ralf S. Engelschall*
18661 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18665 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18666 on the `perl Configure ...` command line. This way one can compile
18667 OpenSSL libraries with Position Independent Code (PIC) which is needed
18668 for linking it into DSOs.
18670 *Ralf S. Engelschall*
18672 * Remarkably, export ciphers were totally broken and no-one had noticed!
18677 * Cleaned up the LICENSE document: The official contact for any license
18678 questions now is the OpenSSL core team under openssl-core@openssl.org.
18679 And add a paragraph about the dual-license situation to make sure people
18680 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18681 to the OpenSSL toolkit.
18683 *Ralf S. Engelschall*
18685 * General source tree makefile cleanups: Made `making xxx in yyy...`
18686 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18687 Additionally cleaned up the `make links` target: Remove unnecessary
18688 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18689 to speed processing and no longer clutter the display with confusing
18690 stuff. Instead only the actually done links are displayed.
18692 *Ralf S. Engelschall*
18694 * Permit null encryption ciphersuites, used for authentication only. It used
18695 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18696 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18701 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18702 signed attributes when verifying signatures (this would break them),
18703 the detached data encoding was wrong and public keys obtained using
18704 X509_get_pubkey() weren't freed.
18708 * Add text documentation for the BUFFER functions. Also added a work around
18709 to a Win95 console bug. This was triggered by the password read stuff: the
18710 last character typed gets carried over to the next fread(). If you were
18711 generating a new cert request using 'req' for example then the last
18712 character of the passphrase would be CR which would then enter the first
18717 * Added the new 'Includes OpenSSL Cryptography Software' button as
18718 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18719 button and can be used by applications based on OpenSSL to show the
18720 relationship to the OpenSSL project.
18722 *Ralf S. Engelschall*
18724 * Remove confusing variables in function signatures in files
18725 ssl/ssl_lib.c and ssl/ssl.h.
18727 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18729 * Don't install bss_file.c under PREFIX/include/
18731 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18733 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18734 functions that return function pointers and has support for NT specific
18735 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18736 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18737 unsigned to signed types: this was killing the Win32 compile.
18741 * Add new certificate file to stack functions,
18742 SSL_add_dir_cert_subjects_to_stack() and
18743 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18744 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18745 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18746 This means that Apache-SSL and similar packages don't have to mess around
18747 to add as many CAs as they want to the preferred list.
18751 * Experiment with doxygen documentation. Currently only partially applied to
18753 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18754 openssl.doxy as the configuration file.
18758 * Get rid of remaining C++-style comments which strict C compilers hate.
18760 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18762 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18763 compiled in by default: it has problems with large keys.
18767 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18768 DH private keys and/or callback functions which directly correspond to
18769 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18770 is needed for applications which have to configure certificates on a
18771 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18773 For the RSA certificate situation is makes no difference, but
18774 for the DSA certificate situation this fixes the "no shared cipher"
18775 problem where the OpenSSL cipher selection procedure failed because the
18776 temporary keys were not overtaken from the context and the API provided
18777 no way to reconfigure them.
18778 The new functions now let applications reconfigure the stuff and they
18779 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18780 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18781 non-public-API function ssl_cert_instantiate() is used as a helper
18782 function and also to reduce code redundancy inside ssl_rsa.c.
18784 *Ralf S. Engelschall*
18786 * Move s_server -dcert and -dkey options out of the undocumented feature
18787 area because they are useful for the DSA situation and should be
18788 recognized by the users.
18790 *Ralf S. Engelschall*
18792 * Fix the cipher decision scheme for export ciphers: the export bits are
18793 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18794 SSL_EXP_MASK. So, the original variable has to be used instead of the
18795 already masked variable.
18797 *Richard Levitte <levitte@stacken.kth.se>*
18799 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18801 *Richard Levitte <levitte@stacken.kth.se>*
18803 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18804 from `int` to `unsigned int` because it is a length and initialized by
18805 EVP_DigestFinal() which expects an `unsigned int *`.
18807 *Richard Levitte <levitte@stacken.kth.se>*
18809 * Don't hard-code path to Perl interpreter on shebang line of Configure
18810 script. Instead use the usual Shell->Perl transition trick.
18812 *Ralf S. Engelschall*
18814 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18815 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18816 -noout -modulus` as it's already the case for `openssl rsa -noout
18817 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18818 currently the public key is printed (a decision which was already done by
18819 `openssl dsa -modulus` in the past) which serves a similar purpose.
18820 Additionally the NO_RSA no longer completely removes the whole -modulus
18821 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18824 *Ralf S. Engelschall*
18826 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18827 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18829 *Arne Ansper <arne@ats.cyber.ee>*
18831 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18832 to be added. Now both 'req' and 'ca' can use new objects defined in the
18837 * Add cool BIO that does syslog (or event log on NT).
18839 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18841 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18842 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18843 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18844 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18848 * Add preliminary config info for new extension code.
18852 * Make RSA_NO_PADDING really use no padding.
18854 *Ulf Moeller <ulf@fitug.de>*
18856 * Generate errors when private/public key check is done.
18860 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18861 for some CRL extensions and new objects added.
18865 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18866 key usage extension and fuller support for authority key id.
18870 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18871 padding method for RSA, which is recommended for new applications in PKCS
18872 #1 v2.0 (RFC 2437, October 1998).
18873 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18874 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18875 against Bleichbacher's attack on RSA.
18876 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18879 * Updates to the new SSL compression code
18881 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18883 * Fix so that the version number in the master secret, when passed
18884 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18885 (because the server will not accept higher), that the version number
18886 is 0x03,0x01, not 0x03,0x00
18888 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18890 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18891 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18892 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18896 * Support for RAW extensions where an arbitrary extension can be
18897 created by including its DER encoding. See `apps/openssl.cnf` for
18902 * Make sure latest Perl versions don't interpret some generated C array
18903 code as Perl array code in the crypto/err/err_genc.pl script.
18905 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18907 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18908 not many people have the assembler. Various Win32 compilation fixes and
18909 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18910 build instructions.
18914 * Modify configure script 'Configure' to automatically create crypto/date.h
18915 file under Win32 and also build pem.h from pem.org. New script
18916 util/mkfiles.pl to create the MINFO file on environments that can't do a
18917 'make files': perl util/mkfiles.pl >MINFO should work.
18921 * Major rework of DES function declarations, in the pursuit of correctness
18922 and purity. As a result, many evil casts evaporated, and some weirdness,
18923 too. You may find this causes warnings in your code. Zapping your evil
18924 casts will probably fix them. Mostly.
18928 * Fix for a typo in asn1.h. Bug fix to object creation script
18929 obj_dat.pl. It considered a zero in an object definition to mean
18930 "end of object": none of the objects in objects.h have any zeros
18931 so it wasn't spotted.
18933 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18935 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18936 Masking (CBCM). In the absence of test vectors, the best I have been able
18937 to do is check that the decrypt undoes the encrypt, so far. Send me test
18938 vectors if you have them.
18942 * Correct calculation of key length for export ciphers (too much space was
18943 allocated for null ciphers). This has not been tested!
18947 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18948 message is now correct (it understands "crypto" and "ssl" on its
18949 command line). There is also now an "update" option. This will update
18950 the util/ssleay.num and util/libeay.num files with any new functions.
18952 perl util/mkdef.pl crypto ssl update
18953 it will update them.
18957 * Overhauled the Perl interface:
18958 - ported BN stuff to OpenSSL's different BN library
18959 - made the perl/ source tree CVS-aware
18960 - renamed the package from SSLeay to OpenSSL (the files still contain
18961 their history because I've copied them in the repository)
18962 - removed obsolete files (the test scripts will be replaced
18963 by better Test::Harness variants in the future)
18965 *Ralf S. Engelschall*
18967 * First cut for a very conservative source tree cleanup:
18968 1. merge various obsolete readme texts into doc/ssleay.txt
18969 where we collect the old documents and readme texts.
18970 2. remove the first part of files where I'm already sure that we no
18971 longer need them because of three reasons: either they are just temporary
18972 files which were left by Eric or they are preserved original files where
18973 I've verified that the diff is also available in the CVS via "cvs diff
18974 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18975 the crypto/md/ stuff).
18977 *Ralf S. Engelschall*
18979 * More extension code. Incomplete support for subject and issuer alt
18980 name, issuer and authority key id. Change the i2v function parameters
18981 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18982 what that's for :-) Fix to ASN1 macro which messed up
18983 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18987 * Preliminary support for ENUMERATED type. This is largely copied from the
18992 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18994 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18996 * Make sure `make rehash` target really finds the `openssl` program.
18998 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19000 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19001 like to hear about it if this slows down other processors.
19005 * Add CygWin32 platform information to Configure script.
19007 *Alan Batie <batie@aahz.jf.intel.com>*
19009 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19011 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19013 * New program nseq to manipulate netscape certificate sequences
19017 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19022 * Fixes to BN code. Previously the default was to define BN_RECURSION
19023 but the BN code had some problems that would cause failures when
19024 doing certificate verification and some other functions.
19026 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19028 * Add ASN1 and PEM code to support netscape certificate sequences.
19032 * Add ASN1 and PEM code to support netscape certificate sequences.
19036 * Add several PKIX and private extended key usage OIDs.
19040 * Modify the 'ca' program to handle the new extension code. Modify
19041 openssl.cnf for new extension format, add comments.
19045 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19046 and add a sample to openssl.cnf so req -x509 now adds appropriate
19051 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19052 error code, add initial support to X509_print() and x509 application.
19056 * Takes a deep breath and start adding X509 V3 extension support code. Add
19057 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19058 stuff is currently isolated and isn't even compiled yet.
19062 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19063 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19064 Removed the versions check from X509 routines when loading extensions:
19065 this allows certain broken certificates that don't set the version
19066 properly to be processed.
19070 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19071 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19072 can still be regenerated with "make depend".
19076 * Spelling mistake in C version of CAST-128.
19078 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19080 * Changes to the error generation code. The perl script err-code.pl
19081 now reads in the old error codes and retains the old numbers, only
19082 adding new ones if necessary. It also only changes the .err files if new
19083 codes are added. The makefiles have been modified to only insert errors
19084 when needed (to avoid needlessly modifying header files). This is done
19085 by only inserting errors if the .err file is newer than the auto generated
19086 C file. To rebuild all the error codes from scratch (the old behaviour)
19087 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19088 or delete all the .err files.
19092 * CAST-128 was incorrectly implemented for short keys. The C version has
19093 been fixed, but is untested. The assembler versions are also fixed, but
19094 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19095 to regenerate it if needed.
19096 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19097 Hagino <itojun@kame.net>*
19099 * File was opened incorrectly in randfile.c.
19101 *Ulf Möller <ulf@fitug.de>*
19103 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19104 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19105 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19106 al: it's just almost always a UTCTime. Note this patch adds new error
19107 codes so do a "make errors" if there are problems.
19111 * Correct Linux 1 recognition in config.
19113 *Ulf Möller <ulf@fitug.de>*
19115 * Remove pointless MD5 hash when using DSA keys in ca.
19117 *Anonymous <nobody@replay.com>*
19119 * Generate an error if given an empty string as a cert directory. Also
19120 generate an error if handed NULL (previously returned 0 to indicate an
19121 error, but didn't set one).
19123 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19125 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19129 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19130 parameters. This was causing a warning which killed off the Win32 compile.
19134 * Remove C++ style comments from crypto/bn/bn_local.h.
19136 *Neil Costigan <neil.costigan@celocom.com>*
19138 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19139 based on a text string, looking up short and long names and finally
19140 "dot" format. The "dot" format stuff didn't work. Added new function
19141 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19142 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19143 OID is not part of the table.
19147 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19148 X509_LOOKUP_by_alias().
19152 * Sort openssl functions by name.
19156 * Get the `gendsa` command working and add it to the `list` command. Remove
19157 encryption from sample DSA keys (in case anyone is interested the password
19162 * Make *all* `*_free` functions accept a NULL pointer.
19164 *Frans Heymans <fheymans@isaserver.be>*
19166 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19169 *Anonymous <nobody@replay.com>*
19171 * s_server should send the CAfile as acceptable CAs, not its own cert.
19173 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19175 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19177 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19179 * Temp key "for export" tests were wrong in s3_srvr.c.
19181 *Anonymous <nobody@replay.com>*
19183 * Add prototype for temp key callback functions
19184 SSL_CTX_set_tmp_{rsa,dh}_callback().
19188 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19189 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19193 * X509_name_add_entry() freed the wrong thing after an error.
19195 *Arne Ansper <arne@ats.cyber.ee>*
19197 * rsa_eay.c would attempt to free a NULL context.
19199 *Arne Ansper <arne@ats.cyber.ee>*
19201 * BIO_s_socket() had a broken should_retry() on Windoze.
19203 *Arne Ansper <arne@ats.cyber.ee>*
19205 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19207 *Arne Ansper <arne@ats.cyber.ee>*
19209 * Make sure the already existing X509_STORE->depth variable is initialized
19210 in X509_STORE_new(), but document the fact that this variable is still
19211 unused in the certificate verification process.
19213 *Ralf S. Engelschall*
19215 * Fix the various library and `apps/` files to free up pkeys obtained from
19216 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19220 * Fix reference counting in X509_PUBKEY_get(). This makes
19221 demos/maurice/example2.c work, amongst others, probably.
19223 *Steve Henson and Ben Laurie*
19225 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19226 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19227 are no longer created. This way we have a single and consistent command
19228 line interface `openssl <command>`, similar to `cvs <command>`.
19230 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19232 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19233 BIT STRING wrapper always have zero unused bits.
19237 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19241 * Make the top-level INSTALL documentation easier to understand.
19245 * Makefiles updated to exit if an error occurs in a sub-directory
19246 make (including if user presses ^C) [Paul Sutton]
19248 * Make Montgomery context stuff explicit in RSA data structure.
19252 * Fix build order of pem and err to allow for generated pem.h.
19256 * Fix renumbering bug in X509_NAME_delete_entry().
19260 * Enhanced the err-ins.pl script so it makes the error library number
19261 global and can add a library name. This is needed for external ASN1 and
19262 other error libraries.
19266 * Fixed sk_insert which never worked properly.
19270 * Fix ASN1 macros so they can handle indefinite length constructed
19271 EXPLICIT tags. Some non standard certificates use these: they can now
19276 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19277 into a single doc/ssleay.txt bundle. This way the information is still
19278 preserved but no longer messes up this directory. Now it's new room for
19279 the new set of documentation files.
19281 *Ralf S. Engelschall*
19283 * SETs were incorrectly DER encoded. This was a major pain, because they
19284 shared code with SEQUENCEs, which aren't coded the same. This means that
19285 almost everything to do with SETs or SEQUENCEs has either changed name or
19286 number of arguments.
19288 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19290 * Fix test data to work with the above.
19294 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19295 was already fixed by Eric for 0.9.1 it seems.
19297 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19299 * Autodetect FreeBSD3.
19303 * Fix various bugs in Configure. This affects the following platforms:
19307 unixware-2.0-pentium
19312 * Eliminate generated files from CVS. Reorder tests to regenerate files
19313 before they are needed.
19317 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19321 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19323 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19324 changed SSLeay to OpenSSL in version strings.
19326 *Ralf S. Engelschall*
19328 * Some fixups to the top-level documents.
19332 * Fixed the nasty bug where rsaref.h was not found under compile-time
19333 because the symlink to include/ was missing.
19335 *Ralf S. Engelschall*
19337 * Incorporated the popular no-RSA/DSA-only patches
19338 which allow to compile an RSA-free SSLeay.
19340 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19342 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19343 when "ssleay" is still not found.
19345 *Ralf S. Engelschall*
19347 * Added more platforms to Configure: Cray T3E, HPUX 11,
19349 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19351 * Updated the README file.
19353 *Ralf S. Engelschall*
19355 * Added various .cvsignore files in the CVS repository subdirs
19356 to make a "cvs update" really silent.
19358 *Ralf S. Engelschall*
19360 * Recompiled the error-definition header files and added
19361 missing symbols to the Win32 linker tables.
19363 *Ralf S. Engelschall*
19365 * Cleaned up the top-level documents;
19366 o new files: CHANGES and LICENSE
19367 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19368 o merged COPYRIGHT into LICENSE
19369 o removed obsolete TODO file
19370 o renamed MICROSOFT to INSTALL.W32
19372 *Ralf S. Engelschall*
19374 * Removed dummy files from the 0.9.1b source tree:
19375 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19376 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19377 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19378 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19379 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19381 *Ralf S. Engelschall*
19383 * Added various platform portability fixes.
19387 * The Genesis of the OpenSSL rpject:
19388 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19389 Young and Tim J. Hudson created while they were working for C2Net until
19392 *The OpenSSL Project*
19394 ### Changes between 0.9.0b and 0.9.1b [not released]
19396 * Updated a few CA certificates under certs/
19400 * Changed some BIGNUM api stuff.
19404 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19405 DGUX x86, Linux Alpha, etc.
19409 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19410 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19415 * Add -strparse option to asn1pars program which parses nested
19418 *Dr Stephen Henson <shenson@bigfoot.com>*
19420 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19424 * DSA fix for "ca" program.
19428 * Added "-genkey" option to "dsaparam" program.
19432 * Added RIPE MD160 (rmd160) message digest.
19436 * Added -a (all) option to "ssleay version" command.
19440 * Added PLATFORM define which is the id given to Configure.
19444 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19448 * Extended the ASN.1 parser routines.
19452 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19456 * Added a BN_CTX to the BN library.
19460 * Fixed the weak key values in DES library
19464 * Changed API in EVP library for cipher aliases.
19468 * Added support for RC2/64bit cipher.
19472 * Converted the lhash library to the crypto/mem.c functions.
19476 * Added more recognized ASN.1 object ids.
19480 * Added more RSA padding checks for SSL/TLS.
19484 * Added BIO proxy/filter functionality.
19488 * Added extra_certs to SSL_CTX which can be used
19489 send extra CA certificates to the client in the CA cert chain sending
19490 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19494 * Now Fortezza is denied in the authentication phase because
19495 this is key exchange mechanism is not supported by SSLeay at all.
19499 * Additional PKCS1 checks.
19503 * Support the string "TLSv1" for all TLS v1 ciphers.
19507 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19508 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19512 * Fixed a few memory leaks.
19516 * Fixed various code and comment typos.
19520 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19521 bytes sent in the client random.
19523 *Edward Bishop <ebishop@spyglass.com>*
19527 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19528 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19529 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19530 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19531 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19532 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19533 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19534 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19535 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19536 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19537 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19538 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19539 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19540 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19541 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19542 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19543 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19544 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19545 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19546 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19547 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19548 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19549 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19550 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19551 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19552 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19553 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19554 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19555 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19556 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19557 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19558 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19559 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19560 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19561 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19562 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19563 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19564 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19565 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19566 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19567 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19568 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19569 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19570 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19571 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19572 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19573 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19574 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19575 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19576 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19577 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19578 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19579 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19580 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19581 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19582 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19583 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19584 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19585 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19586 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19587 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19588 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19589 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19590 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19591 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19592 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19593 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19594 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19595 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19596 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19597 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19598 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19599 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19600 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19601 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19602 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19603 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19604 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19605 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19606 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19607 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19608 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19609 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19610 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19611 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19612 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19613 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19614 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19615 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19616 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19617 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19618 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19619 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19620 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19621 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19622 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19623 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19624 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19625 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19626 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19627 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19628 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19629 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19630 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19631 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19632 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19633 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19634 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19635 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19636 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19637 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19638 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19639 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19640 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19641 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19642 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19643 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19644 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19645 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19646 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19647 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19648 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19649 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19650 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19651 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19652 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19653 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19654 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19655 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19656 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19657 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19658 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19659 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19660 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19661 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19662 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19663 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19664 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19665 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19666 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19667 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19668 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19669 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19670 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19671 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19672 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19673 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19674 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19675 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19676 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19677 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19678 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19679 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19680 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19681 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19682 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19683 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19684 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19685 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19686 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19687 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19688 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19689 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19690 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19691 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655