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 * s_client and s_server apps now explicitly say when the TLS version
28 does not include the renegotiation mechanism. This avoids confusion
29 between that scenario versus when the TLS version includes secure
30 renegotiation but the peer lacks support for it.
34 * The various OBJ_* functions have been made thread safe.
38 * Parallel dual-prime 1536/2048-bit modular exponentiation for
39 AVX512_IFMA capable processors.
41 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
46 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
47 listed here are only a brief description.
48 The migration guide contains more detailed information related to new features,
49 breaking changes, and mappings for the large list of deprecated functions.
51 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
53 ### Changes between 3.0.6 and 3.0.7 [xx XXX xxxx]
55 * Fixed two buffer overflows in punycode decoding functions.
57 A buffer overrun can be triggered in X.509 certificate verification,
58 specifically in name constraint checking. Note that this occurs after
59 certificate chain signature verification and requires either a CA to
60 have signed the malicious certificate or for the application to continue
61 certificate verification despite failure to construct a path to a trusted
64 In a TLS client, this can be triggered by connecting to a malicious
65 server. In a TLS server, this can be triggered if the server requests
66 client authentication and a malicious client connects.
68 An attacker can craft a malicious email address to overflow
69 an arbitrary number of bytes containing the `.` character (decimal 46)
70 on the stack. This buffer overflow could result in a crash (causing a
74 An attacker can craft a malicious email address to overflow four
75 attacker-controlled bytes on the stack. This buffer overflow could
76 result in a crash (causing a denial of service) or potentially remote code
77 execution depending on stack layout for any given platform/compiler.
82 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
83 parameters in OpenSSL code.
84 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
85 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
86 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
87 Using these invalid names may cause algorithms to use slower methods
88 that ignore the CRT parameters.
92 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
97 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
98 data to be signed before signing the certificate.
102 * Added RIPEMD160 to the default provider.
106 * Ensured that the key share group sent or accepted for the key exchange
107 is allowed for the protocol version.
111 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
113 * OpenSSL supports creating a custom cipher via the legacy
114 EVP_CIPHER_meth_new() function and associated function calls. This function
115 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
116 to use the new provider mechanism in order to implement custom ciphers.
118 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
119 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
120 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
121 and decryption initialisation functions). Instead of using the custom cipher
122 directly it incorrectly tries to fetch an equivalent cipher from the
123 available providers. An equivalent cipher is found based on the NID passed to
124 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
125 given cipher. However it is possible for an application to incorrectly pass
126 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
127 is used in this way the OpenSSL encryption/decryption initialisation function
128 will match the NULL cipher as being equivalent and will fetch this from the
129 available providers. This will succeed if the default provider has been
130 loaded (or if a third party provider has been loaded that offers this
131 cipher). Using the NULL cipher means that the plaintext is emitted as the
134 Applications are only affected by this issue if they call
135 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
136 encryption/decryption initialisation function. Applications that only use
137 SSL/TLS are not impacted by this issue.
142 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
147 * Fixed the linux-mips64 Configure target which was missing the
148 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
153 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
158 * Correctly handle a retransmitted ClientHello in DTLS
162 * Fixed detection of ktls support in cross-compile environment on Linux
166 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
171 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
172 report correct results in some cases
176 * Fix UWP builds by defining VirtualLock
180 * For known safe primes use the minimum key length according to RFC 7919.
181 Longer private key sizes unnecessarily raise the cycles needed to compute the
182 shared secret without any increase of the real security. This fixes a
183 regression from 1.1.1 where these shorter keys were generated for the known
188 * Added the loongarch64 target
192 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
193 only passed to the FIPS provider and not to the default or legacy provider.
197 * Fixed reported performance degradation on aarch64. Restored the
198 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
199 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
200 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
201 The new algorithm is still used for 32 bit targets.
205 * Added a missing header for memcmp that caused compilation failure on some
210 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
212 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
213 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
214 This issue makes the RSA implementation with 2048 bit private keys
215 incorrect on such machines and memory corruption will happen during
216 the computation. As a consequence of the memory corruption an attacker
217 may be able to trigger a remote code execution on the machine performing
220 SSL/TLS servers or other servers using 2048 bit RSA private keys running
221 on machines supporting AVX512IFMA instructions of the X86_64 architecture
222 are affected by this issue.
227 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
228 implementation would not encrypt the entirety of the data under some
229 circumstances. This could reveal sixteen bytes of data that was
230 preexisting in the memory that wasn't written. In the special case of
231 "in place" encryption, sixteen bytes of the plaintext would be revealed.
233 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
234 they are both unaffected.
237 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
239 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
241 * In addition to the c_rehash shell command injection identified in
242 CVE-2022-1292, further bugs where the c_rehash script does not
243 properly sanitise shell metacharacters to prevent command injection have been
246 When the CVE-2022-1292 was fixed it was not discovered that there
247 are other places in the script where the file names of certificates
248 being hashed were possibly passed to a command executed through the shell.
250 This script is distributed by some operating systems in a manner where
251 it is automatically executed. On such operating systems, an attacker
252 could execute arbitrary commands with the privileges of the script.
254 Use of the c_rehash script is considered obsolete and should be replaced
255 by the OpenSSL rehash command line tool.
258 *Daniel Fiala, Tomáš Mráz*
260 * Case insensitive string comparison no longer uses locales. It has instead
261 been directly implemented.
265 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
267 * Case insensitive string comparison is reimplemented via new locale-agnostic
268 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
269 comparison. The previous implementation had problems when the Turkish locale
274 * Fixed a bug in the c_rehash script which was not properly sanitising shell
275 metacharacters to prevent command injection. This script is distributed by
276 some operating systems in a manner where it is automatically executed. On
277 such operating systems, an attacker could execute arbitrary commands with the
278 privileges of the script.
280 Use of the c_rehash script is considered obsolete and should be replaced
281 by the OpenSSL rehash command line tool.
286 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
287 certificate on an OCSP response. The bug caused the function in the case
288 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
289 response (meaning a successful verification) even in the case where the
290 response signing certificate fails to verify.
292 It is anticipated that most users of `OCSP_basic_verify` will not use the
293 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
294 a negative value (indicating a fatal error) in the case of a certificate
295 verification failure. The normal expected return value in this case would be
298 This issue also impacts the command line OpenSSL "ocsp" application. When
299 verifying an ocsp response with the "-no_cert_checks" option the command line
300 application will report that the verification is successful even though it
301 has in fact failed. In this case the incorrect successful response will also
302 be accompanied by error messages showing the failure and contradicting the
303 apparently successful result.
308 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
309 AAD data as the MAC key. This made the MAC key trivially predictable.
311 An attacker could exploit this issue by performing a man-in-the-middle attack
312 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
313 that the modified data would still pass the MAC integrity check.
315 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
316 endpoint will always be rejected by the recipient and the connection will
317 fail at that point. Many application protocols require data to be sent from
318 the client to the server first. Therefore, in such a case, only an OpenSSL
319 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
321 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
322 sent in both directions. In this case both clients and servers could be
323 affected, regardless of the application protocol.
325 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
326 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
327 the handshake when using this ciphersuite.
329 The confidentiality of data is not impacted by this issue, i.e. an attacker
330 cannot decrypt data that has been encrypted using this ciphersuite - they can
333 In order for this attack to work both endpoints must legitimately negotiate
334 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
335 OpenSSL 3.0, and is not available within the default provider or the default
336 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
337 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
338 following must have occurred:
340 1) OpenSSL must have been compiled with the (non-default) compile time option
341 enable-weak-ssl-ciphers
343 2) OpenSSL must have had the legacy provider explicitly loaded (either
344 through application code or via configuration)
346 3) The ciphersuite must have been explicitly added to the ciphersuite list
348 4) The libssl security level must have been set to 0 (default is 1)
350 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
352 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
353 others that both endpoints have in common
358 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
359 occuppied by the removed hash table entries.
361 This function is used when decoding certificates or keys. If a long lived
362 process periodically decodes certificates or keys its memory usage will
363 expand without bounds and the process might be terminated by the operating
364 system causing a denial of service. Also traversing the empty hash table
365 entries will take increasingly more time.
367 Typically such long lived processes might be TLS clients or TLS servers
368 configured to accept client certificate authentication.
371 *Hugo Landau, Aliaksei Levin*
373 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
374 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
375 statistics are no longer supported. For compatibility, these statistics are
376 still listed in the output but are now always reported as zero.
380 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
382 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
383 for non-prime moduli.
385 Internally this function is used when parsing certificates that contain
386 elliptic curve public keys in compressed form or explicit elliptic curve
387 parameters with a base point encoded in compressed form.
389 It is possible to trigger the infinite loop by crafting a certificate that
390 has invalid explicit curve parameters.
392 Since certificate parsing happens prior to verification of the certificate
393 signature, any process that parses an externally supplied certificate may thus
394 be subject to a denial of service attack. The infinite loop can also be
395 reached when parsing crafted private keys as they can contain explicit
396 elliptic curve parameters.
398 Thus vulnerable situations include:
400 - TLS clients consuming server certificates
401 - TLS servers consuming client certificates
402 - Hosting providers taking certificates or private keys from customers
403 - Certificate authorities parsing certification requests from subscribers
404 - Anything else which parses ASN.1 elliptic curve parameters
406 Also any other applications that use the BN_mod_sqrt() where the attacker
407 can control the parameter values are vulnerable to this DoS issue.
412 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
413 to the list of ciphersuites providing Perfect Forward Secrecy as
414 required by SECLEVEL >= 3.
416 *Dmitry Belyavskiy, Nicola Tuveri*
418 * Made the AES constant time code for no-asm configurations
419 optional due to the resulting 95% performance degradation.
420 The AES constant time code can be enabled, for no assembly
421 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
425 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
430 * The negative return value handling of the certificate verification callback
431 was reverted. The replacement is to set the verification retry state with
432 the SSL_set_retry_verify() function.
436 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
438 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
439 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
440 verify a certificate supplied by a server. That function may return a
441 negative return value to indicate an internal error (for example out of
442 memory). Such a negative return value is mishandled by OpenSSL and will cause
443 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
444 success and a subsequent call to SSL_get_error() to return the value
445 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
446 returned by OpenSSL if the application has previously called
447 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
448 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
449 totally unexpected and applications may not behave correctly as a result. The
450 exact behaviour will depend on the application but it could result in
451 crashes, infinite loops or other similar incorrect responses.
453 This issue is made more serious in combination with a separate bug in OpenSSL
454 3.0 that will cause X509_verify_cert() to indicate an internal error when
455 processing a certificate chain. This will occur where a certificate does not
456 include the Subject Alternative Name extension but where a Certificate
457 Authority has enforced name constraints. This issue can occur even with valid
463 * Corrected a few file name and file reference bugs in the build,
464 installation and setup scripts, which lead to installation verification
465 failures. Slightly enhanced the installation verification script.
469 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
474 * Fixed PVK encoder to properly query for the passphrase.
478 * Multiple fixes in the OSSL_HTTP API functions.
482 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
483 OSSL_PARAM_INTEGER data type and return error on negative numbers
484 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
485 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
489 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
493 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
497 * Multiple threading fixes.
501 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
505 * Allow fetching an operation from the provider that owns an unexportable key
506 as a fallback if that is still allowed by the property query.
510 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
512 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
517 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
518 S390X capability vector to zero. This simplifies testing of different code
519 paths on S390X architecture.
523 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
524 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
525 SP 800-38D". The communication will fail at this point.
529 * The EC_GROUP_clear_free() function is deprecated as there is nothing
530 confidential in EC_GROUP data.
534 * The byte order mark (BOM) character is ignored if encountered at the
535 beginning of a PEM-formatted file.
539 * Added CMS support for the Russian GOST algorithms.
543 * Due to move of the implementation of cryptographic operations
544 to the providers, validation of various operation parameters can
545 be postponed until the actual operation is executed where previously
546 it happened immediately when an operation parameter was set.
548 For example when setting an unsupported curve with
549 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
550 fail but later keygen operations with the EVP_PKEY_CTX will fail.
552 *OpenSSL team members and many third party contributors*
554 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
555 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
556 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
557 instead to retrieve these algorithms from a provider.
561 * On build targets where the multilib postfix is set in the build
562 configuration the libdir directory was changing based on whether
563 the lib directory with the multilib postfix exists on the system
564 or not. This unpredictable behavior was removed and eventual
565 multilib postfix is now always added to the default libdir. Use
566 `--libdir=lib` to override the libdir if adding the postfix is
571 * The triple DES key wrap functionality now conforms to RFC 3217 but is
572 no longer interoperable with OpenSSL 1.1.1.
576 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
577 function codes, this function can only cause problems for calling
582 * Add a configurable flag to output date formats as ISO 8601. Does not
583 change the default date format.
587 * Version of MSVC earlier than 1300 could get link warnings, which could
588 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
589 Support for this flag has been removed.
593 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
594 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
595 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
596 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
597 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
601 * The signatures of the functions to get and set options on SSL and
602 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
603 Some source code changes may be required.
607 * The public definitions of conf_method_st and conf_st have been
608 deprecated. They will be made opaque in a future release.
610 *Rich Salz and Tomáš Mráz*
612 * Client-initiated renegotiation is disabled by default. To allow it, use
613 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
614 flag, or the "ClientRenegotiation" config parameter as appropriate.
618 * Add "abspath" and "includedir" pragma's to config files, to prevent,
619 or modify relative pathname inclusion.
623 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
624 validated. Please consult the README-FIPS and
625 README-PROVIDERS files, as well as the migration guide.
627 *OpenSSL team members and many third party contributors*
629 * For the key types DH and DHX the allowed settable parameters are now different.
633 * The openssl commands that read keys, certificates, and CRLs now
634 automatically detect the PEM or DER format of the input files.
636 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
638 * Added enhanced PKCS#12 APIs which accept a library context.
642 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
646 * Added support for Kernel TLS (KTLS).
648 *Boris Pismenny, John Baldwin and Andrew Gallatin*
650 * Support for RFC 5746 secure renegotiation is now required by default for
651 SSL or TLS connections to succeed.
655 * The signature of the `copy` functional parameter of the
656 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
657 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
658 the signature of the `pub_decode` functional parameter of the
659 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
660 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
664 * The error return values from some control calls (ctrl) have changed.
668 * A public key check is now performed during EVP_PKEY_derive_set_peer().
672 * Many functions in the EVP_ namespace that are getters of values from
673 implementations or contexts were renamed to include get or get0 in their
674 names. Old names are provided as macro aliases for compatibility and
679 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
680 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
681 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
686 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
689 * The output from the command line applications may have minor
694 * The output from numerous "printing" may have minor changes.
698 * Windows thread synchronization uses read/write primitives (SRWLock) when
699 supported by the OS, otherwise CriticalSection continues to be used.
703 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
704 work on read only BIO source/sinks that do not support these functions.
705 This allows piping or redirection of a file BIO using stdin to be buffered
706 into memory. This is used internally in OSSL_DECODER_from_bio().
710 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
711 this function would return one of the values OSSL_STORE_INFO_NAME,
712 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
713 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
714 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
715 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
716 using this function should be amended to handle the changed return value.
720 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
721 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
722 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
723 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
724 contains more than one certificate identifier: This means that all
725 certificates referenced there MUST be part of the validation chain.
729 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
730 RC5, DESX and DES have been moved to the legacy provider.
734 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
735 RIPEMD-160 have been moved to the legacy provider.
739 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
744 * The deprecated functions EVP_PKEY_get0_RSA(),
745 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
746 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
747 well as the similarly named "get1" functions behave differently in
752 * A number of functions handling low-level keys or engines were deprecated
753 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
754 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
755 EVP_PKEY_get0_siphash().
759 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
760 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
761 will need to load the legacy crypto provider. This includes these PBE
762 algorithms which use this KDF:
763 - NID_pbeWithMD2AndDES_CBC
764 - NID_pbeWithMD5AndDES_CBC
765 - NID_pbeWithSHA1AndRC2_CBC
766 - NID_pbeWithMD2AndRC2_CBC
767 - NID_pbeWithMD5AndRC2_CBC
768 - NID_pbeWithSHA1AndDES_CBC
772 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
773 BIO_debug_callback() functions.
777 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
778 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
782 * The RAND_METHOD APIs have been deprecated.
786 * The SRP APIs have been deprecated.
790 * Add a compile time option to prevent the caching of provider fetched
791 algorithms. This is enabled by including the no-cached-fetch option
792 at configuration time.
796 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
797 count of PKCS12_DEFAULT_ITER.
799 *Tomáš Mráz and Sahana Prasad*
801 * The openssl speed command does not use low-level API calls anymore.
805 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
808 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
810 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
814 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
815 providers may supply their own group implementations (using either the "key
816 exchange" or the "key encapsulation" methods) which will automatically be
817 detected and used by libssl.
819 *Matt Caswell, Nicola Tuveri*
821 * The undocumented function X509_certificate_type() has been deprecated;
825 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
829 * Removed RSA padding mode for SSLv23 (which was only used for
830 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
831 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
836 * Deprecated the obsolete X9.31 RSA key generation related functions.
838 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
839 is not allowed to return a value > 1, this is no more taken as failure.
841 *Viktor Dukhovni and David von Oheimb*
843 * Deprecated the obsolete X9.31 RSA key generation related functions
844 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
845 BN_X931_generate_prime_ex().
849 * The default key generation method for the regular 2-prime RSA keys was
850 changed to the FIPS 186-4 B.3.6 method.
854 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
858 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
862 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
863 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
865 *Rich Salz, Richard Levitte, and David von Oheimb*
867 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
871 * Deprecated `OCSP_parse_url()`.
875 * Validation of SM2 keys has been separated from the validation of regular EC
880 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
881 switches: a validation failure triggers an early exit, returning a failure
882 exit status to the parent process.
886 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
887 to ignore unknown ciphers.
891 * The `-cipher-commands` and `-digest-commands` options
892 of the command line utility `list` have been deprecated.
893 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
897 * Added convenience functions for generating asymmetric key pairs:
898 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
899 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
903 * All of the low level EC_KEY functions have been deprecated.
905 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
907 * Deprecated all the libcrypto and libssl error string loading
912 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
913 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
918 * The `-crypt` option to the `passwd` command line tool has been removed.
922 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
927 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
931 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
932 EVP_PKEY_get1_tls_encodedpoint().
936 * The security callback, which can be customised by application code, supports
937 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
938 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
942 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
943 interface. Their functionality remains unchanged.
947 * Added new option for 'openssl list', '-providers', which will display the
948 list of loaded providers, their names, version and status. It optionally
949 displays their gettable parameters.
953 * Removed EVP_PKEY_set_alias_type().
957 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
958 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
962 * Changed all "STACK" functions to be macros instead of inline functions. Macro
963 parameters are still checked for type safety at compile time via helper
968 * Remove the RAND_DRBG API
970 *Paul Dale and Matthias St. Pierre*
972 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
973 as well as actual hostnames.
977 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
978 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
979 conversely, silently ignore DTLS protocol version bounds when configuring
980 TLS-based contexts. The commands can be repeated to set bounds of both
981 types. The same applies with the corresponding "min_protocol" and
982 "max_protocol" command-line switches, in case some application uses both TLS
985 SSL_CTX instances that are created for a fixed protocol version (e.g.
986 `TLSv1_server_method()`) also silently ignore version bounds. Previously
987 attempts to apply bounds to these protocol versions would result in an
988 error. Now only the "version-flexible" SSL_CTX instances are subject to
989 limits in configuration files in command-line options.
993 * Deprecated the `ENGINE` API. Engines should be replaced with providers
998 * Reworked the recorded ERR codes to make better space for system errors.
999 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1000 given code is a system error (true) or an OpenSSL error (false).
1004 * Reworked the test perl framework to better allow parallel testing.
1006 *Nicola Tuveri and David von Oheimb*
1008 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1009 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1013 * 'Configure' has been changed to figure out the configuration target if
1014 none is given on the command line. Consequently, the 'config' script is
1015 now only a mere wrapper. All documentation is changed to only mention
1018 *Rich Salz and Richard Levitte*
1020 * Added a library context `OSSL_LIB_CTX` that applications as well as
1021 other libraries can use to form a separate context within which
1022 libcrypto operations are performed.
1026 * Added various `_ex` functions to the OpenSSL API that support using
1027 a non-default `OSSL_LIB_CTX`.
1031 * Handshake now fails if Extended Master Secret extension is dropped
1036 * Dropped interactive mode from the `openssl` program.
1040 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1042 *David von Oheimb and Shane Lontis*
1044 * Deprecated `EC_METHOD_get_field_type()`.
1048 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1049 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1050 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1054 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1058 * Add CAdES-BES signature verification support, mostly derived
1059 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1061 *Filipe Raimundo da Silva*
1063 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1067 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1068 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1072 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1076 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1077 EC_KEY_precompute_mult().
1081 * Deprecated EC_POINTs_mul().
1085 * Removed FIPS_mode() and FIPS_mode_set().
1089 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1093 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1094 EC_POINT_get_Jprojective_coordinates_GFp().
1098 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1099 arrays to be more easily constructed via a series of utility functions.
1100 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1101 the various push functions and finally convert to a passable OSSL_PARAM
1102 array using OSSL_PARAM_BLD_to_param().
1106 * The security strength of SHA1 and MD5 based signatures in TLS has been
1111 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1112 contain a provider side internal key.
1116 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1120 * Project text documents not yet having a proper file name extension
1121 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1122 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1126 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1127 have been converted to Markdown with the goal to produce documents
1128 which not only look pretty when viewed online in the browser, but
1129 remain well readable inside a plain text editor.
1131 To achieve this goal, a 'minimalistic' Markdown style has been applied
1132 which avoids formatting elements that interfere too much with the
1133 reading flow in the text file. For example, it
1135 * avoids [ATX headings][] and uses [setext headings][] instead
1136 (which works for `<h1>` and `<h2>` headings only).
1137 * avoids [inline links][] and uses [reference links][] instead.
1138 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1140 [ATX headings]: https://github.github.com/gfm/#atx-headings
1141 [setext headings]: https://github.github.com/gfm/#setext-headings
1142 [inline links]: https://github.github.com/gfm/#inline-link
1143 [reference links]: https://github.github.com/gfm/#reference-link
1144 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1145 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1147 *Matthias St. Pierre*
1149 * The test suite is changed to preserve results of each test recipe.
1150 A new directory test-runs/ with subdirectories named like the
1151 test recipes are created in the build tree for this purpose.
1155 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1156 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1157 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1159 *David von Oheimb, Martin Peylo*
1161 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1162 It supports arbitrary request and response content types, GET redirection,
1163 TLS, connections via HTTP(S) proxies, connections and exchange via
1164 user-defined BIOs (allowing implicit connections), persistent connections,
1165 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1166 The legacy OCSP-focused (and only partly documented) API
1167 is retained for backward compatibility, while most of it is deprecated.
1171 * Added `util/check-format.pl`, a tool for checking adherence to the
1172 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1173 The checks performed are incomplete and yield some false positives.
1174 Still the tool should be useful for detecting most typical glitches.
1178 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1179 If domain name resolution yields multiple IP addresses all of them are tried
1180 after `connect()` failures.
1184 * All of the low level RSA functions have been deprecated.
1188 * X509 certificates signed using SHA1 are no longer allowed at security
1193 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1194 modified to use PKEY APIs. These commands are now in maintenance mode
1195 and no new features will be added to them.
1199 * The command line utility rsautl has been deprecated.
1203 * The command line utilities genrsa and rsa have been modified to use PKEY
1204 APIs. They now write PKCS#8 keys by default. These commands are now in
1205 maintenance mode and no new features will be added to them.
1209 * All of the low level DH functions have been deprecated.
1211 *Paul Dale and Matt Caswell*
1213 * All of the low level DSA functions have been deprecated.
1217 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1218 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1222 * Deprecated low level ECDH and ECDSA functions.
1226 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1230 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1231 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1232 a new formulation to include all the things it can be used for,
1233 as well as words of caution.
1237 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1241 * All of the low level HMAC functions have been deprecated.
1243 *Paul Dale and David von Oheimb*
1245 * Over two thousand fixes were made to the documentation, including:
1246 - Common options (such as -rand/-writerand, TLS version control, etc)
1247 were refactored and point to newly-enhanced descriptions in openssl.pod.
1248 - Added style conformance for all options (with help from Richard Levitte),
1249 documented all reported missing options, added a CI build to check
1250 that all options are documented and that no unimplemented options
1252 - Documented some internals, such as all use of environment variables.
1253 - Addressed all internal broken L<> references.
1257 * All of the low level CMAC functions have been deprecated.
1261 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1262 functions have been deprecated.
1264 *Paul Dale and David von Oheimb*
1266 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1267 set of functions. The documentation mentioned negative values for some
1268 errors, but this was never the case, so the mention of negative values
1271 Code that followed the documentation and thereby check with something
1272 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1276 * All of the low level cipher functions have been deprecated.
1278 *Matt Caswell and Paul Dale*
1280 * Removed include/openssl/opensslconf.h.in and replaced it with
1281 include/openssl/configuration.h.in, which differs in not including
1282 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1283 was added to include both.
1285 This allows internal hacks where one might need to modify the set
1286 of configured macros, for example this if deprecated symbols are
1287 still supposed to be available internally:
1289 #include <openssl/configuration.h>
1291 #undef OPENSSL_NO_DEPRECATED
1292 #define OPENSSL_SUPPRESS_DEPRECATED
1294 #include <openssl/macros.h>
1296 This should not be used by applications that use the exported
1297 symbols, as that will lead to linking errors.
1301 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1302 used in exponentiation with 512-bit moduli. No EC algorithms are
1303 affected. Analysis suggests that attacks against 2-prime RSA1024,
1304 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1305 difficult to perform and are not believed likely. Attacks against DH512
1306 are considered just feasible. However, for an attack the target would
1307 have to re-use the DH512 private key, which is not recommended anyway.
1308 Also applications directly using the low-level API BN_mod_exp may be
1309 affected if they use BN_FLG_CONSTTIME.
1314 * Most memory-debug features have been deprecated, and the functionality
1315 replaced with no-ops.
1319 * Added documentation for the STACK API.
1323 * Introduced a new method type and API, OSSL_ENCODER, to represent
1324 generic encoders. These do the same sort of job that PEM writers
1325 and d2i functions do, but with support for methods supplied by
1326 providers, and the possibility for providers to support other
1331 * Introduced a new method type and API, OSSL_DECODER, to represent
1332 generic decoders. These do the same sort of job that PEM readers
1333 and i2d functions do, but with support for methods supplied by
1334 providers, and the possibility for providers to support other
1339 * Added a .pragma directive to the syntax of configuration files, to
1340 allow varying behavior in a supported and predictable manner.
1341 Currently added pragma:
1345 This allows dollar signs to be a keyword character unless it's
1346 followed by a opening brace or parenthesis. This is useful for
1347 platforms where dollar signs are commonly used in names, such as
1348 volume names and system directory names on VMS.
1352 * Added functionality to create an EVP_PKEY from user data.
1356 * Change the interpretation of the '--api' configuration option to
1357 mean that this is a desired API compatibility level with no
1358 further meaning. The previous interpretation, that this would
1359 also mean to remove all deprecated symbols up to and including
1360 the given version, no requires that 'no-deprecated' is also used
1361 in the configuration.
1363 When building applications, the desired API compatibility level
1364 can be set with the OPENSSL_API_COMPAT macro like before. For
1365 API compatibility version below 3.0, the old style numerical
1366 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1367 For version 3.0 and on, the value is expected to be the decimal
1368 value calculated from the major and minor version like this:
1370 MAJOR * 10000 + MINOR * 100
1374 -DOPENSSL_API_COMPAT=30000 For 3.0
1375 -DOPENSSL_API_COMPAT=30200 For 3.2
1377 To hide declarations that are deprecated up to and including the
1378 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1379 given when building the application as well.
1383 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1384 access to certificate and CRL stores via URIs and OSSL_STORE
1387 This adds the following functions:
1389 - X509_LOOKUP_store()
1390 - X509_STORE_load_file()
1391 - X509_STORE_load_path()
1392 - X509_STORE_load_store()
1393 - SSL_add_store_cert_subjects_to_stack()
1394 - SSL_CTX_set_default_verify_store()
1395 - SSL_CTX_load_verify_file()
1396 - SSL_CTX_load_verify_dir()
1397 - SSL_CTX_load_verify_store()
1401 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1402 The presence of this system service is determined at run-time.
1406 * Added functionality to create an EVP_PKEY context based on data
1407 for methods from providers. This takes an algorithm name and a
1408 property query string and simply stores them, with the intent
1409 that any operation that uses this context will use those strings
1410 to fetch the needed methods implicitly, thereby making the port
1411 of application written for pre-3.0 OpenSSL easier.
1415 * The undocumented function NCONF_WIN32() has been deprecated; for
1416 conversion details see the HISTORY section of doc/man5/config.pod
1420 * Introduced the new functions EVP_DigestSignInit_ex() and
1421 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1422 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1423 pages for further details.
1427 * Over two thousand fixes were made to the documentation, including:
1428 adding missing command flags, better style conformance, documentation
1431 *Rich Salz, Richard Levitte*
1433 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1434 X25519, X448, Ed25519 and Ed448.
1438 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1443 * Deprecated the public definition of `ERR_STATE` as well as the function
1444 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1449 * Added ERR functionality to give callers access to the stored function
1450 names that have replaced the older function code based functions.
1452 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1453 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1454 ERR_peek_error_all() and ERR_peek_last_error_all().
1456 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1457 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1458 ERR_func_error_string().
1462 * Extended testing to be verbose for failing tests only. The make variables
1463 VERBOSE_FAILURE or VF can be used to enable this:
1465 $ make VF=1 test # Unix
1466 $ mms /macro=(VF=1) test ! OpenVMS
1467 $ nmake VF=1 test # Windows
1471 * Added the `-copy_extensions` option to the `x509` command for use with
1472 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1473 all extensions in the request are copied to the certificate or vice versa.
1475 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1477 * Added the `-copy_extensions` option to the `req` command for use with
1478 `-x509`. When given with the `copy` or `copyall` argument,
1479 all extensions in the certification request are copied to the certificate.
1483 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1484 they generate are by default RFC 5280 compliant in the following sense:
1485 There is a subjectKeyIdentifier extension with a hash value of the public key
1486 and for not self-signed certs there is an authorityKeyIdentifier extension
1487 with a keyIdentifier field or issuer information identifying the signing key.
1488 This is done unless some configuration overrides the new default behavior,
1489 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1493 * Added several checks to `X509_verify_cert()` according to requirements in
1494 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1495 (which may be done by using the CLI option `-x509_strict`):
1496 * The basicConstraints of CA certificates must be marked critical.
1497 * CA certificates must explicitly include the keyUsage extension.
1498 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1499 * The issuer name of any certificate must not be empty.
1500 * The subject name of CA certs, certs with keyUsage crlSign,
1501 and certs without subjectAlternativeName must not be empty.
1502 * If a subjectAlternativeName extension is given it must not be empty.
1503 * The signatureAlgorithm field and the cert signature must be consistent.
1504 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1505 must not be marked critical.
1506 * The authorityKeyIdentifier must be given for X.509v3 certs
1507 unless they are self-signed.
1508 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1512 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1513 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1517 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1518 used even when parsing explicit parameters, when loading a encoded key
1519 or calling `EC_GROUP_new_from_ecpkparameters()`/
1520 `EC_GROUP_new_from_ecparameters()`.
1521 This prevents bypass of security hardening and performance gains,
1522 especially for curves with specialized EC_METHODs.
1523 By default, if a key encoded with explicit parameters is loaded and later
1524 encoded, the output is still encoded with explicit parameters, even if
1525 internally a "named" EC_GROUP is used for computation.
1529 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1530 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1531 NULL. After this change, only the cofactor parameter can be NULL. It also
1532 does some minimal sanity checks on the passed order.
1537 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1538 An attack is simple, if the first CMS_recipientInfo is valid but the
1539 second CMS_recipientInfo is chosen ciphertext. If the second
1540 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1541 encryption key will be replaced by garbage, and the message cannot be
1542 decoded, but if the RSA decryption fails, the correct encryption key is
1543 used and the recipient will not notice the attack.
1544 As a work around for this potential attack the length of the decrypted
1545 key must be equal to the cipher default key length, in case the
1546 certifiate is not given and all recipientInfo are tried out.
1547 The old behaviour can be re-enabled in the CMS code by setting the
1548 CMS_DEBUG_DECRYPT flag.
1552 * Early start up entropy quality from the DEVRANDOM seed source has been
1553 improved for older Linux systems. The RAND subsystem will wait for
1554 /dev/random to be producing output before seeding from /dev/urandom.
1555 The seeded state is stored for future library initialisations using
1556 a system global shared memory segment. The shared memory identifier
1557 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1558 the desired value. The default identifier is 114.
1562 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1563 when primes for RSA keys are computed.
1564 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1565 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1566 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1567 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1568 This avoids possible fingerprinting of newly generated RSA modules.
1572 * Correct the extended master secret constant on EBCDIC systems. Without this
1573 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1574 negotiate EMS will fail. Unfortunately this also means that TLS connections
1575 between EBCDIC systems with this fix, and EBCDIC systems without this
1576 fix will fail if they negotiate EMS.
1580 * Changed the library initialisation so that the config file is now loaded
1581 by default. This was already the case for libssl. It now occurs for both
1582 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1583 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1587 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1588 where the former acts as a replacement for `ERR_put_error()`, and the
1589 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1590 `ERR_raise_data()` adds more flexibility by taking a format string and
1591 an arbitrary number of arguments following it, to be processed with
1596 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1597 to check if a named provider is loaded and available. When called, it
1598 will also activate all fallback providers if such are still present.
1602 * Enforce a minimum DH modulus size of 512 bits.
1606 * Changed DH parameters to generate the order q subgroup instead of 2q.
1607 Previously generated DH parameters are still accepted by DH_check
1608 but DH_generate_key works around that by clearing bit 0 of the
1609 private key for those. This avoids leaking bit 0 of the private key.
1613 * Significantly reduce secure memory usage by the randomness pools.
1617 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1622 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1623 algorithms. An implementation of a key exchange algorithm can be obtained
1624 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1625 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1626 the older EVP_PKEY_derive_init() function. See the man pages for the new
1627 functions for further details.
1631 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1635 * Removed the function names from error messages and deprecated the
1640 * Removed NextStep support and the macro OPENSSL_UNISTD
1644 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1645 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1646 Also removed "export var as function" capability; we do not export
1647 variables, only functions.
1651 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1652 an error and 1 indicating success. In previous versions of OpenSSL this
1653 was a void type. If a key was set longer than the maximum possible this
1658 * Support SM2 signing and verification schemes with X509 certificate.
1662 * Use SHA256 as the default digest for TS query in the `ts` app.
1666 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1670 * Default cipher lists/suites are now available via a function, the
1671 #defines are deprecated.
1675 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1676 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1677 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1681 * Join the directories crypto/x509 and crypto/x509v3
1685 * Added command 'openssl kdf' that uses the EVP_KDF API.
1689 * Added command 'openssl mac' that uses the EVP_MAC API.
1693 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1694 as default directories. Also added the command 'openssl info'
1695 for scripting purposes.
1699 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1704 * Add prediction resistance to the DRBG reseeding process.
1708 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1709 mandated by IEEE Std 1619-2018.
1713 * Added newline escaping functionality to a filename when using openssl dgst.
1714 This output format is to replicate the output format found in the `*sum`
1715 checksum programs. This aims to preserve backward compatibility.
1717 *Matt Eaton, Richard Levitte, and Paul Dale*
1719 * Removed the heartbeat message in DTLS feature, as it has very
1720 little usage and doesn't seem to fulfill a valuable purpose.
1721 The configuration option is now deprecated.
1725 * Changed the output of 'openssl {digestname} < file' to display the
1726 digest name in its output.
1730 * Added a new generic trace API which provides support for enabling
1731 instrumentation through trace output.
1733 *Richard Levitte & Matthias St. Pierre*
1735 * Added build tests for C++. These are generated files that only do one
1736 thing, to include one public OpenSSL head file each. This tests that
1737 the public header files can be usefully included in a C++ application.
1739 This test isn't enabled by default. It can be enabled with the option
1740 'enable-buildtest-c++'.
1744 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1748 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1752 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1756 * Added KMAC to EVP_MAC.
1760 * Added property based algorithm implementation selection framework to
1765 * Added SCA hardening for modular field inversion in EC_GROUP through
1766 a new dedicated field_inv() pointer in EC_METHOD.
1767 This also addresses a leakage affecting conversions from projective
1768 to affine coordinates.
1770 *Billy Bob Brumley, Nicola Tuveri*
1772 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1773 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1774 those algorithms that were already supported through the EVP_PKEY API
1775 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1776 and scrypt are now wrappers that call EVP_KDF.
1780 * Build devcrypto engine as a dynamic engine.
1782 *Eneas U de Queiroz*
1784 * Add keyed BLAKE2 to EVP_MAC.
1788 * Fix a bug in the computation of the endpoint-pair shared secret used
1789 by DTLS over SCTP. This breaks interoperability with older versions
1790 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1791 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1792 interoperability with such broken implementations. However, enabling
1793 this switch breaks interoperability with correct implementations.
1795 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1796 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1800 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1804 * Changed the license to the Apache License v2.0.
1808 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1810 - Major releases (indicated by incrementing the MAJOR release number)
1811 may introduce incompatible API/ABI changes.
1812 - Minor releases (indicated by incrementing the MINOR release number)
1813 may introduce new features but retain API/ABI compatibility.
1814 - Patch releases (indicated by incrementing the PATCH number)
1815 are intended for bug fixes and other improvements of existing
1816 features only (like improving performance or adding documentation)
1817 and retain API/ABI compatibility.
1821 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1825 * Remove the 'dist' target and add a tarball building script. The
1826 'dist' target has fallen out of use, and it shouldn't be
1827 necessary to configure just to create a source distribution.
1831 * Recreate the OS390-Unix config target. It no longer relies on a
1832 special script like it did for OpenSSL pre-1.1.0.
1836 * Instead of having the source directories listed in Configure, add
1837 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1842 * Add GMAC to EVP_MAC.
1846 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1850 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1851 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1852 to facilitate the continued use of MACs through raw private keys in
1853 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1857 * Deprecate ECDH_KDF_X9_62().
1861 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1862 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1863 are retained for backwards compatibility.
1867 * AES-XTS mode now enforces that its two keys are different to mitigate
1868 the attacked described in "Efficient Instantiations of Tweakable
1869 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1870 Details of this attack can be obtained from:
1871 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1875 * Rename the object files, i.e. give them other names than in previous
1876 versions. Their names now include the name of the final product, as
1877 well as its type mnemonic (bin, lib, shlib).
1881 * Added new option for 'openssl list', '-objects', which will display the
1882 list of built in objects, i.e. OIDs with names.
1886 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1887 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1892 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1893 improves application performance by removing data copies and providing
1894 applications with zero-copy system calls such as sendfile and splice.
1898 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1902 * `PKCS12_parse` now maintains the order of the parsed certificates
1903 when outputting them via `*ca` (rather than reversing it).
1907 * Deprecated pthread fork support methods.
1911 * Added support for FFDHE key exchange in TLS 1.3.
1915 * Added a new concept for OpenSSL plugability: providers. This
1916 functionality is designed to replace the ENGINE API and ENGINE
1917 implementations, and to be much more dynamic, allowing provider
1918 authors to introduce new algorithms among other things, as long as
1919 there's an API that supports the algorithm type.
1921 With this concept comes a new core API for interaction between
1922 libcrypto and provider implementations. Public libcrypto functions
1923 that want to use providers do so through this core API.
1925 The main documentation for this core API is found in
1926 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1927 refer to other manuals describing the API specific for supported
1928 algorithm types (also called operations).
1935 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1937 * Avoid loading of a dynamic engine twice.
1941 * Prioritise DANE TLSA issuer certs over peer certs
1945 * Fixed random API for MacOS prior to 10.12
1947 These MacOS versions don't support the CommonCrypto APIs
1951 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1953 * Fixed an SM2 Decryption Buffer Overflow.
1955 In order to decrypt SM2 encrypted data an application is expected to
1956 call the API function EVP_PKEY_decrypt(). Typically an application will
1957 call this function twice. The first time, on entry, the "out" parameter
1958 can be NULL and, on exit, the "outlen" parameter is populated with the
1959 buffer size required to hold the decrypted plaintext. The application
1960 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1961 again, but this time passing a non-NULL value for the "out" parameter.
1963 A bug in the implementation of the SM2 decryption code means that the
1964 calculation of the buffer size required to hold the plaintext returned
1965 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1966 size required by the second call. This can lead to a buffer overflow
1967 when EVP_PKEY_decrypt() is called by the application a second time with
1968 a buffer that is too small.
1970 A malicious attacker who is able present SM2 content for decryption to
1971 an application could cause attacker chosen data to overflow the buffer
1972 by up to a maximum of 62 bytes altering the contents of other data held
1973 after the buffer, possibly changing application behaviour or causing
1974 the application to crash. The location of the buffer is application
1975 dependent but is typically heap allocated.
1980 * Fixed various read buffer overruns processing ASN.1 strings
1982 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1983 structure which contains a buffer holding the string data and a field
1984 holding the buffer length. This contrasts with normal C strings which
1985 are repesented as a buffer for the string data which is terminated
1986 with a NUL (0) byte.
1988 Although not a strict requirement, ASN.1 strings that are parsed using
1989 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1990 well as any string whose value has been set with the ASN1_STRING_set()
1991 function will additionally NUL terminate the byte array in the
1992 ASN1_STRING structure.
1994 However, it is possible for applications to directly construct valid
1995 ASN1_STRING structures which do not NUL terminate the byte array by
1996 directly setting the "data" and "length" fields in the ASN1_STRING
1997 array. This can also happen by using the ASN1_STRING_set0() function.
1999 Numerous OpenSSL functions that print ASN.1 data have been found to
2000 assume that the ASN1_STRING byte array will be NUL terminated, even
2001 though this is not guaranteed for strings that have been directly
2002 constructed. Where an application requests an ASN.1 structure to be
2003 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2004 been directly constructed by the application without NUL terminating
2005 the "data" field, then a read buffer overrun can occur.
2007 The same thing can also occur during name constraints processing
2008 of certificates (for example if a certificate has been directly
2009 constructed by the application instead of loading it via the OpenSSL
2010 parsing functions, and the certificate contains non NUL terminated
2011 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2012 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2014 If a malicious actor can cause an application to directly construct an
2015 ASN1_STRING and then process it through one of the affected OpenSSL
2016 functions then this issue could be hit. This might result in a crash
2017 (causing a Denial of Service attack). It could also result in the
2018 disclosure of private memory contents (such as private keys, or
2019 sensitive plaintext).
2024 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2026 * Fixed a problem with verifying a certificate chain when using the
2027 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2028 the certificates present in a certificate chain. It is not set by default.
2030 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2031 the chain that have explicitly encoded elliptic curve parameters was added
2032 as an additional strict check.
2034 An error in the implementation of this check meant that the result of a
2035 previous check to confirm that certificates in the chain are valid CA
2036 certificates was overwritten. This effectively bypasses the check
2037 that non-CA certificates must not be able to issue other certificates.
2039 If a "purpose" has been configured then there is a subsequent opportunity
2040 for checks that the certificate is a valid CA. All of the named "purpose"
2041 values implemented in libcrypto perform this check. Therefore, where
2042 a purpose is set the certificate chain will still be rejected even when the
2043 strict flag has been used. A purpose is set by default in libssl client and
2044 server certificate verification routines, but it can be overridden or
2045 removed by an application.
2047 In order to be affected, an application must explicitly set the
2048 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2049 for the certificate verification or, in the case of TLS client or server
2050 applications, override the default purpose.
2055 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2056 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2057 renegotiation ClientHello omits the signature_algorithms extension (where it
2058 was present in the initial ClientHello), but includes a
2059 signature_algorithms_cert extension then a NULL pointer dereference will
2060 result, leading to a crash and a denial of service attack.
2062 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2063 (which is the default configuration). OpenSSL TLS clients are not impacted by
2067 *Peter Kästle and Samuel Sapalski*
2069 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2071 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2072 create a unique hash value based on the issuer and serial number data
2073 contained within an X509 certificate. However it was failing to correctly
2074 handle any errors that may occur while parsing the issuer field (which might
2075 occur if the issuer field is maliciously constructed). This may subsequently
2076 result in a NULL pointer deref and a crash leading to a potential denial of
2082 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2083 padding mode to correctly check for rollback attacks. This is considered a
2084 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2089 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2090 functions. Previously they could overflow the output length argument in some
2091 cases where the input length is close to the maximum permissable length for
2092 an integer on the platform. In such cases the return value from the function
2093 call would be 1 (indicating success), but the output length value would be
2094 negative. This could cause applications to behave incorrectly or crash.
2099 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2100 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2101 could be exploited in a side channel attack to recover the password. Since
2102 the attack is local host only this is outside of the current OpenSSL
2103 threat model and therefore no CVE is assigned.
2105 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2110 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2112 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2113 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2114 If an attacker can control both items being compared then this could lead
2115 to a possible denial of service attack. OpenSSL itself uses the
2116 GENERAL_NAME_cmp function for two purposes:
2117 1) Comparing CRL distribution point names between an available CRL and a
2118 CRL distribution point embedded in an X509 certificate
2119 2) When verifying that a timestamp response token signer matches the
2120 timestamp authority name (exposed via the API functions
2121 TS_RESP_verify_response and TS_RESP_verify_token)
2126 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2128 * Certificates with explicit curve parameters are now disallowed in
2129 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2133 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2134 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2135 conversely, silently ignore DTLS protocol version bounds when configuring
2136 TLS-based contexts. The commands can be repeated to set bounds of both
2137 types. The same applies with the corresponding "min_protocol" and
2138 "max_protocol" command-line switches, in case some application uses both TLS
2141 SSL_CTX instances that are created for a fixed protocol version (e.g.
2142 TLSv1_server_method()) also silently ignore version bounds. Previously
2143 attempts to apply bounds to these protocol versions would result in an
2144 error. Now only the "version-flexible" SSL_CTX instances are subject to
2145 limits in configuration files in command-line options.
2149 * Handshake now fails if Extended Master Secret extension is dropped
2154 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2156 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2158 * Fixed segmentation fault in SSL_check_chain()
2159 Server or client applications that call the SSL_check_chain() function
2160 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2161 dereference as a result of incorrect handling of the
2162 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2163 or unrecognised signature algorithm is received from the peer. This could
2164 be exploited by a malicious peer in a Denial of Service attack.
2169 * Added AES consttime code for no-asm configurations
2170 an optional constant time support for AES was added
2171 when building openssl for no-asm.
2172 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2173 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2174 At this time this feature is by default disabled.
2175 It will be enabled by default in 3.0.
2179 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2181 * Revert the change of EOF detection while reading in libssl to avoid
2182 regressions in applications depending on the current way of reporting
2183 the EOF. As the existing method is not fully accurate the change to
2184 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2185 branch and will be present in the 3.0 release.
2189 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2190 when primes for RSA keys are computed.
2191 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2192 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2193 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2194 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2195 This avoids possible fingerprinting of newly generated RSA modules.
2199 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2201 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2202 while reading in libssl then we would report an error back to the
2203 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2204 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2205 therefore give a hint as to what went wrong.
2209 * Check that ed25519 and ed448 are allowed by the security level. Previously
2210 signature algorithms not using an MD were not being checked that they were
2211 allowed by the security level.
2215 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2216 was not quite right. The behaviour was not consistent between resumption
2217 and normal handshakes, and also not quite consistent with historical
2218 behaviour. The behaviour in various scenarios has been clarified and
2219 it has been updated to make it match historical behaviour as closely as
2224 * *[VMS only]* The header files that the VMS compilers include automatically,
2225 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2226 that the C++ compiler doesn't understand. This is a shortcoming in the
2227 compiler, but can be worked around with `__cplusplus` guards.
2229 C++ applications that use OpenSSL libraries must be compiled using the
2230 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2231 functions. Otherwise, only functions with symbols of less than 31
2232 characters can be used, as the linker will not be able to successfully
2233 resolve symbols with longer names.
2237 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2238 The presence of this system service is determined at run-time.
2242 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2247 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2249 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2250 number generator (RNG). This was intended to include protection in the
2251 event of a fork() system call in order to ensure that the parent and child
2252 processes did not share the same RNG state. However this protection was not
2253 being used in the default case.
2255 A partial mitigation for this issue is that the output from a high
2256 precision timer is mixed into the RNG state so the likelihood of a parent
2257 and child process sharing state is significantly reduced.
2259 If an application already calls OPENSSL_init_crypto() explicitly using
2260 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2263 *Matthias St. Pierre*
2265 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2266 used even when parsing explicit parameters, when loading a encoded key
2267 or calling `EC_GROUP_new_from_ecpkparameters()`/
2268 `EC_GROUP_new_from_ecparameters()`.
2269 This prevents bypass of security hardening and performance gains,
2270 especially for curves with specialized EC_METHODs.
2271 By default, if a key encoded with explicit parameters is loaded and later
2272 encoded, the output is still encoded with explicit parameters, even if
2273 internally a "named" EC_GROUP is used for computation.
2277 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2278 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2279 NULL. After this change, only the cofactor parameter can be NULL. It also
2280 does some minimal sanity checks on the passed order.
2285 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2286 An attack is simple, if the first CMS_recipientInfo is valid but the
2287 second CMS_recipientInfo is chosen ciphertext. If the second
2288 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2289 encryption key will be replaced by garbage, and the message cannot be
2290 decoded, but if the RSA decryption fails, the correct encryption key is
2291 used and the recipient will not notice the attack.
2292 As a work around for this potential attack the length of the decrypted
2293 key must be equal to the cipher default key length, in case the
2294 certifiate is not given and all recipientInfo are tried out.
2295 The old behaviour can be re-enabled in the CMS code by setting the
2296 CMS_DEBUG_DECRYPT flag.
2301 * Early start up entropy quality from the DEVRANDOM seed source has been
2302 improved for older Linux systems. The RAND subsystem will wait for
2303 /dev/random to be producing output before seeding from /dev/urandom.
2304 The seeded state is stored for future library initialisations using
2305 a system global shared memory segment. The shared memory identifier
2306 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2307 the desired value. The default identifier is 114.
2311 * Correct the extended master secret constant on EBCDIC systems. Without this
2312 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2313 negotiate EMS will fail. Unfortunately this also means that TLS connections
2314 between EBCDIC systems with this fix, and EBCDIC systems without this
2315 fix will fail if they negotiate EMS.
2319 * Use Windows installation paths in the mingw builds
2321 Mingw isn't a POSIX environment per se, which means that Windows
2322 paths should be used for installation.
2327 * Changed DH_check to accept parameters with order q and 2q subgroups.
2328 With order 2q subgroups the bit 0 of the private key is not secret
2329 but DH_generate_key works around that by clearing bit 0 of the
2330 private key for those. This avoids leaking bit 0 of the private key.
2334 * Significantly reduce secure memory usage by the randomness pools.
2338 * Revert the DEVRANDOM_WAIT feature for Linux systems
2340 The DEVRANDOM_WAIT feature added a select() call to wait for the
2341 /dev/random device to become readable before reading from the
2342 /dev/urandom device.
2344 It turned out that this change had negative side effects on
2345 performance which were not acceptable. After some discussion it
2346 was decided to revert this feature and leave it up to the OS
2347 resp. the platform maintainer to ensure a proper initialization
2348 during early boot time.
2350 *Matthias St. Pierre*
2352 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2354 * Add build tests for C++. These are generated files that only do one
2355 thing, to include one public OpenSSL head file each. This tests that
2356 the public header files can be usefully included in a C++ application.
2358 This test isn't enabled by default. It can be enabled with the option
2359 'enable-buildtest-c++'.
2363 * Enable SHA3 pre-hashing for ECDSA and DSA.
2367 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2368 This changes the size when using the `genpkey` command when no size is given.
2369 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2370 generation commands to use 2048 bits by default.
2374 * Reorganize the manual pages to consistently have RETURN VALUES,
2375 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2376 util/fix-doc-nits accordingly.
2378 *Paul Yang, Joshua Lock*
2380 * Add the missing accessor EVP_PKEY_get0_engine()
2384 * Have commands like `s_client` and `s_server` output the signature scheme
2385 along with other cipher suite parameters when debugging.
2389 * Make OPENSSL_config() error agnostic again.
2393 * Do the error handling in RSA decryption constant time.
2397 * Prevent over long nonces in ChaCha20-Poly1305.
2399 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2400 for every encryption operation. RFC 7539 specifies that the nonce value
2401 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2402 and front pads the nonce with 0 bytes if it is less than 12
2403 bytes. However it also incorrectly allows a nonce to be set of up to 16
2404 bytes. In this case only the last 12 bytes are significant and any
2405 additional leading bytes are ignored.
2407 It is a requirement of using this cipher that nonce values are
2408 unique. Messages encrypted using a reused nonce value are susceptible to
2409 serious confidentiality and integrity attacks. If an application changes
2410 the default nonce length to be longer than 12 bytes and then makes a
2411 change to the leading bytes of the nonce expecting the new value to be a
2412 new unique nonce then such an application could inadvertently encrypt
2413 messages with a reused nonce.
2415 Additionally the ignored bytes in a long nonce are not covered by the
2416 integrity guarantee of this cipher. Any application that relies on the
2417 integrity of these ignored leading bytes of a long nonce may be further
2418 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2419 is safe because no such use sets such a long nonce value. However user
2420 applications that use this cipher directly and set a non-default nonce
2421 length to be longer than 12 bytes may be vulnerable.
2423 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2429 * Add DEVRANDOM_WAIT feature for Linux systems
2431 On older Linux systems where the getrandom() system call is not available,
2432 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2433 Contrary to getrandom(), the /dev/urandom device will not block during
2434 early boot when the kernel CSPRNG has not been seeded yet.
2436 To mitigate this known weakness, use select() to wait for /dev/random to
2437 become readable before reading from /dev/urandom.
2439 * Ensure that SM2 only uses SM3 as digest algorithm
2443 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2445 * Change the info callback signals for the start and end of a post-handshake
2446 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2447 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2448 confused by this and assume that a TLSv1.2 renegotiation has started. This
2449 can break KeyUpdate handling. Instead we no longer signal the start and end
2450 of a post handshake message exchange (although the messages themselves are
2451 still signalled). This could break some applications that were expecting
2452 the old signals. However without this KeyUpdate is not usable for many
2457 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2459 * Timing vulnerability in DSA signature generation
2461 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2462 timing side channel attack. An attacker could use variations in the signing
2463 algorithm to recover the private key.
2465 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2470 * Timing vulnerability in ECDSA signature generation
2472 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2473 timing side channel attack. An attacker could use variations in the signing
2474 algorithm to recover the private key.
2476 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2481 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2482 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2483 of two gigabytes and the error handling improved.
2485 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2486 categorized as a normal bug, not a security issue, because the DRBG reseeds
2487 automatically and is fully functional even without additional randomness
2488 provided by the application.
2490 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2492 * Add a new ClientHello callback. Provides a callback interface that gives
2493 the application the ability to adjust the nascent SSL object at the
2494 earliest stage of ClientHello processing, immediately after extensions have
2495 been collected but before they have been processed. In particular, this
2496 callback can adjust the supported TLS versions in response to the contents
2501 * Add SM2 base algorithm support.
2505 * s390x assembly pack: add (improved) hardware-support for the following
2506 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2507 aes-cfb/cfb8, aes-ecb.
2511 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2512 parameter is no longer accepted, as it leads to a corrupt table. NULL
2513 pem_str is reserved for alias entries only.
2517 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2518 step for prime curves. The new implementation is based on formulae from
2519 differential addition-and-doubling in homogeneous projective coordinates
2520 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2521 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2522 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2523 to work in projective coordinates.
2525 *Billy Bob Brumley, Nicola Tuveri*
2527 * Change generating and checking of primes so that the error rate of not
2528 being prime depends on the intended use based on the size of the input.
2529 For larger primes this will result in more rounds of Miller-Rabin.
2530 The maximal error rate for primes with more than 1080 bits is lowered
2533 *Kurt Roeckx, Annie Yousar*
2535 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2539 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2540 moving between systems, and to avoid confusion when a Windows build is
2541 done with mingw vs with MSVC. For POSIX installs, there's still a
2542 symlink or copy named 'tsget' to avoid that confusion as well.
2546 * Revert blinding in ECDSA sign and instead make problematic addition
2547 length-invariant. Switch even to fixed-length Montgomery multiplication.
2551 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2552 step for binary curves. The new implementation is based on formulae from
2553 differential addition-and-doubling in mixed Lopez-Dahab projective
2554 coordinates, modified to independently blind the operands.
2556 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2558 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2559 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2560 EC_METHODs to implement their own specialized "ladder step", to take
2561 advantage of more favorable coordinate systems or more efficient
2562 differential addition-and-doubling algorithms.
2564 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2566 * Modified the random device based seed sources to keep the relevant
2567 file descriptors open rather than reopening them on each access.
2568 This allows such sources to operate in a chroot() jail without
2569 the associated device nodes being available. This behaviour can be
2570 controlled using RAND_keep_random_devices_open().
2574 * Numerous side-channel attack mitigations have been applied. This may have
2575 performance impacts for some algorithms for the benefit of improved
2576 security. Specific changes are noted in this change log by their respective
2581 * AIX shared library support overhaul. Switch to AIX "natural" way of
2582 handling shared libraries, which means collecting shared objects of
2583 different versions and bitnesses in one common archive. This allows to
2584 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2585 doesn't affect the way 3rd party applications are linked, only how
2586 multi-version installation is managed.
2590 * Make ec_group_do_inverse_ord() more robust and available to other
2591 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2592 mitigations are applied to the fallback BN_mod_inverse().
2593 When using this function rather than BN_mod_inverse() directly, new
2594 EC cryptosystem implementations are then safer-by-default.
2598 * Add coordinate blinding for EC_POINT and implement projective
2599 coordinate blinding for generic prime curves as a countermeasure to
2600 chosen point SCA attacks.
2602 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2604 * Add blinding to ECDSA and DSA signatures to protect against side channel
2605 attacks discovered by Keegan Ryan (NCC Group).
2609 * Enforce checking in the `pkeyutl` command to ensure that the input
2610 length does not exceed the maximum supported digest length when performing
2611 a sign, verify or verifyrecover operation.
2615 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2616 I/O in combination with something like select() or poll() will hang. This
2617 can be turned off again using SSL_CTX_clear_mode().
2618 Many applications do not properly handle non-application data records, and
2619 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2620 around the problems in those applications, but can also break some.
2621 It's recommended to read the manpages about SSL_read(), SSL_write(),
2622 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2623 SSL_CTX_set_read_ahead() again.
2627 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2628 now allow empty (zero character) pass phrases.
2632 * Apply blinding to binary field modular inversion and remove patent
2633 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2637 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2638 binary and prime elliptic curves.
2642 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2643 constant time fixed point multiplication.
2647 * Revise elliptic curve scalar multiplication with timing attack
2648 defenses: ec_wNAF_mul redirects to a constant time implementation
2649 when computing fixed point and variable point multiplication (which
2650 in OpenSSL are mostly used with secret scalars in keygen, sign,
2651 ECDH derive operations).
2652 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2655 * Updated CONTRIBUTING
2659 * Updated DRBG / RAND to request nonce and additional low entropy
2660 randomness from the system.
2662 *Matthias St. Pierre*
2664 * Updated 'openssl rehash' to use OpenSSL consistent default.
2668 * Moved the load of the ssl_conf module to libcrypto, which helps
2669 loading engines that libssl uses before libssl is initialised.
2673 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2677 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2679 *Ingo Schwarze, Rich Salz*
2681 * Added output of accepting IP address and port for 'openssl s_server'
2685 * Added a new API for TLSv1.3 ciphersuites:
2686 SSL_CTX_set_ciphersuites()
2687 SSL_set_ciphersuites()
2691 * Memory allocation failures consistently add an error to the error
2696 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2697 in libcrypto when run as setuid/setgid.
2701 * Load any config file by default when libssl is used.
2705 * Added new public header file <openssl/rand_drbg.h> and documentation
2706 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2708 *Matthias St. Pierre*
2710 * QNX support removed (cannot find contributors to get their approval
2711 for the license change).
2715 * TLSv1.3 replay protection for early data has been implemented. See the
2716 SSL_read_early_data() man page for further details.
2720 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2721 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2722 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2723 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2724 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2725 configuration has been separated out. See the ciphers man page or the
2726 SSL_CTX_set_ciphersuites() man page for more information.
2730 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2731 in responder mode now supports the new "-multi" option, which
2732 spawns the specified number of child processes to handle OCSP
2733 requests. The "-timeout" option now also limits the OCSP
2734 responder's patience to wait to receive the full client request
2735 on a newly accepted connection. Child processes are respawned
2736 as needed, and the CA index file is automatically reloaded
2737 when changed. This makes it possible to run the "ocsp" responder
2738 as a long-running service, making the OpenSSL CA somewhat more
2739 feature-complete. In this mode, most diagnostic messages logged
2740 after entering the event loop are logged via syslog(3) rather than
2745 * Added support for X448 and Ed448. Heavily based on original work by
2750 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2751 objects loaded. This adds the functions OSSL_STORE_expect() and
2752 OSSL_STORE_find() as well as needed tools to construct searches and
2753 get the search data out of them.
2757 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2758 version of OpenSSL should review their configuration settings to ensure
2759 that they are still appropriate for TLSv1.3. For further information see:
2760 <https://wiki.openssl.org/index.php/TLS1.3>
2764 * Grand redesign of the OpenSSL random generator
2766 The default RAND method now utilizes an AES-CTR DRBG according to
2767 NIST standard SP 800-90Ar1. The new random generator is essentially
2768 a port of the default random generator from the OpenSSL FIPS 2.0
2769 object module. It is a hybrid deterministic random bit generator
2770 using an AES-CTR bit stream and which seeds and reseeds itself
2771 automatically using trusted system entropy sources.
2773 Some of its new features are:
2774 - Support for multiple DRBG instances with seed chaining.
2775 - The default RAND method makes use of a DRBG.
2776 - There is a public and private DRBG instance.
2777 - The DRBG instances are fork-safe.
2778 - Keep all global DRBG instances on the secure heap if it is enabled.
2779 - The public and private DRBG instance are per thread for lock free
2782 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2784 * Changed Configure so it only says what it does and doesn't dump
2785 so much data. Instead, ./configdata.pm should be used as a script
2786 to display all sorts of configuration data.
2790 * Added processing of "make variables" to Configure.
2794 * Added SHA512/224 and SHA512/256 algorithm support.
2798 * The last traces of Netware support, first removed in 1.1.0, have
2803 * Get rid of Makefile.shared, and in the process, make the processing
2804 of certain files (rc.obj, or the .def/.map/.opt files produced from
2805 the ordinal files) more visible and hopefully easier to trace and
2806 debug (or make silent).
2810 * Make it possible to have environment variable assignments as
2811 arguments to config / Configure.
2815 * Add multi-prime RSA (RFC 8017) support.
2819 * Add SM3 implemented according to GB/T 32905-2016
2820 *Jack Lloyd <jack.lloyd@ribose.com>,*
2821 *Ronald Tse <ronald.tse@ribose.com>,*
2822 *Erick Borsboom <erick.borsboom@ribose.com>*
2824 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2825 as documented in RFC6066.
2826 Based on a patch from Tomasz Moń
2828 *Filipe Raimundo da Silva*
2830 * Add SM4 implemented according to GB/T 32907-2016.
2831 *Jack Lloyd <jack.lloyd@ribose.com>,*
2832 *Ronald Tse <ronald.tse@ribose.com>,*
2833 *Erick Borsboom <erick.borsboom@ribose.com>*
2835 * Reimplement -newreq-nodes and ERR_error_string_n; the
2836 original author does not agree with the license change.
2840 * Add ARIA AEAD TLS support.
2844 * Some macro definitions to support VS6 have been removed. Visual
2845 Studio 6 has not worked since 1.1.0
2849 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2850 without clearing the errors.
2854 * Add "atfork" functions. If building on a system that without
2855 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2856 requirements. The RAND facility now uses/requires this.
2864 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2865 not possible to disable entirely. However, it's still possible to
2866 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2869 To disable, configure with 'no-ui-console'. 'no-ui' is still
2870 possible to use as an alias. Check at compile time with the
2871 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2872 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2876 * Add a STORE module, which implements a uniform and URI based reader of
2877 stores that can contain keys, certificates, CRLs and numerous other
2878 objects. The main API is loosely based on a few stdio functions,
2879 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2880 OSSL_STORE_error and OSSL_STORE_close.
2881 The implementation uses backends called "loaders" to implement arbitrary
2882 URI schemes. There is one built in "loader" for the 'file' scheme.
2886 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2887 then adjusted to work on FreeBSD 8.4 as well.
2888 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2889 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2893 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2894 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2895 error code calls like this:
2897 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2899 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2900 that can be encoded in C. For the foreseeable future, this will only
2903 *Richard Levitte and Tim Hudson*
2905 * Removed BSD cryptodev engine.
2909 * Add a build target 'build_all_generated', to build all generated files
2910 and only that. This can be used to prepare everything that requires
2911 things like perl for a system that lacks perl and then move everything
2912 to that system and do the rest of the build there.
2916 * In the UI interface, make it possible to duplicate the user data. This
2917 can be used by engines that need to retain the data for a longer time
2918 than just the call where this user data is passed.
2922 * Ignore the '-named_curve auto' value for compatibility of applications
2925 *Tomáš Mráz <tmraz@fedoraproject.org>*
2927 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2928 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2929 alerts across multiple records (some of which could be empty). In practice
2930 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2931 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2932 support this at all. Supporting it adds significant complexity to the
2933 record layer, and its removal is unlikely to cause interoperability
2938 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2939 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2940 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2945 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2946 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2948 *Richard Levitte, Andy Polyakov*
2950 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2955 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2956 platform rather than 'mingw'.
2960 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2961 success if they are asked to add an object which already exists
2962 in the store. This change cascades to other functions which load
2963 certificates and CRLs.
2967 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2968 facilitate stack unwinding even from assembly subroutines.
2972 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2973 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2977 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2978 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2979 which is the minimum version we support.
2983 * Certificate time validation (X509_cmp_time) enforces stricter
2984 compliance with RFC 5280. Fractional seconds and timezone offsets
2985 are no longer allowed.
2989 * Add support for ARIA
2993 * s_client will now send the Server Name Indication (SNI) extension by
2994 default unless the new "-noservername" option is used. The server name is
2995 based on the host provided to the "-connect" option unless overridden by
2996 using "-servername".
3000 * Add support for SipHash
3004 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3005 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3006 prevent issues where no progress is being made and the peer continually
3007 sends unrecognised record types, using up resources processing them.
3011 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3012 using the algorithm defined in
3013 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3017 * Heartbeat support has been removed; the ABI is changed for now.
3019 *Richard Levitte, Rich Salz*
3021 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3025 * The RSA "null" method, which was partially supported to avoid patent
3026 issues, has been replaced to always returns NULL.
3033 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3035 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3036 used even when parsing explicit parameters, when loading a encoded key
3037 or calling `EC_GROUP_new_from_ecpkparameters()`/
3038 `EC_GROUP_new_from_ecparameters()`.
3039 This prevents bypass of security hardening and performance gains,
3040 especially for curves with specialized EC_METHODs.
3041 By default, if a key encoded with explicit parameters is loaded and later
3042 encoded, the output is still encoded with explicit parameters, even if
3043 internally a "named" EC_GROUP is used for computation.
3047 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3048 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3049 NULL. After this change, only the cofactor parameter can be NULL. It also
3050 does some minimal sanity checks on the passed order.
3055 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3056 An attack is simple, if the first CMS_recipientInfo is valid but the
3057 second CMS_recipientInfo is chosen ciphertext. If the second
3058 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3059 encryption key will be replaced by garbage, and the message cannot be
3060 decoded, but if the RSA decryption fails, the correct encryption key is
3061 used and the recipient will not notice the attack.
3062 As a work around for this potential attack the length of the decrypted
3063 key must be equal to the cipher default key length, in case the
3064 certifiate is not given and all recipientInfo are tried out.
3065 The old behaviour can be re-enabled in the CMS code by setting the
3066 CMS_DEBUG_DECRYPT flag.
3071 * Use Windows installation paths in the mingw builds
3073 Mingw isn't a POSIX environment per se, which means that Windows
3074 paths should be used for installation.
3079 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3081 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3082 This changes the size when using the `genpkey` command when no size is given.
3083 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3084 generation commands to use 2048 bits by default.
3088 * Prevent over long nonces in ChaCha20-Poly1305.
3090 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3091 for every encryption operation. RFC 7539 specifies that the nonce value
3092 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3093 and front pads the nonce with 0 bytes if it is less than 12
3094 bytes. However it also incorrectly allows a nonce to be set of up to 16
3095 bytes. In this case only the last 12 bytes are significant and any
3096 additional leading bytes are ignored.
3098 It is a requirement of using this cipher that nonce values are
3099 unique. Messages encrypted using a reused nonce value are susceptible to
3100 serious confidentiality and integrity attacks. If an application changes
3101 the default nonce length to be longer than 12 bytes and then makes a
3102 change to the leading bytes of the nonce expecting the new value to be a
3103 new unique nonce then such an application could inadvertently encrypt
3104 messages with a reused nonce.
3106 Additionally the ignored bytes in a long nonce are not covered by the
3107 integrity guarantee of this cipher. Any application that relies on the
3108 integrity of these ignored leading bytes of a long nonce may be further
3109 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3110 is safe because no such use sets such a long nonce value. However user
3111 applications that use this cipher directly and set a non-default nonce
3112 length to be longer than 12 bytes may be vulnerable.
3114 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3120 * Added SCA hardening for modular field inversion in EC_GROUP through
3121 a new dedicated field_inv() pointer in EC_METHOD.
3122 This also addresses a leakage affecting conversions from projective
3123 to affine coordinates.
3125 *Billy Bob Brumley, Nicola Tuveri*
3127 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3128 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3132 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3136 * Remove the 'dist' target and add a tarball building script. The
3137 'dist' target has fallen out of use, and it shouldn't be
3138 necessary to configure just to create a source distribution.
3142 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3144 * Timing vulnerability in DSA signature generation
3146 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3147 timing side channel attack. An attacker could use variations in the signing
3148 algorithm to recover the private key.
3150 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3155 * Timing vulnerability in ECDSA signature generation
3157 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3158 timing side channel attack. An attacker could use variations in the signing
3159 algorithm to recover the private key.
3161 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3166 * Add coordinate blinding for EC_POINT and implement projective
3167 coordinate blinding for generic prime curves as a countermeasure to
3168 chosen point SCA attacks.
3170 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3172 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3174 * Client DoS due to large DH parameter
3176 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3177 malicious server can send a very large prime value to the client. This will
3178 cause the client to spend an unreasonably long period of time generating a
3179 key for this prime resulting in a hang until the client has finished. This
3180 could be exploited in a Denial Of Service attack.
3182 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3187 * Cache timing vulnerability in RSA Key Generation
3189 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3190 a cache timing side channel attack. An attacker with sufficient access to
3191 mount cache timing attacks during the RSA key generation process could
3192 recover the private key.
3194 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3195 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3200 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3201 parameter is no longer accepted, as it leads to a corrupt table. NULL
3202 pem_str is reserved for alias entries only.
3206 * Revert blinding in ECDSA sign and instead make problematic addition
3207 length-invariant. Switch even to fixed-length Montgomery multiplication.
3211 * Change generating and checking of primes so that the error rate of not
3212 being prime depends on the intended use based on the size of the input.
3213 For larger primes this will result in more rounds of Miller-Rabin.
3214 The maximal error rate for primes with more than 1080 bits is lowered
3217 *Kurt Roeckx, Annie Yousar*
3219 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3223 * Add blinding to ECDSA and DSA signatures to protect against side channel
3224 attacks discovered by Keegan Ryan (NCC Group).
3228 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3229 now allow empty (zero character) pass phrases.
3233 * Certificate time validation (X509_cmp_time) enforces stricter
3234 compliance with RFC 5280. Fractional seconds and timezone offsets
3235 are no longer allowed.
3239 * Fixed a text canonicalisation bug in CMS
3241 Where a CMS detached signature is used with text content the text goes
3242 through a canonicalisation process first prior to signing or verifying a
3243 signature. This process strips trailing space at the end of lines, converts
3244 line terminators to CRLF and removes additional trailing line terminators
3245 at the end of a file. A bug in the canonicalisation process meant that
3246 some characters, such as form-feed, were incorrectly treated as whitespace
3247 and removed. This is contrary to the specification (RFC5485). This fix
3248 could mean that detached text data signed with an earlier version of
3249 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3250 signed with a fixed OpenSSL may fail to verify with an earlier version of
3251 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3252 and use the "-binary" flag (for the "cms" command line application) or set
3253 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3257 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3259 * Constructed ASN.1 types with a recursive definition could exceed the stack
3261 Constructed ASN.1 types with a recursive definition (such as can be found
3262 in PKCS7) could eventually exceed the stack given malicious input with
3263 excessive recursion. This could result in a Denial Of Service attack. There
3264 are no such structures used within SSL/TLS that come from untrusted sources
3265 so this is considered safe.
3267 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3273 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3275 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3276 effectively reduced to only comparing the least significant bit of each
3277 byte. This allows an attacker to forge messages that would be considered as
3278 authenticated in an amount of tries lower than that guaranteed by the
3279 security claims of the scheme. The module can only be compiled by the
3280 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3282 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3288 * Add a build target 'build_all_generated', to build all generated files
3289 and only that. This can be used to prepare everything that requires
3290 things like perl for a system that lacks perl and then move everything
3291 to that system and do the rest of the build there.
3295 * Backport SSL_OP_NO_RENGOTIATION
3297 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3298 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3299 changes this is no longer possible in 1.1.0. Therefore the new
3300 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3301 1.1.0 to provide equivalent functionality.
3303 Note that if an application built against 1.1.0h headers (or above) is run
3304 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3305 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3309 * Removed the OS390-Unix config target. It relied on a script that doesn't
3314 * rsaz_1024_mul_avx2 overflow bug on x86_64
3316 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3317 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3318 Analysis suggests that attacks against RSA and DSA as a result of this
3319 defect would be very difficult to perform and are not believed likely.
3320 Attacks against DH1024 are considered just feasible, because most of the
3321 work necessary to deduce information about a private key may be performed
3322 offline. The amount of resources required for such an attack would be
3323 significant. However, for an attack on TLS to be meaningful, the server
3324 would have to share the DH1024 private key among multiple clients, which is
3325 no longer an option since CVE-2016-0701.
3327 This only affects processors that support the AVX2 but not ADX extensions
3328 like Intel Haswell (4th generation).
3330 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3331 was originally found via the OSS-Fuzz project.
3336 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3338 * bn_sqrx8x_internal carry bug on x86_64
3340 There is a carry propagating bug in the x86_64 Montgomery squaring
3341 procedure. No EC algorithms are affected. Analysis suggests that attacks
3342 against RSA and DSA as a result of this defect would be very difficult to
3343 perform and are not believed likely. Attacks against DH are considered just
3344 feasible (although very difficult) because most of the work necessary to
3345 deduce information about a private key may be performed offline. The amount
3346 of resources required for such an attack would be very significant and
3347 likely only accessible to a limited number of attackers. An attacker would
3348 additionally need online access to an unpatched system using the target
3349 private key in a scenario with persistent DH parameters and a private
3350 key that is shared between multiple clients.
3352 This only affects processors that support the BMI1, BMI2 and ADX extensions
3353 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3355 This issue was reported to OpenSSL by the OSS-Fuzz project.
3360 * Malformed X.509 IPAddressFamily could cause OOB read
3362 If an X.509 certificate has a malformed IPAddressFamily extension,
3363 OpenSSL could do a one-byte buffer overread. The most likely result
3364 would be an erroneous display of the certificate in text format.
3366 This issue was reported to OpenSSL by the OSS-Fuzz project.
3371 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3373 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3374 platform rather than 'mingw'.
3378 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3379 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3380 which is the minimum version we support.
3384 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3386 * Encrypt-Then-Mac renegotiation crash
3388 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3389 negotiated where it was not in the original handshake (or vice-versa) then
3390 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3391 and servers are affected.
3393 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3398 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3400 * Truncated packet could crash via OOB read
3402 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3403 cipher is being used, then a truncated packet can cause that host to
3404 perform an out-of-bounds read, usually resulting in a crash.
3406 This issue was reported to OpenSSL by Robert Święcki of Google.
3411 * Bad (EC)DHE parameters cause a client crash
3413 If a malicious server supplies bad parameters for a DHE or ECDHE key
3414 exchange then this can result in the client attempting to dereference a
3415 NULL pointer leading to a client crash. This could be exploited in a Denial
3418 This issue was reported to OpenSSL by Guido Vranken.
3423 * BN_mod_exp may produce incorrect results on x86_64
3425 There is a carry propagating bug in the x86_64 Montgomery squaring
3426 procedure. No EC algorithms are affected. Analysis suggests that attacks
3427 against RSA and DSA as a result of this defect would be very difficult to
3428 perform and are not believed likely. Attacks against DH are considered just
3429 feasible (although very difficult) because most of the work necessary to
3430 deduce information about a private key may be performed offline. The amount
3431 of resources required for such an attack would be very significant and
3432 likely only accessible to a limited number of attackers. An attacker would
3433 additionally need online access to an unpatched system using the target
3434 private key in a scenario with persistent DH parameters and a private
3435 key that is shared between multiple clients. For example this can occur by
3436 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3437 similar to CVE-2015-3193 but must be treated as a separate problem.
3439 This issue was reported to OpenSSL by the OSS-Fuzz project.
3444 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3446 * ChaCha20/Poly1305 heap-buffer-overflow
3448 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3449 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3450 crash. This issue is not considered to be exploitable beyond a DoS.
3452 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3457 * CMS Null dereference
3459 Applications parsing invalid CMS structures can crash with a NULL pointer
3460 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3461 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3462 structure callback if an attempt is made to free certain invalid encodings.
3463 Only CHOICE structures using a callback which do not handle NULL value are
3466 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3471 * Montgomery multiplication may produce incorrect results
3473 There is a carry propagating bug in the Broadwell-specific Montgomery
3474 multiplication procedure that handles input lengths divisible by, but
3475 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3476 and DH private keys are impossible. This is because the subroutine in
3477 question is not used in operations with the private key itself and an input
3478 of the attacker's direct choice. Otherwise the bug can manifest itself as
3479 transient authentication and key negotiation failures or reproducible
3480 erroneous outcome of public-key operations with specially crafted input.
3481 Among EC algorithms only Brainpool P-512 curves are affected and one
3482 presumably can attack ECDH key negotiation. Impact was not analyzed in
3483 detail, because pre-requisites for attack are considered unlikely. Namely
3484 multiple clients have to choose the curve in question and the server has to
3485 share the private key among them, neither of which is default behaviour.
3486 Even then only clients that chose the curve will be affected.
3488 This issue was publicly reported as transient failures and was not
3489 initially recognized as a security issue. Thanks to Richard Morgan for
3490 providing reproducible case.
3495 * Removed automatic addition of RPATH in shared libraries and executables,
3496 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3500 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3502 * Fix Use After Free for large message sizes
3504 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3505 message larger than approx 16k is received then the underlying buffer to
3506 store the incoming message is reallocated and moved. Unfortunately a
3507 dangling pointer to the old location is left which results in an attempt to
3508 write to the previously freed location. This is likely to result in a
3509 crash, however it could potentially lead to execution of arbitrary code.
3511 This issue only affects OpenSSL 1.1.0a.
3513 This issue was reported to OpenSSL by Robert Święcki.
3518 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3520 * OCSP Status Request extension unbounded memory growth
3522 A malicious client can send an excessively large OCSP Status Request
3523 extension. If that client continually requests renegotiation, sending a
3524 large OCSP Status Request extension each time, then there will be unbounded
3525 memory growth on the server. This will eventually lead to a Denial Of
3526 Service attack through memory exhaustion. Servers with a default
3527 configuration are vulnerable even if they do not support OCSP. Builds using
3528 the "no-ocsp" build time option are not affected.
3530 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3535 * SSL_peek() hang on empty record
3537 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3538 sends an empty record. This could be exploited by a malicious peer in a
3539 Denial Of Service attack.
3541 This issue was reported to OpenSSL by Alex Gaynor.
3546 * Excessive allocation of memory in tls_get_message_header() and
3547 dtls1_preprocess_fragment()
3549 A (D)TLS message includes 3 bytes for its length in the header for the
3550 message. This would allow for messages up to 16Mb in length. Messages of
3551 this length are excessive and OpenSSL includes a check to ensure that a
3552 peer is sending reasonably sized messages in order to avoid too much memory
3553 being consumed to service a connection. A flaw in the logic of version
3554 1.1.0 means that memory for the message is allocated too early, prior to
3555 the excessive message length check. Due to way memory is allocated in
3556 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3557 to service a connection. This could lead to a Denial of Service through
3558 memory exhaustion. However, the excessive message length check still takes
3559 place, and this would cause the connection to immediately fail. Assuming
3560 that the application calls SSL_free() on the failed connection in a timely
3561 manner then the 21Mb of allocated memory will then be immediately freed
3562 again. Therefore the excessive memory allocation will be transitory in
3563 nature. This then means that there is only a security impact if:
3565 1) The application does not call SSL_free() in a timely manner in the event
3566 that the connection fails
3568 2) The application is working in a constrained environment where there is
3569 very little free memory
3571 3) The attacker initiates multiple connection attempts such that there are
3572 multiple connections in a state where memory has been allocated for the
3573 connection; SSL_free() has not yet been called; and there is insufficient
3574 memory to service the multiple requests.
3576 Except in the instance of (1) above any Denial Of Service is likely to be
3577 transitory because as soon as the connection fails the memory is
3578 subsequently freed again in the SSL_free() call. However there is an
3579 increased risk during this period of application crashes due to the lack of
3580 memory - which would then mean a more serious Denial of Service.
3582 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3583 (CVE-2016-6307 and CVE-2016-6308)
3587 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3588 had to be removed. Primary reason is that vendor assembler can't
3589 assemble our modules with -KPIC flag. As result it, assembly
3590 support, was not even available as option. But its lack means
3591 lack of side-channel resistant code, which is incompatible with
3592 security by todays standards. Fortunately gcc is readily available
3593 prepackaged option, which we firmly point at...
3597 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3599 * Windows command-line tool supports UTF-8 opt-in option for arguments
3600 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3601 (to any value) allows Windows user to access PKCS#12 file generated
3602 with Windows CryptoAPI and protected with non-ASCII password, as well
3603 as files generated under UTF-8 locale on Linux also protected with
3608 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3609 have been disabled by default and removed from DEFAULT, just like RC4.
3610 See the RC4 item below to re-enable both.
3614 * The method for finding the storage location for the Windows RAND seed file
3615 has changed. First we check %RANDFILE%. If that is not set then we check
3616 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3617 all else fails we fall back to C:\.
3621 * The EVP_EncryptUpdate() function has had its return type changed from void
3622 to int. A return of 0 indicates and error while a return of 1 indicates
3627 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3628 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3629 off the constant time implementation for RSA, DSA and DH have been made
3630 no-ops and deprecated.
3634 * Windows RAND implementation was simplified to only get entropy by
3635 calling CryptGenRandom(). Various other RAND-related tickets
3638 *Joseph Wylie Yandle, Rich Salz*
3640 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3641 and `OPENSSL_LH_`, respectively. The old names are available
3642 with API compatibility. They new names are now completely documented.
3646 * Unify TYPE_up_ref(obj) methods signature.
3647 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3648 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3649 int (instead of void) like all others TYPE_up_ref() methods.
3650 So now these methods also check the return value of CRYPTO_atomic_add(),
3651 and the validity of object reference counter.
3653 *fdasilvayy@gmail.com*
3655 * With Windows Visual Studio builds, the .pdb files are installed
3656 alongside the installed libraries and executables. For a static
3657 library installation, ossl_static.pdb is the associate compiler
3658 generated .pdb file to be used when linking programs.
3662 * Remove openssl.spec. Packaging files belong with the packagers.
3666 * Automatic Darwin/OSX configuration has had a refresh, it will now
3667 recognise x86_64 architectures automatically. You can still decide
3668 to build for a different bitness with the environment variable
3669 KERNEL_BITS (can be 32 or 64), for example:
3671 KERNEL_BITS=32 ./config
3675 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3676 256 bit AES and HMAC with SHA256.
3680 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3684 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3688 * To enable users to have their own config files and build file templates,
3689 Configure looks in the directory indicated by the environment variable
3690 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3691 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3692 name and is used as is.
3696 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3697 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3698 X509_CERT_FILE_CTX was removed.
3702 * "shared" builds are now the default. To create only static libraries use
3703 the "no-shared" Configure option.
3707 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3708 All of these option have not worked for some while and are fundamental
3713 * Make various cleanup routines no-ops and mark them as deprecated. Most
3714 global cleanup functions are no longer required because they are handled
3715 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3716 Explicitly de-initing can cause problems (e.g. where a library that uses
3717 OpenSSL de-inits, but an application is still using it). The affected
3718 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3719 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3720 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3721 COMP_zlib_cleanup().
3725 * --strict-warnings no longer enables runtime debugging options
3726 such as REF_DEBUG. Instead, debug options are automatically
3727 enabled with '--debug' builds.
3729 *Andy Polyakov, Emilia Käsper*
3731 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3732 have been moved out of the public header files. New functions for managing
3733 these have been added.
3737 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3738 objects have been moved out of the public header files. New
3739 functions for managing these have been added.
3743 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3744 have been moved out of the public header files. New functions for managing
3745 these have been added.
3749 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3750 moved out of the public header files. New functions for managing these
3755 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3759 * Removed the mk1mf build scripts.
3763 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3764 it is always safe to #include a header now.
3768 * Removed the aged BC-32 config and all its supporting scripts
3772 * Removed support for Ultrix, Netware, and OS/2.
3776 * Add support for HKDF.
3778 *Alessandro Ghedini*
3780 * Add support for blake2b and blake2s
3784 * Added support for "pipelining". Ciphers that have the
3785 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3786 encryptions/decryptions simultaneously. There are currently no built-in
3787 ciphers with this property but the expectation is that engines will be able
3788 to offer it to significantly improve throughput. Support has been extended
3789 into libssl so that multiple records for a single connection can be
3790 processed in one go (for >=TLS 1.1).
3794 * Added the AFALG engine. This is an async capable engine which is able to
3795 offload work to the Linux kernel. In this initial version it only supports
3796 AES128-CBC. The kernel must be version 4.1.0 or greater.
3800 * OpenSSL now uses a new threading API. It is no longer necessary to
3801 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3802 are two supported threading models: pthreads and windows threads. It is
3803 also possible to configure OpenSSL at compile time for "no-threads". The
3804 old threading API should no longer be used. The functions have been
3805 replaced with "no-op" compatibility macros.
3807 *Alessandro Ghedini, Matt Caswell*
3809 * Modify behavior of ALPN to invoke callback after SNI/servername
3810 callback, such that updates to the SSL_CTX affect ALPN.
3814 * Add SSL_CIPHER queries for authentication and key-exchange.
3818 * Changes to the DEFAULT cipherlist:
3819 - Prefer (EC)DHE handshakes over plain RSA.
3820 - Prefer AEAD ciphers over legacy ciphers.
3821 - Prefer ECDSA over RSA when both certificates are available.
3822 - Prefer TLSv1.2 ciphers/PRF.
3823 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3828 * Change the ECC default curve list to be this, in order: x25519,
3829 secp256r1, secp521r1, secp384r1.
3833 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3834 disabled by default. They can be re-enabled using the
3835 enable-weak-ssl-ciphers option to Configure.
3839 * If the server has ALPN configured, but supports no protocols that the
3840 client advertises, send a fatal "no_application_protocol" alert.
3841 This behaviour is SHALL in RFC 7301, though it isn't universally
3842 implemented by other servers.
3846 * Add X25519 support.
3847 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3848 for public and private key encoding using the format documented in
3849 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3850 key generation and key derivation.
3852 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3857 * Deprecate SRP_VBASE_get_by_user.
3858 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3859 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3860 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3861 seed, even if the seed is configured.
3863 Users should use SRP_VBASE_get1_by_user instead. Note that in
3864 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3865 also that even though configuring the SRP seed attempts to hide
3866 invalid usernames by continuing the handshake with fake
3867 credentials, this behaviour is not constant time and no strong
3868 guarantees are made that the handshake is indistinguishable from
3869 that of a valid user.
3873 * Configuration change; it's now possible to build dynamic engines
3874 without having to build shared libraries and vice versa. This
3875 only applies to the engines in `engines/`, those in `crypto/engine/`
3876 will always be built into libcrypto (i.e. "static").
3878 Building dynamic engines is enabled by default; to disable, use
3879 the configuration option "disable-dynamic-engine".
3881 The only requirements for building dynamic engines are the
3882 presence of the DSO module and building with position independent
3883 code, so they will also automatically be disabled if configuring
3884 with "disable-dso" or "disable-pic".
3886 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3887 are also taken away from openssl/opensslconf.h, as they are
3892 * Configuration change; if there is a known flag to compile
3893 position independent code, it will always be applied on the
3894 libcrypto and libssl object files, and never on the application
3895 object files. This means other libraries that use routines from
3896 libcrypto / libssl can be made into shared libraries regardless
3897 of how OpenSSL was configured.
3899 If this isn't desirable, the configuration options "disable-pic"
3900 or "no-pic" can be used to disable the use of PIC. This will
3901 also disable building shared libraries and dynamic engines.
3905 * Removed JPAKE code. It was experimental and has no wide use.
3909 * The INSTALL_PREFIX Makefile variable has been renamed to
3910 DESTDIR. That makes for less confusion on what this variable
3911 is for. Also, the configuration option --install_prefix is
3916 * Heartbeat for TLS has been removed and is disabled by default
3917 for DTLS; configure with enable-heartbeats. Code that uses the
3918 old #define's might need to be updated.
3920 *Emilia Käsper, Rich Salz*
3922 * Rename REF_CHECK to REF_DEBUG.
3926 * New "unified" build system
3928 The "unified" build system is aimed to be a common system for all
3929 platforms we support. With it comes new support for VMS.
3931 This system builds supports building in a different directory tree
3932 than the source tree. It produces one Makefile (for unix family
3933 or lookalikes), or one descrip.mms (for VMS).
3935 The source of information to make the Makefile / descrip.mms is
3936 small files called 'build.info', holding the necessary
3937 information for each directory with source to compile, and a
3938 template in Configurations, like unix-Makefile.tmpl or
3941 With this change, the library names were also renamed on Windows
3942 and on VMS. They now have names that are closer to the standard
3943 on Unix, and include the major version number, and in certain
3944 cases, the architecture they are built for. See "Notes on shared
3945 libraries" in INSTALL.
3947 We rely heavily on the perl module Text::Template.
3951 * Added support for auto-initialisation and de-initialisation of the library.
3952 OpenSSL no longer requires explicit init or deinit routines to be called,
3953 except in certain circumstances. See the OPENSSL_init_crypto() and
3954 OPENSSL_init_ssl() man pages for further information.
3958 * The arguments to the DTLSv1_listen function have changed. Specifically the
3959 "peer" argument is now expected to be a BIO_ADDR object.
3961 * Rewrite of BIO networking library. The BIO library lacked consistent
3962 support of IPv6, and adding it required some more extensive
3963 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3964 which hold all types of addresses and chains of address information.
3965 It also introduces a new API, with functions like BIO_socket,
3966 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3967 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3968 have been adapted accordingly.
3972 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3977 * CRIME protection: disable compression by default, even if OpenSSL is
3978 compiled with zlib enabled. Applications can still enable compression
3979 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3980 using the SSL_CONF library to configure compression.
3984 * The signature of the session callback configured with
3985 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3986 was explicitly marked as `const unsigned char*` instead of
3991 * Always DPURIFY. Remove the use of uninitialized memory in the
3992 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3996 * Removed many obsolete configuration items, including
3997 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3998 MD2_CHAR, MD2_INT, MD2_LONG
4000 IDEA_SHORT, IDEA_LONG
4001 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4003 *Rich Salz, with advice from Andy Polyakov*
4005 * Many BN internals have been moved to an internal header file.
4007 *Rich Salz with help from Andy Polyakov*
4009 * Configuration and writing out the results from it has changed.
4010 Files such as Makefile include/openssl/opensslconf.h and are now
4011 produced through general templates, such as Makefile.in and
4012 crypto/opensslconf.h.in and some help from the perl module
4015 Also, the center of configuration information is no longer
4016 Makefile. Instead, Configure produces a perl module in
4017 configdata.pm which holds most of the config data (in the hash
4018 table %config), the target data that comes from the target
4019 configuration in one of the `Configurations/*.conf` files (in
4024 * To clarify their intended purposes, the Configure options
4025 --prefix and --openssldir change their semantics, and become more
4026 straightforward and less interdependent.
4028 --prefix shall be used exclusively to give the location INSTALLTOP
4029 where programs, scripts, libraries, include files and manuals are
4030 going to be installed. The default is now /usr/local.
4032 --openssldir shall be used exclusively to give the default
4033 location OPENSSLDIR where certificates, private keys, CRLs are
4034 managed. This is also where the default openssl.cnf gets
4036 If the directory given with this option is a relative path, the
4037 values of both the --prefix value and the --openssldir value will
4038 be combined to become OPENSSLDIR.
4039 The default for --openssldir is INSTALLTOP/ssl.
4041 Anyone who uses --openssldir to specify where OpenSSL is to be
4042 installed MUST change to use --prefix instead.
4046 * The GOST engine was out of date and therefore it has been removed. An up
4047 to date GOST engine is now being maintained in an external repository.
4048 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4049 support for GOST ciphersuites (these are only activated if a GOST engine
4054 * EGD is no longer supported by default; use enable-egd when
4057 *Ben Kaduk and Rich Salz*
4059 * The distribution now has Makefile.in files, which are used to
4060 create Makefile's when Configure is run. *Configure must be run
4061 before trying to build now.*
4065 * The return value for SSL_CIPHER_description() for error conditions
4070 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4072 Obtaining and performing DNSSEC validation of TLSA records is
4073 the application's responsibility. The application provides
4074 the TLSA records of its choice to OpenSSL, and these are then
4075 used to authenticate the peer.
4077 The TLSA records need not even come from DNS. They can, for
4078 example, be used to implement local end-entity certificate or
4079 trust-anchor "pinning", where the "pin" data takes the form
4080 of TLSA records, which can augment or replace verification
4081 based on the usual WebPKI public certification authorities.
4085 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4086 continues to support deprecated interfaces in default builds.
4087 However, applications are strongly advised to compile their
4088 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4089 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4090 or the 1.1.0 releases.
4092 In environments in which all applications have been ported to
4093 not use any deprecated interfaces OpenSSL's Configure script
4094 should be used with the --api=1.1.0 option to entirely remove
4095 support for the deprecated features from the library and
4096 unconditionally disable them in the installed headers.
4097 Essentially the same effect can be achieved with the "no-deprecated"
4098 argument to Configure, except that this will always restrict
4099 the build to just the latest API, rather than a fixed API
4102 As applications are ported to future revisions of the API,
4103 they should update their compile-time OPENSSL_API_COMPAT define
4104 accordingly, but in most cases should be able to continue to
4105 compile with later releases.
4107 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4108 0x10000000L and 0x00908000L, respectively. However those
4109 versions did not support the OPENSSL_API_COMPAT feature, and
4110 so applications are not typically tested for explicit support
4111 of just the undeprecated features of either release.
4115 * Add support for setting the minimum and maximum supported protocol.
4116 It can bet set via the SSL_set_min_proto_version() and
4117 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4118 MaxProtocol. It's recommended to use the new APIs to disable
4119 protocols instead of disabling individual protocols using
4120 SSL_set_options() or SSL_CONF's Protocol. This change also
4121 removes support for disabling TLS 1.2 in the OpenSSL TLS
4122 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4126 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4130 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4131 and integrates ECDSA and ECDH functionality into EC. Implementations can
4132 now redirect key generation and no longer need to convert to or from
4135 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4136 include the ec.h header file instead.
4140 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4141 ciphers who are no longer supported and drops support the ephemeral RSA key
4142 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4146 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4147 opaque. For HMAC_CTX, the following constructors and destructors
4150 HMAC_CTX *HMAC_CTX_new(void);
4151 void HMAC_CTX_free(HMAC_CTX *ctx);
4153 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4154 destroy such methods has been added. See EVP_MD_meth_new(3) and
4155 EVP_CIPHER_meth_new(3) for documentation.
4158 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4159 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4160 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4161 an already created structure.
4162 2) For consistency with the majority of our object creators and
4163 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4164 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4165 for deprecated builds.
4169 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4170 cryptographic operations to be performed asynchronously as long as an
4171 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4172 further details. Libssl has also had this capability integrated with the
4173 introduction of the new mode SSL_MODE_ASYNC and associated error
4174 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4175 pages. This work was developed in partnership with Intel Corp.
4179 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4180 always enabled now. If you want to disable the support you should
4181 exclude it using the list of supported ciphers. This also means that the
4182 "-no_ecdhe" option has been removed from s_server.
4186 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4187 SSL_{CTX_}set1_curves() which can set a list.
4191 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4192 curve you want to support using SSL_{CTX_}set1_curves().
4196 * State machine rewrite. The state machine code has been significantly
4197 refactored in order to remove much duplication of code and solve issues
4198 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4199 further details). This change does have some associated API changes.
4200 Notably the SSL_state() function has been removed and replaced by
4201 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4202 SSL_set_state() has been removed altogether. The previous handshake states
4203 defined in ssl.h and ssl3.h have also been removed.
4207 * All instances of the string "ssleay" in the public API were replaced
4208 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4209 Some error codes related to internal RSA_eay API's were renamed.
4213 * The demo files in crypto/threads were moved to demo/threads.
4217 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4220 *Matt Caswell, Rich Salz*
4222 * New ASN.1 embed macro.
4224 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4225 structure is not allocated: it is part of the parent. That is instead of
4233 This reduces memory fragmentation and make it impossible to accidentally
4234 set a mandatory field to NULL.
4236 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4237 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4238 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4243 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4247 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4248 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4249 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4250 DES and RC4 ciphersuites.
4254 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4255 This changes the decoding behaviour for some invalid messages,
4256 though the change is mostly in the more lenient direction, and
4257 legacy behaviour is preserved as much as possible.
4261 * Fix no-stdio build.
4262 *David Woodhouse <David.Woodhouse@intel.com> and also*
4263 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4265 * New testing framework
4266 The testing framework has been largely rewritten and is now using
4267 perl and the perl modules Test::Harness and an extended variant of
4268 Test::More called OpenSSL::Test to do its work. All test scripts in
4269 test/ have been rewritten into test recipes, and all direct calls to
4270 executables in test/Makefile have become individual recipes using the
4271 simplified testing OpenSSL::Test::Simple.
4273 For documentation on our testing modules, do:
4275 perldoc test/testlib/OpenSSL/Test/Simple.pm
4276 perldoc test/testlib/OpenSSL/Test.pm
4280 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4281 are used; the latter aborts on memory leaks (usually checked on exit).
4282 Some undocumented "set malloc, etc., hooks" functions were removed
4283 and others were changed. All are now documented.
4287 * In DSA_generate_parameters_ex, if the provided seed is too short,
4290 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4292 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4293 from RFC4279, RFC4785, RFC5487, RFC5489.
4295 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4296 original RSA_PSK patch.
4300 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4301 era flag was never set throughout the codebase (only read). Also removed
4302 SSL3_FLAGS_POP_BUFFER which was only used if
4303 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4307 * Changed the default name options in the "ca", "crl", "req" and "x509"
4308 to be "oneline" instead of "compat".
4312 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4313 not aware of clients that still exhibit this bug, and the workaround
4314 hasn't been working properly for a while.
4318 * The return type of BIO_number_read() and BIO_number_written() as well as
4319 the corresponding num_read and num_write members in the BIO structure has
4320 changed from unsigned long to uint64_t. On platforms where an unsigned
4321 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4326 * Given the pervasive nature of TLS extensions it is inadvisable to run
4327 OpenSSL without support for them. It also means that maintaining
4328 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4329 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4333 * Removed support for the two export grade static DH ciphersuites
4334 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4335 were newly added (along with a number of other static DH ciphersuites) to
4336 1.0.2. However the two export ones have *never* worked since they were
4337 introduced. It seems strange in any case to be adding new export
4338 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4342 * Version negotiation has been rewritten. In particular SSLv23_method(),
4343 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4344 and turned into macros which simply call the new preferred function names
4345 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4346 should use the new names instead. Also as part of this change the ssl23.h
4347 header file has been removed.
4351 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4352 code and the associated standard is no longer considered fit-for-purpose.
4356 * RT2547 was closed. When generating a private key, try to make the
4357 output file readable only by the owner. This behavior change might
4358 be noticeable when interacting with other software.
4360 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4365 * Added HTTP GET support to the ocsp command.
4369 * Changed default digest for the dgst and enc commands from MD5 to
4374 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4378 * Added support for TLS extended master secret from
4379 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4380 initial patch which was a great help during development.
4384 * All libssl internal structures have been removed from the public header
4385 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4386 now redundant). Users should not attempt to access internal structures
4387 directly. Instead they should use the provided API functions.
4391 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4392 Access to deprecated functions can be re-enabled by running config with
4393 "enable-deprecated". In addition applications wishing to use deprecated
4394 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4395 will, by default, disable some transitive includes that previously existed
4396 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4400 * Added support for OCB mode. OpenSSL has been granted a patent license
4401 compatible with the OpenSSL license for use of OCB. Details are available
4402 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4403 for OCB can be removed by calling config with no-ocb.
4407 * SSLv2 support has been removed. It still supports receiving a SSLv2
4408 compatible client hello.
4412 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4413 done while fixing the error code for the key-too-small case.
4415 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4417 * CA.sh has been removed; use CA.pl instead.
4421 * Removed old DES API.
4425 * Remove various unsupported platforms:
4431 Sinix/ReliantUNIX RM400
4436 16-bit platforms such as WIN16
4440 * Clean up OPENSSL_NO_xxx #define's
4441 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4442 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4443 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4444 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4445 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4446 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4447 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4448 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4449 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4450 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4454 * Cleaned up dead code
4455 Remove all but one '#ifdef undef' which is to be looked at.
4459 * Clean up calling of xxx_free routines.
4460 Just like free(), fix most of the xxx_free routines to accept
4461 NULL. Remove the non-null checks from callers. Save much code.
4465 * Add secure heap for storage of private keys (when possible).
4466 Add BIO_s_secmem(), CBIGNUM, etc.
4467 Contributed by Akamai Technologies under our Corporate CLA.
4471 * Experimental support for a new, fast, unbiased prime candidate generator,
4472 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4474 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4476 * New output format NSS in the sess_id command line tool. This allows
4477 exporting the session id and the master key in NSS keylog format.
4479 *Martin Kaiser <martin@kaiser.cx>*
4481 * Harmonize version and its documentation. -f flag is used to display
4484 *mancha <mancha1@zoho.com>*
4486 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4487 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4489 *mancha <mancha1@zoho.com>*
4491 * Fix some double frees. These are not thought to be exploitable.
4493 *mancha <mancha1@zoho.com>*
4495 * A missing bounds check in the handling of the TLS heartbeat extension
4496 can be used to reveal up to 64k of memory to a connected client or
4499 Thanks for Neel Mehta of Google Security for discovering this bug and to
4500 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4501 preparing the fix ([CVE-2014-0160])
4503 *Adam Langley, Bodo Moeller*
4505 * Fix for the attack described in the paper "Recovering OpenSSL
4506 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4507 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4508 <http://eprint.iacr.org/2014/140>
4510 Thanks to Yuval Yarom and Naomi Benger for discovering this
4511 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4513 *Yuval Yarom and Naomi Benger*
4515 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4516 this fixes a limitation in previous versions of OpenSSL.
4520 * Experimental encrypt-then-mac support.
4522 Experimental support for encrypt then mac from
4523 draft-gutmann-tls-encrypt-then-mac-02.txt
4525 To enable it set the appropriate extension number (0x42 for the test
4526 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4528 For non-compliant peers (i.e. just about everything) this should have no
4531 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4535 * Add EVP support for key wrapping algorithms, to avoid problems with
4536 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4537 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4538 algorithms and include tests cases.
4542 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4547 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4548 MGF1 digest and OAEP label.
4552 * Make openssl verify return errors.
4554 *Chris Palmer <palmer@google.com> and Ben Laurie*
4556 * New function ASN1_TIME_diff to calculate the difference between two
4557 ASN1_TIME structures or one structure and the current time.
4561 * Update fips_test_suite to support multiple command line options. New
4562 test to induce all self test errors in sequence and check expected
4567 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4568 sign or verify all in one operation.
4572 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4573 test programs and fips_test_suite. Includes functionality to parse
4574 the minimal script output of fipsalgest.pl directly.
4578 * Add authorisation parameter to FIPS_module_mode_set().
4582 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4586 * Use separate DRBG fields for internal and external flags. New function
4587 FIPS_drbg_health_check() to perform on demand health checking. Add
4588 generation tests to fips_test_suite with reduced health check interval to
4589 demonstrate periodic health checking. Add "nodh" option to
4590 fips_test_suite to skip very slow DH test.
4594 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4599 * More extensive health check for DRBG checking many more failure modes.
4600 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4601 combination: call this in fips_test_suite.
4605 * Add support for canonical generation of DSA parameter 'g'. See
4608 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4609 POST to handle HMAC cases.
4613 * Add functions FIPS_module_version() and FIPS_module_version_text()
4614 to return numerical and string versions of the FIPS module number.
4618 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4619 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4620 outside the validated module in the FIPS capable OpenSSL.
4624 * Minor change to DRBG entropy callback semantics. In some cases
4625 there is no multiple of the block length between min_len and
4626 max_len. Allow the callback to return more than max_len bytes
4627 of entropy but discard any extra: it is the callback's responsibility
4628 to ensure that the extra data discarded does not impact the
4629 requested amount of entropy.
4633 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4634 information in FIPS186-3, SP800-57 and SP800-131A.
4638 * CCM support via EVP. Interface is very similar to GCM case except we
4639 must supply all data in one chunk (i.e. no update, final) and the
4640 message length must be supplied if AAD is used. Add algorithm test
4645 * Initial version of POST overhaul. Add POST callback to allow the status
4646 of POST to be monitored and/or failures induced. Modify fips_test_suite
4647 to use callback. Always run all selftests even if one fails.
4651 * XTS support including algorithm test driver in the fips_gcmtest program.
4652 Note: this does increase the maximum key length from 32 to 64 bytes but
4653 there should be no binary compatibility issues as existing applications
4654 will never use XTS mode.
4658 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4659 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4660 performs algorithm blocking for unapproved PRNG types. Also do not
4661 set PRNG type in FIPS_mode_set(): leave this to the application.
4662 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4663 the standard OpenSSL PRNG: set additional data to a date time vector.
4667 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4668 This shouldn't present any incompatibility problems because applications
4669 shouldn't be using these directly and any that are will need to rethink
4670 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4674 * Extensive self tests and health checking required by SP800-90 DRBG.
4675 Remove strength parameter from FIPS_drbg_instantiate and always
4676 instantiate at maximum supported strength.
4680 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4684 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4688 * New function DH_compute_key_padded() to compute a DH key and pad with
4689 leading zeroes if needed: this complies with SP800-56A et al.
4693 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4694 anything, incomplete, subject to change and largely untested at present.
4698 * Modify fipscanisteronly build option to only build the necessary object
4699 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4703 * Add experimental option FIPSSYMS to give all symbols in
4704 fipscanister.o and FIPS or fips prefix. This will avoid
4705 conflicts with future versions of OpenSSL. Add perl script
4706 util/fipsas.pl to preprocess assembly language source files
4707 and rename any affected symbols.
4711 * Add selftest checks and algorithm block of non-fips algorithms in
4712 FIPS mode. Remove DES2 from selftests.
4716 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4717 return internal method without any ENGINE dependencies. Add new
4718 tiny fips sign and verify functions.
4722 * New build option no-ec2m to disable characteristic 2 code.
4726 * New build option "fipscanisteronly". This only builds fipscanister.o
4727 and (currently) associated fips utilities. Uses the file Makefile.fips
4728 instead of Makefile.org as the prototype.
4732 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4733 Update fips_gcmtest to use IV generator.
4737 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4738 setting output buffer to NULL. The `*Final` function must be
4739 called although it will not retrieve any additional data. The tag
4740 can be set or retrieved with a ctrl. The IV length is by default 12
4741 bytes (96 bits) but can be set to an alternative value. If the IV
4742 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4747 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4748 underlying do_cipher function handles all cipher semantics itself
4749 including padding and finalisation. This is useful if (for example)
4750 an ENGINE cipher handles block padding itself. The behaviour of
4751 do_cipher is subtly changed if this flag is set: the return value
4752 is the number of characters written to the output buffer (zero is
4753 no longer an error code) or a negative error code. Also if the
4754 input buffer is NULL and length 0 finalisation should be performed.
4758 * If a candidate issuer certificate is already part of the constructed
4759 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4763 * Improve forward-security support: add functions
4765 void SSL_CTX_set_not_resumable_session_callback(
4766 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4767 void SSL_set_not_resumable_session_callback(
4768 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4770 for use by SSL/TLS servers; the callback function will be called whenever a
4771 new session is created, and gets to decide whether the session may be
4772 cached to make it resumable (return 0) or not (return 1). (As by the
4773 SSL/TLS protocol specifications, the session_id sent by the server will be
4774 empty to indicate that the session is not resumable; also, the server will
4775 not generate RFC 4507 (RFC 5077) session tickets.)
4777 A simple reasonable callback implementation is to return is_forward_secure.
4778 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4779 by the SSL/TLS server library, indicating whether it can provide forward
4782 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4784 * New -verify_name option in command line utilities to set verification
4789 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4790 Add CMAC pkey methods.
4794 * Experimental renegotiation in s_server -www mode. If the client
4795 browses /reneg connection is renegotiated. If /renegcert it is
4796 renegotiated requesting a certificate.
4800 * Add an "external" session cache for debugging purposes to s_server. This
4801 should help trace issues which normally are only apparent in deployed
4802 multi-process servers.
4806 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4807 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4808 BIO_set_cipher() and some obscure PEM functions were changed so they
4809 can now return an error. The RAND changes required a change to the
4810 RAND_METHOD structure.
4814 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4815 a gcc attribute to warn if the result of a function is ignored. This
4816 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4817 whose return value is often ignored.
4821 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4822 These allow SCTs (signed certificate timestamps) to be requested and
4823 validated when establishing a connection.
4825 *Rob Percival <robpercival@google.com>*
4830 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4832 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4833 used even when parsing explicit parameters, when loading a encoded key
4834 or calling `EC_GROUP_new_from_ecpkparameters()`/
4835 `EC_GROUP_new_from_ecparameters()`.
4836 This prevents bypass of security hardening and performance gains,
4837 especially for curves with specialized EC_METHODs.
4838 By default, if a key encoded with explicit parameters is loaded and later
4839 encoded, the output is still encoded with explicit parameters, even if
4840 internally a "named" EC_GROUP is used for computation.
4844 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4845 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4846 NULL. After this change, only the cofactor parameter can be NULL. It also
4847 does some minimal sanity checks on the passed order.
4852 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4853 An attack is simple, if the first CMS_recipientInfo is valid but the
4854 second CMS_recipientInfo is chosen ciphertext. If the second
4855 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4856 encryption key will be replaced by garbage, and the message cannot be
4857 decoded, but if the RSA decryption fails, the correct encryption key is
4858 used and the recipient will not notice the attack.
4859 As a work around for this potential attack the length of the decrypted
4860 key must be equal to the cipher default key length, in case the
4861 certifiate is not given and all recipientInfo are tried out.
4862 The old behaviour can be re-enabled in the CMS code by setting the
4863 CMS_DEBUG_DECRYPT flag.
4868 * Document issue with installation paths in diverse Windows builds
4870 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4871 binaries and run-time config file.
4876 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4878 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4879 This changes the size when using the `genpkey` command when no size is given.
4880 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4881 generation commands to use 2048 bits by default.
4885 * Add FIPS support for Android Arm 64-bit
4887 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4888 Module in Version 2.0.10. For some reason, the corresponding target
4889 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4890 built with FIPS support on Android Arm 64-bit. This omission has been
4893 *Matthias St. Pierre*
4895 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4897 * 0-byte record padding oracle
4899 If an application encounters a fatal protocol error and then calls
4900 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4901 then OpenSSL can respond differently to the calling application if a 0 byte
4902 record is received with invalid padding compared to if a 0 byte record is
4903 received with an invalid MAC. If the application then behaves differently
4904 based on that in a way that is detectable to the remote peer, then this
4905 amounts to a padding oracle that could be used to decrypt data.
4907 In order for this to be exploitable "non-stitched" ciphersuites must be in
4908 use. Stitched ciphersuites are optimised implementations of certain
4909 commonly used ciphersuites. Also the application must call SSL_shutdown()
4910 twice even if a protocol error has occurred (applications should not do
4911 this but some do anyway).
4913 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4914 Aviram, with additional investigation by Steven Collison and Andrew
4915 Hourselt. It was reported to OpenSSL on 10th December 2018.
4920 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4924 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4926 * Microarchitecture timing vulnerability in ECC scalar multiplication
4928 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4929 shown to be vulnerable to a microarchitecture timing side channel attack.
4930 An attacker with sufficient access to mount local timing attacks during
4931 ECDSA signature generation could recover the private key.
4933 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4934 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4940 * Timing vulnerability in DSA signature generation
4942 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4943 timing side channel attack. An attacker could use variations in the signing
4944 algorithm to recover the private key.
4946 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4951 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4952 Module, accidentally introduced while backporting security fixes from the
4953 development branch and hindering the use of ECC in FIPS mode.
4957 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4959 * Client DoS due to large DH parameter
4961 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4962 malicious server can send a very large prime value to the client. This will
4963 cause the client to spend an unreasonably long period of time generating a
4964 key for this prime resulting in a hang until the client has finished. This
4965 could be exploited in a Denial Of Service attack.
4967 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4972 * Cache timing vulnerability in RSA Key Generation
4974 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4975 a cache timing side channel attack. An attacker with sufficient access to
4976 mount cache timing attacks during the RSA key generation process could
4977 recover the private key.
4979 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4980 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4985 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4986 parameter is no longer accepted, as it leads to a corrupt table. NULL
4987 pem_str is reserved for alias entries only.
4991 * Revert blinding in ECDSA sign and instead make problematic addition
4992 length-invariant. Switch even to fixed-length Montgomery multiplication.
4996 * Change generating and checking of primes so that the error rate of not
4997 being prime depends on the intended use based on the size of the input.
4998 For larger primes this will result in more rounds of Miller-Rabin.
4999 The maximal error rate for primes with more than 1080 bits is lowered
5002 *Kurt Roeckx, Annie Yousar*
5004 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5008 * Add blinding to ECDSA and DSA signatures to protect against side channel
5009 attacks discovered by Keegan Ryan (NCC Group).
5013 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5014 now allow empty (zero character) pass phrases.
5018 * Certificate time validation (X509_cmp_time) enforces stricter
5019 compliance with RFC 5280. Fractional seconds and timezone offsets
5020 are no longer allowed.
5024 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5026 * Constructed ASN.1 types with a recursive definition could exceed the stack
5028 Constructed ASN.1 types with a recursive definition (such as can be found
5029 in PKCS7) could eventually exceed the stack given malicious input with
5030 excessive recursion. This could result in a Denial Of Service attack. There
5031 are no such structures used within SSL/TLS that come from untrusted sources
5032 so this is considered safe.
5034 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5040 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5042 * Read/write after SSL object in error state
5044 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5045 mechanism. The intent was that if a fatal error occurred during a handshake
5046 then OpenSSL would move into the error state and would immediately fail if
5047 you attempted to continue the handshake. This works as designed for the
5048 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5049 SSL_connect()), however due to a bug it does not work correctly if
5050 SSL_read() or SSL_write() is called directly. In that scenario, if the
5051 handshake fails then a fatal error will be returned in the initial function
5052 call. If SSL_read()/SSL_write() is subsequently called by the application
5053 for the same SSL object then it will succeed and the data is passed without
5054 being decrypted/encrypted directly from the SSL/TLS record layer.
5056 In order to exploit this issue an application bug would have to be present
5057 that resulted in a call to SSL_read()/SSL_write() being issued after having
5058 already received a fatal error.
5060 This issue was reported to OpenSSL by David Benjamin (Google).
5065 * rsaz_1024_mul_avx2 overflow bug on x86_64
5067 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5068 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5069 Analysis suggests that attacks against RSA and DSA as a result of this
5070 defect would be very difficult to perform and are not believed likely.
5071 Attacks against DH1024 are considered just feasible, because most of the
5072 work necessary to deduce information about a private key may be performed
5073 offline. The amount of resources required for such an attack would be
5074 significant. However, for an attack on TLS to be meaningful, the server
5075 would have to share the DH1024 private key among multiple clients, which is
5076 no longer an option since CVE-2016-0701.
5078 This only affects processors that support the AVX2 but not ADX extensions
5079 like Intel Haswell (4th generation).
5081 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5082 was originally found via the OSS-Fuzz project.
5087 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5089 * bn_sqrx8x_internal carry bug on x86_64
5091 There is a carry propagating bug in the x86_64 Montgomery squaring
5092 procedure. No EC algorithms are affected. Analysis suggests that attacks
5093 against RSA and DSA as a result of this defect would be very difficult to
5094 perform and are not believed likely. Attacks against DH are considered just
5095 feasible (although very difficult) because most of the work necessary to
5096 deduce information about a private key may be performed offline. The amount
5097 of resources required for such an attack would be very significant and
5098 likely only accessible to a limited number of attackers. An attacker would
5099 additionally need online access to an unpatched system using the target
5100 private key in a scenario with persistent DH parameters and a private
5101 key that is shared between multiple clients.
5103 This only affects processors that support the BMI1, BMI2 and ADX extensions
5104 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5106 This issue was reported to OpenSSL by the OSS-Fuzz project.
5111 * Malformed X.509 IPAddressFamily could cause OOB read
5113 If an X.509 certificate has a malformed IPAddressFamily extension,
5114 OpenSSL could do a one-byte buffer overread. The most likely result
5115 would be an erroneous display of the certificate in text format.
5117 This issue was reported to OpenSSL by the OSS-Fuzz project.
5121 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5123 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5124 platform rather than 'mingw'.
5128 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5130 * Truncated packet could crash via OOB read
5132 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5133 cipher is being used, then a truncated packet can cause that host to
5134 perform an out-of-bounds read, usually resulting in a crash.
5136 This issue was reported to OpenSSL by Robert Święcki of Google.
5141 * BN_mod_exp may produce incorrect results on x86_64
5143 There is a carry propagating bug in the x86_64 Montgomery squaring
5144 procedure. No EC algorithms are affected. Analysis suggests that attacks
5145 against RSA and DSA as a result of this defect would be very difficult to
5146 perform and are not believed likely. Attacks against DH are considered just
5147 feasible (although very difficult) because most of the work necessary to
5148 deduce information about a private key may be performed offline. The amount
5149 of resources required for such an attack would be very significant and
5150 likely only accessible to a limited number of attackers. An attacker would
5151 additionally need online access to an unpatched system using the target
5152 private key in a scenario with persistent DH parameters and a private
5153 key that is shared between multiple clients. For example this can occur by
5154 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5155 similar to CVE-2015-3193 but must be treated as a separate problem.
5157 This issue was reported to OpenSSL by the OSS-Fuzz project.
5162 * Montgomery multiplication may produce incorrect results
5164 There is a carry propagating bug in the Broadwell-specific Montgomery
5165 multiplication procedure that handles input lengths divisible by, but
5166 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5167 and DH private keys are impossible. This is because the subroutine in
5168 question is not used in operations with the private key itself and an input
5169 of the attacker's direct choice. Otherwise the bug can manifest itself as
5170 transient authentication and key negotiation failures or reproducible
5171 erroneous outcome of public-key operations with specially crafted input.
5172 Among EC algorithms only Brainpool P-512 curves are affected and one
5173 presumably can attack ECDH key negotiation. Impact was not analyzed in
5174 detail, because pre-requisites for attack are considered unlikely. Namely
5175 multiple clients have to choose the curve in question and the server has to
5176 share the private key among them, neither of which is default behaviour.
5177 Even then only clients that chose the curve will be affected.
5179 This issue was publicly reported as transient failures and was not
5180 initially recognized as a security issue. Thanks to Richard Morgan for
5181 providing reproducible case.
5186 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5187 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5188 prevent issues where no progress is being made and the peer continually
5189 sends unrecognised record types, using up resources processing them.
5193 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5195 * Missing CRL sanity check
5197 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5198 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5199 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5201 This issue only affects the OpenSSL 1.0.2i
5206 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5208 * OCSP Status Request extension unbounded memory growth
5210 A malicious client can send an excessively large OCSP Status Request
5211 extension. If that client continually requests renegotiation, sending a
5212 large OCSP Status Request extension each time, then there will be unbounded
5213 memory growth on the server. This will eventually lead to a Denial Of
5214 Service attack through memory exhaustion. Servers with a default
5215 configuration are vulnerable even if they do not support OCSP. Builds using
5216 the "no-ocsp" build time option are not affected.
5218 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5223 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5226 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5232 * OOB write in MDC2_Update()
5234 An overflow can occur in MDC2_Update() either if called directly or
5235 through the EVP_DigestUpdate() function using MDC2. If an attacker
5236 is able to supply very large amounts of input data after a previous
5237 call to EVP_EncryptUpdate() with a partial block then a length check
5238 can overflow resulting in a heap corruption.
5240 The amount of data needed is comparable to SIZE_MAX which is impractical
5243 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5248 * Malformed SHA512 ticket DoS
5250 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5251 DoS attack where a malformed ticket will result in an OOB read which will
5254 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5255 a custom server callback and ticket lookup mechanism.
5257 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5262 * OOB write in BN_bn2dec()
5264 The function BN_bn2dec() does not check the return value of BN_div_word().
5265 This can cause an OOB write if an application uses this function with an
5266 overly large BIGNUM. This could be a problem if an overly large certificate
5267 or CRL is printed out from an untrusted source. TLS is not affected because
5268 record limits will reject an oversized certificate before it is parsed.
5270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5275 * OOB read in TS_OBJ_print_bio()
5277 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5278 the total length the OID text representation would use and not the amount
5279 of data written. This will result in OOB reads when large OIDs are
5282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5287 * Pointer arithmetic undefined behaviour
5289 Avoid some undefined pointer arithmetic
5291 A common idiom in the codebase is to check limits in the following manner:
5294 Where "p" points to some malloc'd data of SIZE bytes and
5297 "len" here could be from some externally supplied data (e.g. from a TLS
5300 The rules of C pointer arithmetic are such that "p + len" is only well
5301 defined where len <= SIZE. Therefore the above idiom is actually
5302 undefined behaviour.
5304 For example this could cause problems if some malloc implementation
5305 provides an address for "p" such that "p + len" actually overflows for
5306 values of len that are too big and therefore p + len < limit.
5308 This issue was reported to OpenSSL by Guido Vranken
5313 * Constant time flag not preserved in DSA signing
5315 Operations in the DSA signing algorithm should run in constant time in
5316 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5317 implementation means that a non-constant time codepath is followed for
5318 certain operations. This has been demonstrated through a cache-timing
5319 attack to be sufficient for an attacker to recover the private DSA key.
5321 This issue was reported by César Pereida (Aalto University), Billy Brumley
5322 (Tampere University of Technology), and Yuval Yarom (The University of
5323 Adelaide and NICTA).
5328 * DTLS buffered message DoS
5330 In a DTLS connection where handshake messages are delivered out-of-order
5331 those messages that OpenSSL is not yet ready to process will be buffered
5332 for later use. Under certain circumstances, a flaw in the logic means that
5333 those messages do not get removed from the buffer even though the handshake
5334 has been completed. An attacker could force up to approx. 15 messages to
5335 remain in the buffer when they are no longer required. These messages will
5336 be cleared when the DTLS connection is closed. The default maximum size for
5337 a message is 100k. Therefore the attacker could force an additional 1500k
5338 to be consumed per connection. By opening many simulataneous connections an
5339 attacker could cause a DoS attack through memory exhaustion.
5341 This issue was reported to OpenSSL by Quan Luo.
5346 * DTLS replay protection DoS
5348 A flaw in the DTLS replay attack protection mechanism means that records
5349 that arrive for future epochs update the replay protection "window" before
5350 the MAC for the record has been validated. This could be exploited by an
5351 attacker by sending a record for the next epoch (which does not have to
5352 decrypt or have a valid MAC), with a very large sequence number. This means
5353 that all subsequent legitimate packets are dropped causing a denial of
5354 service for a specific DTLS connection.
5356 This issue was reported to OpenSSL by the OCAP audit team.
5361 * Certificate message OOB reads
5363 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5364 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5365 theoretical DoS risk but this has not been observed in practice on common
5368 The messages affected are client certificate, client certificate request
5369 and server certificate. As a result the attack can only be performed
5370 against a client or a server which enables client authentication.
5372 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5377 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5379 * Prevent padding oracle in AES-NI CBC MAC check
5381 A MITM attacker can use a padding oracle attack to decrypt traffic
5382 when the connection uses an AES CBC cipher and the server support
5385 This issue was introduced as part of the fix for Lucky 13 padding
5386 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5387 constant time by making sure that always the same bytes are read and
5388 compared against either the MAC or padding bytes. But it no longer
5389 checked that there was enough data to have both the MAC and padding
5392 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5396 * Fix EVP_EncodeUpdate overflow
5398 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5399 Base64 encoding of binary data. If an attacker is able to supply very large
5400 amounts of input data then a length check can overflow resulting in a heap
5403 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5404 the `PEM_write_bio*` family of functions. These are mainly used within the
5405 OpenSSL command line applications, so any application which processes data
5406 from an untrusted source and outputs it as a PEM file should be considered
5407 vulnerable to this issue. User applications that call these APIs directly
5408 with large amounts of untrusted data may also be vulnerable.
5410 This issue was reported by Guido Vranken.
5415 * Fix EVP_EncryptUpdate overflow
5417 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5418 is able to supply very large amounts of input data after a previous call to
5419 EVP_EncryptUpdate() with a partial block then a length check can overflow
5420 resulting in a heap corruption. Following an analysis of all OpenSSL
5421 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5422 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5423 the first called function after an EVP_EncryptInit(), and therefore that
5424 specific call must be safe. The second form is where the length passed to
5425 EVP_EncryptUpdate() can be seen from the code to be some small value and
5426 therefore there is no possibility of an overflow. Since all instances are
5427 one of these two forms, it is believed that there can be no overflows in
5428 internal code due to this problem. It should be noted that
5429 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5430 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5431 of these calls have also been analysed too and it is believed there are no
5432 instances in internal usage where an overflow could occur.
5434 This issue was reported by Guido Vranken.
5439 * Prevent ASN.1 BIO excessive memory allocation
5441 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5442 a short invalid encoding can cause allocation of large amounts of memory
5443 potentially consuming excessive resources or exhausting memory.
5445 Any application parsing untrusted data through d2i BIO functions is
5446 affected. The memory based functions such as d2i_X509() are *not* affected.
5447 Since the memory based functions are used by the TLS library, TLS
5448 applications are not affected.
5450 This issue was reported by Brian Carpenter.
5457 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5458 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5459 in arbitrary stack data being returned in the buffer.
5461 This issue was reported by Guido Vranken.
5466 * Modify behavior of ALPN to invoke callback after SNI/servername
5467 callback, such that updates to the SSL_CTX affect ALPN.
5471 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5476 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5477 methods are enabled and ssl2 is disabled the methods return NULL.
5481 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5483 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5484 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5485 provide any "EXPORT" or "LOW" strength ciphers.
5489 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5490 is by default disabled at build-time. Builds that are not configured with
5491 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5492 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5493 will need to explicitly call either of:
5495 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5497 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5499 as appropriate. Even if either of those is used, or the application
5500 explicitly uses the version-specific SSLv2_method() or its client and
5501 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5502 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5503 ciphers, and SSLv2 56-bit DES are no longer available.
5508 * Fix a double-free in DSA code
5510 A double free bug was discovered when OpenSSL parses malformed DSA private
5511 keys and could lead to a DoS attack or memory corruption for applications
5512 that receive DSA private keys from untrusted sources. This scenario is
5515 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5521 * Disable SRP fake user seed to address a server memory leak.
5523 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5525 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5526 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5527 was changed to ignore the "fake user" SRP seed, even if the seed
5530 Users should use SRP_VBASE_get1_by_user instead. Note that in
5531 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5532 also that even though configuring the SRP seed attempts to hide
5533 invalid usernames by continuing the handshake with fake
5534 credentials, this behaviour is not constant time and no strong
5535 guarantees are made that the handshake is indistinguishable from
5536 that of a valid user.
5541 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5543 In the BN_hex2bn function the number of hex digits is calculated using an
5544 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5545 large values of `i` this can result in `bn_expand` not allocating any
5546 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5547 field as NULL leading to a subsequent NULL ptr deref. For very large values
5548 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5549 In this case memory is allocated to the internal BIGNUM data field, but it
5550 is insufficiently sized leading to heap corruption. A similar issue exists
5551 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5552 is ever called by user applications with very large untrusted hex/dec data.
5553 This is anticipated to be a rare occurrence.
5555 All OpenSSL internal usage of these functions use data that is not expected
5556 to be untrusted, e.g. config file data or application command line
5557 arguments. If user developed applications generate config file data based
5558 on untrusted data then it is possible that this could also lead to security
5559 consequences. This is also anticipated to be rare.
5561 This issue was reported to OpenSSL by Guido Vranken.
5566 * Fix memory issues in `BIO_*printf` functions
5568 The internal `fmtstr` function used in processing a "%s" format string in
5569 the `BIO_*printf` functions could overflow while calculating the length of a
5570 string and cause an OOB read when printing very long strings.
5572 Additionally the internal `doapr_outch` function can attempt to write to an
5573 OOB memory location (at an offset from the NULL pointer) in the event of a
5574 memory allocation failure. In 1.0.2 and below this could be caused where
5575 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5576 could be in processing a very long "%s" format string. Memory leaks can
5579 The first issue may mask the second issue dependent on compiler behaviour.
5580 These problems could enable attacks where large amounts of untrusted data
5581 is passed to the `BIO_*printf` functions. If applications use these functions
5582 in this way then they could be vulnerable. OpenSSL itself uses these
5583 functions when printing out human-readable dumps of ASN.1 data. Therefore
5584 applications that print this data could be vulnerable if the data is from
5585 untrusted sources. OpenSSL command line applications could also be
5586 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5587 as command line arguments.
5589 Libssl is not considered directly vulnerable. Additionally certificates etc
5590 received via remote connections via libssl are also unlikely to be able to
5591 trigger these issues because of message size limits enforced within libssl.
5593 This issue was reported to OpenSSL Guido Vranken.
5598 * Side channel attack on modular exponentiation
5600 A side-channel attack was found which makes use of cache-bank conflicts on
5601 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5602 of RSA keys. The ability to exploit this issue is limited as it relies on
5603 an attacker who has control of code in a thread running on the same
5604 hyper-threaded core as the victim thread which is performing decryptions.
5606 This issue was reported to OpenSSL by Yuval Yarom, The University of
5607 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5608 Nadia Heninger, University of Pennsylvania with more information at
5609 <http://cachebleed.info>.
5614 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5615 if no keysize is specified with default_bits. This fixes an
5616 omission in an earlier change that changed all RSA/DSA key generation
5617 commands to use 2048 bits by default.
5621 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5623 * DH small subgroups
5625 Historically OpenSSL only ever generated DH parameters based on "safe"
5626 primes. More recently (in version 1.0.2) support was provided for
5627 generating X9.42 style parameter files such as those required for RFC 5114
5628 support. The primes used in such files may not be "safe". Where an
5629 application is using DH configured with parameters based on primes that are
5630 not "safe" then an attacker could use this fact to find a peer's private
5631 DH exponent. This attack requires that the attacker complete multiple
5632 handshakes in which the peer uses the same private DH exponent. For example
5633 this could be used to discover a TLS server's private DH exponent if it's
5634 reusing the private DH exponent or it's using a static DH ciphersuite.
5636 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5637 TLS. It is not on by default. If the option is not set then the server
5638 reuses the same private DH exponent for the life of the server process and
5639 would be vulnerable to this attack. It is believed that many popular
5640 applications do set this option and would therefore not be at risk.
5642 The fix for this issue adds an additional check where a "q" parameter is
5643 available (as is the case in X9.42 based parameters). This detects the
5644 only known attack, and is the only possible defense for static DH
5645 ciphersuites. This could have some performance impact.
5647 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5648 default and cannot be disabled. This could have some performance impact.
5650 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5655 * SSLv2 doesn't block disabled ciphers
5657 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5658 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5659 been disabled, provided that the SSLv2 protocol was not also disabled via
5662 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5663 and Sebastian Schinzel.
5668 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5670 * BN_mod_exp may produce incorrect results on x86_64
5672 There is a carry propagating bug in the x86_64 Montgomery squaring
5673 procedure. No EC algorithms are affected. Analysis suggests that attacks
5674 against RSA and DSA as a result of this defect would be very difficult to
5675 perform and are not believed likely. Attacks against DH are considered just
5676 feasible (although very difficult) because most of the work necessary to
5677 deduce information about a private key may be performed offline. The amount
5678 of resources required for such an attack would be very significant and
5679 likely only accessible to a limited number of attackers. An attacker would
5680 additionally need online access to an unpatched system using the target
5681 private key in a scenario with persistent DH parameters and a private
5682 key that is shared between multiple clients. For example this can occur by
5683 default in OpenSSL DHE based SSL/TLS ciphersuites.
5685 This issue was reported to OpenSSL by Hanno Böck.
5690 * Certificate verify crash with missing PSS parameter
5692 The signature verification routines will crash with a NULL pointer
5693 dereference if presented with an ASN.1 signature using the RSA PSS
5694 algorithm and absent mask generation function parameter. Since these
5695 routines are used to verify certificate signature algorithms this can be
5696 used to crash any certificate verification operation and exploited in a
5697 DoS attack. Any application which performs certificate verification is
5698 vulnerable including OpenSSL clients and servers which enable client
5701 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5706 * X509_ATTRIBUTE memory leak
5708 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5709 memory. This structure is used by the PKCS#7 and CMS routines so any
5710 application which reads PKCS#7 or CMS data from untrusted sources is
5711 affected. SSL/TLS is not affected.
5713 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5719 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5720 This changes the decoding behaviour for some invalid messages,
5721 though the change is mostly in the more lenient direction, and
5722 legacy behaviour is preserved as much as possible.
5726 * In DSA_generate_parameters_ex, if the provided seed is too short,
5729 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5731 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5733 * Alternate chains certificate forgery
5735 During certificate verification, OpenSSL will attempt to find an
5736 alternative certificate chain if the first attempt to build such a chain
5737 fails. An error in the implementation of this logic can mean that an
5738 attacker could cause certain checks on untrusted certificates to be
5739 bypassed, such as the CA flag, enabling them to use a valid leaf
5740 certificate to act as a CA and "issue" an invalid certificate.
5742 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5747 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5749 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5750 incompatibility in the handling of HMAC. The previous ABI has now been
5755 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5757 * Malformed ECParameters causes infinite loop
5759 When processing an ECParameters structure OpenSSL enters an infinite loop
5760 if the curve specified is over a specially malformed binary polynomial
5763 This can be used to perform denial of service against any
5764 system which processes public keys, certificate requests or
5765 certificates. This includes TLS clients and TLS servers with
5766 client authentication enabled.
5768 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5773 * Exploitable out-of-bounds read in X509_cmp_time
5775 X509_cmp_time does not properly check the length of the ASN1_TIME
5776 string and can read a few bytes out of bounds. In addition,
5777 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5780 An attacker can use this to craft malformed certificates and CRLs of
5781 various sizes and potentially cause a segmentation fault, resulting in
5782 a DoS on applications that verify certificates or CRLs. TLS clients
5783 that verify CRLs are affected. TLS clients and servers with client
5784 authentication enabled may be affected if they use custom verification
5787 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5788 independently by Hanno Böck.
5793 * PKCS7 crash with missing EnvelopedContent
5795 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5796 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5797 with missing content and trigger a NULL pointer dereference on parsing.
5799 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5800 structures from untrusted sources are affected. OpenSSL clients and
5801 servers are not affected.
5803 This issue was reported to OpenSSL by Michal Zalewski (Google).
5808 * CMS verify infinite loop with unknown hash function
5810 When verifying a signedData message the CMS code can enter an infinite loop
5811 if presented with an unknown hash function OID. This can be used to perform
5812 denial of service against any system which verifies signedData messages using
5814 This issue was reported to OpenSSL by Johannes Bauer.
5819 * Race condition handling NewSessionTicket
5821 If a NewSessionTicket is received by a multi-threaded client when attempting to
5822 reuse a previous ticket then a race condition can occur potentially leading to
5823 a double free of the ticket data.
5828 * Only support 256-bit or stronger elliptic curves with the
5829 'ecdh_auto' setting (server) or by default (client). Of supported
5830 curves, prefer P-256 (both).
5834 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5836 * ClientHello sigalgs DoS fix
5838 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5839 invalid signature algorithms extension a NULL pointer dereference will
5840 occur. This can be exploited in a DoS attack against the server.
5842 This issue was was reported to OpenSSL by David Ramos of Stanford
5846 *Stephen Henson and Matt Caswell*
5848 * Multiblock corrupted pointer fix
5850 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5851 feature only applies on 64 bit x86 architecture platforms that support AES
5852 NI instructions. A defect in the implementation of "multiblock" can cause
5853 OpenSSL's internal write buffer to become incorrectly set to NULL when
5854 using non-blocking IO. Typically, when the user application is using a
5855 socket BIO for writing, this will only result in a failed connection.
5856 However if some other BIO is used then it is likely that a segmentation
5857 fault will be triggered, thus enabling a potential DoS attack.
5859 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5864 * Segmentation fault in DTLSv1_listen fix
5866 The DTLSv1_listen function is intended to be stateless and processes the
5867 initial ClientHello from many peers. It is common for user code to loop
5868 over the call to DTLSv1_listen until a valid ClientHello is received with
5869 an associated cookie. A defect in the implementation of DTLSv1_listen means
5870 that state is preserved in the SSL object from one invocation to the next
5871 that can lead to a segmentation fault. Errors processing the initial
5872 ClientHello can trigger this scenario. An example of such an error could be
5873 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5876 This issue was reported to OpenSSL by Per Allansson.
5881 * Segmentation fault in ASN1_TYPE_cmp fix
5883 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5884 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5885 certificate signature algorithm consistency this can be used to crash any
5886 certificate verification operation and exploited in a DoS attack. Any
5887 application which performs certificate verification is vulnerable including
5888 OpenSSL clients and servers which enable client authentication.
5893 * Segmentation fault for invalid PSS parameters fix
5895 The signature verification routines will crash with a NULL pointer
5896 dereference if presented with an ASN.1 signature using the RSA PSS
5897 algorithm and invalid parameters. Since these routines are used to verify
5898 certificate signature algorithms this can be used to crash any
5899 certificate verification operation and exploited in a DoS attack. Any
5900 application which performs certificate verification is vulnerable including
5901 OpenSSL clients and servers which enable client authentication.
5903 This issue was was reported to OpenSSL by Brian Carpenter.
5908 * ASN.1 structure reuse memory corruption fix
5910 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5911 memory corruption via an invalid write. Such reuse is and has been
5912 strongly discouraged and is believed to be rare.
5914 Applications that parse structures containing CHOICE or ANY DEFINED BY
5915 components may be affected. Certificate parsing (d2i_X509 and related
5916 functions) are however not affected. OpenSSL clients and servers are
5922 * PKCS7 NULL pointer dereferences fix
5924 The PKCS#7 parsing code does not handle missing outer ContentInfo
5925 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5926 missing content and trigger a NULL pointer dereference on parsing.
5928 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5929 otherwise parse PKCS#7 structures from untrusted sources are
5930 affected. OpenSSL clients and servers are not affected.
5932 This issue was reported to OpenSSL by Michal Zalewski (Google).
5937 * DoS via reachable assert in SSLv2 servers fix
5939 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5940 servers that both support SSLv2 and enable export cipher suites by sending
5941 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5943 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5944 (OpenSSL development team).
5949 * Empty CKE with client auth and DHE fix
5951 If client auth is used then a server can seg fault in the event of a DHE
5952 ciphersuite being selected and a zero length ClientKeyExchange message
5953 being sent by the client. This could be exploited in a DoS attack.
5958 * Handshake with unseeded PRNG fix
5960 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5961 with an unseeded PRNG. The conditions are:
5962 - The client is on a platform where the PRNG has not been seeded
5963 automatically, and the user has not seeded manually
5964 - A protocol specific client method version has been used (i.e. not
5965 SSL_client_methodv23)
5966 - A ciphersuite is used that does not require additional random data from
5967 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5969 If the handshake succeeds then the client random that has been used will
5970 have been generated from a PRNG with insufficient entropy and therefore the
5971 output may be predictable.
5973 For example using the following command with an unseeded openssl will
5974 succeed on an unpatched platform:
5976 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5981 * Use After Free following d2i_ECPrivatekey error fix
5983 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5984 could cause a use after free condition. This, in turn, could cause a double
5985 free in several private key parsing functions (such as d2i_PrivateKey
5986 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5987 for applications that receive EC private keys from untrusted
5988 sources. This scenario is considered rare.
5990 This issue was discovered by the BoringSSL project and fixed in their
5996 * X509_to_X509_REQ NULL pointer deref fix
5998 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5999 the certificate key is invalid. This function is rarely used in practice.
6001 This issue was discovered by Brian Carpenter.
6006 * Removed the export ciphers from the DEFAULT ciphers
6010 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6012 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6013 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6014 So far those who have to target multiple platforms would compromise
6015 and argue that binary targeting say ARMv5 would still execute on
6016 ARMv8. "Universal" build resolves this compromise by providing
6017 near-optimal performance even on newer platforms.
6021 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6022 (other platforms pending).
6024 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6026 * Add support for the SignedCertificateTimestampList certificate and
6027 OCSP response extensions from RFC6962.
6031 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6032 for corner cases. (Certain input points at infinity could lead to
6033 bogus results, with non-infinity inputs mapped to infinity too.)
6037 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6038 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6039 common cases are optimized and there still is room for further
6040 improvements. Vector Permutation AES for Altivec is also added.
6044 * Add support for little-endian ppc64 Linux target.
6046 *Marcelo Cerri (IBM)*
6048 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6049 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6050 are optimized and there still is room for further improvements.
6051 Both 32- and 64-bit modes are supported.
6053 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6055 * Improved ARMv7 NEON support.
6059 * Support for SPARC Architecture 2011 crypto extensions, first
6060 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6061 SHA256/512, MD5, GHASH and modular exponentiation.
6063 *Andy Polyakov, David Miller*
6065 * Accelerated modular exponentiation for Intel processors, a.k.a.
6068 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6070 * Support for new and upcoming Intel processors, including AVX2,
6071 BMI and SHA ISA extensions. This includes additional "stitched"
6072 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6075 This work was sponsored by Intel Corp.
6079 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6080 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6081 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6085 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6086 this fixes a limitation in previous versions of OpenSSL.
6090 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6091 MGF1 digest and OAEP label.
6095 * Add EVP support for key wrapping algorithms, to avoid problems with
6096 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6097 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6098 algorithms and include tests cases.
6102 * Add functions to allocate and set the fields of an ECDSA_METHOD
6105 *Douglas E. Engert, Steve Henson*
6107 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6108 difference in days and seconds between two tm or ASN1_TIME structures.
6112 * Add -rev test option to s_server to just reverse order of characters
6113 received by client and send back to server. Also prints an abbreviated
6114 summary of the connection parameters.
6118 * New option -brief for s_client and s_server to print out a brief summary
6119 of connection parameters.
6123 * Add callbacks for arbitrary TLS extensions.
6125 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6127 * New option -crl_download in several openssl utilities to download CRLs
6128 from CRLDP extension in certificates.
6132 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6136 * New function X509_CRL_diff to generate a delta CRL from the difference
6137 of two full CRLs. Add support to "crl" utility.
6141 * New functions to set lookup_crls function and to retrieve
6142 X509_STORE from X509_STORE_CTX.
6146 * Print out deprecated issuer and subject unique ID fields in
6151 * Extend OCSP I/O functions so they can be used for simple general purpose
6152 HTTP as well as OCSP. New wrapper function which can be used to download
6153 CRLs using the OCSP API.
6157 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6161 * `SSL_CONF*` functions. These provide a common framework for application
6162 configuration using configuration files or command lines.
6166 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6167 message callback and prints the results. Needs compile time option
6168 "enable-ssl-trace". New options to s_client and s_server to enable
6173 * New ctrl and macro to retrieve supported points extensions.
6174 Print out extension in s_server and s_client.
6178 * New functions to retrieve certificate signature and signature
6183 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6188 * New Suite B modes for TLS code. These use and enforce the requirements
6189 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6190 only use Suite B curves. The Suite B modes can be set by using the
6191 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6195 * New chain verification flags for Suite B levels of security. Check
6196 algorithms are acceptable when flags are set in X509_verify_cert.
6200 * Make tls1_check_chain return a set of flags indicating checks passed
6201 by a certificate chain. Add additional tests to handle client
6202 certificates: checks for matching certificate type and issuer name
6207 * If an attempt is made to use a signature algorithm not in the peer
6208 preference list abort the handshake. If client has no suitable
6209 signature algorithms in response to a certificate request do not
6210 use the certificate.
6214 * If server EC tmp key is not in client preference list abort handshake.
6218 * Add support for certificate stores in CERT structure. This makes it
6219 possible to have different stores per SSL structure or one store in
6220 the parent SSL_CTX. Include distinct stores for certificate chain
6221 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6222 to build and store a certificate chain in CERT structure: returning
6223 an error if the chain cannot be built: this will allow applications
6224 to test if a chain is correctly configured.
6226 Note: if the CERT based stores are not set then the parent SSL_CTX
6227 store is used to retain compatibility with existing behaviour.
6231 * New function ssl_set_client_disabled to set a ciphersuite disabled
6232 mask based on the current session, check mask when sending client
6233 hello and checking the requested ciphersuite.
6237 * New ctrls to retrieve and set certificate types in a certificate
6238 request message. Print out received values in s_client. If certificate
6239 types is not set with custom values set sensible values based on
6240 supported signature algorithms.
6244 * Support for distinct client and server supported signature algorithms.
6248 * Add certificate callback. If set this is called whenever a certificate
6249 is required by client or server. An application can decide which
6250 certificate chain to present based on arbitrary criteria: for example
6251 supported signature algorithms. Add very simple example to s_server.
6252 This fixes many of the problems and restrictions of the existing client
6253 certificate callback: for example you can now clear an existing
6254 certificate and specify the whole chain.
6258 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6259 the certificate can be used for (if anything). Set valid_flags field
6260 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6261 to have similar checks in it.
6263 Add new "cert_flags" field to CERT structure and include a "strict mode".
6264 This enforces some TLS certificate requirements (such as only permitting
6265 certificate signature algorithms contained in the supported algorithms
6266 extension) which some implementations ignore: this option should be used
6267 with caution as it could cause interoperability issues.
6271 * Update and tidy signature algorithm extension processing. Work out
6272 shared signature algorithms based on preferences and peer algorithms
6273 and print them out in s_client and s_server. Abort handshake if no
6274 shared signature algorithms.
6278 * Add new functions to allow customised supported signature algorithms
6279 for SSL and SSL_CTX structures. Add options to s_client and s_server
6284 * New function SSL_certs_clear() to delete all references to certificates
6285 from an SSL structure. Before this once a certificate had been added
6286 it couldn't be removed.
6290 * Integrate hostname, email address and IP address checking with certificate
6291 verification. New verify options supporting checking in openssl utility.
6295 * Fixes and wildcard matching support to hostname and email checking
6296 functions. Add manual page.
6298 *Florian Weimer (Red Hat Product Security Team)*
6300 * New functions to check a hostname email or IP address against a
6301 certificate. Add options x509 utility to print results of checks against
6306 * Fix OCSP checking.
6308 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6310 * Initial experimental support for explicitly trusted non-root CAs.
6311 OpenSSL still tries to build a complete chain to a root but if an
6312 intermediate CA has a trust setting included that is used. The first
6313 setting is used: whether to trust (e.g., -addtrust option to the x509
6318 * Add -trusted_first option which attempts to find certificates in the
6319 trusted store even if an untrusted chain is also supplied.
6323 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6324 platform support for Linux and Android.
6328 * Support for linux-x32, ILP32 environment in x86_64 framework.
6332 * Experimental multi-implementation support for FIPS capable OpenSSL.
6333 When in FIPS mode the approved implementations are used as normal,
6334 when not in FIPS mode the internal unapproved versions are used instead.
6335 This means that the FIPS capable OpenSSL isn't forced to use the
6336 (often lower performance) FIPS implementations outside FIPS mode.
6340 * Transparently support X9.42 DH parameters when calling
6341 PEM_read_bio_DHparameters. This means existing applications can handle
6342 the new parameter format automatically.
6346 * Initial experimental support for X9.42 DH parameter format: mainly
6347 to support use of 'q' parameter for RFC5114 parameters.
6351 * Add DH parameters from RFC5114 including test data to dhtest.
6355 * Support for automatic EC temporary key parameter selection. If enabled
6356 the most preferred EC parameters are automatically used instead of
6357 hardcoded fixed parameters. Now a server just has to call:
6358 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6359 support ECDH and use the most appropriate parameters.
6363 * Enhance and tidy EC curve and point format TLS extension code. Use
6364 static structures instead of allocation if default values are used.
6365 New ctrls to set curves we wish to support and to retrieve shared curves.
6366 Print out shared curves in s_server. New options to s_server and s_client
6367 to set list of supported curves.
6371 * New ctrls to retrieve supported signature algorithms and
6372 supported curve values as an array of NIDs. Extend openssl utility
6373 to print out received values.
6377 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6378 between NIDs and the more common NIST names such as "P-256". Enhance
6379 ecparam utility and ECC method to recognise the NIST names for curves.
6383 * Enhance SSL/TLS certificate chain handling to support different
6384 chains for each certificate instead of one chain in the parent SSL_CTX.
6388 * Support for fixed DH ciphersuite client authentication: where both
6389 server and client use DH certificates with common parameters.
6393 * Support for fixed DH ciphersuites: those requiring DH server
6398 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6400 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6401 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6402 X509_CINF_get_signature were reverted post internal team review.
6407 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6409 * OCSP Status Request extension unbounded memory growth
6411 A malicious client can send an excessively large OCSP Status Request
6412 extension. If that client continually requests renegotiation, sending a
6413 large OCSP Status Request extension each time, then there will be unbounded
6414 memory growth on the server. This will eventually lead to a Denial Of
6415 Service attack through memory exhaustion. Servers with a default
6416 configuration are vulnerable even if they do not support OCSP. Builds using
6417 the "no-ocsp" build time option are not affected.
6419 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6424 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6427 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6433 * OOB write in MDC2_Update()
6435 An overflow can occur in MDC2_Update() either if called directly or
6436 through the EVP_DigestUpdate() function using MDC2. If an attacker
6437 is able to supply very large amounts of input data after a previous
6438 call to EVP_EncryptUpdate() with a partial block then a length check
6439 can overflow resulting in a heap corruption.
6441 The amount of data needed is comparable to SIZE_MAX which is impractical
6444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6449 * Malformed SHA512 ticket DoS
6451 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6452 DoS attack where a malformed ticket will result in an OOB read which will
6455 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6456 a custom server callback and ticket lookup mechanism.
6458 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6463 * OOB write in BN_bn2dec()
6465 The function BN_bn2dec() does not check the return value of BN_div_word().
6466 This can cause an OOB write if an application uses this function with an
6467 overly large BIGNUM. This could be a problem if an overly large certificate
6468 or CRL is printed out from an untrusted source. TLS is not affected because
6469 record limits will reject an oversized certificate before it is parsed.
6471 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6476 * OOB read in TS_OBJ_print_bio()
6478 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6479 the total length the OID text representation would use and not the amount
6480 of data written. This will result in OOB reads when large OIDs are
6483 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6488 * Pointer arithmetic undefined behaviour
6490 Avoid some undefined pointer arithmetic
6492 A common idiom in the codebase is to check limits in the following manner:
6495 Where "p" points to some malloc'd data of SIZE bytes and
6498 "len" here could be from some externally supplied data (e.g. from a TLS
6501 The rules of C pointer arithmetic are such that "p + len" is only well
6502 defined where len <= SIZE. Therefore the above idiom is actually
6503 undefined behaviour.
6505 For example this could cause problems if some malloc implementation
6506 provides an address for "p" such that "p + len" actually overflows for
6507 values of len that are too big and therefore p + len < limit.
6509 This issue was reported to OpenSSL by Guido Vranken
6514 * Constant time flag not preserved in DSA signing
6516 Operations in the DSA signing algorithm should run in constant time in
6517 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6518 implementation means that a non-constant time codepath is followed for
6519 certain operations. This has been demonstrated through a cache-timing
6520 attack to be sufficient for an attacker to recover the private DSA key.
6522 This issue was reported by César Pereida (Aalto University), Billy Brumley
6523 (Tampere University of Technology), and Yuval Yarom (The University of
6524 Adelaide and NICTA).
6529 * DTLS buffered message DoS
6531 In a DTLS connection where handshake messages are delivered out-of-order
6532 those messages that OpenSSL is not yet ready to process will be buffered
6533 for later use. Under certain circumstances, a flaw in the logic means that
6534 those messages do not get removed from the buffer even though the handshake
6535 has been completed. An attacker could force up to approx. 15 messages to
6536 remain in the buffer when they are no longer required. These messages will
6537 be cleared when the DTLS connection is closed. The default maximum size for
6538 a message is 100k. Therefore the attacker could force an additional 1500k
6539 to be consumed per connection. By opening many simulataneous connections an
6540 attacker could cause a DoS attack through memory exhaustion.
6542 This issue was reported to OpenSSL by Quan Luo.
6547 * DTLS replay protection DoS
6549 A flaw in the DTLS replay attack protection mechanism means that records
6550 that arrive for future epochs update the replay protection "window" before
6551 the MAC for the record has been validated. This could be exploited by an
6552 attacker by sending a record for the next epoch (which does not have to
6553 decrypt or have a valid MAC), with a very large sequence number. This means
6554 that all subsequent legitimate packets are dropped causing a denial of
6555 service for a specific DTLS connection.
6557 This issue was reported to OpenSSL by the OCAP audit team.
6562 * Certificate message OOB reads
6564 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6565 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6566 theoretical DoS risk but this has not been observed in practice on common
6569 The messages affected are client certificate, client certificate request
6570 and server certificate. As a result the attack can only be performed
6571 against a client or a server which enables client authentication.
6573 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6578 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6580 * Prevent padding oracle in AES-NI CBC MAC check
6582 A MITM attacker can use a padding oracle attack to decrypt traffic
6583 when the connection uses an AES CBC cipher and the server support
6586 This issue was introduced as part of the fix for Lucky 13 padding
6587 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6588 constant time by making sure that always the same bytes are read and
6589 compared against either the MAC or padding bytes. But it no longer
6590 checked that there was enough data to have both the MAC and padding
6593 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6598 * Fix EVP_EncodeUpdate overflow
6600 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6601 Base64 encoding of binary data. If an attacker is able to supply very large
6602 amounts of input data then a length check can overflow resulting in a heap
6605 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6606 the `PEM_write_bio*` family of functions. These are mainly used within the
6607 OpenSSL command line applications, so any application which processes data
6608 from an untrusted source and outputs it as a PEM file should be considered
6609 vulnerable to this issue. User applications that call these APIs directly
6610 with large amounts of untrusted data may also be vulnerable.
6612 This issue was reported by Guido Vranken.
6617 * Fix EVP_EncryptUpdate overflow
6619 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6620 is able to supply very large amounts of input data after a previous call to
6621 EVP_EncryptUpdate() with a partial block then a length check can overflow
6622 resulting in a heap corruption. Following an analysis of all OpenSSL
6623 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6624 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6625 the first called function after an EVP_EncryptInit(), and therefore that
6626 specific call must be safe. The second form is where the length passed to
6627 EVP_EncryptUpdate() can be seen from the code to be some small value and
6628 therefore there is no possibility of an overflow. Since all instances are
6629 one of these two forms, it is believed that there can be no overflows in
6630 internal code due to this problem. It should be noted that
6631 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6632 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6633 of these calls have also been analysed too and it is believed there are no
6634 instances in internal usage where an overflow could occur.
6636 This issue was reported by Guido Vranken.
6641 * Prevent ASN.1 BIO excessive memory allocation
6643 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6644 a short invalid encoding can casuse allocation of large amounts of memory
6645 potentially consuming excessive resources or exhausting memory.
6647 Any application parsing untrusted data through d2i BIO functions is
6648 affected. The memory based functions such as d2i_X509() are *not* affected.
6649 Since the memory based functions are used by the TLS library, TLS
6650 applications are not affected.
6652 This issue was reported by Brian Carpenter.
6659 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6660 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6661 in arbitrary stack data being returned in the buffer.
6663 This issue was reported by Guido Vranken.
6668 * Modify behavior of ALPN to invoke callback after SNI/servername
6669 callback, such that updates to the SSL_CTX affect ALPN.
6673 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6678 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6679 methods are enabled and ssl2 is disabled the methods return NULL.
6683 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6685 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6686 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6687 provide any "EXPORT" or "LOW" strength ciphers.
6691 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6692 is by default disabled at build-time. Builds that are not configured with
6693 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6694 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6695 will need to explicitly call either of:
6697 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6699 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6701 as appropriate. Even if either of those is used, or the application
6702 explicitly uses the version-specific SSLv2_method() or its client and
6703 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6704 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6705 ciphers, and SSLv2 56-bit DES are no longer available.
6710 * Fix a double-free in DSA code
6712 A double free bug was discovered when OpenSSL parses malformed DSA private
6713 keys and could lead to a DoS attack or memory corruption for applications
6714 that receive DSA private keys from untrusted sources. This scenario is
6717 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6723 * Disable SRP fake user seed to address a server memory leak.
6725 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6727 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6728 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6729 was changed to ignore the "fake user" SRP seed, even if the seed
6732 Users should use SRP_VBASE_get1_by_user instead. Note that in
6733 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6734 also that even though configuring the SRP seed attempts to hide
6735 invalid usernames by continuing the handshake with fake
6736 credentials, this behaviour is not constant time and no strong
6737 guarantees are made that the handshake is indistinguishable from
6738 that of a valid user.
6743 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6745 In the BN_hex2bn function the number of hex digits is calculated using an
6746 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6747 large values of `i` this can result in `bn_expand` not allocating any
6748 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6749 field as NULL leading to a subsequent NULL ptr deref. For very large values
6750 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6751 In this case memory is allocated to the internal BIGNUM data field, but it
6752 is insufficiently sized leading to heap corruption. A similar issue exists
6753 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6754 is ever called by user applications with very large untrusted hex/dec data.
6755 This is anticipated to be a rare occurrence.
6757 All OpenSSL internal usage of these functions use data that is not expected
6758 to be untrusted, e.g. config file data or application command line
6759 arguments. If user developed applications generate config file data based
6760 on untrusted data then it is possible that this could also lead to security
6761 consequences. This is also anticipated to be rare.
6763 This issue was reported to OpenSSL by Guido Vranken.
6768 * Fix memory issues in `BIO_*printf` functions
6770 The internal `fmtstr` function used in processing a "%s" format string in
6771 the `BIO_*printf` functions could overflow while calculating the length of a
6772 string and cause an OOB read when printing very long strings.
6774 Additionally the internal `doapr_outch` function can attempt to write to an
6775 OOB memory location (at an offset from the NULL pointer) in the event of a
6776 memory allocation failure. In 1.0.2 and below this could be caused where
6777 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6778 could be in processing a very long "%s" format string. Memory leaks can
6781 The first issue may mask the second issue dependent on compiler behaviour.
6782 These problems could enable attacks where large amounts of untrusted data
6783 is passed to the `BIO_*printf` functions. If applications use these functions
6784 in this way then they could be vulnerable. OpenSSL itself uses these
6785 functions when printing out human-readable dumps of ASN.1 data. Therefore
6786 applications that print this data could be vulnerable if the data is from
6787 untrusted sources. OpenSSL command line applications could also be
6788 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6789 as command line arguments.
6791 Libssl is not considered directly vulnerable. Additionally certificates etc
6792 received via remote connections via libssl are also unlikely to be able to
6793 trigger these issues because of message size limits enforced within libssl.
6795 This issue was reported to OpenSSL Guido Vranken.
6800 * Side channel attack on modular exponentiation
6802 A side-channel attack was found which makes use of cache-bank conflicts on
6803 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6804 of RSA keys. The ability to exploit this issue is limited as it relies on
6805 an attacker who has control of code in a thread running on the same
6806 hyper-threaded core as the victim thread which is performing decryptions.
6808 This issue was reported to OpenSSL by Yuval Yarom, The University of
6809 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6810 Nadia Heninger, University of Pennsylvania with more information at
6811 <http://cachebleed.info>.
6816 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6817 if no keysize is specified with default_bits. This fixes an
6818 omission in an earlier change that changed all RSA/DSA key generation
6819 commands to use 2048 bits by default.
6823 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6825 * Protection for DH small subgroup attacks
6827 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6828 switched on by default and cannot be disabled. This could have some
6833 * SSLv2 doesn't block disabled ciphers
6835 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6836 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6837 been disabled, provided that the SSLv2 protocol was not also disabled via
6840 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6841 and Sebastian Schinzel.
6846 * Reject DH handshakes with parameters shorter than 1024 bits.
6850 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6852 * Certificate verify crash with missing PSS parameter
6854 The signature verification routines will crash with a NULL pointer
6855 dereference if presented with an ASN.1 signature using the RSA PSS
6856 algorithm and absent mask generation function parameter. Since these
6857 routines are used to verify certificate signature algorithms this can be
6858 used to crash any certificate verification operation and exploited in a
6859 DoS attack. Any application which performs certificate verification is
6860 vulnerable including OpenSSL clients and servers which enable client
6863 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6868 * X509_ATTRIBUTE memory leak
6870 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6871 memory. This structure is used by the PKCS#7 and CMS routines so any
6872 application which reads PKCS#7 or CMS data from untrusted sources is
6873 affected. SSL/TLS is not affected.
6875 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6881 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6882 This changes the decoding behaviour for some invalid messages,
6883 though the change is mostly in the more lenient direction, and
6884 legacy behaviour is preserved as much as possible.
6888 * In DSA_generate_parameters_ex, if the provided seed is too short,
6889 use a random seed, as already documented.
6891 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6893 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6895 * Alternate chains certificate forgery
6897 During certificate verfification, OpenSSL will attempt to find an
6898 alternative certificate chain if the first attempt to build such a chain
6899 fails. An error in the implementation of this logic can mean that an
6900 attacker could cause certain checks on untrusted certificates to be
6901 bypassed, such as the CA flag, enabling them to use a valid leaf
6902 certificate to act as a CA and "issue" an invalid certificate.
6904 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6910 * Race condition handling PSK identify hint
6912 If PSK identity hints are received by a multi-threaded client then
6913 the values are wrongly updated in the parent SSL_CTX structure. This can
6914 result in a race condition potentially leading to a double free of the
6920 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6922 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6923 incompatibility in the handling of HMAC. The previous ABI has now been
6926 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6928 * Malformed ECParameters causes infinite loop
6930 When processing an ECParameters structure OpenSSL enters an infinite loop
6931 if the curve specified is over a specially malformed binary polynomial
6934 This can be used to perform denial of service against any
6935 system which processes public keys, certificate requests or
6936 certificates. This includes TLS clients and TLS servers with
6937 client authentication enabled.
6939 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6944 * Exploitable out-of-bounds read in X509_cmp_time
6946 X509_cmp_time does not properly check the length of the ASN1_TIME
6947 string and can read a few bytes out of bounds. In addition,
6948 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6951 An attacker can use this to craft malformed certificates and CRLs of
6952 various sizes and potentially cause a segmentation fault, resulting in
6953 a DoS on applications that verify certificates or CRLs. TLS clients
6954 that verify CRLs are affected. TLS clients and servers with client
6955 authentication enabled may be affected if they use custom verification
6958 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6959 independently by Hanno Böck.
6964 * PKCS7 crash with missing EnvelopedContent
6966 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6967 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6968 with missing content and trigger a NULL pointer dereference on parsing.
6970 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6971 structures from untrusted sources are affected. OpenSSL clients and
6972 servers are not affected.
6974 This issue was reported to OpenSSL by Michal Zalewski (Google).
6979 * CMS verify infinite loop with unknown hash function
6981 When verifying a signedData message the CMS code can enter an infinite loop
6982 if presented with an unknown hash function OID. This can be used to perform
6983 denial of service against any system which verifies signedData messages using
6985 This issue was reported to OpenSSL by Johannes Bauer.
6990 * Race condition handling NewSessionTicket
6992 If a NewSessionTicket is received by a multi-threaded client when attempting to
6993 reuse a previous ticket then a race condition can occur potentially leading to
6994 a double free of the ticket data.
6999 * Reject DH handshakes with parameters shorter than 768 bits.
7001 *Kurt Roeckx and Emilia Kasper*
7003 * dhparam: generate 2048-bit parameters by default.
7005 *Kurt Roeckx and Emilia Kasper*
7007 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7009 * Segmentation fault in ASN1_TYPE_cmp fix
7011 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7012 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7013 certificate signature algorithm consistency this can be used to crash any
7014 certificate verification operation and exploited in a DoS attack. Any
7015 application which performs certificate verification is vulnerable including
7016 OpenSSL clients and servers which enable client authentication.
7021 * ASN.1 structure reuse memory corruption fix
7023 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7024 memory corruption via an invalid write. Such reuse is and has been
7025 strongly discouraged and is believed to be rare.
7027 Applications that parse structures containing CHOICE or ANY DEFINED BY
7028 components may be affected. Certificate parsing (d2i_X509 and related
7029 functions) are however not affected. OpenSSL clients and servers are
7035 * PKCS7 NULL pointer dereferences fix
7037 The PKCS#7 parsing code does not handle missing outer ContentInfo
7038 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7039 missing content and trigger a NULL pointer dereference on parsing.
7041 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7042 otherwise parse PKCS#7 structures from untrusted sources are
7043 affected. OpenSSL clients and servers are not affected.
7045 This issue was reported to OpenSSL by Michal Zalewski (Google).
7050 * DoS via reachable assert in SSLv2 servers fix
7052 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7053 servers that both support SSLv2 and enable export cipher suites by sending
7054 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7056 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7057 (OpenSSL development team).
7062 * Use After Free following d2i_ECPrivatekey error fix
7064 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7065 could cause a use after free condition. This, in turn, could cause a double
7066 free in several private key parsing functions (such as d2i_PrivateKey
7067 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7068 for applications that receive EC private keys from untrusted
7069 sources. This scenario is considered rare.
7071 This issue was discovered by the BoringSSL project and fixed in their
7077 * X509_to_X509_REQ NULL pointer deref fix
7079 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7080 the certificate key is invalid. This function is rarely used in practice.
7082 This issue was discovered by Brian Carpenter.
7087 * Removed the export ciphers from the DEFAULT ciphers
7091 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7093 * Build fixes for the Windows and OpenVMS platforms
7095 *Matt Caswell and Richard Levitte*
7097 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7099 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7100 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7101 dereference. This could lead to a Denial Of Service attack. Thanks to
7102 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7107 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7108 dtls1_buffer_record function under certain conditions. In particular this
7109 could occur if an attacker sent repeated DTLS records with the same
7110 sequence number but for the next epoch. The memory leak could be exploited
7111 by an attacker in a Denial of Service attack through memory exhaustion.
7112 Thanks to Chris Mueller for reporting this issue.
7117 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7118 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7119 method would be set to NULL which could later result in a NULL pointer
7120 dereference. Thanks to Frank Schmirler for reporting this issue.
7125 * Abort handshake if server key exchange message is omitted for ephemeral
7128 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7129 reporting this issue.
7134 * Remove non-export ephemeral RSA code on client and server. This code
7135 violated the TLS standard by allowing the use of temporary RSA keys in
7136 non-export ciphersuites and could be used by a server to effectively
7137 downgrade the RSA key length used to a value smaller than the server
7138 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7139 INRIA or reporting this issue.
7144 * Fixed issue where DH client certificates are accepted without verification.
7145 An OpenSSL server will accept a DH certificate for client authentication
7146 without the certificate verify message. This effectively allows a client to
7147 authenticate without the use of a private key. This only affects servers
7148 which trust a client certificate authority which issues certificates
7149 containing DH keys: these are extremely rare and hardly ever encountered.
7150 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7156 * Ensure that the session ID context of an SSL is updated when its
7157 SSL_CTX is updated via SSL_set_SSL_CTX.
7159 The session ID context is typically set from the parent SSL_CTX,
7160 and can vary with the CTX.
7164 * Fix various certificate fingerprint issues.
7166 By using non-DER or invalid encodings outside the signed portion of a
7167 certificate the fingerprint can be changed without breaking the signature.
7168 Although no details of the signed portion of the certificate can be changed
7169 this can cause problems with some applications: e.g. those using the
7170 certificate fingerprint for blacklists.
7172 1. Reject signatures with non zero unused bits.
7174 If the BIT STRING containing the signature has non zero unused bits reject
7175 the signature. All current signature algorithms require zero unused bits.
7177 2. Check certificate algorithm consistency.
7179 Check the AlgorithmIdentifier inside TBS matches the one in the
7180 certificate signature. NB: this will result in signature failure
7181 errors for some broken certificates.
7183 Thanks to Konrad Kraszewski from Google for reporting this issue.
7185 3. Check DSA/ECDSA signatures use DER.
7187 Re-encode DSA/ECDSA signatures and compare with the original received
7188 signature. Return an error if there is a mismatch.
7190 This will reject various cases including garbage after signature
7191 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7192 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7193 (negative or with leading zeroes).
7195 Further analysis was conducted and fixes were developed by Stephen Henson
7196 of the OpenSSL core team.
7202 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7203 results on some platforms, including x86_64. This bug occurs at random
7204 with a very low probability, and is not known to be exploitable in any
7205 way, though its exact impact is difficult to determine. Thanks to Pieter
7206 Wuille (Blockstream) who reported this issue and also suggested an initial
7207 fix. Further analysis was conducted by the OpenSSL development team and
7208 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7209 the OpenSSL core team.
7214 * Do not resume sessions on the server if the negotiated protocol
7215 version does not match the session's version. Resuming with a different
7216 version, while not strictly forbidden by the RFC, is of questionable
7217 sanity and breaks all known clients.
7219 *David Benjamin, Emilia Käsper*
7221 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7222 early CCS messages during renegotiation. (Note that because
7223 renegotiation is encrypted, this early CCS was not exploitable.)
7227 * Tighten client-side session ticket handling during renegotiation:
7228 ensure that the client only accepts a session ticket if the server sends
7229 the extension anew in the ServerHello. Previously, a TLS client would
7230 reuse the old extension state and thus accept a session ticket if one was
7231 announced in the initial ServerHello.
7233 Similarly, ensure that the client requires a session ticket if one
7234 was advertised in the ServerHello. Previously, a TLS client would
7235 ignore a missing NewSessionTicket message.
7239 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7243 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7244 sends a carefully crafted handshake message, to cause OpenSSL to fail
7245 to free up to 64k of memory causing a memory leak. This could be
7246 exploited in a Denial Of Service attack. This issue affects OpenSSL
7247 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7248 whether SRTP is used or configured. Implementations of OpenSSL that
7249 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7251 The fix was developed by the OpenSSL team.
7256 * Session Ticket Memory Leak.
7258 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7259 integrity of that ticket is first verified. In the event of a session
7260 ticket integrity check failing, OpenSSL will fail to free memory
7261 causing a memory leak. By sending a large number of invalid session
7262 tickets an attacker could exploit this issue in a Denial Of Service
7268 * Build option no-ssl3 is incomplete.
7270 When OpenSSL is configured with "no-ssl3" as a build option, servers
7271 could accept and complete a SSL 3.0 handshake, and clients could be
7272 configured to send them.
7275 *Akamai and the OpenSSL team*
7277 * Add support for TLS_FALLBACK_SCSV.
7278 Client applications doing fallback retries should call
7279 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7282 *Adam Langley, Bodo Moeller*
7284 * Add additional DigestInfo checks.
7286 Re-encode DigestInto in DER and check against the original when
7287 verifying RSA signature: this will reject any improperly encoded
7288 DigestInfo structures.
7290 Note: this is a precautionary measure and no attacks are currently known.
7294 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7296 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7297 SRP code can be overrun an internal buffer. Add sanity check that
7298 g, A, B < N to SRP code.
7300 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7301 Group for discovering this issue.
7306 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7307 TLS 1.0 instead of higher protocol versions when the ClientHello message
7308 is badly fragmented. This allows a man-in-the-middle attacker to force a
7309 downgrade to TLS 1.0 even if both the server and the client support a
7310 higher protocol version, by modifying the client's TLS records.
7312 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7313 researching this issue.
7318 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7319 to a denial of service attack. A malicious server can crash the client
7320 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7321 ciphersuite and sending carefully crafted handshake messages.
7323 Thanks to Felix Gröbert (Google) for discovering and researching this
7329 * By sending carefully crafted DTLS packets an attacker could cause openssl
7330 to leak memory. This can be exploited through a Denial of Service attack.
7331 Thanks to Adam Langley for discovering and researching this issue.
7336 * An attacker can force openssl to consume large amounts of memory whilst
7337 processing DTLS handshake messages. This can be exploited through a
7338 Denial of Service attack.
7339 Thanks to Adam Langley for discovering and researching this issue.
7344 * An attacker can force an error condition which causes openssl to crash
7345 whilst processing DTLS packets due to memory being freed twice. This
7346 can be exploited through a Denial of Service attack.
7347 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7353 * If a multithreaded client connects to a malicious server using a resumed
7354 session and the server sends an ec point format extension it could write
7355 up to 255 bytes to freed memory.
7357 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7363 * A malicious server can crash an OpenSSL client with a null pointer
7364 dereference (read) by specifying an SRP ciphersuite even though it was not
7365 properly negotiated with the client. This can be exploited through a
7366 Denial of Service attack.
7368 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7369 discovering and researching this issue.
7374 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7375 X509_name_oneline, X509_name_print_ex et al. to leak some information
7376 from the stack. Applications may be affected if they echo pretty printing
7377 output to the attacker.
7379 Thanks to Ivan Fratric (Google) for discovering this issue.
7382 *Emilia Käsper, and Steve Henson*
7384 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7385 for corner cases. (Certain input points at infinity could lead to
7386 bogus results, with non-infinity inputs mapped to infinity too.)
7390 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7392 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7393 handshake can force the use of weak keying material in OpenSSL
7394 SSL/TLS clients and servers.
7396 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7397 researching this issue. ([CVE-2014-0224])
7399 *KIKUCHI Masashi, Steve Henson*
7401 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7402 OpenSSL DTLS client the code can be made to recurse eventually crashing
7405 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7408 *Imre Rad, Steve Henson*
7410 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7411 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7412 client or server. This is potentially exploitable to run arbitrary
7413 code on a vulnerable client or server.
7415 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7417 *Jüri Aedla, Steve Henson*
7419 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7420 are subject to a denial of service attack.
7422 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7423 this issue. ([CVE-2014-3470])
7425 *Felix Gröbert, Ivan Fratric, Steve Henson*
7427 * Harmonize version and its documentation. -f flag is used to display
7430 *mancha <mancha1@zoho.com>*
7432 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7433 in i2d_ECPrivateKey.
7435 *mancha <mancha1@zoho.com>*
7437 * Fix some double frees. These are not thought to be exploitable.
7439 *mancha <mancha1@zoho.com>*
7441 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7443 * A missing bounds check in the handling of the TLS heartbeat extension
7444 can be used to reveal up to 64k of memory to a connected client or
7447 Thanks for Neel Mehta of Google Security for discovering this bug and to
7448 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7449 preparing the fix ([CVE-2014-0160])
7451 *Adam Langley, Bodo Moeller*
7453 * Fix for the attack described in the paper "Recovering OpenSSL
7454 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7455 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7456 <http://eprint.iacr.org/2014/140>
7458 Thanks to Yuval Yarom and Naomi Benger for discovering this
7459 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7461 *Yuval Yarom and Naomi Benger*
7463 * TLS pad extension: draft-agl-tls-padding-03
7465 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7466 TLS client Hello record length value would otherwise be > 255 and
7467 less that 512 pad with a dummy extension containing zeroes so it
7468 is at least 512 bytes long.
7470 *Adam Langley, Steve Henson*
7472 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7474 * Fix for TLS record tampering bug. A carefully crafted invalid
7475 handshake could crash OpenSSL with a NULL pointer exception.
7476 Thanks to Anton Johansson for reporting this issues.
7479 * Keep original DTLS digest and encryption contexts in retransmission
7480 structures so we can use the previous session parameters if they need
7481 to be resent. ([CVE-2013-6450])
7485 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7486 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7487 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7488 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7489 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7490 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7492 *Rob Stradling, Adam Langley*
7494 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7496 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7497 supporting platforms or when small records were transferred.
7499 *Andy Polyakov, Steve Henson*
7501 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7503 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7505 This addresses the flaw in CBC record processing discovered by
7506 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7507 at: <http://www.isg.rhul.ac.uk/tls/>
7509 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7510 Security Group at Royal Holloway, University of London
7511 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7512 Emilia Käsper for the initial patch.
7515 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7517 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7518 ciphersuites which can be exploited in a denial of service attack.
7519 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7520 and detecting this bug and to Wolfgang Ettlinger
7521 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7526 * Return an error when checking OCSP signatures when key is NULL.
7527 This fixes a DoS attack. ([CVE-2013-0166])
7531 * Make openssl verify return errors.
7533 *Chris Palmer <palmer@google.com> and Ben Laurie*
7535 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7536 the right response is stapled. Also change SSL_get_certificate()
7537 so it returns the certificate actually sent.
7538 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7540 *Rob Stradling <rob.stradling@comodo.com>*
7542 * Fix possible deadlock when decoding public keys.
7546 * Don't use TLS 1.0 record version number in initial client hello
7551 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7553 * Sanity check record length before skipping explicit IV in TLS
7554 1.2, 1.1 and DTLS to fix DoS attack.
7556 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7557 fuzzing as a service testing platform.
7562 * Initialise tkeylen properly when encrypting CMS messages.
7563 Thanks to Solar Designer of Openwall for reporting this issue.
7567 * In FIPS mode don't try to use composite ciphers as they are not
7572 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7574 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7575 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7576 mean any application compiled against OpenSSL 1.0.0 headers setting
7577 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7578 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7579 0x10000000L Any application which was previously compiled against
7580 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7581 will need to be recompiled as a result. Letting be results in
7582 inability to disable specifically TLS 1.1 and in client context,
7583 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7587 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7588 disable just protocol X, but all protocols above X *if* there are
7589 protocols *below* X still enabled. In more practical terms it means
7590 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7591 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7592 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7597 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7599 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7600 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7601 in CRYPTO_realloc_clean.
7603 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7604 issue and to Adam Langley <agl@chromium.org> for fixing it.
7607 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7609 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7613 * Workarounds for some broken servers that "hang" if a client hello
7614 record length exceeds 255 bytes.
7616 1. Do not use record version number > TLS 1.0 in initial client
7617 hello: some (but not all) hanging servers will now work.
7618 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7619 the number of ciphers sent in the client hello. This should be
7620 set to an even number, such as 50, for example by passing:
7621 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7622 Most broken servers should now work.
7623 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7624 TLS 1.2 client support entirely.
7628 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7632 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7634 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7635 STRING form instead of a DigestInfo.
7639 * The format used for MDC2 RSA signatures is inconsistent between EVP
7640 and the RSA_sign/RSA_verify functions. This was made more apparent when
7641 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7642 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7643 the correct format in RSA_verify so both forms transparently work.
7647 * Some servers which support TLS 1.0 can choke if we initially indicate
7648 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7649 encrypted premaster secret. As a workaround use the maximum permitted
7650 client version in client hello, this should keep such servers happy
7651 and still work with previous versions of OpenSSL.
7655 * Add support for TLS/DTLS heartbeats.
7657 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7659 * Add support for SCTP.
7661 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7663 * Improved PRNG seeding for VOS.
7665 *Paul Green <Paul.Green@stratus.com>*
7667 * Extensive assembler packs updates, most notably:
7669 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7670 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7671 - x86_64: bit-sliced AES implementation;
7672 - ARM: NEON support, contemporary platforms optimizations;
7673 - s390x: z196 support;
7674 - `*`: GHASH and GF(2^m) multiplication implementations;
7678 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7679 (removal of unnecessary code)
7681 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7683 * Add TLS key material exporter from RFC 5705.
7687 * Add DTLS-SRTP negotiation from RFC 5764.
7691 * Add Next Protocol Negotiation,
7692 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7693 disabled with a no-npn flag to config or Configure. Code donated
7696 *Adam Langley <agl@google.com> and Ben Laurie*
7698 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7699 NIST-P256, NIST-P521, with constant-time single point multiplication on
7700 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7701 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7702 Code made available under Apache License version 2.0.
7704 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7705 line to include this in your build of OpenSSL, and run "make depend" (or
7706 "make update"). This enables the following EC_METHODs:
7708 EC_GFp_nistp224_method()
7709 EC_GFp_nistp256_method()
7710 EC_GFp_nistp521_method()
7712 EC_GROUP_new_by_curve_name() will automatically use these (while
7713 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7716 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7718 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7719 all platforms. Move ssize_t definition from e_os.h to the public
7720 header file e_os2.h as it now appears in public header file cms.h
7724 * New -sigopt option to the ca, req and x509 utilities. Additional
7725 signature parameters can be passed using this option and in
7730 * Add RSA PSS signing function. This will generate and set the
7731 appropriate AlgorithmIdentifiers for PSS based on those in the
7732 corresponding EVP_MD_CTX structure. No application support yet.
7736 * Support for companion algorithm specific ASN1 signing routines.
7737 New function ASN1_item_sign_ctx() signs a pre-initialised
7738 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7739 the appropriate parameters.
7743 * Add new algorithm specific ASN1 verification initialisation function
7744 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7745 handling will be the same no matter what EVP_PKEY_METHOD is used.
7746 Add a PSS handler to support verification of PSS signatures: checked
7747 against a number of sample certificates.
7751 * Add signature printing for PSS. Add PSS OIDs.
7753 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7755 * Add algorithm specific signature printing. An individual ASN1 method
7756 can now print out signatures instead of the standard hex dump.
7758 More complex signatures (e.g. PSS) can print out more meaningful
7759 information. Include DSA version that prints out the signature
7764 * Password based recipient info support for CMS library: implementing
7769 * Split password based encryption into PBES2 and PBKDF2 functions. This
7770 neatly separates the code into cipher and PBE sections and is required
7771 for some algorithms that split PBES2 into separate pieces (such as
7772 password based CMS).
7776 * Session-handling fixes:
7777 - Fix handling of connections that are resuming with a session ID,
7778 but also support Session Tickets.
7779 - Fix a bug that suppressed issuing of a new ticket if the client
7780 presented a ticket with an expired session.
7781 - Try to set the ticket lifetime hint to something reasonable.
7782 - Make tickets shorter by excluding irrelevant information.
7783 - On the client side, don't ignore renewed tickets.
7785 *Adam Langley, Bodo Moeller (Google)*
7787 * Fix PSK session representation.
7791 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7793 This work was sponsored by Intel.
7797 * Add GCM support to TLS library. Some custom code is needed to split
7798 the IV between the fixed (from PRF) and explicit (from TLS record)
7799 portions. This adds all GCM ciphersuites supported by RFC5288 and
7800 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7801 add a special AESGCM string for GCM only.
7805 * Expand range of ctrls for AES GCM. Permit setting invocation
7806 field on decrypt and retrieval of invocation field only on encrypt.
7810 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7811 As required by RFC5289 these ciphersuites cannot be used if for
7812 versions of TLS earlier than 1.2.
7816 * For FIPS capable OpenSSL interpret a NULL default public key method
7817 as unset and return the appropriate default but do *not* set the default.
7818 This means we can return the appropriate method in applications that
7819 switch between FIPS and non-FIPS modes.
7823 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7824 ENGINE is used then we cannot handle that in the FIPS module so we
7825 keep original code iff non-FIPS operations are allowed.
7829 * Add -attime option to openssl utilities.
7831 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7833 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7837 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7838 FIPS EC methods unconditionally for now.
7842 * New build option no-ec2m to disable characteristic 2 code.
7846 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7847 all cases can be covered as some introduce binary incompatibilities.
7851 * Redirect RSA operations to FIPS module including keygen,
7852 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7856 * Add similar low-level API blocking to ciphers.
7860 * low-level digest APIs are not approved in FIPS mode: any attempt
7861 to use these will cause a fatal error. Applications that *really* want
7862 to use them can use the `private_*` version instead.
7866 * Redirect cipher operations to FIPS module for FIPS builds.
7870 * Redirect digest operations to FIPS module for FIPS builds.
7874 * Update build system to add "fips" flag which will link in fipscanister.o
7875 for static and shared library builds embedding a signature if needed.
7879 * Output TLS supported curves in preference order instead of numerical
7880 order. This is currently hardcoded for the highest order curves first.
7881 This should be configurable so applications can judge speed vs strength.
7885 * Add TLS v1.2 server support for client authentication.
7889 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7894 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7895 FIPS modules versions.
7899 * Add TLS v1.2 client side support for client authentication. Keep cache
7900 of handshake records longer as we don't know the hash algorithm to use
7901 until after the certificate request message is received.
7905 * Initial TLS v1.2 client support. Add a default signature algorithms
7906 extension including all the algorithms we support. Parse new signature
7907 format in client key exchange. Relax some ECC signing restrictions for
7908 TLS v1.2 as indicated in RFC5246.
7912 * Add server support for TLS v1.2 signature algorithms extension. Switch
7913 to new signature format when needed using client digest preference.
7914 All server ciphersuites should now work correctly in TLS v1.2. No client
7915 support yet and no support for client certificates.
7919 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7920 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7921 ciphersuites. At present only RSA key exchange ciphersuites work with
7922 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7923 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7924 and version checking.
7928 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7929 with this defined it will not be affected by any changes to ssl internal
7930 structures. Add several utility functions to allow openssl application
7931 to work with OPENSSL_NO_SSL_INTERN defined.
7935 * A long standing patch to add support for SRP from EdelWeb (Peter
7936 Sylvester and Christophe Renou) was integrated.
7937 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7938 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7941 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7945 * Permit abbreviated handshakes when renegotiating using the function
7946 SSL_renegotiate_abbreviated().
7948 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7950 * Add call to ENGINE_register_all_complete() to
7951 ENGINE_load_builtin_engines(), so some implementations get used
7952 automatically instead of needing explicit application support.
7956 * Add support for TLS key exporter as described in RFC5705.
7958 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7960 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7961 a few changes are required:
7963 Add SSL_OP_NO_TLSv1_1 flag.
7964 Add TLSv1_1 methods.
7965 Update version checking logic to handle version 1.1.
7966 Add explicit IV handling (ported from DTLS code).
7967 Add command line options to s_client/s_server.
7974 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7976 * X509_ATTRIBUTE memory leak
7978 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7979 memory. This structure is used by the PKCS#7 and CMS routines so any
7980 application which reads PKCS#7 or CMS data from untrusted sources is
7981 affected. SSL/TLS is not affected.
7983 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7989 * Race condition handling PSK identify hint
7991 If PSK identity hints are received by a multi-threaded client then
7992 the values are wrongly updated in the parent SSL_CTX structure. This can
7993 result in a race condition potentially leading to a double free of the
7999 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8001 * Malformed ECParameters causes infinite loop
8003 When processing an ECParameters structure OpenSSL enters an infinite loop
8004 if the curve specified is over a specially malformed binary polynomial
8007 This can be used to perform denial of service against any
8008 system which processes public keys, certificate requests or
8009 certificates. This includes TLS clients and TLS servers with
8010 client authentication enabled.
8012 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8017 * Exploitable out-of-bounds read in X509_cmp_time
8019 X509_cmp_time does not properly check the length of the ASN1_TIME
8020 string and can read a few bytes out of bounds. In addition,
8021 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8024 An attacker can use this to craft malformed certificates and CRLs of
8025 various sizes and potentially cause a segmentation fault, resulting in
8026 a DoS on applications that verify certificates or CRLs. TLS clients
8027 that verify CRLs are affected. TLS clients and servers with client
8028 authentication enabled may be affected if they use custom verification
8031 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8032 independently by Hanno Böck.
8037 * PKCS7 crash with missing EnvelopedContent
8039 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8040 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8041 with missing content and trigger a NULL pointer dereference on parsing.
8043 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8044 structures from untrusted sources are affected. OpenSSL clients and
8045 servers are not affected.
8047 This issue was reported to OpenSSL by Michal Zalewski (Google).
8052 * CMS verify infinite loop with unknown hash function
8054 When verifying a signedData message the CMS code can enter an infinite loop
8055 if presented with an unknown hash function OID. This can be used to perform
8056 denial of service against any system which verifies signedData messages using
8058 This issue was reported to OpenSSL by Johannes Bauer.
8063 * Race condition handling NewSessionTicket
8065 If a NewSessionTicket is received by a multi-threaded client when attempting to
8066 reuse a previous ticket then a race condition can occur potentially leading to
8067 a double free of the ticket data.
8072 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8074 * Segmentation fault in ASN1_TYPE_cmp fix
8076 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8077 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8078 certificate signature algorithm consistency this can be used to crash any
8079 certificate verification operation and exploited in a DoS attack. Any
8080 application which performs certificate verification is vulnerable including
8081 OpenSSL clients and servers which enable client authentication.
8086 * ASN.1 structure reuse memory corruption fix
8088 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8089 memory corruption via an invalid write. Such reuse is and has been
8090 strongly discouraged and is believed to be rare.
8092 Applications that parse structures containing CHOICE or ANY DEFINED BY
8093 components may be affected. Certificate parsing (d2i_X509 and related
8094 functions) are however not affected. OpenSSL clients and servers are
8100 * PKCS7 NULL pointer dereferences fix
8102 The PKCS#7 parsing code does not handle missing outer ContentInfo
8103 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8104 missing content and trigger a NULL pointer dereference on parsing.
8106 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8107 otherwise parse PKCS#7 structures from untrusted sources are
8108 affected. OpenSSL clients and servers are not affected.
8110 This issue was reported to OpenSSL by Michal Zalewski (Google).
8115 * DoS via reachable assert in SSLv2 servers fix
8117 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8118 servers that both support SSLv2 and enable export cipher suites by sending
8119 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8121 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8122 (OpenSSL development team).
8127 * Use After Free following d2i_ECPrivatekey error fix
8129 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8130 could cause a use after free condition. This, in turn, could cause a double
8131 free in several private key parsing functions (such as d2i_PrivateKey
8132 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8133 for applications that receive EC private keys from untrusted
8134 sources. This scenario is considered rare.
8136 This issue was discovered by the BoringSSL project and fixed in their
8142 * X509_to_X509_REQ NULL pointer deref fix
8144 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8145 the certificate key is invalid. This function is rarely used in practice.
8147 This issue was discovered by Brian Carpenter.
8152 * Removed the export ciphers from the DEFAULT ciphers
8156 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8158 * Build fixes for the Windows and OpenVMS platforms
8160 *Matt Caswell and Richard Levitte*
8162 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8164 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8165 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8166 dereference. This could lead to a Denial Of Service attack. Thanks to
8167 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8172 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8173 dtls1_buffer_record function under certain conditions. In particular this
8174 could occur if an attacker sent repeated DTLS records with the same
8175 sequence number but for the next epoch. The memory leak could be exploited
8176 by an attacker in a Denial of Service attack through memory exhaustion.
8177 Thanks to Chris Mueller for reporting this issue.
8182 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8183 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8184 method would be set to NULL which could later result in a NULL pointer
8185 dereference. Thanks to Frank Schmirler for reporting this issue.
8190 * Abort handshake if server key exchange message is omitted for ephemeral
8193 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8194 reporting this issue.
8199 * Remove non-export ephemeral RSA code on client and server. This code
8200 violated the TLS standard by allowing the use of temporary RSA keys in
8201 non-export ciphersuites and could be used by a server to effectively
8202 downgrade the RSA key length used to a value smaller than the server
8203 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8204 INRIA or reporting this issue.
8209 * Fixed issue where DH client certificates are accepted without verification.
8210 An OpenSSL server will accept a DH certificate for client authentication
8211 without the certificate verify message. This effectively allows a client to
8212 authenticate without the use of a private key. This only affects servers
8213 which trust a client certificate authority which issues certificates
8214 containing DH keys: these are extremely rare and hardly ever encountered.
8215 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8221 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8222 results on some platforms, including x86_64. This bug occurs at random
8223 with a very low probability, and is not known to be exploitable in any
8224 way, though its exact impact is difficult to determine. Thanks to Pieter
8225 Wuille (Blockstream) who reported this issue and also suggested an initial
8226 fix. Further analysis was conducted by the OpenSSL development team and
8227 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8228 the OpenSSL core team.
8233 * Fix various certificate fingerprint issues.
8235 By using non-DER or invalid encodings outside the signed portion of a
8236 certificate the fingerprint can be changed without breaking the signature.
8237 Although no details of the signed portion of the certificate can be changed
8238 this can cause problems with some applications: e.g. those using the
8239 certificate fingerprint for blacklists.
8241 1. Reject signatures with non zero unused bits.
8243 If the BIT STRING containing the signature has non zero unused bits reject
8244 the signature. All current signature algorithms require zero unused bits.
8246 2. Check certificate algorithm consistency.
8248 Check the AlgorithmIdentifier inside TBS matches the one in the
8249 certificate signature. NB: this will result in signature failure
8250 errors for some broken certificates.
8252 Thanks to Konrad Kraszewski from Google for reporting this issue.
8254 3. Check DSA/ECDSA signatures use DER.
8256 Reencode DSA/ECDSA signatures and compare with the original received
8257 signature. Return an error if there is a mismatch.
8259 This will reject various cases including garbage after signature
8260 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8261 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8262 (negative or with leading zeroes).
8264 Further analysis was conducted and fixes were developed by Stephen Henson
8265 of the OpenSSL core team.
8271 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8273 * Session Ticket Memory Leak.
8275 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8276 integrity of that ticket is first verified. In the event of a session
8277 ticket integrity check failing, OpenSSL will fail to free memory
8278 causing a memory leak. By sending a large number of invalid session
8279 tickets an attacker could exploit this issue in a Denial Of Service
8285 * Build option no-ssl3 is incomplete.
8287 When OpenSSL is configured with "no-ssl3" as a build option, servers
8288 could accept and complete a SSL 3.0 handshake, and clients could be
8289 configured to send them.
8292 *Akamai and the OpenSSL team*
8294 * Add support for TLS_FALLBACK_SCSV.
8295 Client applications doing fallback retries should call
8296 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8299 *Adam Langley, Bodo Moeller*
8301 * Add additional DigestInfo checks.
8303 Reencode DigestInto in DER and check against the original when
8304 verifying RSA signature: this will reject any improperly encoded
8305 DigestInfo structures.
8307 Note: this is a precautionary measure and no attacks are currently known.
8311 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8313 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8314 to a denial of service attack. A malicious server can crash the client
8315 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8316 ciphersuite and sending carefully crafted handshake messages.
8318 Thanks to Felix Gröbert (Google) for discovering and researching this
8324 * By sending carefully crafted DTLS packets an attacker could cause openssl
8325 to leak memory. This can be exploited through a Denial of Service attack.
8326 Thanks to Adam Langley for discovering and researching this issue.
8331 * An attacker can force openssl to consume large amounts of memory whilst
8332 processing DTLS handshake messages. This can be exploited through a
8333 Denial of Service attack.
8334 Thanks to Adam Langley for discovering and researching this issue.
8339 * An attacker can force an error condition which causes openssl to crash
8340 whilst processing DTLS packets due to memory being freed twice. This
8341 can be exploited through a Denial of Service attack.
8342 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8348 * If a multithreaded client connects to a malicious server using a resumed
8349 session and the server sends an ec point format extension it could write
8350 up to 255 bytes to freed memory.
8352 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8358 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8359 X509_name_oneline, X509_name_print_ex et al. to leak some information
8360 from the stack. Applications may be affected if they echo pretty printing
8361 output to the attacker.
8363 Thanks to Ivan Fratric (Google) for discovering this issue.
8366 *Emilia Käsper, and Steve Henson*
8368 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8369 for corner cases. (Certain input points at infinity could lead to
8370 bogus results, with non-infinity inputs mapped to infinity too.)
8374 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8376 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8377 handshake can force the use of weak keying material in OpenSSL
8378 SSL/TLS clients and servers.
8380 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8381 researching this issue. ([CVE-2014-0224])
8383 *KIKUCHI Masashi, Steve Henson*
8385 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8386 OpenSSL DTLS client the code can be made to recurse eventually crashing
8389 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8392 *Imre Rad, Steve Henson*
8394 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8395 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8396 client or server. This is potentially exploitable to run arbitrary
8397 code on a vulnerable client or server.
8399 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8401 *Jüri Aedla, Steve Henson*
8403 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8404 are subject to a denial of service attack.
8406 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8407 this issue. ([CVE-2014-3470])
8409 *Felix Gröbert, Ivan Fratric, Steve Henson*
8411 * Harmonize version and its documentation. -f flag is used to display
8414 *mancha <mancha1@zoho.com>*
8416 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8417 in i2d_ECPrivateKey.
8419 *mancha <mancha1@zoho.com>*
8421 * Fix some double frees. These are not thought to be exploitable.
8423 *mancha <mancha1@zoho.com>*
8425 * Fix for the attack described in the paper "Recovering OpenSSL
8426 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8427 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8428 <http://eprint.iacr.org/2014/140>
8430 Thanks to Yuval Yarom and Naomi Benger for discovering this
8431 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8433 *Yuval Yarom and Naomi Benger*
8435 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8437 * Keep original DTLS digest and encryption contexts in retransmission
8438 structures so we can use the previous session parameters if they need
8439 to be resent. ([CVE-2013-6450])
8443 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8444 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8445 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8446 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8447 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8448 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8450 *Rob Stradling, Adam Langley*
8452 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8454 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8456 This addresses the flaw in CBC record processing discovered by
8457 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8458 at: <http://www.isg.rhul.ac.uk/tls/>
8460 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8461 Security Group at Royal Holloway, University of London
8462 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8463 Emilia Käsper for the initial patch.
8466 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8468 * Return an error when checking OCSP signatures when key is NULL.
8469 This fixes a DoS attack. ([CVE-2013-0166])
8473 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8474 the right response is stapled. Also change SSL_get_certificate()
8475 so it returns the certificate actually sent.
8476 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8477 (This is a backport)
8479 *Rob Stradling <rob.stradling@comodo.com>*
8481 * Fix possible deadlock when decoding public keys.
8485 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8487 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8490 * Sanity check record length before skipping explicit IV in DTLS
8493 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8494 fuzzing as a service testing platform.
8499 * Initialise tkeylen properly when encrypting CMS messages.
8500 Thanks to Solar Designer of Openwall for reporting this issue.
8504 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8506 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8507 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8508 in CRYPTO_realloc_clean.
8510 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8511 issue and to Adam Langley <agl@chromium.org> for fixing it.
8514 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8516 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8518 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8519 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8520 content decryption and always return the same error. Note: this attack
8521 needs on average 2^20 messages so it only affects automated senders. The
8522 old behaviour can be re-enabled in the CMS code by setting the
8523 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8524 an MMA defence is not necessary.
8525 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8526 this issue. ([CVE-2012-0884])
8530 * Fix CVE-2011-4619: make sure we really are receiving a
8531 client hello before rejecting multiple SGC restarts. Thanks to
8532 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8536 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8538 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8539 Thanks to Antonio Martin, Enterprise Secure Access Research and
8540 Development, Cisco Systems, Inc. for discovering this bug and
8541 preparing a fix. ([CVE-2012-0050])
8545 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8547 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8548 of the Vaudenay padding oracle attack on CBC mode encryption
8549 which enables an efficient plaintext recovery attack against
8550 the OpenSSL implementation of DTLS. Their attack exploits timing
8551 differences arising during decryption processing. A research
8552 paper describing this attack can be found at:
8553 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8554 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8555 Security Group at Royal Holloway, University of London
8556 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8557 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8558 for preparing the fix. ([CVE-2011-4108])
8560 *Robin Seggelmann, Michael Tuexen*
8562 * Clear bytes used for block padding of SSL 3.0 records.
8565 *Adam Langley (Google)*
8567 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8568 Kadianakis <desnacked@gmail.com> for discovering this issue and
8569 Adam Langley for preparing the fix. ([CVE-2011-4619])
8571 *Adam Langley (Google)*
8573 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8575 *Andrey Kulikov <amdeich@gmail.com>*
8577 * Prevent malformed RFC3779 data triggering an assertion failure.
8578 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8579 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8581 *Rob Austein <sra@hactrn.net>*
8583 * Improved PRNG seeding for VOS.
8585 *Paul Green <Paul.Green@stratus.com>*
8587 * Fix ssl_ciph.c set-up race.
8589 *Adam Langley (Google)*
8591 * Fix spurious failures in ecdsatest.c.
8593 *Emilia Käsper (Google)*
8595 * Fix the BIO_f_buffer() implementation (which was mixing different
8596 interpretations of the `..._len` fields).
8598 *Adam Langley (Google)*
8600 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8601 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8602 threads won't reuse the same blinding coefficients.
8604 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8605 lock to call BN_BLINDING_invert_ex, and avoids one use of
8606 BN_BLINDING_update for each BN_BLINDING structure (previously,
8607 the last update always remained unused).
8609 *Emilia Käsper (Google)*
8611 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8613 *Bob Buckholz (Google)*
8615 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8617 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8618 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8620 *Kaspar Brand <ossl@velox.ch>*
8622 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8623 for multi-threaded use of ECDH. ([CVE-2011-3210])
8625 *Adam Langley (Google)*
8627 * Fix x509_name_ex_d2i memory leak on bad inputs.
8631 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8632 signature public key algorithm by using OID xref utilities instead.
8633 Before this you could only use some ECC ciphersuites with SHA1 only.
8637 * Add protection against ECDSA timing attacks as mentioned in the paper
8638 by Billy Bob Brumley and Nicola Tuveri, see:
8639 <http://eprint.iacr.org/2011/232.pdf>
8641 *Billy Bob Brumley and Nicola Tuveri*
8643 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8645 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8647 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8649 * Fix bug in string printing code: if *any* escaping is enabled we must
8650 escape the escape character (backslash) or the resulting string is
8655 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8657 * Disable code workaround for ancient and obsolete Netscape browsers
8658 and servers: an attacker can use it in a ciphersuite downgrade attack.
8659 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8663 * Fixed J-PAKE implementation error, originally discovered by
8664 Sebastien Martini, further info and confirmation from Stefan
8665 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8669 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8671 * Fix extension code to avoid race conditions which can result in a buffer
8672 overrun vulnerability: resumed sessions must not be modified as they can
8673 be shared by multiple threads. CVE-2010-3864
8677 * Fix WIN32 build system to correctly link an ENGINE directory into
8682 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8684 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8687 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8689 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8691 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8692 context. The operation can be customised via the ctrl mechanism in
8693 case ENGINEs want to include additional functionality.
8697 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8701 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8702 output hashes compatible with older versions of OpenSSL.
8704 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8706 * Fix compression algorithm handling: if resuming a session use the
8707 compression algorithm of the resumed session instead of determining
8708 it from client hello again. Don't allow server to change algorithm.
8712 * Add load_crls() function to commands tidying load_certs() too. Add option
8713 to verify utility to allow additional CRLs to be included.
8717 * Update OCSP request code to permit adding custom headers to the request:
8718 some responders need this.
8722 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8725 *Julia Lawall <julia@diku.dk>*
8727 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8728 needlessly dereferenced structures, used obsolete functions and
8729 didn't handle all updated verify codes correctly.
8733 * Disable MD2 in the default configuration.
8737 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8738 indicate the initial BIO being pushed or popped. This makes it possible
8739 to determine whether the BIO is the one explicitly called or as a result
8740 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8741 it handles reference counts correctly and doesn't zero out the I/O bio
8742 when it is not being explicitly popped. WARNING: applications which
8743 included workarounds for the old buggy behaviour will need to be modified
8744 or they could free up already freed BIOs.
8748 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8749 renaming to all platforms (within the 0.9.8 branch, this was
8750 done conditionally on Netware platforms to avoid a name clash).
8752 *Guenter <lists@gknw.net>*
8754 * Add ECDHE and PSK support to DTLS.
8756 *Michael Tuexen <tuexen@fh-muenster.de>*
8758 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8763 * Add "missing" function EVP_MD_flags() (without this the only way to
8764 retrieve a digest flags is by accessing the structure directly. Update
8765 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8766 or cipher is registered as in the "from" argument. Print out all
8767 registered digests in the dgst usage message instead of manually
8768 attempting to work them out.
8772 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8773 this allows the use of compression and extensions. Change default cipher
8774 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8775 by default unless an application cipher string requests it.
8779 * Alter match criteria in PKCS12_parse(). It used to try to use local
8780 key ids to find matching certificates and keys but some PKCS#12 files
8781 don't follow the (somewhat unwritten) rules and this strategy fails.
8782 Now just gather all certificates together and the first private key
8783 then look for the first certificate that matches the key.
8787 * Support use of registered digest and cipher names for dgst and cipher
8788 commands instead of having to add each one as a special case. So now
8795 openssl dgst -sha256 foo
8797 and this works for ENGINE based algorithms too.
8801 * Update Gost ENGINE to support parameter files.
8803 *Victor B. Wagner <vitus@cryptocom.ru>*
8805 * Support GeneralizedTime in ca utility.
8807 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8809 * Enhance the hash format used for certificate directory links. The new
8810 form uses the canonical encoding (meaning equivalent names will work
8811 even if they aren't identical) and uses SHA1 instead of MD5. This form
8812 is incompatible with the older format and as a result c_rehash should
8813 be used to rebuild symbolic links.
8817 * Make PKCS#8 the default write format for private keys, replacing the
8818 traditional format. This form is standardised, more secure and doesn't
8819 include an implicit MD5 dependency.
8823 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8824 committed to OpenSSL should pass this lot as a minimum.
8828 * Add session ticket override functionality for use by EAP-FAST.
8830 *Jouni Malinen <j@w1.fi>*
8832 * Modify HMAC functions to return a value. Since these can be implemented
8833 in an ENGINE errors can occur.
8837 * Type-checked OBJ_bsearch_ex.
8841 * Type-checked OBJ_bsearch. Also some constification necessitated
8842 by type-checking. Still to come: TXT_DB, bsearch(?),
8843 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8848 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8849 seconds to a tm structure directly, instead of going through OS
8850 specific date routines. This avoids any issues with OS routines such
8851 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8852 and X509_time_adj_ex() to cover the extended range. The existing
8853 X509_time_adj() is still usable and will no longer have any date issues.
8857 * Delta CRL support. New use deltas option which will attempt to locate
8858 and search any appropriate delta CRLs available.
8860 This work was sponsored by Google.
8864 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8865 code and add additional score elements. Validate alternate CRL paths
8866 as part of the CRL checking and indicate a new error "CRL path validation
8867 error" in this case. Applications wanting additional details can use
8868 the verify callback and check the new "parent" field. If this is not
8869 NULL CRL path validation is taking place. Existing applications won't
8870 see this because it requires extended CRL support which is off by
8873 This work was sponsored by Google.
8877 * Support for freshest CRL extension.
8879 This work was sponsored by Google.
8883 * Initial indirect CRL support. Currently only supported in the CRLs
8884 passed directly and not via lookup. Process certificate issuer
8885 CRL entry extension and lookup CRL entries by bother issuer name
8886 and serial number. Check and process CRL issuer entry in IDP extension.
8888 This work was sponsored by Google.
8892 * Add support for distinct certificate and CRL paths. The CRL issuer
8893 certificate is validated separately in this case. Only enabled if
8894 an extended CRL support flag is set: this flag will enable additional
8895 CRL functionality in future.
8897 This work was sponsored by Google.
8901 * Add support for policy mappings extension.
8903 This work was sponsored by Google.
8907 * Fixes to pathlength constraint, self issued certificate handling,
8908 policy processing to align with RFC3280 and PKITS tests.
8910 This work was sponsored by Google.
8914 * Support for name constraints certificate extension. DN, email, DNS
8915 and URI types are currently supported.
8917 This work was sponsored by Google.
8921 * To cater for systems that provide a pointer-based thread ID rather
8922 than numeric, deprecate the current numeric thread ID mechanism and
8923 replace it with a structure and associated callback type. This
8924 mechanism allows a numeric "hash" to be extracted from a thread ID in
8925 either case, and on platforms where pointers are larger than 'long',
8926 mixing is done to help ensure the numeric 'hash' is usable even if it
8927 can't be guaranteed unique. The default mechanism is to use "&errno"
8928 as a pointer-based thread ID to distinguish between threads.
8930 Applications that want to provide their own thread IDs should now use
8931 CRYPTO_THREADID_set_callback() to register a callback that will call
8932 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8934 Note that ERR_remove_state() is now deprecated, because it is tied
8935 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8936 to free the current thread's error state should be replaced by
8937 ERR_remove_thread_state(NULL).
8939 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8940 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8941 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8942 application was previously providing a numeric thread callback that
8943 was inappropriate for distinguishing threads, then uniqueness might
8944 have been obtained with &errno that happened immediately in the
8945 intermediate development versions of OpenSSL; this is no longer the
8946 case, the numeric thread callback will now override the automatic use
8949 *Geoff Thorpe, with help from Bodo Moeller*
8951 * Initial support for different CRL issuing certificates. This covers a
8952 simple case where the self issued certificates in the chain exist and
8953 the real CRL issuer is higher in the existing chain.
8955 This work was sponsored by Google.
8959 * Removed effectively defunct crypto/store from the build.
8963 * Revamp of STACK to provide stronger type-checking. Still to come:
8964 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8965 ASN1_STRING, CONF_VALUE.
8969 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8970 RAM on SSL connections. This option can save about 34k per idle SSL.
8974 * Revamp of LHASH to provide stronger type-checking. Still to come:
8975 STACK, TXT_DB, bsearch, qsort.
8979 * Initial support for Cryptographic Message Syntax (aka CMS) based
8980 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8981 support for data, signedData, compressedData, digestedData and
8982 encryptedData, envelopedData types included. Scripts to check against
8983 RFC4134 examples draft and interop and consistency checks of many
8984 content types and variants.
8988 * Add options to enc utility to support use of zlib compression BIO.
8992 * Extend mk1mf to support importing of options and assembly language
8993 files from Configure script, currently only included in VC-WIN32.
8994 The assembly language rules can now optionally generate the source
8995 files from the associated perl scripts.
8999 * Implement remaining functionality needed to support GOST ciphersuites.
9000 Interop testing has been performed using CryptoPro implementations.
9002 *Victor B. Wagner <vitus@cryptocom.ru>*
9004 * s390x assembler pack.
9008 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9013 * Implement Opaque PRF Input TLS extension as specified in
9014 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9015 official specification yet and no extension type assignment by
9016 IANA exists, this extension (for now) will have to be explicitly
9017 enabled when building OpenSSL by providing the extension number
9018 to use. For example, specify an option
9020 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9022 to the "config" or "Configure" script to enable the extension,
9023 assuming extension number 0x9527 (which is a completely arbitrary
9024 and unofficial assignment based on the MD5 hash of the Internet
9025 Draft). Note that by doing so, you potentially lose
9026 interoperability with other TLS implementations since these might
9027 be using the same extension number for other purposes.
9029 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9030 opaque PRF input value to use in the handshake. This will create
9031 an internal copy of the length-'len' string at 'src', and will
9032 return non-zero for success.
9034 To get more control and flexibility, provide a callback function
9037 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9038 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9042 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9045 Callback function 'cb' will be called in handshakes, and is
9046 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9047 Argument 'arg' is for application purposes (the value as given to
9048 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9049 be provided to the callback function). The callback function
9050 has to return non-zero to report success: usually 1 to use opaque
9051 PRF input just if possible, or 2 to enforce use of the opaque PRF
9052 input. In the latter case, the library will abort the handshake
9053 if opaque PRF input is not successfully negotiated.
9055 Arguments 'peerinput' and 'len' given to the callback function
9056 will always be NULL and 0 in the case of a client. A server will
9057 see the client's opaque PRF input through these variables if
9058 available (NULL and 0 otherwise). Note that if the server
9059 provides an opaque PRF input, the length must be the same as the
9060 length of the client's opaque PRF input.
9062 Note that the callback function will only be called when creating
9063 a new session (session resumption can resume whatever was
9064 previously negotiated), and will not be called in SSL 2.0
9065 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9066 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9067 for applications that need to enforce opaque PRF input.
9071 * Update ssl code to support digests other than SHA1+MD5 for handshake
9074 *Victor B. Wagner <vitus@cryptocom.ru>*
9076 * Add RFC4507 support to OpenSSL. This includes the corrections in
9077 RFC4507bis. The encrypted ticket format is an encrypted encoded
9078 SSL_SESSION structure, that way new session features are automatically
9081 If a client application caches session in an SSL_SESSION structure
9082 support is transparent because tickets are now stored in the encoded
9085 The SSL_CTX structure automatically generates keys for ticket
9086 protection in servers so again support should be possible
9087 with no application modification.
9089 If a client or server wishes to disable RFC4507 support then the option
9090 SSL_OP_NO_TICKET can be set.
9092 Add a TLS extension debugging callback to allow the contents of any client
9093 or server extensions to be examined.
9095 This work was sponsored by Google.
9099 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9100 OpenSSL should now compile cleanly on gcc 4.2
9102 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9104 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9105 support including streaming MAC support: this is required for GOST
9106 ciphersuite support.
9108 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9110 * Add option -stream to use PKCS#7 streaming in smime utility. New
9111 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9112 to output in BER and PEM format.
9116 * Experimental support for use of HMAC via EVP_PKEY interface. This
9117 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9118 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9119 ENGINE support for HMAC keys which are unextractable. New -mac and
9120 -macopt options to dgst utility.
9124 * New option -sigopt to dgst utility. Update dgst to use
9125 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9126 alternative signing parameters such as X9.31 or PSS in the dgst
9131 * Change ssl_cipher_apply_rule(), the internal function that does
9132 the work each time a ciphersuite string requests enabling
9133 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9134 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9135 the order of disabled ciphersuites such that those ciphersuites
9136 that most recently went from enabled to disabled not only stay
9137 in order with respect to each other, but also have higher priority
9138 than other disabled ciphersuites the next time ciphersuites are
9141 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9142 the same ciphersuites as with "HIGH" alone, but in a specific
9143 order where the PSK ciphersuites come first (since they are the
9144 most recently disabled ciphersuites when "HIGH" is parsed).
9146 Also, change ssl_create_cipher_list() (using this new
9147 functionality) such that between otherwise identical
9148 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9153 * Change ssl_create_cipher_list() so that it automatically
9154 arranges the ciphersuites in reasonable order before starting
9155 to process the rule string. Thus, the definition for "DEFAULT"
9156 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9157 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9158 This makes it much easier to arrive at a reasonable default order
9159 in applications for which anonymous ciphers are OK (meaning
9160 that you can't actually use DEFAULT).
9162 *Bodo Moeller; suggested by Victor Duchovni*
9164 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9165 processing) into multiple integers instead of setting
9166 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9167 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9168 (These masks as well as the individual bit definitions are hidden
9169 away into the non-exported interface ssl/ssl_locl.h, so this
9170 change to the definition of the SSL_CIPHER structure shouldn't
9171 affect applications.) This give us more bits for each of these
9172 categories, so there is no longer a need to coagulate AES128 and
9173 AES256 into a single algorithm bit, and to coagulate Camellia128
9174 and Camellia256 into a single algorithm bit, which has led to all
9177 Thus, among other things, the kludge introduced in 0.9.7m and
9178 0.9.8e for masking out AES256 independently of AES128 or masking
9179 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9181 With the change, we also introduce new ciphersuite aliases that
9182 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9187 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9188 Use the leftmost N bytes of the signature input if the input is
9189 larger than the prime q (with N being the size in bytes of q).
9193 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9194 it yet and it is largely untested.
9198 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9202 * Initial incomplete changes to avoid need for function casts in OpenSSL
9203 some compilers (gcc 4.2 and later) reject their use. Safestack is
9204 reimplemented. Update ASN1 to avoid use of legacy functions.
9208 * Win32/64 targets are linked with Winsock2.
9212 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9213 to external functions. This can be used to increase CRL handling
9214 efficiency especially when CRLs are very large by (for example) storing
9215 the CRL revoked certificates in a database.
9219 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9220 new CRLs added to a directory can be used. New command line option
9221 -verify_return_error to s_client and s_server. This causes real errors
9222 to be returned by the verify callback instead of carrying on no matter
9223 what. This reflects the way a "real world" verify callback would behave.
9227 * GOST engine, supporting several GOST algorithms and public key formats.
9228 Kindly donated by Cryptocom.
9232 * Partial support for Issuing Distribution Point CRL extension. CRLs
9233 partitioned by DP are handled but no indirect CRL or reason partitioning
9234 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9235 selected via a scoring technique which handles IDP and AKID in CRLs.
9239 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9240 will ultimately be used for all verify operations: this will remove the
9241 X509_STORE dependency on certificate verification and allow alternative
9242 lookup methods. X509_STORE based implementations of these two callbacks.
9246 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9247 Modify get_crl() to find a valid (unexpired) CRL if possible.
9251 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9252 this would be called X509_CRL_cmp() but that name is already used by
9253 a function that just compares CRL issuer names. Cache several CRL
9254 extensions in X509_CRL structure and cache CRLDP in X509.
9258 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9259 this maps equivalent X509_NAME structures into a consistent structure.
9260 Name comparison can then be performed rapidly using memcmp().
9264 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9269 * Allow digests to supply their own micalg string for S/MIME type using
9270 the ctrl EVP_MD_CTRL_MICALG.
9274 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9275 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9276 ctrl. It can then customise the structure before and/or after signing
9281 * New function OBJ_add_sigid() to allow application defined signature OIDs
9282 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9283 to free up any added signature OIDs.
9287 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9288 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9289 digest and cipher tables. New options added to openssl utility:
9290 list-message-digest-algorithms and list-cipher-algorithms.
9294 * Change the array representation of binary polynomials: the list
9295 of degrees of non-zero coefficients is now terminated with -1.
9296 Previously it was terminated with 0, which was also part of the
9297 value; thus, the array representation was not applicable to
9298 polynomials where t^0 has coefficient zero. This change makes
9299 the array representation useful in a more general context.
9303 * Various modifications and fixes to SSL/TLS cipher string
9304 handling. For ECC, the code now distinguishes between fixed ECDH
9305 with RSA certificates on the one hand and with ECDSA certificates
9306 on the other hand, since these are separate ciphersuites. The
9307 unused code for Fortezza ciphersuites has been removed.
9309 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9310 (not "ECDHE"). For consistency with the code for DH
9311 certificates, use of ECDH certificates is now considered ECDH
9312 authentication, not RSA or ECDSA authentication (the latter is
9313 merely the CA's signing algorithm and not actively used in the
9316 The temporary ciphersuite alias "ECCdraft" is no longer
9317 available, and ECC ciphersuites are no longer excluded from "ALL"
9318 and "DEFAULT". The following aliases now exist for RFC 4492
9319 ciphersuites, most of these by analogy with the DH case:
9321 kECDHr - ECDH cert, signed with RSA
9322 kECDHe - ECDH cert, signed with ECDSA
9323 kECDH - ECDH cert (signed with either RSA or ECDSA)
9324 kEECDH - ephemeral ECDH
9325 ECDH - ECDH cert or ephemeral ECDH
9331 AECDH - anonymous ECDH
9332 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9336 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9337 Use correct micalg parameters depending on digest(s) in signed message.
9341 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9342 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9346 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9347 an engine to register a method. Add ENGINE lookups for methods and
9348 functional reference processing.
9352 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9353 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9358 * New -resign option to smime utility. This adds one or more signers
9359 to an existing PKCS#7 signedData structure. Also -md option to use an
9360 alternative message digest algorithm for signing.
9364 * Tidy up PKCS#7 routines and add new functions to make it easier to
9365 create PKCS7 structures containing multiple signers. Update smime
9366 application to support multiple signers.
9370 * New -macalg option to pkcs12 utility to allow setting of an alternative
9375 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9376 Reorganize PBE internals to lookup from a static table using NIDs,
9377 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9378 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9379 PRF which will be automatically used with PBES2.
9383 * Replace the algorithm specific calls to generate keys in "req" with the
9388 * Update PKCS#7 enveloped data routines to use new API. This is now
9389 supported by any public key method supporting the encrypt operation. A
9390 ctrl is added to allow the public key algorithm to examine or modify
9391 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9396 * Add a ctrl to asn1 method to allow a public key algorithm to express
9397 a default digest type to use. In most cases this will be SHA1 but some
9398 algorithms (such as GOST) need to specify an alternative digest. The
9399 return value indicates how strong the preference is 1 means optional and
9400 2 is mandatory (that is it is the only supported type). Modify
9401 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9402 use the default md. Update openssl utilities to use the default digest
9403 type for signing if it is not explicitly indicated.
9407 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9408 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9409 signing method from the key type. This effectively removes the link
9410 between digests and public key types.
9414 * Add an OID cross reference table and utility functions. Its purpose is to
9415 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9416 rsaEncryption. This will allow some of the algorithm specific hackery
9417 needed to use the correct OID to be removed.
9421 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9422 structures for PKCS7_sign(). They are now set up by the relevant public
9427 * Add provisional EC pkey method with support for ECDSA and ECDH.
9431 * Add support for key derivation (agreement) in the API, DH method and
9436 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9437 public and private key formats. As a side effect these add additional
9438 command line functionality not previously available: DSA signatures can be
9439 generated and verified using pkeyutl and DH key support and generation in
9446 *Oliver Tappe <zooey@hirschkaefer.de>*
9448 * New make target "install_html_docs" installs HTML renditions of the
9451 *Oliver Tappe <zooey@hirschkaefer.de>*
9453 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9454 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9455 support key and parameter generation and add initial key generation
9456 functionality for RSA.
9460 * Add functions for main EVP_PKEY_method operations. The undocumented
9461 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9462 `EVP_PKEY_{encrypt,decrypt}_old`.
9466 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9467 key API, doesn't do much yet.
9471 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9472 public key algorithms. New option to openssl utility:
9473 "list-public-key-algorithms" to print out info.
9477 * Implement the Supported Elliptic Curves Extension for
9478 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9482 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9483 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9487 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9488 utilities such as rsa, dsa, dsaparam etc except they process any key
9493 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9494 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9495 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9500 * Initial support for pluggable public key ASN1.
9501 De-spaghettify the public key ASN1 handling. Move public and private
9502 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9503 algorithm specific handling to a single module within the relevant
9504 algorithm directory. Add functions to allow (near) opaque processing
9505 of public and private key structures.
9509 * Implement the Supported Point Formats Extension for
9510 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9514 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9515 for the psk identity [hint] and the psk callback functions to the
9516 SSL_SESSION, SSL and SSL_CTX structure.
9519 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9523 SSL_CTX_use_psk_identity_hint
9524 SSL_get_psk_identity_hint
9525 SSL_get_psk_identity
9526 SSL_use_psk_identity_hint
9528 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9530 * Add RFC 3161 compliant time stamp request creation, response generation
9531 and response verification functionality.
9533 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9535 * Add initial support for TLS extensions, specifically for the server_name
9536 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9537 have new members for a host name. The SSL data structure has an
9538 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9539 stored in that context to allow for session resumption, even after the
9540 SSL has been switched to a new SSL_CTX in reaction to a client's
9541 server_name extension.
9543 New functions (subject to change):
9545 SSL_get_servername()
9546 SSL_get_servername_type()
9549 New CTRL codes and macros (subject to change):
9551 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9552 - SSL_CTX_set_tlsext_servername_callback()
9553 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9554 - SSL_CTX_set_tlsext_servername_arg()
9555 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9557 openssl s_client has a new '-servername ...' option.
9559 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9560 '-key2 ...', '-servername_fatal' (subject to change). This allows
9561 testing the HostName extension for a specific single host name ('-cert'
9562 and '-key' remain fallbacks for handshakes without HostName
9563 negotiation). If the unrecognized_name alert has to be sent, this by
9564 default is a warning; it becomes fatal with the '-servername_fatal'
9567 *Peter Sylvester, Remy Allais, Christophe Renou*
9569 * Whirlpool hash implementation is added.
9573 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9574 bn(64,32). Because of instruction set limitations it doesn't have
9575 any negative impact on performance. This was done mostly in order
9576 to make it possible to share assembler modules, such as bn_mul_mont
9577 implementations, between 32- and 64-bit builds without hassle.
9581 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9582 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9587 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9588 dedicated Montgomery multiplication procedure, is introduced.
9589 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9590 "64-bit" performance on certain 32-bit targets.
9594 * New option SSL_OP_NO_COMP to disable use of compression selectively
9595 in SSL structures. New SSL ctrl to set maximum send fragment size.
9596 Save memory by setting the I/O buffer sizes dynamically instead of
9597 using the maximum available value.
9601 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9602 in addition to the text details.
9606 * Very, very preliminary EXPERIMENTAL support for printing of general
9607 ASN1 structures. This currently produces rather ugly output and doesn't
9608 handle several customised structures at all.
9612 * Integrated support for PVK file format and some related formats such
9613 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9614 these in the 'rsa' and 'dsa' utilities.
9618 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9622 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9623 place for the (very old) "NETSCAPE" format certificates which are now
9624 handled using new ASN1 code equivalents.
9628 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9629 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9630 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9634 * Modify CRL distribution points extension code to print out previously
9635 unsupported fields. Enhance extension setting code to allow setting of
9640 * Add print and set support for Issuing Distribution Point CRL extension.
9644 * Change 'Configure' script to enable Camellia by default.
9651 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9653 * When rejecting SSL/TLS records due to an incorrect version number, never
9654 update s->server with a new major version number. As of
9655 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9656 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9657 the previous behavior could result in a read attempt at NULL when
9658 receiving specific incorrect SSL/TLS records once record payload
9659 protection is active. ([CVE-2010-0740])
9661 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9663 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9664 could be crashed if the relevant tables were not present (e.g. chrooted).
9666 *Tomas Hoger <thoger@redhat.com>*
9668 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9670 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9672 *Martin Olsson, Neel Mehta*
9674 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9675 accommodate for stack sorting, always a write lock!).
9679 * On some versions of WIN32 Heap32Next is very slow. This can cause
9680 excessive delays in the RAND_poll(): over a minute. As a workaround
9681 include a time check in the inner Heap32Next loop too.
9685 * The code that handled flushing of data in SSL/TLS originally used the
9686 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9687 the problem outlined in PR#1949. The fix suggested there however can
9688 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9689 of Apache). So instead simplify the code to flush unconditionally.
9690 This should be fine since flushing with no data to flush is a no op.
9694 * Handle TLS versions 2.0 and later properly and correctly use the
9695 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9696 off ancient servers have a habit of sticking around for a while...
9700 * Modify compression code so it frees up structures without using the
9701 ex_data callbacks. This works around a problem where some applications
9702 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9703 restarting) then use compression (e.g. SSL with compression) later.
9704 This results in significant per-connection memory leaks and
9705 has caused some security issues including CVE-2008-1678 and
9710 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9711 change when encrypting or decrypting.
9715 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9716 connect and renegotiate with servers which do not support RI.
9717 Until RI is more widely deployed this option is enabled by default.
9721 * Add "missing" ssl ctrls to clear options and mode.
9725 * If client attempts to renegotiate and doesn't support RI respond with
9726 a no_renegotiation alert as required by RFC5746. Some renegotiating
9727 TLS clients will continue a connection gracefully when they receive
9728 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9729 waiting for a server hello which it will never receive. Now we treat a
9730 received no_renegotiation alert as a fatal error. This is because
9731 applications requesting a renegotiation might well expect it to succeed
9732 and would have no code in place to handle the server denying it so the
9733 only safe thing to do is to terminate the connection.
9737 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9738 peer supports secure renegotiation and 0 otherwise. Print out peer
9739 renegotiation support in s_client/s_server.
9743 * Replace the highly broken and deprecated SPKAC certification method with
9744 the updated NID creation version. This should correctly handle UTF8.
9748 * Implement RFC5746. Re-enable renegotiation but require the extension
9749 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9750 turns out to be a bad idea. It has been replaced by
9751 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9752 SSL_CTX_set_options(). This is really not recommended unless you
9753 know what you are doing.
9755 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9757 * Fixes to stateless session resumption handling. Use initial_ctx when
9758 issuing and attempting to decrypt tickets in case it has changed during
9759 servername handling. Use a non-zero length session ID when attempting
9760 stateless session resumption: this makes it possible to determine if
9761 a resumption has occurred immediately after receiving server hello
9762 (several places in OpenSSL subtly assume this) instead of later in
9767 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9768 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9769 fixes for a few places where the return code is not checked
9772 *Julia Lawall <julia@diku.dk>*
9774 * Add --strict-warnings option to Configure script to include devteam
9775 warnings in other configurations.
9779 * Add support for --libdir option and LIBDIR variable in makefiles. This
9780 makes it possible to install openssl libraries in locations which
9781 have names other than "lib", for example "/usr/lib64" which some
9784 *Steve Henson, based on patch from Jeremy Utley*
9786 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9787 X690 8.9.12 and can produce some misleading textual output of OIDs.
9789 *Steve Henson, reported by Dan Kaminsky*
9791 * Delete MD2 from algorithm tables. This follows the recommendation in
9792 several standards that it is not used in new applications due to
9793 several cryptographic weaknesses. For binary compatibility reasons
9794 the MD2 API is still compiled in by default.
9798 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9803 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9804 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9807 *Guenter <lists@gknw.net>*
9809 * Fix the server certificate chain building code to use X509_verify_cert(),
9810 it used to have an ad-hoc builder which was unable to cope with anything
9811 other than a simple chain.
9813 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9815 * Don't check self signed certificate signatures in X509_verify_cert()
9816 by default (a flag can override this): it just wastes time without
9817 adding any security. As a useful side effect self signed root CAs
9818 with non-FIPS digests are now usable in FIPS mode.
9822 * In dtls1_process_out_of_seq_message() the check if the current message
9823 is already buffered was missing. For every new message was memory
9824 allocated, allowing an attacker to perform an denial of service attack
9825 with sending out of seq handshake messages until there is no memory
9826 left. Additionally every future message was buffered, even if the
9827 sequence number made no sense and would be part of another handshake.
9828 So only messages with sequence numbers less than 10 in advance will be
9829 buffered. ([CVE-2009-1378])
9831 *Robin Seggelmann, discovered by Daniel Mentz*
9833 * Records are buffered if they arrive with a future epoch to be
9834 processed after finishing the corresponding handshake. There is
9835 currently no limitation to this buffer allowing an attacker to perform
9836 a DOS attack with sending records with future epochs until there is no
9837 memory left. This patch adds the pqueue_size() function to determine
9838 the size of a buffer and limits the record buffer to 100 entries.
9841 *Robin Seggelmann, discovered by Daniel Mentz*
9843 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9844 parent structure is freed. ([CVE-2009-1379])
9848 * Handle non-blocking I/O properly in SSL_shutdown() call.
9850 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9852 * Add `2.5.4.*` OIDs
9854 *Ilya O. <vrghost@gmail.com>*
9856 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9858 * Disable renegotiation completely - this fixes a severe security
9859 problem ([CVE-2009-3555]) at the cost of breaking all
9860 renegotiation. Renegotiation can be re-enabled by setting
9861 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9862 run-time. This is really not recommended unless you know what
9867 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9869 * Don't set val to NULL when freeing up structures, it is freed up by
9870 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9871 zeroing past the valid field. ([CVE-2009-0789])
9873 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9875 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9876 checked correctly. This would allow some invalid signed attributes to
9877 appear to verify correctly. ([CVE-2009-0591])
9879 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9881 * Reject UniversalString and BMPString types with invalid lengths. This
9882 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9883 a legal length. ([CVE-2009-0590])
9887 * Set S/MIME signing as the default purpose rather than setting it
9888 unconditionally. This allows applications to override it at the store
9893 * Permit restricted recursion of ASN1 strings. This is needed in practice
9894 to handle some structures.
9898 * Improve efficiency of mem_gets: don't search whole buffer each time
9901 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9903 * New -hex option for openssl rand.
9907 * Print out UTF8String and NumericString when parsing ASN1.
9911 * Support NumericString type for name components.
9915 * Allow CC in the environment to override the automatically chosen
9916 compiler. Note that nothing is done to ensure flags work with the
9921 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9923 * Properly check EVP_VerifyFinal() and similar return values
9926 *Ben Laurie, Bodo Moeller, Google Security Team*
9928 * Enable TLS extensions by default.
9932 * Allow the CHIL engine to be loaded, whether the application is
9933 multithreaded or not. (This does not release the developer from the
9934 obligation to set up the dynamic locking callbacks.)
9936 *Sander Temme <sander@temme.net>*
9938 * Use correct exit code if there is an error in dgst command.
9940 *Steve Henson; problem pointed out by Roland Dirlewanger*
9942 * Tweak Configure so that you need to say "experimental-jpake" to enable
9943 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9947 * Add experimental JPAKE support, including demo authentication in
9948 s_client and s_server.
9952 * Set the comparison function in v3_addr_canonize().
9954 *Rob Austein <sra@hactrn.net>*
9956 * Add support for XMPP STARTTLS in s_client.
9958 *Philip Paeps <philip@freebsd.org>*
9960 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9961 to ensure that even with this option, only ciphersuites in the
9962 server's preference list will be accepted. (Note that the option
9963 applies only when resuming a session, so the earlier behavior was
9964 just about the algorithm choice for symmetric cryptography.)
9968 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9970 * Fix NULL pointer dereference if a DTLS server received
9971 ChangeCipherSpec as first record ([CVE-2009-1386]).
9975 * Fix a state transition in s3_srvr.c and d1_srvr.c
9976 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9980 * The fix in 0.9.8c that supposedly got rid of unsafe
9981 double-checked locking was incomplete for RSA blinding,
9982 addressing just one layer of what turns out to have been
9983 doubly unsafe triple-checked locking.
9985 So now fix this for real by retiring the MONT_HELPER macro
9986 in crypto/rsa/rsa_eay.c.
9988 *Bodo Moeller; problem pointed out by Marius Schilder*
9990 * Various precautionary measures:
9992 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9994 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9995 (NB: This would require knowledge of the secret session ticket key
9996 to exploit, in which case you'd be SOL either way.)
9998 - Change bn_nist.c so that it will properly handle input BIGNUMs
9999 outside the expected range.
10001 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10004 *Neel Mehta, Bodo Moeller*
10006 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10007 the load fails. Useful for distros.
10009 *Ben Laurie and the FreeBSD team*
10011 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10015 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10019 * Expand ENGINE to support engine supplied SSL client certificate functions.
10021 This work was sponsored by Logica.
10025 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10026 keystores. Support for SSL/TLS client authentication too.
10027 Not compiled unless enable-capieng specified to Configure.
10029 This work was sponsored by Logica.
10033 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10034 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10035 attribute creation routines such as certificate requests and PKCS#12
10040 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10042 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10043 handshake which could lead to a client crash as found using the
10044 Codenomicon TLS test suite ([CVE-2008-1672])
10046 *Steve Henson, Mark Cox*
10048 * Fix double free in TLS server name extensions which could lead to
10049 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10053 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10055 Clear the error queue to ensure that error entries left from
10056 older function calls do not interfere with the correct operation.
10058 *Lutz Jaenicke, Erik de Castro Lopo*
10060 * Remove root CA certificates of commercial CAs:
10062 The OpenSSL project does not recommend any specific CA and does not
10063 have any policy with respect to including or excluding any CA.
10064 Therefore it does not make any sense to ship an arbitrary selection
10065 of root CA certificates with the OpenSSL software.
10069 * RSA OAEP patches to fix two separate invalid memory reads.
10070 The first one involves inputs when 'lzero' is greater than
10071 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10072 before the beginning of from). The second one involves inputs where
10073 the 'db' section contains nothing but zeroes (there is a one-byte
10074 invalid read after the end of 'db').
10076 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10078 * Partial backport from 0.9.9-dev:
10080 Introduce bn_mul_mont (dedicated Montgomery multiplication
10081 procedure) as a candidate for BIGNUM assembler implementation.
10082 While 0.9.9-dev uses assembler for various architectures, only
10083 x86_64 is available by default here in the 0.9.8 branch, and
10084 32-bit x86 is available through a compile-time setting.
10086 To try the 32-bit x86 assembler implementation, use Configure
10087 option "enable-montasm" (which exists only for this backport).
10089 As "enable-montasm" for 32-bit x86 disclaims code stability
10090 anyway, in this constellation we activate additional code
10091 backported from 0.9.9-dev for further performance improvements,
10092 namely BN_from_montgomery_word. (To enable this otherwise,
10093 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10095 *Andy Polyakov (backport partially by Bodo Moeller)*
10097 * Add TLS session ticket callback. This allows an application to set
10098 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10099 values. This is useful for key rollover for example where several key
10100 sets may exist with different names.
10104 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10105 This was broken until now in 0.9.8 releases, such that the only way
10106 a registered ENGINE could be used (assuming it initialises
10107 successfully on the host) was to explicitly set it as the default
10108 for the relevant algorithms. This is in contradiction with 0.9.7
10109 behaviour and the documentation. With this fix, when an ENGINE is
10110 registered into a given algorithm's table of implementations, the
10111 'uptodate' flag is reset so that auto-discovery will be used next
10112 time a new context for that algorithm attempts to select an
10115 *Ian Lister (tweaked by Geoff Thorpe)*
10117 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10118 implementation in the following ways:
10120 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10123 Lack of BER streaming support means one pass streaming processing is
10124 only supported if data is detached: setting the streaming flag is
10125 ignored for embedded content.
10127 CMS support is disabled by default and must be explicitly enabled
10128 with the enable-cms configuration option.
10132 * Update the GMP engine glue to do direct copies between BIGNUM and
10133 mpz_t when openssl and GMP use the same limb size. Otherwise the
10134 existing "conversion via a text string export" trick is still used.
10136 *Paul Sheer <paulsheer@gmail.com>*
10138 * Zlib compression BIO. This is a filter BIO which compressed and
10139 uncompresses any data passed through it.
10143 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10144 RFC3394 compatible AES key wrapping.
10148 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10149 sets string data without copying. X509_ALGOR_set0() and
10150 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10151 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10152 from an X509_ATTRIBUTE structure optionally checking it occurs only
10153 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10158 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10159 to get the expected BN_FLG_CONSTTIME behavior.
10161 *Bodo Moeller (Google)*
10165 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10166 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10167 - added some more tests to do_tests.pl
10168 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10169 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10170 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10171 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10172 - various changes to netware.pl to enable gcc-cross builds on Win32
10174 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10175 - various changes to fix missing prototype warnings
10176 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10177 - added AES, WHIRLPOOL and CPUID assembler code to build files
10178 - added missing AES assembler make rules to mk1mf.pl
10179 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10181 *Guenter Knauf <eflash@gmx.net>*
10183 * Implement certificate status request TLS extension defined in RFC3546.
10184 A client can set the appropriate parameters and receive the encoded
10185 OCSP response via a callback. A server can query the supplied parameters
10186 and set the encoded OCSP response in the callback. Add simplified examples
10187 to s_client and s_server.
10191 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10193 * Fix various bugs:
10194 + Binary incompatibility of ssl_ctx_st structure
10195 + DTLS interoperation with non-compliant servers
10196 + Don't call get_session_cb() without proposed session
10197 + Fix ia64 assembler code
10199 *Andy Polyakov, Steve Henson*
10201 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10203 * DTLS Handshake overhaul. There were longstanding issues with
10204 OpenSSL DTLS implementation, which were making it impossible for
10205 RFC 4347 compliant client to communicate with OpenSSL server.
10206 Unfortunately just fixing these incompatibilities would "cut off"
10207 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10208 server keeps tolerating non RFC compliant syntax. The opposite is
10209 not true, 0.9.8f client can not communicate with earlier server.
10210 This update even addresses CVE-2007-4995.
10214 * Changes to avoid need for function casts in OpenSSL: some compilers
10215 (gcc 4.2 and later) reject their use.
10216 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10219 * Add RFC4507 support to OpenSSL. This includes the corrections in
10220 RFC4507bis. The encrypted ticket format is an encrypted encoded
10221 SSL_SESSION structure, that way new session features are automatically
10224 If a client application caches session in an SSL_SESSION structure
10225 support is transparent because tickets are now stored in the encoded
10228 The SSL_CTX structure automatically generates keys for ticket
10229 protection in servers so again support should be possible
10230 with no application modification.
10232 If a client or server wishes to disable RFC4507 support then the option
10233 SSL_OP_NO_TICKET can be set.
10235 Add a TLS extension debugging callback to allow the contents of any client
10236 or server extensions to be examined.
10238 This work was sponsored by Google.
10242 * Add initial support for TLS extensions, specifically for the server_name
10243 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10244 have new members for a host name. The SSL data structure has an
10245 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10246 stored in that context to allow for session resumption, even after the
10247 SSL has been switched to a new SSL_CTX in reaction to a client's
10248 server_name extension.
10250 New functions (subject to change):
10252 SSL_get_servername()
10253 SSL_get_servername_type()
10256 New CTRL codes and macros (subject to change):
10258 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10259 - SSL_CTX_set_tlsext_servername_callback()
10260 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10261 - SSL_CTX_set_tlsext_servername_arg()
10262 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10264 openssl s_client has a new '-servername ...' option.
10266 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10267 '-key2 ...', '-servername_fatal' (subject to change). This allows
10268 testing the HostName extension for a specific single host name ('-cert'
10269 and '-key' remain fallbacks for handshakes without HostName
10270 negotiation). If the unrecognized_name alert has to be sent, this by
10271 default is a warning; it becomes fatal with the '-servername_fatal'
10274 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10276 * Add AES and SSE2 assembly language support to VC++ build.
10280 * Mitigate attack on final subtraction in Montgomery reduction.
10284 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10285 (which previously caused an internal error).
10289 * Squeeze another 10% out of IGE mode when in != out.
10293 * AES IGE mode speedup.
10295 *Dean Gaudet (Google)*
10297 * Add the Korean symmetric 128-bit cipher SEED (see
10298 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10299 add SEED ciphersuites from RFC 4162:
10301 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10302 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10303 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10304 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10306 To minimize changes between patchlevels in the OpenSSL 0.9.8
10307 series, SEED remains excluded from compilation unless OpenSSL
10308 is configured with 'enable-seed'.
10310 *KISA, Bodo Moeller*
10312 * Mitigate branch prediction attacks, which can be practical if a
10313 single processor is shared, allowing a spy process to extract
10314 information. For detailed background information, see
10315 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10316 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10317 and Necessary Software Countermeasures"). The core of the change
10318 are new versions BN_div_no_branch() and
10319 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10320 respectively, which are slower, but avoid the security-relevant
10321 conditional branches. These are automatically called by BN_div()
10322 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10323 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10324 remove a conditional branch.
10326 BN_FLG_CONSTTIME is the new name for the previous
10327 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10328 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10329 in the exponent causes BN_mod_exp_mont() to use the alternative
10330 implementation in BN_mod_exp_mont_consttime().) The old name
10331 remains as a deprecated alias.
10333 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10334 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10335 constant-time implementations for more than just exponentiation.
10336 Here too the old name is kept as a deprecated alias.
10338 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10339 the BN_BLINDING structure gets an independent copy of the
10340 modulus. This means that the previous `BIGNUM *m` argument to
10341 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10342 essentially becomes `const BIGNUM *m`, although we can't actually
10343 change this in the header file before 0.9.9. It allows
10344 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10345 enable BN_FLG_CONSTTIME.
10347 *Matthew D Wood (Intel Corp)*
10349 * In the SSL/TLS server implementation, be strict about session ID
10350 context matching (which matters if an application uses a single
10351 external cache for different purposes). Previously,
10352 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10353 set. This did ensure strict client verification, but meant that,
10354 with applications using a single external cache for quite
10355 different requirements, clients could circumvent ciphersuite
10356 restrictions for a given session ID context by starting a session
10357 in a different context.
10361 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10362 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10363 authentication-only ciphersuites.
10367 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10368 not complete and could lead to a possible single byte overflow
10369 ([CVE-2007-5135]) [Ben Laurie]
10371 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10373 * Since AES128 and AES256 (and similarly Camellia128 and
10374 Camellia256) share a single mask bit in the logic of
10375 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10376 kludge to work properly if AES128 is available and AES256 isn't
10377 (or if Camellia128 is available and Camellia256 isn't).
10381 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10382 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10383 When a point or a seed is encoded in a BIT STRING, we need to
10384 prevent the removal of trailing zero bits to get the proper DER
10385 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10386 of a NamedBitList, for which trailing 0 bits need to be removed.)
10390 * Have SSL/TLS server implementation tolerate "mismatched" record
10391 protocol version while receiving ClientHello even if the
10392 ClientHello is fragmented. (The server can't insist on the
10393 particular protocol version it has chosen before the ServerHello
10394 message has informed the client about his choice.)
10398 * Add RFC 3779 support.
10400 *Rob Austein for ARIN, Ben Laurie*
10402 * Load error codes if they are not already present instead of using a
10403 static variable. This allows them to be cleanly unloaded and reloaded.
10404 Improve header file function name parsing.
10408 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10409 or CAPABILITY handshake as required by RFCs.
10411 *Goetz Babin-Ebell*
10413 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10415 * Introduce limits to prevent malicious keys being able to
10416 cause a denial of service. ([CVE-2006-2940])
10418 *Steve Henson, Bodo Moeller*
10420 * Fix ASN.1 parsing of certain invalid structures that can result
10421 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10423 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10424 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10426 * Fix SSL client code which could crash if connecting to a
10427 malicious SSLv2 server. ([CVE-2006-4343])
10429 *Tavis Ormandy and Will Drewry, Google Security Team*
10431 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10432 match only those. Before that, "AES256-SHA" would be interpreted
10433 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10434 the same strength classification in 0.9.7h) as we currently only
10435 have a single AES bit in the ciphersuite description bitmap.
10436 That change, however, also applied to ciphersuite strings such as
10437 "RC4-MD5" that intentionally matched multiple ciphersuites --
10438 namely, SSL 2.0 ciphersuites in addition to the more common ones
10439 from SSL 3.0/TLS 1.0.
10441 So we change the selection algorithm again: Naming an explicit
10442 ciphersuite selects this one ciphersuite, and any other similar
10443 ciphersuite (same bitmap) from *other* protocol versions.
10444 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10445 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10447 Since SSL 2.0 does not have any ciphersuites for which the
10448 128/256 bit distinction would be relevant, this works for now.
10449 The proper fix will be to use different bits for AES128 and
10450 AES256, which would have avoided the problems from the beginning;
10451 however, bits are scarce, so we can only do this in a new release
10452 (not just a patchlevel) when we can change the SSL_CIPHER
10453 definition to split the single 'unsigned long mask' bitmap into
10454 multiple values to extend the available space.
10458 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10460 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10461 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10463 * Add AES IGE and biIGE modes.
10467 * Change the Unix randomness entropy gathering to use poll() when
10468 possible instead of select(), since the latter has some
10469 undesirable limitations.
10471 *Darryl Miles via Richard Levitte and Bodo Moeller*
10473 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10474 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10475 cannot be implicitly activated as part of, e.g., the "AES" alias.
10476 However, please upgrade to OpenSSL 0.9.9[-dev] for
10477 non-experimental use of the ECC ciphersuites to get TLS extension
10478 support, which is required for curve and point format negotiation
10479 to avoid potential handshake problems.
10483 * Disable rogue ciphersuites:
10485 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10486 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10487 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10489 The latter two were purportedly from
10490 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10493 Also deactivate the remaining ciphersuites from
10494 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10495 unofficial, and the ID has long expired.
10499 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10500 dual-core machines) and other potential thread-safety issues.
10504 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10505 versions), which is now available for royalty-free use
10506 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10507 Also, add Camellia TLS ciphersuites from RFC 4132.
10509 To minimize changes between patchlevels in the OpenSSL 0.9.8
10510 series, Camellia remains excluded from compilation unless OpenSSL
10511 is configured with 'enable-camellia'.
10515 * Disable the padding bug check when compression is in use. The padding
10516 bug check assumes the first packet is of even length, this is not
10517 necessarily true if compression is enabled and can result in false
10518 positives causing handshake failure. The actual bug test is ancient
10519 code so it is hoped that implementations will either have fixed it by
10520 now or any which still have the bug do not support compression.
10524 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10526 * When applying a cipher rule check to see if string match is an explicit
10527 cipher suite and only match that one cipher suite if it is.
10531 * Link in manifests for VC++ if needed.
10533 *Austin Ziegler <halostatue@gmail.com>*
10535 * Update support for ECC-based TLS ciphersuites according to
10536 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10537 TLS extensions, which are supported starting with the 0.9.9
10538 branch, not in the OpenSSL 0.9.8 branch).
10542 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10543 opaque EVP_CIPHER_CTX handling.
10547 * Fixes and enhancements to zlib compression code. We now only use
10548 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10549 to conform with the standards mentioned here:
10550 <http://www.zlib.net/DLL_FAQ.txt>
10551 Static zlib linking now works on Windows and the new --with-zlib-include
10552 --with-zlib-lib options to Configure can be used to supply the location
10553 of the headers and library. Gracefully handle case where zlib library
10558 * Several fixes and enhancements to the OID generation code. The old code
10559 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10560 handle numbers larger than ULONG_MAX, truncated printing and had a
10561 non standard OBJ_obj2txt() behaviour.
10565 * Add support for building of engines under engine/ as shared libraries
10566 under VC++ build system.
10570 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10571 Hopefully, we will not see any false combination of paths any more.
10575 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10577 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10578 (part of SSL_OP_ALL). This option used to disable the
10579 countermeasure against man-in-the-middle protocol-version
10580 rollback in the SSL 2.0 server implementation, which is a bad
10581 idea. ([CVE-2005-2969])
10583 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10584 for Information Security, National Institute of Advanced Industrial
10585 Science and Technology [AIST], Japan)*
10587 * Add two function to clear and return the verify parameter flags.
10591 * Keep cipherlists sorted in the source instead of sorting them at
10592 runtime, thus removing the need for a lock.
10596 * Avoid some small subgroup attacks in Diffie-Hellman.
10598 *Nick Mathewson and Ben Laurie*
10600 * Add functions for well-known primes.
10604 * Extended Windows CE support.
10606 *Satoshi Nakamura and Andy Polyakov*
10608 * Initialize SSL_METHOD structures at compile time instead of during
10609 runtime, thus removing the need for a lock.
10613 * Make PKCS7_decrypt() work even if no certificate is supplied by
10614 attempting to decrypt each encrypted key in turn. Add support to
10619 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10621 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10624 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10628 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10629 key into the same file any more.
10633 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10637 * Add -utf8 command line and config file option to 'ca'.
10639 *Stefan <stf@udoma.org*
10641 * Removed the macro des_crypt(), as it seems to conflict with some
10642 libraries. Use DES_crypt().
10646 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10647 involves renaming the source and generated shared-libs for
10648 both. The engines will accept the corrected or legacy ids
10649 ('ncipher' and '4758_cca' respectively) when binding. NB,
10650 this only applies when building 'shared'.
10652 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10654 * Add attribute functions to EVP_PKEY structure. Modify
10655 PKCS12_create() to recognize a CSP name attribute and
10656 use it. Make -CSP option work again in pkcs12 utility.
10660 * Add new functionality to the bn blinding code:
10661 - automatic re-creation of the BN_BLINDING parameters after
10662 a fixed number of uses (currently 32)
10663 - add new function for parameter creation
10664 - introduce flags to control the update behaviour of the
10665 BN_BLINDING parameters
10666 - hide BN_BLINDING structure
10667 Add a second BN_BLINDING slot to the RSA structure to improve
10668 performance when a single RSA object is shared among several
10673 * Add support for DTLS.
10675 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10677 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10678 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10682 * Remove buggy and incomplete DH cert support from
10683 ssl/ssl_rsa.c and ssl/s3_both.c
10687 * Use SHA-1 instead of MD5 as the default digest algorithm for
10688 the `apps/openssl` commands.
10692 * Compile clean with "-Wall -Wmissing-prototypes
10693 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10694 DEBUG_SAFESTACK must also be set.
10698 * Change ./Configure so that certain algorithms can be disabled by default.
10699 The new counterpiece to "no-xxx" is "enable-xxx".
10701 The patented RC5 and MDC2 algorithms will now be disabled unless
10702 "enable-rc5" and "enable-mdc2", respectively, are specified.
10704 (IDEA remains enabled despite being patented. This is because IDEA
10705 is frequently required for interoperability, and there is no license
10706 fee for non-commercial use. As before, "no-idea" can be used to
10707 avoid this algorithm.)
10711 * Add processing of proxy certificates (see RFC 3820). This work was
10712 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10713 EGEE (Enabling Grids for E-science in Europe).
10717 * RC4 performance overhaul on modern architectures/implementations, such
10718 as Intel P4, IA-64 and AMD64.
10722 * New utility extract-section.pl. This can be used specify an alternative
10723 section number in a pod file instead of having to treat each file as
10724 a separate case in Makefile. This can be done by adding two lines to the
10727 =for comment openssl_section:XXX
10729 The blank line is mandatory.
10733 * New arguments -certform, -keyform and -pass for s_client and s_server
10734 to allow alternative format key and certificate files and passphrase
10739 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10740 update associated structures and add various utility functions.
10742 Add new policy related verify parameters, include policy checking in
10743 standard verify code. Enhance 'smime' application with extra parameters
10744 to support policy checking and print out.
10748 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10749 Nehemiah processors. These extensions support AES encryption in hardware
10750 as well as RNG (though RNG support is currently disabled).
10752 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10754 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10758 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10760 *Andy Polyakov and a number of other people*
10762 * Improved PowerPC platform support. Most notably BIGNUM assembler
10763 implementation contributed by IBM.
10765 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10767 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10768 exponent rather than 'unsigned long'. There is a corresponding change to
10769 the new 'rsa_keygen' element of the RSA_METHOD structure.
10771 *Jelte Jansen, Geoff Thorpe*
10773 * Functionality for creating the initial serial number file is now
10774 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10776 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10777 number file to 1, which is bound to cause problems. To avoid
10778 the problems while respecting compatibility between different 0.9.7
10779 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10780 CA.pl for serial number initialization. With the new release 0.9.8,
10781 we can fix the problem directly in the 'ca' utility.)
10785 * Reduced header interdependencies by declaring more opaque objects in
10786 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10787 give fewer recursive includes, which could break lazy source code - so
10788 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10789 developers should define this symbol when building and using openssl to
10790 ensure they track the recommended behaviour, interfaces, [etc], but
10791 backwards-compatible behaviour prevails when this isn't defined.
10795 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10799 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10800 This will generate a random key of the appropriate length based on the
10801 cipher context. The EVP_CIPHER can provide its own random key generation
10802 routine to support keys of a specific form. This is used in the des and
10803 3des routines to generate a key of the correct parity. Update S/MIME
10804 code to use new functions and hence generate correct parity DES keys.
10805 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10806 valid (weak or incorrect parity).
10810 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10811 as looking them up. This is useful when the verified structure may contain
10812 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10813 present unless the new PKCS7_NO_CRL flag is asserted.
10817 * Extend ASN1 oid configuration module. It now additionally accepts the
10820 shortName = some long name, 1.2.3.4
10824 * Reimplemented the BN_CTX implementation. There is now no more static
10825 limitation on the number of variables it can handle nor the depth of the
10826 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10827 information can now expand as required, and rather than having a single
10828 static array of bignums, BN_CTX now uses a linked-list of such arrays
10829 allowing it to expand on demand whilst maintaining the usefulness of
10830 BN_CTX's "bundling".
10834 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10835 to allow all RSA operations to function using a single BN_CTX.
10839 * Preliminary support for certificate policy evaluation and checking. This
10840 is initially intended to pass the tests outlined in "Conformance Testing
10841 of Relying Party Client Certificate Path Processing Logic" v1.07.
10845 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10846 remained unused and not that useful. A variety of other little bignum
10847 tweaks and fixes have also been made continuing on from the audit (see
10852 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10853 associated ASN1, EVP and SSL functions and old ASN1 macros.
10857 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10858 and this should never fail. So the return value from the use of
10859 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10860 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10864 * BN_CTX_get() should return zero-valued bignums, providing the same
10865 initialised value as BN_new().
10867 *Geoff Thorpe, suggested by Ulf Möller*
10869 * Support for inhibitAnyPolicy certificate extension.
10873 * An audit of the BIGNUM code is underway, for which debugging code is
10874 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10875 is considered valid when processing BIGNUMs, and causes execution to
10876 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10877 further steps are taken to deliberately pollute unused data in BIGNUM
10878 structures to try and expose faulty code further on. For now, openssl will
10879 (in its default mode of operation) continue to tolerate the inconsistent
10880 forms that it has tolerated in the past, but authors and packagers should
10881 consider trying openssl and their own applications when compiled with
10882 these debugging symbols defined. It will help highlight potential bugs in
10883 their own code, and will improve the test coverage for OpenSSL itself. At
10884 some point, these tighter rules will become openssl's default to improve
10885 maintainability, though the assert()s and other overheads will remain only
10886 in debugging configurations. See bn.h for more details.
10888 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10890 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10891 that can only be obtained through BN_CTX_new() (which implicitly
10892 initialises it). The presence of this function only made it possible
10893 to overwrite an existing structure (and cause memory leaks).
10897 * Because of the callback-based approach for implementing LHASH as a
10898 template type, lh_insert() adds opaque objects to hash-tables and
10899 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10900 to clean up those corresponding objects before destroying the hash table
10901 (and losing the object pointers). So some over-zealous constifications in
10902 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10903 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10904 prototyped to have "const" restrictions on the object pointers they are
10905 given (and so aren't required to cast them away any more).
10909 * The tmdiff.h API was so ugly and minimal that our own timing utility
10910 (speed) prefers to use its own implementation. The two implementations
10911 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10912 its object type properly exposed (MS_TM) instead of casting to/from
10913 `char *`. This may still change yet if someone realises MS_TM and
10915 aren't necessarily the greatest nomenclatures - but this is what was used
10916 internally to the implementation so I've used that for now.
10920 * Ensure that deprecated functions do not get compiled when
10921 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10922 the self-tests were still using deprecated key-generation functions so
10923 these have been updated also.
10927 * Reorganise PKCS#7 code to separate the digest location functionality
10928 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10929 New function PKCS7_set_digest() to set the digest type for PKCS#7
10930 digestedData type. Add additional code to correctly generate the
10931 digestedData type and add support for this type in PKCS7 initialization
10936 * New function PKCS7_set0_type_other() this initializes a PKCS7
10937 structure of type "other".
10941 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10942 sure the loop does correctly stop and breaking ("division by zero")
10943 modulus operations are not performed. The (pre-generated) prime
10944 table crypto/bn/bn_prime.h was already correct, but it could not be
10945 re-generated on some platforms because of the "division by zero"
10946 situation in the script.
10948 *Ralf S. Engelschall*
10950 * Update support for ECC-based TLS ciphersuites according to
10951 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10952 SHA-1 now is only used for "small" curves (where the
10953 representation of a field element takes up to 24 bytes); for
10954 larger curves, the field element resulting from ECDH is directly
10955 used as premaster secret.
10957 *Douglas Stebila (Sun Microsystems Laboratories)*
10959 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10960 curve secp160r1 to the tests.
10962 *Douglas Stebila (Sun Microsystems Laboratories)*
10964 * Add the possibility to load symbols globally with DSO.
10966 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10968 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10969 control of the error stack.
10973 * Add support for STORE in ENGINE.
10977 * Add the STORE type. The intention is to provide a common interface
10978 to certificate and key stores, be they simple file-based stores, or
10979 HSM-type store, or LDAP stores, or...
10980 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10984 * Add a generic structure called OPENSSL_ITEM. This can be used to
10985 pass a list of arguments to any function as well as provide a way
10986 for a function to pass data back to the caller.
10990 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10991 works like BUF_strdup() but can be used to duplicate a portion of
10992 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10997 * Add the function sk_find_ex() which works like sk_find(), but will
10998 return an index to an element even if an exact match couldn't be
10999 found. The index is guaranteed to point at the element where the
11000 searched-for key would be inserted to preserve sorting order.
11004 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11005 takes an extra flags argument for optional functionality. Currently,
11006 the following flags are defined:
11008 OBJ_BSEARCH_VALUE_ON_NOMATCH
11009 This one gets OBJ_bsearch_ex() to return a pointer to the first
11010 element where the comparing function returns a negative or zero
11013 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11014 This one gets OBJ_bsearch_ex() to return a pointer to the first
11015 element where the comparing function returns zero. This is useful
11016 if there are more than one element where the comparing function
11021 * Make it possible to create self-signed certificates with 'openssl ca'
11022 in such a way that the self-signed certificate becomes part of the
11023 CA database and uses the same mechanisms for serial number generation
11024 as all other certificate signing. The new flag '-selfsign' enables
11025 this functionality. Adapt CA.sh and CA.pl.in.
11029 * Add functionality to check the public key of a certificate request
11030 against a given private. This is useful to check that a certificate
11031 request can be signed by that key (self-signing).
11035 * Make it possible to have multiple active certificates with the same
11036 subject in the CA index file. This is done only if the keyword
11037 'unique_subject' is set to 'no' in the main CA section (default
11038 if 'CA_default') of the configuration file. The value is saved
11039 with the database itself in a separate index attribute file,
11040 named like the index file with '.attr' appended to the name.
11044 * Generate multi-valued AVAs using '+' notation in config files for
11049 * Support for nameConstraints certificate extension.
11053 * Support for policyConstraints certificate extension.
11057 * Support for policyMappings certificate extension.
11061 * Make sure the default DSA_METHOD implementation only uses its
11062 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11063 and change its own handlers to be NULL so as to remove unnecessary
11064 indirection. This lets alternative implementations fallback to the
11065 default implementation more easily.
11069 * Support for directoryName in GeneralName related extensions
11074 * Make it possible to link applications using Makefile.shared.
11075 Make that possible even when linking against static libraries!
11079 * Support for single pass processing for S/MIME signing. This now
11080 means that S/MIME signing can be done from a pipe, in addition
11081 cleartext signing (multipart/signed type) is effectively streaming
11082 and the signed data does not need to be all held in memory.
11084 This is done with a new flag PKCS7_STREAM. When this flag is set
11085 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11086 is done after the data is output (and digests calculated) in
11087 SMIME_write_PKCS7().
11091 * Add full support for -rpath/-R, both in shared libraries and
11092 applications, at least on the platforms where it's known how
11097 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11098 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11099 will now compute a table of multiples of the generator that
11100 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11101 faster (notably in the case of a single point multiplication,
11102 scalar * generator).
11104 *Nils Larsch, Bodo Moeller*
11106 * IPv6 support for certificate extensions. The various extensions
11107 which use the IP:a.b.c.d can now take IPv6 addresses using the
11108 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11113 * Added an ENGINE that implements RSA by performing private key
11114 exponentiations with the GMP library. The conversions to and from
11115 GMP's mpz_t format aren't optimised nor are any montgomery forms
11116 cached, and on x86 it appears OpenSSL's own performance has caught up.
11117 However there are likely to be other architectures where GMP could
11118 provide a boost. This ENGINE is not built in by default, but it can be
11119 specified at Configure time and should be accompanied by the necessary
11120 linker additions, eg;
11121 ./config -DOPENSSL_USE_GMP -lgmp
11125 * "openssl engine" will not display ENGINE/DSO load failure errors when
11126 testing availability of engines with "-t" - the old behaviour is
11127 produced by increasing the feature's verbosity with "-tt".
11131 * ECDSA routines: under certain error conditions uninitialized BN objects
11132 could be freed. Solution: make sure initialization is performed early
11133 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11138 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11139 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11140 software implementations. For DSA and DH, parameter generation can
11141 also be overridden by providing the appropriate method callbacks.
11145 * Change the "progress" mechanism used in key-generation and
11146 primality testing to functions that take a new BN_GENCB pointer in
11147 place of callback/argument pairs. The new API functions have `_ex`
11148 postfixes and the older functions are reimplemented as wrappers for
11149 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11150 declarations of the old functions to help (graceful) attempts to
11151 migrate to the new functions. Also, the new key-generation API
11152 functions operate on a caller-supplied key-structure and return
11153 success/failure rather than returning a key or NULL - this is to
11154 help make "keygen" another member function of RSA_METHOD etc.
11156 Example for using the new callback interface:
11158 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11159 void *my_arg = ...;
11162 BN_GENCB_set(&my_cb, my_callback, my_arg);
11164 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11165 /* For the meaning of a, b in calls to my_callback(), see the
11166 * documentation of the function that calls the callback.
11167 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11168 * my_callback should return 1 if it wants BN_is_prime_ex()
11169 * to continue, or 0 to stop.
11174 * Change the ZLIB compression method to be stateful, and make it
11175 available to TLS with the number defined in
11176 draft-ietf-tls-compression-04.txt.
11180 * Add the ASN.1 structures and functions for CertificatePair, which
11181 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11183 CertificatePair ::= SEQUENCE {
11184 forward [0] Certificate OPTIONAL,
11185 reverse [1] Certificate OPTIONAL,
11186 -- at least one of the pair shall be present -- }
11188 Also implement the PEM functions to read and write certificate
11189 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11191 This needed to be defined, mostly for the sake of the LDAP
11192 attribute crossCertificatePair, but may prove useful elsewhere as
11197 * Make it possible to inhibit symlinking of shared libraries in
11198 Makefile.shared, for Cygwin's sake.
11202 * Extend the BIGNUM API by creating a function
11203 void BN_set_negative(BIGNUM *a, int neg);
11204 and a macro that behave like
11205 int BN_is_negative(const BIGNUM *a);
11207 to avoid the need to access 'a->neg' directly in applications.
11211 * Implement fast modular reduction for pseudo-Mersenne primes
11212 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11213 EC_GROUP_new_curve_GFp() will now automatically use this
11216 *Nils Larsch <nla@trustcenter.de>*
11218 * Add new lock type (CRYPTO_LOCK_BN).
11222 * Change the ENGINE framework to automatically load engines
11223 dynamically from specific directories unless they could be
11224 found to already be built in or loaded. Move all the
11225 current engines except for the cryptodev one to a new
11226 directory engines/.
11227 The engines in engines/ are built as shared libraries if
11228 the "shared" options was given to ./Configure or ./config.
11229 Otherwise, they are inserted in libcrypto.a.
11230 /usr/local/ssl/engines is the default directory for dynamic
11231 engines, but that can be overridden at configure time through
11232 the usual use of --prefix and/or --openssldir, and at run
11233 time with the environment variable OPENSSL_ENGINES.
11235 *Geoff Thorpe and Richard Levitte*
11237 * Add Makefile.shared, a helper makefile to build shared
11238 libraries. Adapt Makefile.org.
11242 * Add version info to Win32 DLLs.
11244 *Peter 'Luna' Runestig" <peter@runestig.com>*
11246 * Add new 'medium level' PKCS#12 API. Certificates and keys
11247 can be added using this API to created arbitrary PKCS#12
11248 files while avoiding the low-level API.
11250 New options to PKCS12_create(), key or cert can be NULL and
11251 will then be omitted from the output file. The encryption
11252 algorithm NIDs can be set to -1 for no encryption, the mac
11253 iteration count can be set to 0 to omit the mac.
11255 Enhance pkcs12 utility by making the -nokeys and -nocerts
11256 options work when creating a PKCS#12 file. New option -nomac
11257 to omit the mac, NONE can be set for an encryption algorithm.
11258 New code is modified to use the enhanced PKCS12_create()
11259 instead of the low-level API.
11263 * Extend ASN1 encoder to support indefinite length constructed
11264 encoding. This can output sequences tags and octet strings in
11265 this form. Modify pk7_asn1.c to support indefinite length
11266 encoding. This is experimental and needs additional code to
11267 be useful, such as an ASN1 bio and some enhanced streaming
11270 Extend template encode functionality so that tagging is passed
11271 down to the template encoder.
11275 * Let 'openssl req' fail if an argument to '-newkey' is not
11276 recognized instead of using RSA as a default.
11280 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11281 As these are not official, they are not included in "ALL";
11282 the "ECCdraft" ciphersuite group alias can be used to select them.
11284 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11286 * Add ECDH engine support.
11288 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11290 * Add ECDH in new directory crypto/ecdh/.
11292 *Douglas Stebila (Sun Microsystems Laboratories)*
11294 * Let BN_rand_range() abort with an error after 100 iterations
11295 without success (which indicates a broken PRNG).
11299 * Change BN_mod_sqrt() so that it verifies that the input value
11300 is really the square of the return value. (Previously,
11301 BN_mod_sqrt would show GIGO behaviour.)
11305 * Add named elliptic curves over binary fields from X9.62, SECG,
11306 and WAP/WTLS; add OIDs that were still missing.
11308 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11310 * Extend the EC library for elliptic curves over binary fields
11311 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11314 EC_GF2m_simple_method
11318 EC_GROUP_new_curve_GF2m
11319 EC_GROUP_set_curve_GF2m
11320 EC_GROUP_get_curve_GF2m
11321 EC_POINT_set_affine_coordinates_GF2m
11322 EC_POINT_get_affine_coordinates_GF2m
11323 EC_POINT_set_compressed_coordinates_GF2m
11325 Point compression for binary fields is disabled by default for
11326 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11329 As binary polynomials are represented as BIGNUMs, various members
11330 of the EC_GROUP and EC_POINT data structures can be shared
11331 between the implementations for prime fields and binary fields;
11332 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11333 are essentially identical to their `..._GFp` counterparts.
11334 (For simplicity, the `..._GFp` prefix has been dropped from
11335 various internal method names.)
11337 An internal 'field_div' method (similar to 'field_mul' and
11338 'field_sqr') has been added; this is used only for binary fields.
11340 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11342 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11343 through methods ('mul', 'precompute_mult').
11345 The generic implementations (now internally called 'ec_wNAF_mul'
11346 and 'ec_wNAF_precomputed_mult') remain the default if these
11347 methods are undefined.
11349 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11351 * New function EC_GROUP_get_degree, which is defined through
11352 EC_METHOD. For curves over prime fields, this returns the bit
11353 length of the modulus.
11355 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11357 * New functions EC_GROUP_dup, EC_POINT_dup.
11358 (These simply call ..._new and ..._copy).
11360 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11362 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11363 Polynomials are represented as BIGNUMs (where the sign bit is not
11364 used) in the following functions [macros]:
11367 BN_GF2m_sub [= BN_GF2m_add]
11368 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11369 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11370 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11372 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11373 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11374 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11375 BN_GF2m_cmp [= BN_ucmp]
11377 (Note that only the 'mod' functions are actually for fields GF(2^m).
11378 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11380 For some functions, an the irreducible polynomial defining a
11381 field can be given as an 'unsigned int[]' with strictly
11382 decreasing elements giving the indices of those bits that are set;
11383 i.e., p[] represents the polynomial
11384 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11386 p[0] > p[1] > ... > p[k] = 0.
11387 This applies to the following functions:
11390 BN_GF2m_mod_mul_arr
11391 BN_GF2m_mod_sqr_arr
11392 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11393 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11394 BN_GF2m_mod_exp_arr
11395 BN_GF2m_mod_sqrt_arr
11396 BN_GF2m_mod_solve_quad_arr
11400 Conversion can be performed by the following functions:
11405 bntest.c has additional tests for binary polynomial arithmetic.
11407 Two implementations for BN_GF2m_mod_div() are available.
11408 The default algorithm simply uses BN_GF2m_mod_inv() and
11409 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11410 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11411 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11413 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11415 * Add new error code 'ERR_R_DISABLED' that can be used when some
11416 functionality is disabled at compile-time.
11418 *Douglas Stebila <douglas.stebila@sun.com>*
11420 * Change default behaviour of 'openssl asn1parse' so that more
11421 information is visible when viewing, e.g., a certificate:
11423 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11424 mode the content of non-printable OCTET STRINGs is output in a
11425 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11426 avoid the appearance of a printable string.
11428 *Nils Larsch <nla@trustcenter.de>*
11430 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11432 EC_GROUP_set_asn1_flag()
11433 EC_GROUP_get_asn1_flag()
11434 EC_GROUP_set_point_conversion_form()
11435 EC_GROUP_get_point_conversion_form()
11436 These control ASN1 encoding details:
11437 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11438 has been set to OPENSSL_EC_NAMED_CURVE.
11439 - Points are encoded in uncompressed form by default; options for
11440 asn1_for are as for point2oct, namely
11441 POINT_CONVERSION_COMPRESSED
11442 POINT_CONVERSION_UNCOMPRESSED
11443 POINT_CONVERSION_HYBRID
11445 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11447 EC_GROUP_set_seed()
11448 EC_GROUP_get0_seed()
11449 EC_GROUP_get_seed_len()
11450 This is used only for ASN1 purposes (so far).
11452 *Nils Larsch <nla@trustcenter.de>*
11454 * Add 'field_type' member to EC_METHOD, which holds the NID
11455 of the appropriate field type OID. The new function
11456 EC_METHOD_get_field_type() returns this value.
11458 *Nils Larsch <nla@trustcenter.de>*
11461 EC_POINT_point2bn()
11462 EC_POINT_bn2point()
11463 EC_POINT_point2hex()
11464 EC_POINT_hex2point()
11465 providing useful interfaces to EC_POINT_point2oct() and
11466 EC_POINT_oct2point().
11468 *Nils Larsch <nla@trustcenter.de>*
11470 * Change internals of the EC library so that the functions
11471 EC_GROUP_set_generator()
11472 EC_GROUP_get_generator()
11473 EC_GROUP_get_order()
11474 EC_GROUP_get_cofactor()
11475 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11476 to methods, which would lead to unnecessary code duplication when
11477 adding different types of curves.
11479 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11481 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11482 arithmetic, and such that modified wNAFs are generated
11483 (which avoid length expansion in many cases).
11487 * Add a function EC_GROUP_check_discriminant() (defined via
11488 EC_METHOD) that verifies that the curve discriminant is non-zero.
11490 Add a function EC_GROUP_check() that makes some sanity tests
11491 on a EC_GROUP, its generator and order. This includes
11492 EC_GROUP_check_discriminant().
11494 *Nils Larsch <nla@trustcenter.de>*
11496 * Add ECDSA in new directory crypto/ecdsa/.
11498 Add applications 'openssl ecparam' and 'openssl ecdsa'
11499 (these are based on 'openssl dsaparam' and 'openssl dsa').
11501 ECDSA support is also included in various other files across the
11502 library. Most notably,
11503 - 'openssl req' now has a '-newkey ecdsa:file' option;
11504 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11505 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11506 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11507 them suitable for ECDSA where domain parameters must be
11508 extracted before the specific public key;
11509 - ECDSA engine support has been added.
11511 *Nils Larsch <nla@trustcenter.de>*
11513 * Include some named elliptic curves, and add OIDs from X9.62,
11514 SECG, and WAP/WTLS. Each curve can be obtained from the new
11516 EC_GROUP_new_by_curve_name(),
11517 and the list of available named curves can be obtained with
11518 EC_get_builtin_curves().
11519 Also add a 'curve_name' member to EC_GROUP objects, which can be
11521 EC_GROUP_set_curve_name()
11522 EC_GROUP_get_curve_name()
11524 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11526 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11527 was actually never needed) and in BN_mul(). The removal in BN_mul()
11528 required a small change in bn_mul_part_recursive() and the addition
11529 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11530 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11531 bn_sub_words() and bn_add_words() except they take arrays with
11536 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11538 * Cleanse PEM buffers before freeing them since they may contain
11541 *Benjamin Bennett <ben@psc.edu>*
11543 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11544 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11545 authentication-only ciphersuites.
11549 * Since AES128 and AES256 share a single mask bit in the logic of
11550 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11551 kludge to work properly if AES128 is available and AES256 isn't.
11555 * Expand security boundary to match 1.1.1 module.
11559 * Remove redundant features: hash file source, editing of test vectors
11560 modify fipsld to use external fips_premain.c signature.
11564 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11565 run algorithm test programs.
11569 * Make algorithm test programs more tolerant of whitespace.
11573 * Have SSL/TLS server implementation tolerate "mismatched" record
11574 protocol version while receiving ClientHello even if the
11575 ClientHello is fragmented. (The server can't insist on the
11576 particular protocol version it has chosen before the ServerHello
11577 message has informed the client about his choice.)
11581 * Load error codes if they are not already present instead of using a
11582 static variable. This allows them to be cleanly unloaded and reloaded.
11586 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11588 * Introduce limits to prevent malicious keys being able to
11589 cause a denial of service. ([CVE-2006-2940])
11591 *Steve Henson, Bodo Moeller*
11593 * Fix ASN.1 parsing of certain invalid structures that can result
11594 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11596 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11597 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11599 * Fix SSL client code which could crash if connecting to a
11600 malicious SSLv2 server. ([CVE-2006-4343])
11602 *Tavis Ormandy and Will Drewry, Google Security Team*
11604 * Change ciphersuite string processing so that an explicit
11605 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11606 will no longer include "AES128-SHA"), and any other similar
11607 ciphersuite (same bitmap) from *other* protocol versions (so that
11608 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11609 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11610 changes from 0.9.8b and 0.9.8d.
11614 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11616 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11617 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11619 * Change the Unix randomness entropy gathering to use poll() when
11620 possible instead of select(), since the latter has some
11621 undesirable limitations.
11623 *Darryl Miles via Richard Levitte and Bodo Moeller*
11625 * Disable rogue ciphersuites:
11627 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11628 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11629 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11631 The latter two were purportedly from
11632 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11635 Also deactivate the remaining ciphersuites from
11636 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11637 unofficial, and the ID has long expired.
11641 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11642 dual-core machines) and other potential thread-safety issues.
11646 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11648 * Adapt fipsld and the build system to link against the validated FIPS
11649 module in FIPS mode.
11653 * Fixes for VC++ 2005 build under Windows.
11657 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11658 from a Windows bash shell such as MSYS. It is autodetected from the
11659 "config" script when run from a VC++ environment. Modify standard VC++
11660 build to use fipscanister.o from the GNU make build.
11664 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11666 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11667 The value now differs depending on if you build for FIPS or not.
11668 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11669 safely run with a non-FIPSed libcrypto, as it may crash because of
11670 the difference induced by this change.
11674 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11676 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11677 (part of SSL_OP_ALL). This option used to disable the
11678 countermeasure against man-in-the-middle protocol-version
11679 rollback in the SSL 2.0 server implementation, which is a bad
11680 idea. ([CVE-2005-2969])
11682 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11683 for Information Security, National Institute of Advanced Industrial
11684 Science and Technology [AIST, Japan)]*
11686 * Minimal support for X9.31 signatures and PSS padding modes. This is
11687 mainly for FIPS compliance and not fully integrated at this stage.
11691 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11692 the exponentiation using a fixed-length exponent. (Otherwise,
11693 the information leaked through timing could expose the secret key
11694 after many signatures; cf. Bleichenbacher's attack on DSA with
11699 * Make a new fixed-window mod_exp implementation the default for
11700 RSA, DSA, and DH private-key operations so that the sequence of
11701 squares and multiplies and the memory access pattern are
11702 independent of the particular secret key. This will mitigate
11703 cache-timing and potential related attacks.
11705 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11706 and this is automatically used by BN_mod_exp_mont() if the new flag
11707 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11708 will use this BN flag for private exponents unless the flag
11709 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11710 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11712 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11714 * Change the client implementation for SSLv23_method() and
11715 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11716 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11717 (Previously, the SSL 2.0 backwards compatible Client Hello
11718 message format would be used even with SSL_OP_NO_SSLv2.)
11722 * Add support for smime-type MIME parameter in S/MIME messages which some
11727 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11728 a threadsafe manner. Modify rsa code to use new function and add calls
11729 to dsa and dh code (which had race conditions before).
11733 * Include the fixed error library code in the C error file definitions
11734 instead of fixing them up at runtime. This keeps the error code
11735 structures constant.
11739 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11741 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11744 * Fixes for newer kerberos headers. NB: the casts are needed because
11745 the 'length' field is signed on one version and unsigned on another
11746 with no (?) obvious way to tell the difference, without these VC++
11747 complains. Also the "definition" of FAR (blank) is no longer included
11748 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11749 some needed definitions.
11753 * Undo Cygwin change.
11757 * Added support for proxy certificates according to RFC 3820.
11758 Because they may be a security thread to unaware applications,
11759 they must be explicitly allowed in run-time. See
11760 docs/HOWTO/proxy_certificates.txt for further information.
11764 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11766 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11767 server and client random values. Previously
11768 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11769 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11771 This change has negligible security impact because:
11773 1. Server and client random values still have 24 bytes of pseudo random
11776 2. Server and client random values are sent in the clear in the initial
11779 3. The master secret is derived using the premaster secret (48 bytes in
11780 size for static RSA ciphersuites) as well as client server and random
11783 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11786 *Stephen Henson, reported by UK NISCC*
11788 * Use Windows randomness collection on Cygwin.
11792 * Fix hang in EGD/PRNGD query when communication socket is closed
11793 prematurely by EGD/PRNGD.
11795 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11797 * Prompt for pass phrases when appropriate for PKCS12 input format.
11801 * Back-port of selected performance improvements from development
11802 branch, as well as improved support for PowerPC platforms.
11806 * Add lots of checks for memory allocation failure, error codes to indicate
11807 failure and freeing up memory if a failure occurs.
11809 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11811 * Add new -passin argument to dgst.
11815 * Perform some character comparisons of different types in X509_NAME_cmp:
11816 this is needed for some certificates that re-encode DNs into UTF8Strings
11817 (in violation of RFC3280) and can't or won't issue name rollover
11822 * Make an explicit check during certificate validation to see that
11823 the CA setting in each certificate on the chain is correct. As a
11824 side effect always do the following basic checks on extensions,
11825 not just when there's an associated purpose to the check:
11827 - if there is an unhandled critical extension (unless the user
11828 has chosen to ignore this fault)
11829 - if the path length has been exceeded (if one is set at all)
11830 - that certain extensions fit the associated purpose (if one has
11835 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11837 * Avoid a race condition when CRLs are checked in a multi threaded
11838 environment. This would happen due to the reordering of the revoked
11839 entries during signature checking and serial number lookup. Now the
11840 encoding is cached and the serial number sort performed under a lock.
11841 Add new STACK function sk_is_sorted().
11845 * Add Delta CRL to the extension code.
11849 * Various fixes to s3_pkt.c so alerts are sent properly.
11851 *David Holmes <d.holmes@f5.com>*
11853 * Reduce the chances of duplicate issuer name and serial numbers (in
11854 violation of RFC3280) using the OpenSSL certificate creation utilities.
11855 This is done by creating a random 64 bit value for the initial serial
11856 number when a serial number file is created or when a self signed
11857 certificate is created using 'openssl req -x509'. The initial serial
11858 number file is created using 'openssl x509 -next_serial' in CA.pl
11859 rather than being initialized to 1.
11863 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11865 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11866 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11868 *Joe Orton, Steve Henson*
11870 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11873 *Joe Orton, Steve Henson*
11875 * Make it possible to have multiple active certificates with the same
11876 subject in the CA index file. This is done only if the keyword
11877 'unique_subject' is set to 'no' in the main CA section (default
11878 if 'CA_default') of the configuration file. The value is saved
11879 with the database itself in a separate index attribute file,
11880 named like the index file with '.attr' appended to the name.
11884 * X509 verify fixes. Disable broken certificate workarounds when
11885 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11886 keyUsage extension present. Don't accept CRLs with unhandled critical
11887 extensions: since verify currently doesn't process CRL extensions this
11888 rejects a CRL with *any* critical extensions. Add new verify error codes
11893 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11894 A clarification of RFC2560 will require the use of OCTET STRINGs and
11895 some implementations cannot handle the current raw format. Since OpenSSL
11896 copies and compares OCSP nonces as opaque blobs without any attempt at
11897 parsing them this should not create any compatibility issues.
11901 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11902 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11903 this HMAC (and other) operations are several times slower than OpenSSL
11908 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11910 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11912 * Use the correct content when signing type "other".
11916 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11918 * Fix various bugs revealed by running the NISCC test suite:
11920 Stop out of bounds reads in the ASN1 code when presented with
11921 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11923 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11925 If verify callback ignores invalid public key errors don't try to check
11926 certificate signature with the NULL public key.
11930 * New -ignore_err option in ocsp application to stop the server
11931 exiting on the first error in a request.
11935 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11936 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11941 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11942 extra data after the compression methods not only for TLS 1.0
11943 but also for SSL 3.0 (as required by the specification).
11945 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11947 * Change X509_certificate_type() to mark the key as exported/exportable
11948 when it's 512 *bits* long, not 512 bytes.
11952 * Change AES_cbc_encrypt() so it outputs exact multiple of
11953 blocks during encryption.
11957 * Various fixes to base64 BIO and non blocking I/O. On write
11958 flushes were not handled properly if the BIO retried. On read
11959 data was not being buffered properly and had various logic bugs.
11960 This also affects blocking I/O when the data being decoded is a
11965 * Various S/MIME bugfixes and compatibility changes:
11966 output correct application/pkcs7 MIME type if
11967 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11968 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11969 of files as .eml work). Correctly handle very long lines in MIME
11974 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11976 * Countermeasure against the Klima-Pokorny-Rosa extension of
11977 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11978 a protocol version number mismatch like a decryption error
11979 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11983 * Turn on RSA blinding by default in the default implementation
11984 to avoid a timing attack. Applications that don't want it can call
11985 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11986 They would be ill-advised to do so in most cases.
11988 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11990 * Change RSA blinding code so that it works when the PRNG is not
11991 seeded (in this case, the secret RSA exponent is abused as
11992 an unpredictable seed -- if it is not unpredictable, there
11993 is no point in blinding anyway). Make RSA blinding thread-safe
11994 by remembering the creator's thread ID in rsa->blinding and
11995 having all other threads use local one-time blinding factors
11996 (this requires more computation than sharing rsa->blinding, but
11997 avoids excessive locking; and if an RSA object is not shared
11998 between threads, blinding will still be very fast).
12002 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12003 ENGINE as defaults for all supported algorithms irrespective of
12004 the 'flags' parameter. 'flags' is now honoured, so applications
12005 should make sure they are passing it correctly.
12009 * Target "mingw" now allows native Windows code to be generated in
12010 the Cygwin environment as well as with the MinGW compiler.
12014 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12016 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12017 via timing by performing a MAC computation even if incorrect
12018 block cipher padding has been found. This is a countermeasure
12019 against active attacks where the attacker has to distinguish
12020 between bad padding and a MAC verification error. ([CVE-2003-0078])
12022 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12023 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12024 Martin Vuagnoux (EPFL, Ilion)*
12026 * Make the no-err option work as intended. The intention with no-err
12027 is not to have the whole error stack handling routines removed from
12028 libcrypto, it's only intended to remove all the function name and
12029 reason texts, thereby removing some of the footprint that may not
12030 be interesting if those errors aren't displayed anyway.
12032 NOTE: it's still possible for any application or module to have its
12033 own set of error texts inserted. The routines are there, just not
12034 used by default when no-err is given.
12038 * Add support for FreeBSD on IA64.
12040 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12042 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12043 Kerberos function mit_des_cbc_cksum(). Before this change,
12044 the value returned by DES_cbc_cksum() was like the one from
12045 mit_des_cbc_cksum(), except the bytes were swapped.
12047 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12049 * Allow an application to disable the automatic SSL chain building.
12050 Before this a rather primitive chain build was always performed in
12051 ssl3_output_cert_chain(): an application had no way to send the
12052 correct chain if the automatic operation produced an incorrect result.
12054 Now the chain builder is disabled if either:
12056 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12058 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12060 The reasoning behind this is that an application would not want the
12061 auto chain building to take place if extra chain certificates are
12062 present and it might also want a means of sending no additional
12063 certificates (for example the chain has two certificates and the
12068 * Add the possibility to build without the ENGINE framework.
12070 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12072 * Under Win32 gmtime() can return NULL: check return value in
12073 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12077 * DSA routines: under certain error conditions uninitialized BN objects
12078 could be freed. Solution: make sure initialization is performed early
12079 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12080 Nils Larsch <nla@trustcenter.de> via PR#459)
12084 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12085 checked on reconnect on the client side, therefore session resumption
12086 could still fail with a "ssl session id is different" error. This
12087 behaviour is masked when SSL_OP_ALL is used due to
12088 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12089 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12090 followup to PR #377.
12094 * IA-32 assembler support enhancements: unified ELF targets, support
12095 for SCO/Caldera platforms, fix for Cygwin shared build.
12099 * Add support for FreeBSD on sparc64. As a consequence, support for
12100 FreeBSD on non-x86 processors is separate from x86 processors on
12101 the config script, much like the NetBSD support.
12103 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12105 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12107 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12110 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12111 code (06) was taken as the first octet of the session ID and the last
12112 octet was ignored consequently. As a result SSLv2 client side session
12113 caching could not have worked due to the session ID mismatch between
12115 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12120 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12121 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12126 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12127 seems that in spite of existing for more than a year, many application
12128 author have done nothing to provide the necessary callbacks, which
12129 means that this particular engine will not work properly anywhere.
12130 This is a very unfortunate situation which forces us, in the name
12131 of usability, to give the hw_ncipher.c a static lock, which is part
12133 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12134 appear in 0.9.8 or later. We EXPECT application authors to have
12135 dealt properly with this when 0.9.8 is released (unless we actually
12136 make such changes in the libcrypto locking code that changes will
12137 have to be made anyway).
12141 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12142 octets have been read, EOF or an error occurs. Without this change
12143 some truncated ASN1 structures will not produce an error.
12147 * Disable Heimdal support, since it hasn't been fully implemented.
12148 Still give the possibility to force the use of Heimdal, but with
12149 warnings and a request that patches get sent to openssl-dev.
12153 * Add the VC-CE target, introduce the WINCE sysname, and add
12154 INSTALL.WCE and appropriate conditionals to make it build.
12156 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12158 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12159 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12160 edit numbers of the version.
12162 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12164 * Introduce safe string copy and catenation functions
12165 (BUF_strlcpy() and BUF_strlcat()).
12167 *Ben Laurie (CHATS) and Richard Levitte*
12169 * Avoid using fixed-size buffers for one-line DNs.
12171 *Ben Laurie (CHATS)*
12173 * Add BUF_MEM_grow_clean() to avoid information leakage when
12174 resizing buffers containing secrets, and use where appropriate.
12176 *Ben Laurie (CHATS)*
12178 * Avoid using fixed size buffers for configuration file location.
12180 *Ben Laurie (CHATS)*
12182 * Avoid filename truncation for various CA files.
12184 *Ben Laurie (CHATS)*
12186 * Use sizeof in preference to magic numbers.
12188 *Ben Laurie (CHATS)*
12190 * Avoid filename truncation in cert requests.
12192 *Ben Laurie (CHATS)*
12194 * Add assertions to check for (supposedly impossible) buffer
12197 *Ben Laurie (CHATS)*
12199 * Don't cache truncated DNS entries in the local cache (this could
12200 potentially lead to a spoofing attack).
12202 *Ben Laurie (CHATS)*
12204 * Fix various buffers to be large enough for hex/decimal
12205 representations in a platform independent manner.
12207 *Ben Laurie (CHATS)*
12209 * Add CRYPTO_realloc_clean() to avoid information leakage when
12210 resizing buffers containing secrets, and use where appropriate.
12212 *Ben Laurie (CHATS)*
12214 * Add BIO_indent() to avoid much slightly worrying code to do
12217 *Ben Laurie (CHATS)*
12219 * Convert sprintf()/BIO_puts() to BIO_printf().
12221 *Ben Laurie (CHATS)*
12223 * buffer_gets() could terminate with the buffer only half
12226 *Ben Laurie (CHATS)*
12228 * Add assertions to prevent user-supplied crypto functions from
12229 overflowing internal buffers by having large block sizes, etc.
12231 *Ben Laurie (CHATS)*
12233 * New OPENSSL_assert() macro (similar to assert(), but enabled
12236 *Ben Laurie (CHATS)*
12238 * Eliminate unused copy of key in RC4.
12240 *Ben Laurie (CHATS)*
12242 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12244 *Ben Laurie (CHATS)*
12246 * Fix off-by-one error in EGD path.
12248 *Ben Laurie (CHATS)*
12250 * If RANDFILE path is too long, ignore instead of truncating.
12252 *Ben Laurie (CHATS)*
12254 * Eliminate unused and incorrectly sized X.509 structure
12257 *Ben Laurie (CHATS)*
12259 * Eliminate unused and dangerous function knumber().
12261 *Ben Laurie (CHATS)*
12263 * Eliminate unused and dangerous structure, KSSL_ERR.
12265 *Ben Laurie (CHATS)*
12267 * Protect against overlong session ID context length in an encoded
12268 session object. Since these are local, this does not appear to be
12271 *Ben Laurie (CHATS)*
12273 * Change from security patch (see 0.9.6e below) that did not affect
12274 the 0.9.6 release series:
12276 Remote buffer overflow in SSL3 protocol - an attacker could
12277 supply an oversized master key in Kerberos-enabled versions.
12280 *Ben Laurie (CHATS)*
12282 * Change the SSL kerb5 codes to match RFC 2712.
12286 * Make -nameopt work fully for req and add -reqopt switch.
12288 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12290 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12292 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12294 * Make sure tests can be performed even if the corresponding algorithms
12295 have been removed entirely. This was also the last step to make
12296 OpenSSL compilable with DJGPP under all reasonable conditions.
12298 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12300 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12301 to allow version independent disabling of normally unselected ciphers,
12302 which may be activated as a side-effect of selecting a single cipher.
12304 (E.g., cipher list string "RSA" enables ciphersuites that are left
12305 out of "ALL" because they do not provide symmetric encryption.
12306 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12308 *Lutz Jaenicke, Bodo Moeller*
12310 * Add appropriate support for separate platform-dependent build
12311 directories. The recommended way to make a platform-dependent
12312 build directory is the following (tested on Linux), maybe with
12315 # Place yourself outside of the OpenSSL source tree. In
12316 # this example, the environment variable OPENSSL_SOURCE
12317 # is assumed to contain the absolute OpenSSL source directory.
12318 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12319 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12320 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12321 mkdir -p `dirname $F`
12322 ln -s $OPENSSL_SOURCE/$F $F
12325 To be absolutely sure not to disturb the source tree, a "make clean"
12326 is a good thing. If it isn't successful, don't worry about it,
12327 it probably means the source directory is very clean.
12331 * Make sure any ENGINE control commands make local copies of string
12332 pointers passed to them whenever necessary. Otherwise it is possible
12333 the caller may have overwritten (or deallocated) the original string
12334 data when a later ENGINE operation tries to use the stored values.
12336 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12338 * Improve diagnostics in file reading and command-line digests.
12340 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12342 * Add AES modes CFB and OFB to the object database. Correct an
12343 error in AES-CFB decryption.
12347 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12348 allows existing EVP_CIPHER_CTX structures to be reused after
12349 calling `EVP_*Final()`. This behaviour is used by encryption
12350 BIOs and some applications. This has the side effect that
12351 applications must explicitly clean up cipher contexts with
12352 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12356 * Check the values of dna and dnb in bn_mul_recursive before calling
12357 bn_mul_comba (a non zero value means the a or b arrays do not contain
12358 n2 elements) and fallback to bn_mul_normal if either is not zero.
12362 * Fix escaping of non-ASCII characters when using the -subj option
12363 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12367 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12368 form for "surname", serialNumber has no short form.
12369 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12370 therefore remove "mail" short name for "internet 7".
12371 The OID for unique identifiers in X509 certificates is
12372 x500UniqueIdentifier, not uniqueIdentifier.
12373 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12377 * Add an "init" command to the ENGINE config module and auto initialize
12378 ENGINEs. Without any "init" command the ENGINE will be initialized
12379 after all ctrl commands have been executed on it. If init=1 the
12380 ENGINE is initialized at that point (ctrls before that point are run
12381 on the uninitialized ENGINE and after on the initialized one). If
12382 init=0 then the ENGINE will not be initialized at all.
12386 * Fix the 'app_verify_callback' interface so that the user-defined
12387 argument is actually passed to the callback: In the
12388 SSL_CTX_set_cert_verify_callback() prototype, the callback
12389 declaration has been changed from
12392 int (*cb)(X509_STORE_CTX *,void *);
12393 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12394 i=s->ctx->app_verify_callback(&ctx)
12395 has been changed into
12396 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12398 To update applications using SSL_CTX_set_cert_verify_callback(),
12399 a dummy argument can be added to their callback functions.
12401 *D. K. Smetters <smetters@parc.xerox.com>*
12403 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12405 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12407 * Add and OPENSSL_LOAD_CONF define which will cause
12408 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12409 This allows older applications to transparently support certain
12410 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12411 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12412 load the config file and OPENSSL_add_all_algorithms_conf() which will
12413 always load it have also been added.
12417 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12418 Adjust NIDs and EVP layer.
12420 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12422 * Config modules support in openssl utility.
12424 Most commands now load modules from the config file,
12425 though in a few (such as version) this isn't done
12426 because it couldn't be used for anything.
12428 In the case of ca and req the config file used is
12429 the same as the utility itself: that is the -config
12430 command line option can be used to specify an
12435 * Move default behaviour from OPENSSL_config(). If appname is NULL
12436 use "openssl_conf" if filename is NULL use default openssl config file.
12440 * Add an argument to OPENSSL_config() to allow the use of an alternative
12441 config section name. Add a new flag to tolerate a missing config file
12442 and move code to CONF_modules_load_file().
12446 * Support for crypto accelerator cards from Accelerated Encryption
12447 Processing, www.aep.ie. (Use engine 'aep')
12448 The support was copied from 0.9.6c [engine] and adapted/corrected
12449 to work with the new engine framework.
12451 *AEP Inc. and Richard Levitte*
12453 * Support for SureWare crypto accelerator cards from Baltimore
12454 Technologies. (Use engine 'sureware')
12455 The support was copied from 0.9.6c [engine] and adapted
12456 to work with the new engine framework.
12460 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12461 make the newer ENGINE framework commands for the CHIL engine work.
12463 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12465 * Make it possible to produce shared libraries on ReliantUNIX.
12467 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12469 * Add the configuration target debug-linux-ppro.
12470 Make 'openssl rsa' use the general key loading routines
12471 implemented in `apps.c`, and make those routines able to
12472 handle the key format FORMAT_NETSCAPE and the variant
12475 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12477 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12479 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12481 * Add -keyform to rsautl, and document -engine.
12483 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12485 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12486 BIO_R_NO_SUCH_FILE error code rather than the generic
12487 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12491 * Add new functions
12492 ERR_peek_last_error
12493 ERR_peek_last_error_line
12494 ERR_peek_last_error_line_data.
12495 These are similar to
12497 ERR_peek_error_line
12498 ERR_peek_error_line_data,
12499 but report on the latest error recorded rather than the first one
12500 still in the error queue.
12502 *Ben Laurie, Bodo Moeller*
12504 * default_algorithms option in ENGINE config module. This allows things
12506 default_algorithms = ALL
12507 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12511 * Preliminary ENGINE config module.
12515 * New experimental application configuration code.
12519 * Change the AES code to follow the same name structure as all other
12520 symmetric ciphers, and behave the same way. Move everything to
12521 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12523 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12525 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12527 *Ben Laurie and Theo de Raadt*
12529 * Add option to output public keys in req command.
12531 *Massimiliano Pala madwolf@openca.org*
12533 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12534 (up to about 10% better than before for P-192 and P-224).
12538 * New functions/macros
12540 SSL_CTX_set_msg_callback(ctx, cb)
12541 SSL_CTX_set_msg_callback_arg(ctx, arg)
12542 SSL_set_msg_callback(ssl, cb)
12543 SSL_set_msg_callback_arg(ssl, arg)
12545 to request calling a callback function
12547 void cb(int write_p, int version, int content_type,
12548 const void *buf, size_t len, SSL *ssl, void *arg)
12550 whenever a protocol message has been completely received
12551 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12552 protocol version according to which the SSL library interprets
12553 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12554 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12555 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12556 specification (change_cipher_spec(20), alert(21), handshake(22)).
12557 'buf' and 'len' point to the actual message, 'ssl' to the
12558 SSL object, and 'arg' is the application-defined value set by
12559 SSL[_CTX]_set_msg_callback_arg().
12561 'openssl s_client' and 'openssl s_server' have new '-msg' options
12562 to enable a callback that displays all protocol messages.
12566 * Change the shared library support so shared libraries are built as
12567 soon as the corresponding static library is finished, and thereby get
12568 openssl and the test programs linked against the shared library.
12569 This still only happens when the keyword "shard" has been given to
12570 the configuration scripts.
12572 NOTE: shared library support is still an experimental thing, and
12573 backward binary compatibility is still not guaranteed.
12575 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12577 * Add support for Subject Information Access extension.
12579 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12581 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12582 additional bytes when new memory had to be allocated, not just
12583 when reusing an existing buffer.
12587 * New command line and configuration option 'utf8' for the req command.
12588 This allows field values to be specified as UTF8 strings.
12592 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12593 runs for the former and machine-readable output for the latter.
12597 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12598 of the e-mail address in the DN (i.e., it will go into a certificate
12599 extension only). The new configuration file option 'email_in_dn = no'
12600 has the same effect.
12602 *Massimiliano Pala madwolf@openca.org*
12604 * Change all functions with names starting with `des_` to be starting
12605 with `DES_` instead. Add wrappers that are compatible with libdes,
12606 but are named `_ossl_old_des_*`. Finally, add macros that map the
12607 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12608 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12609 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12612 Since we provide two compatibility mappings, the user needs to
12613 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12614 compatibility is desired. The default (i.e., when that macro
12615 isn't defined) is OpenSSL 0.9.6c compatibility.
12617 There are also macros that enable and disable the support of old
12618 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12619 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12620 are defined, the default will apply: to support the old des routines.
12622 In either case, one must include openssl/des.h to get the correct
12623 definitions. Do not try to just include openssl/des_old.h, that
12626 NOTE: This is a major break of an old API into a new one. Software
12627 authors are encouraged to switch to the `DES_` style functions. Some
12628 time in the future, des_old.h and the libdes compatibility functions
12629 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12630 default), and then completely removed.
12634 * Test for certificates which contain unsupported critical extensions.
12635 If such a certificate is found during a verify operation it is
12636 rejected by default: this behaviour can be overridden by either
12637 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12638 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12639 X509_supported_extension() has also been added which returns 1 if a
12640 particular extension is supported.
12644 * Modify the behaviour of EVP cipher functions in similar way to digests
12645 to retain compatibility with existing code.
12649 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12650 compatibility with existing code. In particular the 'ctx' parameter does
12651 not have to be to be initialized before the call to EVP_DigestInit() and
12652 it is tidied up after a call to EVP_DigestFinal(). New function
12653 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12654 EVP_MD_CTX_copy() changed to not require the destination to be
12655 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12656 requires the destination to be valid.
12658 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12659 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12663 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12664 so that complete 'Handshake' protocol structures are kept in memory
12665 instead of overwriting 'msg_type' and 'length' with 'body' data.
12669 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12671 *Massimo Santin via Richard Levitte*
12673 * Major restructuring to the underlying ENGINE code. This includes
12674 reduction of linker bloat, separation of pure "ENGINE" manipulation
12675 (initialisation, etc) from functionality dealing with implementations
12676 of specific crypto interfaces. This change also introduces integrated
12677 support for symmetric ciphers and digest implementations - so ENGINEs
12678 can now accelerate these by providing EVP_CIPHER and EVP_MD
12679 implementations of their own. This is detailed in
12680 [crypto/engine/README.md](crypto/engine/README.md)
12681 as it couldn't be adequately described here. However, there are a few
12682 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12683 were changed in the original introduction of ENGINE code have now
12684 reverted back - the hooking from this code to ENGINE is now a good
12685 deal more passive and at run-time, operations deal directly with
12686 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12687 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12688 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12689 they were not being used by the framework as there is no concept of a
12690 BIGNUM_METHOD and they could not be generalised to the new
12691 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12692 ENGINE_cpy() has been removed as it cannot be consistently defined in
12697 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12701 * Change mkdef.pl to sort symbols that get the same entry number,
12702 and make sure the automatically generated functions `ERR_load_*`
12703 become part of libeay.num as well.
12707 * New function SSL_renegotiate_pending(). This returns true once
12708 renegotiation has been requested (either SSL_renegotiate() call
12709 or HelloRequest/ClientHello received from the peer) and becomes
12710 false once a handshake has been completed.
12711 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12712 sends a HelloRequest, but does not ensure that a handshake takes
12713 place. SSL_renegotiate_pending() is useful for checking if the
12714 client has followed the request.)
12718 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12719 By default, clients may request session resumption even during
12720 renegotiation (if session ID contexts permit); with this option,
12721 session resumption is possible only in the first handshake.
12723 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12724 more bits available for options that should not be part of
12725 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12729 * Add some demos for certificate and certificate request creation.
12733 * Make maximum certificate chain size accepted from the peer application
12734 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12735 "Douglas E. Engert" <deengert@anl.gov>.
12739 * Add support for shared libraries for Unixware-7
12740 (Boyd Lynn Gerber <gerberb@zenez.com>).
12744 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12745 be done prior to destruction. Use this to unload error strings from
12746 ENGINEs that load their own error strings. NB: This adds two new API
12747 functions to "get" and "set" this destroy handler in an ENGINE.
12751 * Alter all existing ENGINE implementations (except "openssl" and
12752 "openbsd") to dynamically instantiate their own error strings. This
12753 makes them more flexible to be built both as statically-linked ENGINEs
12754 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12755 Also, add stub code to each that makes building them as self-contained
12756 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12760 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12761 implementations into applications that are completely implemented in
12762 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12763 commands that can be used to configure what shared-library to load and
12764 to control aspects of the way it is handled. Also, made an update to
12765 the [README-Engine.md](README-Engine.md) file
12766 that brings its information up-to-date and
12767 provides some information and instructions on the "dynamic" ENGINE
12768 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12772 * Make it possible to unload ranges of ERR strings with a new
12773 "ERR_unload_strings" function.
12777 * Add a copy() function to EVP_MD.
12781 * Make EVP_MD routines take a context pointer instead of just the
12782 md_data void pointer.
12786 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12787 that the digest can only process a single chunk of data
12788 (typically because it is provided by a piece of
12789 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12790 is only going to provide a single chunk of data, and hence the
12791 framework needn't accumulate the data for oneshot drivers.
12795 * As with "ERR", make it possible to replace the underlying "ex_data"
12796 functions. This change also alters the storage and management of global
12797 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12798 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12799 index counters. The API functions that use this state have been changed
12800 to take a "class_index" rather than pointers to the class's local STACK
12801 and counter, and there is now an API function to dynamically create new
12802 classes. This centralisation allows us to (a) plug a lot of the
12803 thread-safety problems that existed, and (b) makes it possible to clean
12804 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12805 such data would previously have always leaked in application code and
12806 workarounds were in place to make the memory debugging turn a blind eye
12807 to it. Application code that doesn't use this new function will still
12808 leak as before, but their memory debugging output will announce it now
12809 rather than letting it slide.
12811 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12812 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12813 has a return value to indicate success or failure.
12817 * Make it possible to replace the underlying "ERR" functions such that the
12818 global state (2 LHASH tables and 2 locks) is only used by the "default"
12819 implementation. This change also adds two functions to "get" and "set"
12820 the implementation prior to it being automatically set the first time
12821 any other ERR function takes place. Ie. an application can call "get",
12822 pass the return value to a module it has just loaded, and that module
12823 can call its own "set" function using that value. This means the
12824 module's "ERR" operations will use (and modify) the error state in the
12825 application and not in its own statically linked copy of OpenSSL code.
12829 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12830 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12831 the operation, and provides a more encapsulated way for external code
12832 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12833 to use these functions rather than manually incrementing the counts.
12835 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12839 * Add EVP test program.
12843 * Add symmetric cipher support to ENGINE. Expect the API to change!
12847 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12848 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12849 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12850 These allow a CRL to be built without having to access X509_CRL fields
12851 directly. Modify 'ca' application to use new functions.
12855 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12856 bug workarounds. Rollback attack detection is a security feature.
12857 The problem will only arise on OpenSSL servers when TLSv1 is not
12858 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12859 Software authors not wanting to support TLSv1 will have special reasons
12860 for their choice and can explicitly enable this option.
12862 *Bodo Moeller, Lutz Jaenicke*
12864 * Rationalise EVP so it can be extended: don't include a union of
12865 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12866 (similar to those existing for EVP_CIPHER_CTX).
12871 EVP_MD_CTX_init(&md); /* new function call */
12872 EVP_DigestInit(&md, EVP_sha1());
12873 EVP_DigestUpdate(&md, in, len);
12874 EVP_DigestFinal(&md, out, NULL);
12875 EVP_MD_CTX_cleanup(&md); /* new function call */
12879 * Make DES key schedule conform to the usual scheme, as well as
12880 correcting its structure. This means that calls to DES functions
12881 now have to pass a pointer to a des_key_schedule instead of a
12882 plain des_key_schedule (which was actually always a pointer
12885 des_key_schedule ks;
12887 des_set_key_checked(..., &ks);
12888 des_ncbc_encrypt(..., &ks, ...);
12890 (Note that a later change renames 'des_...' into 'DES_...'.)
12894 * Initial reduction of linker bloat: the use of some functions, such as
12895 PEM causes large amounts of unused functions to be linked in due to
12896 poor organisation. For example pem_all.c contains every PEM function
12897 which has a knock on effect of linking in large amounts of (unused)
12898 ASN1 code. Grouping together similar functions and splitting unrelated
12899 functions prevents this.
12903 * Cleanup of EVP macros.
12907 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12908 correct `_ecb suffix`.
12912 * Add initial OCSP responder support to ocsp application. The
12913 revocation information is handled using the text based index
12914 use by the ca application. The responder can either handle
12915 requests generated internally, supplied in files (for example
12916 via a CGI script) or using an internal minimal server.
12920 * Add configuration choices to get zlib compression for TLS.
12924 * Changes to Kerberos SSL for RFC 2712 compliance:
12925 1. Implemented real KerberosWrapper, instead of just using
12926 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12927 2. Implemented optional authenticator field of KerberosWrapper.
12929 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12930 and authenticator structs; see crypto/krb5/.
12932 Generalized Kerberos calls to support multiple Kerberos libraries.
12933 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12934 via Richard Levitte*
12936 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12937 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12938 values for each of the key sizes rather than having just
12939 parameters (and 'speed' generating keys each time).
12943 * Speed up EVP routines.
12946 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12947 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12948 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12949 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12951 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12952 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12953 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12956 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12958 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12962 * Added the OS2-EMX target.
12964 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12966 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12967 New functions to support `NCONF` routines in extension code.
12968 New function `CONF_set_nconf()`
12969 to allow functions which take an `NCONF` to also handle the old `LHASH`
12970 structure: this means that the old `CONF` compatible routines can be
12971 retained (in particular w.rt. extensions) without having to duplicate the
12972 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12976 * Enhance the general user interface with mechanisms for inner control
12977 and with possibilities to have yes/no kind of prompts.
12981 * Change all calls to low-level digest routines in the library and
12982 applications to use EVP. Add missing calls to HMAC_cleanup() and
12983 don't assume HMAC_CTX can be copied using memcpy().
12985 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12987 * Add the possibility to control engines through control names but with
12988 arbitrary arguments instead of just a string.
12989 Change the key loaders to take a UI_METHOD instead of a callback
12990 function pointer. NOTE: this breaks binary compatibility with earlier
12991 versions of OpenSSL [engine].
12992 Adapt the nCipher code for these new conditions and add a card insertion
12997 * Enhance the general user interface with mechanisms to better support
12998 dialog box interfaces, application-defined prompts, the possibility
12999 to use defaults (for example default passwords from somewhere else)
13000 and interrupts/cancellations.
13004 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13005 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13009 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13010 tidy up some unnecessarily weird code in 'sk_new()').
13012 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13014 * Change the key loading routines for ENGINEs to use the same kind
13015 callback (pem_password_cb) as all other routines that need this
13020 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13021 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13022 than this minimum value is recommended.
13026 * New random seeder for OpenVMS, using the system process statistics
13027 that are easily reachable.
13031 * Windows apparently can't transparently handle global
13032 variables defined in DLLs. Initialisations such as:
13034 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13036 won't compile. This is used by the any applications that need to
13037 declare their own ASN1 modules. This was fixed by adding the option
13038 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13039 needed for static libraries under Win32.
13043 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13044 setting of purpose and trust fields. New X509_STORE trust and
13045 purpose functions and tidy up setting in other SSL functions.
13049 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13050 structure. These are inherited by X509_STORE_CTX when it is
13051 initialised. This allows various defaults to be set in the
13052 X509_STORE structure (such as flags for CRL checking and custom
13053 purpose or trust settings) for functions which only use X509_STORE_CTX
13054 internally such as S/MIME.
13056 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13057 trust settings if they are not set in X509_STORE. This allows X509_STORE
13058 purposes and trust (in S/MIME for example) to override any set by default.
13060 Add command line options for CRL checking to smime, s_client and s_server
13065 * Initial CRL based revocation checking. If the CRL checking flag(s)
13066 are set then the CRL is looked up in the X509_STORE structure and
13067 its validity and signature checked, then if the certificate is found
13068 in the CRL the verify fails with a revoked error.
13070 Various new CRL related callbacks added to X509_STORE_CTX structure.
13072 Command line options added to 'verify' application to support this.
13074 This needs some additional work, such as being able to handle multiple
13075 CRLs with different times, extension based lookup (rather than just
13076 by subject name) and ultimately more complete V2 CRL extension
13081 * Add a general user interface API (crypto/ui/). This is designed
13082 to replace things like des_read_password and friends (backward
13083 compatibility functions using this new API are provided).
13084 The purpose is to remove prompting functions from the DES code
13085 section as well as provide for prompting through dialog boxes in
13086 a window system and the like.
13090 * Add "ex_data" support to ENGINE so implementations can add state at a
13091 per-structure level rather than having to store it globally.
13095 * Make it possible for ENGINE structures to be copied when retrieved by
13096 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13097 This causes the "original" ENGINE structure to act like a template,
13098 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13099 operational state can be localised to each ENGINE structure, despite the
13100 fact they all share the same "methods". New ENGINE structures returned in
13101 this case have no functional references and the return value is the single
13102 structural reference. This matches the single structural reference returned
13103 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13108 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13109 needs to match any other type at all we need to manually clear the
13114 * Changes to the "openssl engine" utility to include;
13115 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13116 about an ENGINE's available control commands.
13117 - executing control commands from command line arguments using the
13118 '-pre' and '-post' switches. '-post' is only used if '-t' is
13119 specified and the ENGINE is successfully initialised. The syntax for
13120 the individual commands are colon-separated, for example;
13121 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13125 * New dynamic control command support for ENGINEs. ENGINEs can now
13126 declare their own commands (numbers), names (strings), descriptions,
13127 and input types for run-time discovery by calling applications. A
13128 subset of these commands are implicitly classed as "executable"
13129 depending on their input type, and only these can be invoked through
13130 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13131 can be based on user input, config files, etc). The distinction is
13132 that "executable" commands cannot return anything other than a boolean
13133 result and can only support numeric or string input, whereas some
13134 discoverable commands may only be for direct use through
13135 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13136 pointers, or other custom uses. The "executable" commands are to
13137 support parameterisations of ENGINE behaviour that can be
13138 unambiguously defined by ENGINEs and used consistently across any
13139 OpenSSL-based application. Commands have been added to all the
13140 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13141 control over shared-library paths without source code alterations.
13145 * Changed all ENGINE implementations to dynamically allocate their
13146 ENGINEs rather than declaring them statically. Apart from this being
13147 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13148 this also allows the implementations to compile without using the
13149 internal engine_int.h header.
13153 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13154 'const' value. Any code that should be able to modify a RAND_METHOD
13155 should already have non-const pointers to it (ie. they should only
13156 modify their own ones).
13160 * Made a variety of little tweaks to the ENGINE code.
13161 - "atalla" and "ubsec" string definitions were moved from header files
13162 to C code. "nuron" string definitions were placed in variables
13163 rather than hard-coded - allowing parameterisation of these values
13164 later on via ctrl() commands.
13165 - Removed unused "#if 0"'d code.
13166 - Fixed engine list iteration code so it uses ENGINE_free() to release
13167 structural references.
13168 - Constified the RAND_METHOD element of ENGINE structures.
13169 - Constified various get/set functions as appropriate and added
13170 missing functions (including a catch-all ENGINE_cpy that duplicates
13171 all ENGINE values onto a new ENGINE except reference counts/state).
13172 - Removed NULL parameter checks in get/set functions. Setting a method
13173 or function to NULL is a way of cancelling out a previously set
13174 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13175 and doesn't justify the extra error symbols and code.
13176 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13177 flags from engine_int.h to engine.h.
13178 - Changed prototypes for ENGINE handler functions (init(), finish(),
13179 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13183 * Implement binary inversion algorithm for BN_mod_inverse in addition
13184 to the algorithm using long division. The binary algorithm can be
13185 used only if the modulus is odd. On 32-bit systems, it is faster
13186 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13187 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13188 up to 450 bits. In 64-bit environments, the binary algorithm
13189 appears to be advantageous for much longer moduli; here we use it
13190 for moduli up to 2048 bits.
13194 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13195 could not support the combine flag in choice fields.
13199 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13200 extensions from a certificate request to the certificate.
13204 * Allow multiple 'certopt' and 'nameopt' options to be separated
13205 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13206 file: this allows the display of the certificate about to be
13207 signed to be customised, to allow certain fields to be included
13208 or excluded and extension details. The old system didn't display
13209 multicharacter strings properly, omitted fields not in the policy
13210 and couldn't display additional details such as extensions.
13214 * Function EC_POINTs_mul for multiple scalar multiplication
13215 of an arbitrary number of elliptic curve points
13216 \sum scalars[i]*points[i],
13217 optionally including the generator defined for the EC_GROUP:
13218 scalar*generator + \sum scalars[i]*points[i].
13220 EC_POINT_mul is a simple wrapper function for the typical case
13221 that the point list has just one item (besides the optional
13226 * First EC_METHODs for curves over GF(p):
13228 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13229 operations and provides various method functions that can also
13230 operate with faster implementations of modular arithmetic.
13232 EC_GFp_mont_method() reuses most functions that are part of
13233 EC_GFp_simple_method, but uses Montgomery arithmetic.
13235 *Bodo Moeller; point addition and point doubling
13236 implementation directly derived from source code provided by
13237 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13239 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13240 crypto/ec/ec_lib.c):
13242 Curves are EC_GROUP objects (with an optional group generator)
13243 based on EC_METHODs that are built into the library.
13245 Points are EC_POINT objects based on EC_GROUP objects.
13247 Most of the framework would be able to handle curves over arbitrary
13248 finite fields, but as there are no obvious types for fields other
13249 than GF(p), some functions are limited to that for now.
13253 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13254 that the file contains a complete HTTP response.
13258 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13259 change the def and num file printf format specifier from "%-40sXXX"
13260 to "%-39s XXX". The latter will always guarantee a space after the
13261 field while the former will cause them to run together if the field
13262 is 40 of more characters long.
13266 * Constify the cipher and digest 'method' functions and structures
13267 and modify related functions to take constant EVP_MD and EVP_CIPHER
13272 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13273 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13277 * Modify `EVP_Digest*()` routines so they now return values. Although the
13278 internal software routines can never fail additional hardware versions
13283 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13285 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13286 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13289 ERR_R_NESTED_ASN1_ERROR
13291 ERR_R_MISSING_ASN1_EOS
13292 were 4 .. 9, conflicting with
13293 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13295 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13296 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13298 Add new error code 'ERR_R_INTERNAL_ERROR'.
13302 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13307 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13308 sets the subject name for a new request or supersedes the
13309 subject name in a given request. Formats that can be parsed are
13310 'CN=Some Name, OU=myOU, C=IT'
13312 'CN=Some Name/OU=myOU/C=IT'.
13314 Add options '-batch' and '-verbose' to 'openssl req'.
13316 *Massimiliano Pala <madwolf@hackmasters.net>*
13318 * Introduce the possibility to access global variables through
13319 functions on platform were that's the best way to handle exporting
13320 global variables in shared libraries. To enable this functionality,
13321 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13322 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13323 is normally done by Configure or something similar).
13325 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13326 in the source file (foo.c) like this:
13328 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13329 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13331 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13332 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13334 OPENSSL_DECLARE_GLOBAL(int,foo);
13335 #define foo OPENSSL_GLOBAL_REF(foo)
13336 OPENSSL_DECLARE_GLOBAL(double,bar);
13337 #define bar OPENSSL_GLOBAL_REF(bar)
13339 The #defines are very important, and therefore so is including the
13340 header file everywhere where the defined globals are used.
13342 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13343 of ASN.1 items, but that structure is a bit different.
13345 The largest change is in util/mkdef.pl which has been enhanced with
13346 better and easier to understand logic to choose which symbols should
13347 go into the Windows .def files as well as a number of fixes and code
13348 cleanup (among others, algorithm keywords are now sorted
13349 lexicographically to avoid constant rewrites).
13353 * In BN_div() keep a copy of the sign of 'num' before writing the
13354 result to 'rm' because if rm==num the value will be overwritten
13355 and produce the wrong result if 'num' is negative: this caused
13356 problems with BN_mod() and BN_nnmod().
13360 * Function OCSP_request_verify(). This checks the signature on an
13361 OCSP request and verifies the signer certificate. The signer
13362 certificate is just checked for a generic purpose and OCSP request
13367 * Add OCSP_check_validity() function to check the validity of OCSP
13368 responses. OCSP responses are prepared in real time and may only
13369 be a few seconds old. Simply checking that the current time lies
13370 between thisUpdate and nextUpdate max reject otherwise valid responses
13371 caused by either OCSP responder or client clock inaccuracy. Instead
13372 we allow thisUpdate and nextUpdate to fall within a certain period of
13373 the current time. The age of the response can also optionally be
13374 checked. Two new options -validity_period and -status_age added to
13379 * If signature or public key algorithm is unrecognized print out its
13380 OID rather that just UNKNOWN.
13384 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13385 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13386 ID to be generated from the issuer certificate alone which can then be
13387 passed to OCSP_id_issuer_cmp().
13391 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13392 ASN1 modules to export functions returning ASN1_ITEM pointers
13393 instead of the ASN1_ITEM structures themselves. This adds several
13394 new macros which allow the underlying ASN1 function/structure to
13395 be accessed transparently. As a result code should not use ASN1_ITEM
13396 references directly (such as &X509_it) but instead use the relevant
13397 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13398 use of the new ASN1 code on platforms where exporting structures
13399 is problematical (for example in shared libraries) but exporting
13400 functions returning pointers to structures is not.
13404 * Add support for overriding the generation of SSL/TLS session IDs.
13405 These callbacks can be registered either in an SSL_CTX or per SSL.
13406 The purpose of this is to allow applications to control, if they wish,
13407 the arbitrary values chosen for use as session IDs, particularly as it
13408 can be useful for session caching in multiple-server environments. A
13409 command-line switch for testing this (and any client code that wishes
13410 to use such a feature) has been added to "s_server".
13412 *Geoff Thorpe, Lutz Jaenicke*
13414 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13415 of the form `#if defined(...) || defined(...) || ...` and
13416 `#if !defined(...) && !defined(...) && ...`. This also avoids
13417 the growing number of special cases it was previously handling.
13421 * Make all configuration macros available for application by making
13422 sure they are available in opensslconf.h, by giving them names starting
13423 with `OPENSSL_` to avoid conflicts with other packages and by making
13424 sure e_os2.h will cover all platform-specific cases together with
13426 Additionally, it is now possible to define configuration/platform-
13427 specific names (called "system identities"). In the C code, these
13428 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13429 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13430 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13435 * New option -set_serial to 'req' and 'x509' this allows the serial
13436 number to use to be specified on the command line. Previously self
13437 signed certificates were hard coded with serial number 0 and the
13438 CA options of 'x509' had to use a serial number in a file which was
13443 * New options to 'ca' utility to support V2 CRL entry extensions.
13444 Currently CRL reason, invalidity date and hold instruction are
13445 supported. Add new CRL extensions to V3 code and some new objects.
13449 * New function EVP_CIPHER_CTX_set_padding() this is used to
13450 disable standard block padding (aka PKCS#5 padding) in the EVP
13451 API, which was previously mandatory. This means that the data is
13452 not padded in any way and so the total length much be a multiple
13453 of the block size, otherwise an error occurs.
13457 * Initial (incomplete) OCSP SSL support.
13461 * New function OCSP_parse_url(). This splits up a URL into its host,
13462 port and path components: primarily to parse OCSP URLs. New -url
13463 option to ocsp utility.
13467 * New nonce behavior. The return value of OCSP_check_nonce() now
13468 reflects the various checks performed. Applications can decide
13469 whether to tolerate certain situations such as an absent nonce
13470 in a response when one was present in a request: the ocsp application
13471 just prints out a warning. New function OCSP_add1_basic_nonce()
13472 this is to allow responders to include a nonce in a response even if
13473 the request is nonce-less.
13477 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13478 skipped when using openssl x509 multiple times on a single input file,
13479 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13483 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13484 set string type: to handle setting ASN1_TIME structures. Fix ca
13485 utility to correctly initialize revocation date of CRLs.
13489 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13490 the clients preferred ciphersuites and rather use its own preferences.
13491 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13492 Internet Explorer by ensuring unchanged hash method during stepup.
13493 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13497 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13498 to aes and add a new 'exist' option to print out symbols that don't
13503 * Additional options to ocsp utility to allow flags to be set and
13504 additional certificates supplied.
13508 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13509 OCSP client a number of certificate to only verify the response
13514 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13515 handle the new API. Currently only ECB, CBC modes supported. Add new
13518 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13519 Encryption Standard (AES) Ciphersuites for Transport Layer
13520 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13521 not enabled by default and were not part of the "ALL" ciphersuite
13522 alias because they were not yet official; they could be
13523 explicitly requested by specifying the "AESdraft" ciphersuite
13524 group alias. In the final release of OpenSSL 0.9.7, the group
13525 alias is called "AES" and is part of "ALL".)
13527 *Ben Laurie, Steve Henson, Bodo Moeller*
13529 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13530 request to response.
13534 * Functions for OCSP responders. OCSP_request_onereq_count(),
13535 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13536 extract information from a certificate request. OCSP_response_create()
13537 creates a response and optionally adds a basic response structure.
13538 OCSP_basic_add1_status() adds a complete single response to a basic
13539 response and returns the OCSP_SINGLERESP structure just added (to allow
13540 extensions to be included for example). OCSP_basic_add1_cert() adds a
13541 certificate to a basic response and OCSP_basic_sign() signs a basic
13542 response with various flags. New helper functions ASN1_TIME_check()
13543 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13544 (converts ASN1_TIME to GeneralizedTime).
13548 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13549 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13550 structure from a certificate. X509_pubkey_digest() digests the public_key
13551 contents: this is used in various key identifiers.
13555 * Make sk_sort() tolerate a NULL argument.
13557 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13559 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13560 passed by the function are trusted implicitly. If any of them signed the
13561 response then it is assumed to be valid and is not verified.
13565 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13566 to data. This was previously part of the PKCS7 ASN1 code. This
13567 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13568 *Steve Henson, reported by Kenneth R. Robinette
13569 <support@securenetterm.com>*
13571 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13572 routines: without these tracing memory leaks is very painful.
13573 Fix leaks in PKCS12 and PKCS7 routines.
13577 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13578 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13579 effectively meant GeneralizedTime would never be used. Now it
13580 is initialised to -1 but X509_time_adj() now has to check the value
13581 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13582 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13583 *Steve Henson, reported by Kenneth R. Robinette
13584 <support@securenetterm.com>*
13586 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13587 result in a zero length in the ASN1_INTEGER structure which was
13588 not consistent with the structure when d2i_ASN1_INTEGER() was used
13589 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13590 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13591 where it did not print out a minus for negative ASN1_INTEGER.
13595 * Add summary printout to ocsp utility. The various functions which
13596 convert status values to strings have been renamed to:
13597 OCSP_response_status_str(), OCSP_cert_status_str() and
13598 OCSP_crl_reason_str() and are no longer static. New options
13599 to verify nonce values and to disable verification. OCSP response
13600 printout format cleaned up.
13604 * Add additional OCSP certificate checks. These are those specified
13605 in RFC2560. This consists of two separate checks: the CA of the
13606 certificate being checked must either be the OCSP signer certificate
13607 or the issuer of the OCSP signer certificate. In the latter case the
13608 OCSP signer certificate must contain the OCSP signing extended key
13609 usage. This check is performed by attempting to match the OCSP
13610 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13611 in the OCSP_CERTID structures of the response.
13615 * Initial OCSP certificate verification added to OCSP_basic_verify()
13616 and related routines. This uses the standard OpenSSL certificate
13617 verify routines to perform initial checks (just CA validity) and
13618 to obtain the certificate chain. Then additional checks will be
13619 performed on the chain. Currently the root CA is checked to see
13620 if it is explicitly trusted for OCSP signing. This is used to set
13621 a root CA as a global signing root: that is any certificate that
13622 chains to that CA is an acceptable OCSP signing certificate.
13626 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13627 extensions from a separate configuration file.
13628 As when reading extensions from the main configuration file,
13629 the '-extensions ...' option may be used for specifying the
13632 *Massimiliano Pala <madwolf@comune.modena.it>*
13634 * New OCSP utility. Allows OCSP requests to be generated or
13635 read. The request can be sent to a responder and the output
13636 parsed, outputted or printed in text form. Not complete yet:
13637 still needs to check the OCSP response validity.
13641 * New subcommands for 'openssl ca':
13642 `openssl ca -status <serial>` prints the status of the cert with
13643 the given serial number (according to the index file).
13644 `openssl ca -updatedb` updates the expiry status of certificates
13647 *Massimiliano Pala <madwolf@comune.modena.it>*
13649 * New '-newreq-nodes' command option to CA.pl. This is like
13650 '-newreq', but calls 'openssl req' with the '-nodes' option
13651 so that the resulting key is not encrypted.
13653 *Damien Miller <djm@mindrot.org>*
13655 * New configuration for the GNU Hurd.
13657 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13659 * Initial code to implement OCSP basic response verify. This
13660 is currently incomplete. Currently just finds the signer's
13661 certificate and verifies the signature on the response.
13665 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13666 value of OPENSSLDIR. This is available via the new '-d' option
13667 to 'openssl version', and is also included in 'openssl version -a'.
13671 * Allowing defining memory allocation callbacks that will be given
13672 file name and line number information in additional arguments
13673 (a `const char*` and an int). The basic functionality remains, as
13674 well as the original possibility to just replace malloc(),
13675 realloc() and free() by functions that do not know about these
13676 additional arguments. To register and find out the current
13677 settings for extended allocation functions, the following
13678 functions are provided:
13680 CRYPTO_set_mem_ex_functions
13681 CRYPTO_set_locked_mem_ex_functions
13682 CRYPTO_get_mem_ex_functions
13683 CRYPTO_get_locked_mem_ex_functions
13685 These work the same way as CRYPTO_set_mem_functions and friends.
13686 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13687 extended allocation function is enabled.
13688 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13689 a conventional allocation function is enabled.
13691 *Richard Levitte, Bodo Moeller*
13693 * Finish off removing the remaining LHASH function pointer casts.
13694 There should no longer be any prototype-casting required when using
13695 the LHASH abstraction, and any casts that remain are "bugs". See
13696 the callback types and macros at the head of lhash.h for details
13697 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13701 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13702 If /dev/[u]random devices are not available or do not return enough
13703 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13705 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13706 /etc/entropy will be queried once each in this sequence, querying stops
13707 when enough entropy was collected without querying more sockets.
13711 * Change the Unix RAND_poll() variant to be able to poll several
13712 random devices, as specified by DEVRANDOM, until a sufficient amount
13713 of data has been collected. We spend at most 10 ms on each file
13714 (select timeout) and read in non-blocking mode. DEVRANDOM now
13715 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13716 (previously it was just the string "/dev/urandom"), so on typical
13717 platforms the 10 ms delay will never occur.
13718 Also separate out the Unix variant to its own file, rand_unix.c.
13719 For VMS, there's a currently-empty rand_vms.c.
13723 * Move OCSP client related routines to ocsp_cl.c. These
13724 provide utility functions which an application needing
13725 to issue a request to an OCSP responder and analyse the
13726 response will typically need: as opposed to those which an
13727 OCSP responder itself would need which will be added later.
13729 OCSP_request_sign() signs an OCSP request with an API similar
13730 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13731 response. OCSP_response_get1_basic() extracts basic response
13732 from response. OCSP_resp_find_status(): finds and extracts status
13733 information from an OCSP_CERTID structure (which will be created
13734 when the request structure is built). These are built from lower
13735 level functions which work on OCSP_SINGLERESP structures but
13736 won't normally be used unless the application wishes to examine
13737 extensions in the OCSP response for example.
13739 Replace nonce routines with a pair of functions.
13740 OCSP_request_add1_nonce() adds a nonce value and optionally
13741 generates a random value. OCSP_check_nonce() checks the
13742 validity of the nonce in an OCSP response.
13746 * Change function OCSP_request_add() to OCSP_request_add0_id().
13747 This doesn't copy the supplied OCSP_CERTID and avoids the
13748 need to free up the newly created id. Change return type
13749 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13750 This can then be used to add extensions to the request.
13751 Deleted OCSP_request_new(), since most of its functionality
13752 is now in OCSP_REQUEST_new() (and the case insensitive name
13753 clash) apart from the ability to set the request name which
13754 will be added elsewhere.
13758 * Update OCSP API. Remove obsolete extensions argument from
13759 various functions. Extensions are now handled using the new
13760 OCSP extension code. New simple OCSP HTTP function which
13761 can be used to send requests and parse the response.
13765 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13766 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13767 uses the special reorder version of SET OF to sort the attributes
13768 and reorder them to match the encoded order. This resolves a long
13769 standing problem: a verify on a PKCS7 structure just after signing
13770 it used to fail because the attribute order did not match the
13771 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13772 it uses the received order. This is necessary to tolerate some broken
13773 software that does not order SET OF. This is handled by encoding
13774 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13775 to produce the required SET OF.
13779 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13780 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13781 files to get correct declarations of the ASN.1 item variables.
13785 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13786 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13787 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13788 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13789 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13790 ASN1_ITEM and no wrapper functions.
13794 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13795 replace the old function pointer based I/O routines. Change most of
13796 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13800 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13801 lines, recognize more "algorithms" that can be deselected, and make
13802 it complain about algorithm deselection that isn't recognised.
13806 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13807 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13808 to use new functions. Add NO_ASN1_OLD which can be set to remove
13809 some old style ASN1 functions: this can be used to determine if old
13810 code will still work when these eventually go away.
13814 * New extension functions for OCSP structures, these follow the
13815 same conventions as certificates and CRLs.
13819 * New function X509V3_add1_i2d(). This automatically encodes and
13820 adds an extension. Its behaviour can be customised with various
13821 flags to append, replace or delete. Various wrappers added for
13822 certificates and CRLs.
13826 * Fix to avoid calling the underlying ASN1 print routine when
13827 an extension cannot be parsed. Correct a typo in the
13828 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13832 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13833 entries for variables.
13837 * Add functionality to `apps/openssl.c` for detecting locking
13838 problems: As the program is single-threaded, all we have
13839 to do is register a locking callback using an array for
13840 storing which locks are currently held by the program.
13844 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13845 SSL_get_ex_data_X509_STORE_idx(), which is used in
13846 ssl_verify_cert_chain() and thus can be called at any time
13847 during TLS/SSL handshakes so that thread-safety is essential.
13848 Unfortunately, the ex_data design is not at all suited
13849 for multi-threaded use, so it probably should be abolished.
13853 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13855 *Broadcom, tweaked and integrated by Geoff Thorpe*
13857 * Move common extension printing code to new function
13858 X509V3_print_extensions(). Reorganise OCSP print routines and
13859 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13863 * New function X509_signature_print() to remove duplication in some
13868 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13869 set (this was treated exactly the same as SET OF previously). This
13870 is used to reorder the STACK representing the structure to match the
13871 encoding. This will be used to get round a problem where a PKCS7
13872 structure which was signed could not be verified because the STACK
13873 order did not reflect the encoded order.
13877 * Reimplement the OCSP ASN1 module using the new code.
13881 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13882 for its ASN1 operations. The old style function pointers still exist
13883 for now but they will eventually go away.
13887 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13888 completely replaces the old ASN1 functionality with a table driven
13889 encoder and decoder which interprets an ASN1_ITEM structure describing
13890 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13891 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13892 has also been converted to the new form.
13896 * Change BN_mod_exp_recp so that negative moduli are tolerated
13897 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13898 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13899 for negative moduli.
13903 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13904 of not touching the result's sign bit.
13908 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13913 * Changed the LHASH code to use prototypes for callbacks, and created
13914 macros to declare and implement thin (optionally static) functions
13915 that provide type-safety and avoid function pointer casting for the
13916 type-specific callbacks.
13920 * Added Kerberos Cipher Suites to be used with TLS, as written in
13922 *Veers Staats <staatsvr@asc.hpc.mil>,
13923 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13925 * Reformat the FAQ so the different questions and answers can be divided
13926 in sections depending on the subject.
13930 * Have the zlib compression code load ZLIB.DLL dynamically under
13935 * New function BN_mod_sqrt for computing square roots modulo a prime
13936 (using the probabilistic Tonelli-Shanks algorithm unless
13937 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13938 be handled deterministically).
13940 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13942 * Make BN_mod_inverse faster by explicitly handling small quotients
13943 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13944 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13948 * New function BN_kronecker.
13952 * Fix BN_gcd so that it works on negative inputs; the result is
13953 positive unless both parameters are zero.
13954 Previously something reasonably close to an infinite loop was
13955 possible because numbers could be growing instead of shrinking
13956 in the implementation of Euclid's algorithm.
13960 * Fix BN_is_word() and BN_is_one() macros to take into account the
13961 sign of the number in question.
13963 Fix BN_is_word(a,w) to work correctly for w == 0.
13965 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13966 because its test if the absolute value of 'a' equals 'w'.
13967 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13968 it exists mostly for use in the implementations of BN_is_zero(),
13969 BN_is_one(), and BN_is_word().
13973 * New function BN_swap.
13977 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13978 the exponentiation functions are more likely to produce reasonable
13979 results on negative inputs.
13983 * Change BN_mod_mul so that the result is always non-negative.
13984 Previously, it could be negative if one of the factors was negative;
13985 I don't think anyone really wanted that behaviour.
13989 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13990 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13991 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13992 and add new functions:
14001 BN_mod_lshift1_quick
14003 BN_mod_lshift_quick
14005 These functions always generate non-negative results.
14007 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14008 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14010 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14011 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14012 be reduced modulo `m`.
14014 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14017 The following entry accidentally appeared in the CHANGES file
14018 distributed with OpenSSL 0.9.7. The modifications described in
14019 it do *not* apply to OpenSSL 0.9.7.
14021 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14022 was actually never needed) and in BN_mul(). The removal in BN_mul()
14023 required a small change in bn_mul_part_recursive() and the addition
14024 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14025 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14026 bn_sub_words() and bn_add_words() except they take arrays with
14032 * In 'openssl passwd', verify passwords read from the terminal
14033 unless the '-salt' option is used (which usually means that
14034 verification would just waste user's time since the resulting
14035 hash is going to be compared with some given password hash)
14036 or the new '-noverify' option is used.
14038 This is an incompatible change, but it does not affect
14039 non-interactive use of 'openssl passwd' (passwords on the command
14040 line, '-stdin' option, '-in ...' option) and thus should not
14041 cause any problems.
14045 * Remove all references to RSAref, since there's no more need for it.
14049 * Make DSO load along a path given through an environment variable
14050 (SHLIB_PATH) with shl_load().
14054 * Constify the ENGINE code as a result of BIGNUM constification.
14055 Also constify the RSA code and most things related to it. In a
14056 few places, most notable in the depth of the ASN.1 code, ugly
14057 casts back to non-const were required (to be solved at a later
14062 * Make it so the openssl application has all engines loaded by default.
14066 * Constify the BIGNUM routines a little more.
14070 * Add the following functions:
14072 ENGINE_load_cswift()
14074 ENGINE_load_atalla()
14075 ENGINE_load_nuron()
14076 ENGINE_load_builtin_engines()
14078 That way, an application can itself choose if external engines that
14079 are built-in in OpenSSL shall ever be used or not. The benefit is
14080 that applications won't have to be linked with libdl or other dso
14081 libraries unless it's really needed.
14083 Changed 'openssl engine' to load all engines on demand.
14084 Changed the engine header files to avoid the duplication of some
14085 declarations (they differed!).
14089 * 'openssl engine' can now list capabilities.
14093 * Better error reporting in 'openssl engine'.
14097 * Never call load_dh_param(NULL) in s_server.
14101 * Add engine application. It can currently list engines by name and
14102 identity, and test if they are actually available.
14106 * Improve RPM specification file by forcing symbolic linking and making
14107 sure the installed documentation is also owned by root.root.
14109 *Damien Miller <djm@mindrot.org>*
14111 * Give the OpenSSL applications more possibilities to make use of
14112 keys (public as well as private) handled by engines.
14116 * Add OCSP code that comes from CertCo.
14120 * Add VMS support for the Rijndael code.
14124 * Added untested support for Nuron crypto accelerator.
14128 * Add support for external cryptographic devices. This code was
14129 previously distributed separately as the "engine" branch.
14131 *Geoff Thorpe, Richard Levitte*
14133 * Rework the filename-translation in the DSO code. It is now possible to
14134 have far greater control over how a "name" is turned into a filename
14135 depending on the operating environment and any oddities about the
14136 different shared library filenames on each system.
14140 * Support threads on FreeBSD-elf in Configure.
14144 * Fix for SHA1 assembly problem with MASM: it produces
14145 warnings about corrupt line number information when assembling
14146 with debugging information. This is caused by the overlapping
14149 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14152 NCONF_get_number() has no error checking at all. As a replacement,
14153 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14154 promoted strongly. The old NCONF_get_number is kept around for
14155 binary backward compatibility.
14156 Make it possible for methods to load from something other than a BIO,
14157 by providing a function pointer that is given a name instead of a BIO.
14158 For example, this could be used to load configuration data from an
14163 * Fix for non blocking accept BIOs. Added new I/O special reason
14164 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14165 with non blocking I/O was not possible because no retry code was
14166 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14171 * Added the beginnings of Rijndael support.
14175 * Fix for bug in DirectoryString mask setting. Add support for
14176 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14177 to allow certificate printing to more controllable, additional
14178 'certopt' option to 'x509' to allow new printing options to be
14183 * Clean old EAY MD5 hack from e_os.h.
14187 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14189 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14190 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14192 *Joe Orton, Steve Henson*
14194 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14196 * Fix additional bug revealed by the NISCC test suite:
14198 Stop bug triggering large recursion when presented with
14199 certain ASN.1 tags ([CVE-2003-0851])
14203 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14205 * Fix various bugs revealed by running the NISCC test suite:
14207 Stop out of bounds reads in the ASN1 code when presented with
14208 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14210 If verify callback ignores invalid public key errors don't try to check
14211 certificate signature with the NULL public key.
14215 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14216 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14221 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14222 extra data after the compression methods not only for TLS 1.0
14223 but also for SSL 3.0 (as required by the specification).
14225 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14227 * Change X509_certificate_type() to mark the key as exported/exportable
14228 when it's 512 *bits* long, not 512 bytes.
14232 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14234 * Countermeasure against the Klima-Pokorny-Rosa extension of
14235 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14236 a protocol version number mismatch like a decryption error
14237 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14241 * Turn on RSA blinding by default in the default implementation
14242 to avoid a timing attack. Applications that don't want it can call
14243 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14244 They would be ill-advised to do so in most cases.
14246 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14248 * Change RSA blinding code so that it works when the PRNG is not
14249 seeded (in this case, the secret RSA exponent is abused as
14250 an unpredictable seed -- if it is not unpredictable, there
14251 is no point in blinding anyway). Make RSA blinding thread-safe
14252 by remembering the creator's thread ID in rsa->blinding and
14253 having all other threads use local one-time blinding factors
14254 (this requires more computation than sharing rsa->blinding, but
14255 avoids excessive locking; and if an RSA object is not shared
14256 between threads, blinding will still be very fast).
14260 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14262 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14263 via timing by performing a MAC computation even if incorrect
14264 block cipher padding has been found. This is a countermeasure
14265 against active attacks where the attacker has to distinguish
14266 between bad padding and a MAC verification error. ([CVE-2003-0078])
14268 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14269 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14270 Martin Vuagnoux (EPFL, Ilion)*
14272 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14274 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14275 memory from its contents. This is done with a counter that will
14276 place alternating values in each byte. This can be used to solve
14277 two issues: 1) the removal of calls to memset() by highly optimizing
14278 compilers, and 2) cleansing with other values than 0, since those can
14279 be read through on certain media, for example a swap space on disk.
14283 * Bugfix: client side session caching did not work with external caching,
14284 because the session->cipher setting was not restored when reloading
14285 from the external cache. This problem was masked, when
14286 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14287 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14291 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14292 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14294 *Zeev Lieber <zeev-l@yahoo.com>*
14296 * Undo an undocumented change introduced in 0.9.6e which caused
14297 repeated calls to OpenSSL_add_all_ciphers() and
14298 OpenSSL_add_all_digests() to be ignored, even after calling
14303 * Change the default configuration reader to deal with last line not
14304 being properly terminated.
14308 * Change X509_NAME_cmp() so it applies the special rules on handling
14309 DN values that are of type PrintableString, as well as RDNs of type
14310 emailAddress where the value has the type ia5String.
14312 *stefank@valicert.com via Richard Levitte*
14314 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14315 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14316 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14317 the bitwise-OR of the two for use by the majority of applications
14318 wanting this behaviour, and update the docs. The documented
14319 behaviour and actual behaviour were inconsistent and had been
14320 changing anyway, so this is more a bug-fix than a behavioural
14323 *Geoff Thorpe, diagnosed by Nadav Har'El*
14325 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14326 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14330 * Fix initialization code race conditions in
14331 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14332 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14333 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14334 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14335 ssl2_get_cipher_by_char(),
14336 ssl3_get_cipher_by_char().
14338 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14340 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14341 the cached sessions are flushed, as the remove_cb() might use ex_data
14342 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14343 (see [openssl.org #212]).
14345 *Geoff Thorpe, Lutz Jaenicke*
14347 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14348 length, instead of the encoding length to d2i_ASN1_OBJECT.
14352 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14354 * [In 0.9.6g-engine release:]
14355 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14357 *Lynn Gazis <lgazis@rainbow.com>*
14359 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14361 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14362 and get fix the header length calculation.
14363 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14364 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14366 * Use proper error handling instead of 'assertions' in buffer
14367 overflow checks added in 0.9.6e. This prevents DoS (the
14368 assertions could call abort()).
14370 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14372 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14374 * Add various sanity checks to asn1_get_length() to reject
14375 the ASN1 length bytes if they exceed sizeof(long), will appear
14376 negative or the content length exceeds the length of the
14379 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14381 * Fix cipher selection routines: ciphers without encryption had no flags
14382 for the cipher strength set and where therefore not handled correctly
14383 by the selection routines (PR #130).
14387 * Fix EVP_dsa_sha macro.
14392 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14393 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14394 that was added in OpenSSL 0.9.6d.
14396 As the countermeasure turned out to be incompatible with some
14397 broken SSL implementations, the new option is part of SSL_OP_ALL.
14398 SSL_OP_ALL is usually employed when compatibility with weird SSL
14399 implementations is desired (e.g. '-bugs' option to 's_client' and
14400 's_server'), so the new option is automatically set in many
14405 * Changes in security patch:
14407 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14408 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14409 Air Force Materiel Command, USAF, under agreement number
14412 * Add various sanity checks to asn1_get_length() to reject
14413 the ASN1 length bytes if they exceed sizeof(long), will appear
14414 negative or the content length exceeds the length of the
14415 supplied buffer. ([CVE-2002-0659])
14417 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14419 * Assertions for various potential buffer overflows, not known to
14420 happen in practice.
14422 *Ben Laurie (CHATS)*
14424 * Various temporary buffers to hold ASCII versions of integers were
14425 too small for 64 bit platforms. ([CVE-2002-0655])
14426 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14428 * Remote buffer overflow in SSL3 protocol - an attacker could
14429 supply an oversized session ID to a client. ([CVE-2002-0656])
14431 *Ben Laurie (CHATS)*
14433 * Remote buffer overflow in SSL2 protocol - an attacker could
14434 supply an oversized client master key. ([CVE-2002-0656])
14436 *Ben Laurie (CHATS)*
14438 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14440 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14441 encoded as NULL) with id-dsa-with-sha1.
14443 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14445 * Check various `X509_...()` return values in `apps/req.c`.
14447 *Nils Larsch <nla@trustcenter.de>*
14449 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14450 an end-of-file condition would erroneously be flagged, when the CRLF
14451 was just at the end of a processed block. The bug was discovered when
14452 processing data through a buffering memory BIO handing the data to a
14453 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14454 <ptsekov@syntrex.com> and Nedelcho Stanev.
14458 * Implement a countermeasure against a vulnerability recently found
14459 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14460 before application data chunks to avoid the use of known IVs
14461 with data potentially chosen by the attacker.
14465 * Fix length checks in ssl3_get_client_hello().
14469 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14470 to prevent ssl3_read_internal() from incorrectly assuming that
14471 ssl3_read_bytes() found application data while handshake
14472 processing was enabled when in fact s->s3->in_read_app_data was
14473 merely automatically cleared during the initial handshake.
14475 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14477 * Fix object definitions for Private and Enterprise: they were not
14478 recognized in their shortname (=lowercase) representation. Extend
14479 obj_dat.pl to issue an error when using undefined keywords instead
14480 of silently ignoring the problem (Svenning Sorensen
14481 <sss@sss.dnsalias.net>).
14485 * Fix DH_generate_parameters() so that it works for 'non-standard'
14486 generators, i.e. generators other than 2 and 5. (Previously, the
14487 code did not properly initialise the 'add' and 'rem' values to
14488 BN_generate_prime().)
14490 In the new general case, we do not insist that 'generator' is
14491 actually a primitive root: This requirement is rather pointless;
14492 a generator of the order-q subgroup is just as good, if not
14497 * Map new X509 verification errors to alerts. Discovered and submitted by
14498 Tom Wu <tom@arcot.com>.
14502 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14503 returning non-zero before the data has been completely received
14504 when using non-blocking I/O.
14506 *Bodo Moeller; problem pointed out by John Hughes*
14508 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14510 *Ben Laurie, Lutz Jaenicke*
14512 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14513 Yoram Zahavi <YoramZ@gilian.com>).
14517 * Add information about CygWin 1.3 and on, and preserve proper
14518 configuration for the versions before that.
14520 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14522 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14523 check whether we deal with a copy of a session and do not delete from
14524 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14525 <izhar@checkpoint.com>.
14529 * Do not store session data into the internal session cache, if it
14530 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14531 flag is set). Proposed by Aslam <aslam@funk.com>.
14535 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14540 * [In 0.9.6d-engine release:]
14541 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14543 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14545 * Add the configuration target linux-s390x.
14547 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14549 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14550 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14551 variable as an indication that a ClientHello message has been
14552 received. As the flag value will be lost between multiple
14553 invocations of ssl3_accept when using non-blocking I/O, the
14554 function may not be aware that a handshake has actually taken
14555 place, thus preventing a new session from being added to the
14558 To avoid this problem, we now set s->new_session to 2 instead of
14559 using a local variable.
14561 *Lutz Jaenicke, Bodo Moeller*
14563 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14564 if the SSL_R_LENGTH_MISMATCH error is detected.
14566 *Geoff Thorpe, Bodo Moeller*
14568 * New 'shared_ldflag' column in Configure platform table.
14572 * Fix EVP_CIPHER_mode macro.
14574 *"Dan S. Camper" <dan@bti.net>*
14576 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14577 type, we must throw them away by setting rr->length to 0.
14579 *D P Chang <dpc@qualys.com>*
14581 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14583 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14584 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14585 worked incorrectly for those cases where range = `10..._2` and
14586 `3*range` is two bits longer than range.)
14590 * Only add signing time to PKCS7 structures if it is not already
14595 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14596 OBJ_ld_ce should be OBJ_id_ce.
14597 Also some ip-pda OIDs in crypto/objects/objects.txt were
14598 incorrect (cf. RFC 3039).
14600 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14602 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14603 returns early because it has nothing to do.
14605 *Andy Schneider <andy.schneider@bjss.co.uk>*
14607 * [In 0.9.6c-engine release:]
14608 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14610 *Andy Schneider <andy.schneider@bjss.co.uk>*
14612 * [In 0.9.6c-engine release:]
14613 Add support for Cryptographic Appliance's keyserver technology.
14614 (Use engine 'keyclient')
14616 *Cryptographic Appliances and Geoff Thorpe*
14618 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14619 is called via tools/c89.sh because arguments have to be
14620 rearranged (all '-L' options must appear before the first object
14623 *Richard Shapiro <rshapiro@abinitio.com>*
14625 * [In 0.9.6c-engine release:]
14626 Add support for Broadcom crypto accelerator cards, backported
14629 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14631 * [In 0.9.6c-engine release:]
14632 Add support for SureWare crypto accelerator cards from
14633 Baltimore Technologies. (Use engine 'sureware')
14635 *Baltimore Technologies and Mark Cox*
14637 * [In 0.9.6c-engine release:]
14638 Add support for crypto accelerator cards from Accelerated
14639 Encryption Processing, www.aep.ie. (Use engine 'aep')
14641 *AEP Inc. and Mark Cox*
14643 * Add a configuration entry for gcc on UnixWare.
14645 *Gary Benson <gbenson@redhat.com>*
14647 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14648 messages are stored in a single piece (fixed-length part and
14649 variable-length part combined) and fix various bugs found on the way.
14653 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14654 instead. BIO_gethostbyname() does not know what timeouts are
14655 appropriate, so entries would stay in cache even when they have
14657 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14659 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14660 faced with a pathologically small ClientHello fragment that does
14661 not contain client_version: Instead of aborting with an error,
14662 simply choose the highest available protocol version (i.e.,
14663 TLS 1.0 unless it is disabled). In practice, ClientHello
14664 messages are never sent like this, but this change gives us
14665 strictly correct behaviour at least for TLS.
14669 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14670 never resets s->method to s->ctx->method when called from within
14671 one of the SSL handshake functions.
14673 *Bodo Moeller; problem pointed out by Niko Baric*
14675 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14676 (sent using the client's version number) if client_version is
14677 smaller than the protocol version in use. Also change
14678 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14679 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14680 the client will at least see that alert.
14684 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14689 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14690 client receives HelloRequest while in a handshake.
14692 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14694 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14695 should end in 'break', not 'goto end' which circumvents various
14696 cleanups done in state SSL_ST_OK. But session related stuff
14697 must be disabled for SSL_ST_OK in the case that we just sent a
14700 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14701 before just sending a HelloRequest.
14703 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14705 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14706 reveal whether illegal block cipher padding was found or a MAC
14707 verification error occurred. (Neither SSLerr() codes nor alerts
14708 are directly visible to potential attackers, but the information
14709 may leak via logfiles.)
14711 Similar changes are not required for the SSL 2.0 implementation
14712 because the number of padding bytes is sent in clear for SSL 2.0,
14713 and the extra bytes are just ignored. However ssl/s2_pkt.c
14714 failed to verify that the purported number of padding bytes is in
14719 * Add OpenUNIX-8 support including shared libraries
14720 (Boyd Lynn Gerber <gerberb@zenez.com>).
14724 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14725 'wristwatch attack' using huge encoding parameters (cf.
14726 James H. Manger's CRYPTO 2001 paper). Note that the
14727 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14728 encoding parameters and hence was not vulnerable.
14732 * BN_sqr() bug fix.
14734 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14736 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14737 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14738 followed by modular reduction.
14740 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14742 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14743 equivalent based on BN_pseudo_rand() instead of BN_rand().
14747 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14748 This function was broken, as the check for a new client hello message
14749 to handle SGC did not allow these large messages.
14750 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14754 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14758 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14759 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14763 * Rework the configuration and shared library support for Tru64 Unix.
14764 The configuration part makes use of modern compiler features and
14765 still retains old compiler behavior for those that run older versions
14766 of the OS. The shared library support part includes a variant that
14767 uses the RPATH feature, and is available through the special
14768 configuration target "alpha-cc-rpath", which will never be selected
14771 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14773 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14774 with the same message size as in ssl3_get_certificate_request().
14775 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14776 messages might inadvertently be reject as too long.
14778 *Petr Lampa <lampa@fee.vutbr.cz>*
14780 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14784 * Modified SSL library such that the verify_callback that has been set
14785 specifically for an SSL object with SSL_set_verify() is actually being
14786 used. Before the change, a verify_callback set with this function was
14787 ignored and the verify_callback() set in the SSL_CTX at the time of
14788 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14789 to allow the necessary settings.
14793 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14794 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14795 done automatically (in contradiction to the requirements of the C
14796 standard). This made problems when used from OpenSSH.
14800 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14801 dh->length and always used
14803 BN_rand_range(priv_key, dh->p).
14805 BN_rand_range() is not necessary for Diffie-Hellman, and this
14806 specific range makes Diffie-Hellman unnecessarily inefficient if
14807 dh->length (recommended exponent length) is much smaller than the
14808 length of dh->p. We could use BN_rand_range() if the order of
14809 the subgroup was stored in the DH structure, but we only have
14814 BN_rand(priv_key, l, ...)
14816 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14823 RSA_eay_public_encrypt
14824 RSA_eay_private_decrypt
14825 RSA_eay_private_encrypt (signing)
14826 RSA_eay_public_decrypt (signature verification)
14828 (default implementations for RSA_public_encrypt,
14829 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14830 always reject numbers >= n.
14834 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14835 to synchronize access to 'locking_thread'. This is necessary on
14836 systems where access to 'locking_thread' (an 'unsigned long'
14837 variable) is not atomic.
14841 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14842 *before* setting the 'crypto_lock_rand' flag. The previous code had
14843 a race condition if 0 is a valid thread ID.
14845 *Travis Vitek <vitek@roguewave.com>*
14847 * Add support for shared libraries under Irix.
14849 *Albert Chin-A-Young <china@thewrittenword.com>*
14851 * Add configuration option to build on Linux on both big-endian and
14852 little-endian MIPS.
14854 *Ralf Baechle <ralf@uni-koblenz.de>*
14856 * Add the possibility to create shared libraries on HP-UX.
14860 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14862 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14863 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14864 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14865 PRNG state recovery was possible based on the output of
14866 one PRNG request appropriately sized to gain knowledge on
14867 'md' followed by enough consecutive 1-byte PRNG requests
14868 to traverse all of 'state'.
14870 1. When updating 'md_local' (the current thread's copy of 'md')
14871 during PRNG output generation, hash all of the previous
14872 'md_local' value, not just the half used for PRNG output.
14874 2. Make the number of bytes from 'state' included into the hash
14875 independent from the number of PRNG bytes requested.
14877 The first measure alone would be sufficient to avoid
14878 Markku-Juhani's attack. (Actually it had never occurred
14879 to me that the half of 'md_local' used for chaining was the
14880 half from which PRNG output bytes were taken -- I had always
14881 assumed that the secret half would be used.) The second
14882 measure makes sure that additional data from 'state' is never
14883 mixed into 'md_local' in small portions; this heuristically
14884 further strengthens the PRNG.
14888 * Fix crypto/bn/asm/mips3.s.
14892 * When only the key is given to "enc", the IV is undefined. Print out
14893 an error message in this case.
14897 * Handle special case when X509_NAME is empty in X509 printing routines.
14901 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14902 positive and less than q.
14906 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14907 used: it isn't thread safe and the add_lock_callback should handle
14910 *Paul Rose <Paul.Rose@bridge.com>*
14912 * Verify that incoming data obeys the block size in
14913 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14919 *Ulf Möller, Bodo Möller*
14921 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14922 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14923 when fixing the server behaviour for backwards-compatible 'client
14924 hello' messages. (Note that the attack is impractical against
14925 SSL 3.0 and TLS 1.0 anyway because length and version checking
14926 means that the probability of guessing a valid ciphertext is
14927 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14930 Before 0.9.5, the countermeasure (hide the error by generating a
14931 random 'decryption result') did not work properly because
14932 ERR_clear_error() was missing, meaning that SSL_get_error() would
14933 detect the supposedly ignored error.
14935 Both problems are now fixed.
14939 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14940 (previously it was 1024).
14944 * Fix for compatibility mode trust settings: ignore trust settings
14945 unless some valid trust or reject settings are present.
14949 * Fix for blowfish EVP: its a variable length cipher.
14953 * Fix various bugs related to DSA S/MIME verification. Handle missing
14954 parameters in DSA public key structures and return an error in the
14955 DSA routines if parameters are absent.
14959 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14960 in the current directory if neither $RANDFILE nor $HOME was set.
14961 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14962 caused some confusion to Windows users who haven't defined $HOME.
14963 Thus RAND_file_name() is changed again: e_os.h can define a
14964 DEFAULT_HOME, which will be used if $HOME is not set.
14965 For Windows, we use "C:"; on other platforms, we still require
14966 environment variables.
14968 * Move 'if (!initialized) RAND_poll()' into regions protected by
14969 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14970 having multiple threads call RAND_poll() concurrently.
14974 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14975 combination of a flag and a thread ID variable.
14976 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14977 flag), *other* threads can enter ssleay_add_bytes without obeying
14978 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14979 that they do not hold after the first thread unsets add_do_not_lock).
14983 * Change bctest again: '-x' expressions are not available in all
14984 versions of 'test'.
14988 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14990 * Fix a couple of memory leaks in PKCS7_dataDecode()
14992 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14994 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14995 the default extension for executables, if any. Also, make the perl
14996 scripts that use symlink() to test if it really exists and use "cp"
14997 if it doesn't. All this made OpenSSL compilable and installable in
15002 * Fix for asn1_GetSequence() for indefinite length constructed data.
15003 If SEQUENCE is length is indefinite just set c->slen to the total
15004 amount of data available.
15006 *Steve Henson, reported by shige@FreeBSD.org*
15008 *This change does not apply to 0.9.7.*
15010 * Change bctest to avoid here-documents inside command substitution
15011 (workaround for FreeBSD /bin/sh bug).
15012 For compatibility with Ultrix, avoid shell functions (introduced
15013 in the bctest version that searches along $PATH).
15017 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15018 with des_encrypt() defined on some operating systems, like Solaris
15023 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15024 On the Importance of Eliminating Errors in Cryptographic
15025 Computations, J. Cryptology 14 (2001) 2, 101-119,
15026 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15030 * MIPS assembler BIGNUM division bug fix.
15034 * Disabled incorrect Alpha assembler code.
15038 * Fix PKCS#7 decode routines so they correctly update the length
15039 after reading an EOC for the EXPLICIT tag.
15043 *This change does not apply to 0.9.7.*
15045 * Fix bug in PKCS#12 key generation routines. This was triggered
15046 if a 3DES key was generated with a 0 initial byte. Include
15047 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15048 (but broken) behaviour.
15052 * Enhance bctest to search for a working bc along $PATH and print
15055 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15057 * Fix memory leaks in err.c: free err_data string if necessary;
15058 don't write to the wrong index in ERR_set_error_data.
15062 * Implement ssl23_peek (analogous to ssl23_read), which previously
15067 * Replace rdtsc with `_emit` statements for VC++ version 5.
15069 *Jeremy Cooper <jeremy@baymoo.org>*
15071 * Make it possible to reuse SSLv2 sessions.
15075 * In copy_email() check for >= 0 as a return value for
15076 X509_NAME_get_index_by_NID() since 0 is a valid index.
15078 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15080 * Avoid coredump with unsupported or invalid public keys by checking if
15081 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15082 PKCS7_verify() fails with non detached data.
15086 * Don't use getenv in library functions when run as setuid/setgid.
15087 New function OPENSSL_issetugid().
15091 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15092 due to incorrect handling of multi-threading:
15094 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15096 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15098 3. Count how many times MemCheck_off() has been called so that
15099 nested use can be treated correctly. This also avoids
15100 inband-signalling in the previous code (which relied on the
15101 assumption that thread ID 0 is impossible).
15105 * Add "-rand" option also to s_client and s_server.
15109 * Fix CPU detection on Irix 6.x.
15110 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15111 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15113 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15118 *This change does not apply to 0.9.7.*
15120 * Use the cached encoding of an X509_NAME structure rather than
15121 copying it. This is apparently the reason for the libsafe "errors"
15122 but the code is actually correct.
15126 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15127 Bleichenbacher's DSA attack.
15128 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15129 to be set and top=0 forces the highest bit to be set; top=-1 is new
15130 and leaves the highest bit random.
15132 *Ulf Moeller, Bodo Moeller*
15134 * In the `NCONF_...`-based implementations for `CONF_...` queries
15135 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15136 a temporary CONF structure with the data component set to NULL
15137 (which gives segmentation faults in lh_retrieve).
15138 Instead, use NULL for the CONF pointer in CONF_get_string and
15139 CONF_get_number (which may use environment variables) and directly
15140 return NULL from CONF_get_section.
15144 * Fix potential buffer overrun for EBCDIC.
15148 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15149 keyUsage if basicConstraints absent for a CA.
15153 * Make SMIME_write_PKCS7() write mail header values with a format that
15154 is more generally accepted (no spaces before the semicolon), since
15155 some programs can't parse those values properly otherwise. Also make
15156 sure BIO's that break lines after each write do not create invalid
15161 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15162 macros previously used would not encode an empty SEQUENCE OF
15163 and break the signature.
15167 *This change does not apply to 0.9.7.*
15169 * Zero the premaster secret after deriving the master secret in
15174 * Add some EVP_add_digest_alias registrations (as found in
15175 OpenSSL_add_all_digests()) to SSL_library_init()
15176 aka OpenSSL_add_ssl_algorithms(). This provides improved
15177 compatibility with peers using X.509 certificates
15178 with unconventional AlgorithmIdentifier OIDs.
15182 * Fix for Irix with NO_ASM.
15184 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15186 * ./config script fixes.
15188 *Ulf Moeller, Richard Levitte*
15190 * Fix 'openssl passwd -1'.
15194 * Change PKCS12_key_gen_asc() so it can cope with non null
15195 terminated strings whose length is passed in the passlen
15196 parameter, for example from PEM callbacks. This was done
15197 by adding an extra length parameter to asc2uni().
15199 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15201 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15202 call failed, free the DSA structure.
15206 * Fix to uni2asc() to cope with zero length Unicode strings.
15207 These are present in some PKCS#12 files.
15211 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15212 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15213 when writing a 32767 byte record.
15215 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15217 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15218 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15220 (RSA objects have a reference count access to which is protected
15221 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15222 so they are meant to be shared between threads.)
15223 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15224 "Reddie, Steven" <Steven.Reddie@ca.com>*
15226 * Fix a deadlock in CRYPTO_mem_leaks().
15230 * Use better test patterns in bntest.
15234 * rand_win.c fix for Borland C.
15238 * BN_rshift bugfix for n == 0.
15242 * Add a 'bctest' script that checks for some known 'bc' bugs
15243 so that 'make test' does not abort just because 'bc' is broken.
15247 * Store verify_result within SSL_SESSION also for client side to
15248 avoid potential security hole. (Re-used sessions on the client side
15249 always resulted in verify_result==X509_V_OK, not using the original
15250 result of the server certificate verification.)
15254 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15255 SSL3_RT_APPLICATION_DATA, return 0.
15256 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15261 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15262 releases, have been re-implemented by renaming the previous
15263 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15264 and ssl3_read_internal, respectively, and adding 'peek' parameters
15265 to them. The new ssl[23]_{read,peek} functions are calls to
15266 ssl[23]_read_internal with the 'peek' flag set appropriately.
15267 A 'peek' parameter has also been added to ssl3_read_bytes, which
15268 does the actual work for ssl3_read_internal.
15272 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15273 the method-specific "init()" handler. Also clean up ex_data after
15274 calling the method-specific "finish()" handler. Previously, this was
15275 happening the other way round.
15279 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15280 The previous value, 12, was not always sufficient for BN_mod_exp().
15284 * Make sure that shared libraries get the internal name engine with
15285 the full version number and not just 0. This should mark the
15286 shared libraries as not backward compatible. Of course, this should
15287 be changed again when we can guarantee backward binary compatibility.
15291 * Fix typo in get_cert_by_subject() in by_dir.c
15293 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15295 * Rework the system to generate shared libraries:
15297 - Make note of the expected extension for the shared libraries and
15298 if there is a need for symbolic links from for example libcrypto.so.0
15299 to libcrypto.so.0.9.7. There is extended info in Configure for
15302 - Make as few rebuilds of the shared libraries as possible.
15304 - Still avoid linking the OpenSSL programs with the shared libraries.
15306 - When installing, install the shared libraries separately from the
15311 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15313 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15314 and not in SSL_clear because the latter is also used by the
15315 accept/connect functions; previously, the settings made by
15316 SSL_set_read_ahead would be lost during the handshake.
15318 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15320 * Correct util/mkdef.pl to be selective about disabled algorithms.
15321 Previously, it would create entries for disabled algorithms no
15326 * Added several new manual pages for SSL_* function.
15330 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15332 * In ssl23_get_client_hello, generate an error message when faced
15333 with an initial SSL 3.0/TLS record that is too small to contain the
15334 first two bytes of the ClientHello message, i.e. client_version.
15335 (Note that this is a pathologic case that probably has never happened
15336 in real life.) The previous approach was to use the version number
15337 from the record header as a substitute; but our protocol choice
15338 should not depend on that one because it is not authenticated
15339 by the Finished messages.
15343 * More robust randomness gathering functions for Windows.
15345 *Jeffrey Altman <jaltman@columbia.edu>*
15347 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15348 not set then we don't setup the error code for issuer check errors
15349 to avoid possibly overwriting other errors which the callback does
15350 handle. If an application does set the flag then we assume it knows
15351 what it is doing and can handle the new informational codes
15356 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15357 a general "ANY" type, as such it should be able to decode anything
15358 including tagged types. However it didn't check the class so it would
15359 wrongly interpret tagged types in the same way as their universal
15360 counterpart and unknown types were just rejected. Changed so that the
15361 tagged and unknown types are handled in the same way as a SEQUENCE:
15362 that is the encoding is stored intact. There is also a new type
15363 "V_ASN1_OTHER" which is used when the class is not universal, in this
15364 case we have no idea what the actual type is so we just lump them all
15369 * On VMS, stdout may very well lead to a file that is written to
15370 in a record-oriented fashion. That means that every write() will
15371 write a separate record, which will be read separately by the
15372 programs trying to read from it. This can be very confusing.
15374 The solution is to put a BIO filter in the way that will buffer
15375 text until a linefeed is reached, and then write everything a
15376 line at a time, so every record written will be an actual line,
15377 not chunks of lines and not (usually doesn't happen, but I've
15378 seen it once) several lines in one record. BIO_f_linebuffer() is
15381 Currently, it's a VMS-only method, because that's where it has
15382 been tested well enough.
15386 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15387 it can return incorrect results.
15388 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15389 but it was in 0.9.6-beta[12].)
15393 * Disable the check for content being present when verifying detached
15394 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15395 include zero length content when signing messages.
15399 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15400 BIO_ctrl (for BIO pairs).
15404 * Add DSO method for VMS.
15408 * Bug fix: Montgomery multiplication could produce results with the
15413 * Add RPM specification openssl.spec and modify it to build three
15414 packages. The default package contains applications, application
15415 documentation and run-time libraries. The devel package contains
15416 include files, static libraries and function documentation. The
15417 doc package contains the contents of the doc directory. The original
15418 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15422 * Add a large number of documentation files for many SSL routines.
15424 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15426 * Add a configuration entry for Sony News 4.
15428 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15430 * Don't set the two most significant bits to one when generating a
15431 random number < q in the DSA library.
15435 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15436 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15437 the underlying transport is blocking) if a handshake took place.
15438 (The default behaviour is needed by applications such as s_client
15439 and s_server that use select() to determine when to use SSL_read;
15440 but for applications that know in advance when to expect data, it
15441 just makes things more complicated.)
15445 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15450 * Add a few more EBCDIC conditionals that make `req` and `x509`
15451 work better on such systems.
15453 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15455 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15456 Update PKCS12_parse() so it copies the friendlyName and the
15457 keyid to the certificates aux info.
15461 * Fix bug in PKCS7_verify() which caused an infinite loop
15462 if there was more than one signature.
15464 *Sven Uszpelkat <su@celocom.de>*
15466 * Major change in util/mkdef.pl to include extra information
15467 about each symbol, as well as presenting variables as well
15468 as functions. This change means that there's n more need
15469 to rebuild the .num files when some algorithms are excluded.
15473 * Allow the verify time to be set by an application,
15474 rather than always using the current time.
15478 * Phase 2 verify code reorganisation. The certificate
15479 verify code now looks up an issuer certificate by a
15480 number of criteria: subject name, authority key id
15481 and key usage. It also verifies self signed certificates
15482 by the same criteria. The main comparison function is
15483 X509_check_issued() which performs these checks.
15485 Lot of changes were necessary in order to support this
15486 without completely rewriting the lookup code.
15488 Authority and subject key identifier are now cached.
15490 The LHASH 'certs' is X509_STORE has now been replaced
15491 by a STACK_OF(X509_OBJECT). This is mainly because an
15492 LHASH can't store or retrieve multiple objects with
15493 the same hash value.
15495 As a result various functions (which were all internal
15496 use only) have changed to handle the new X509_STORE
15497 structure. This will break anything that messed round
15498 with X509_STORE internally.
15500 The functions X509_STORE_add_cert() now checks for an
15501 exact match, rather than just subject name.
15503 The X509_STORE API doesn't directly support the retrieval
15504 of multiple certificates matching a given criteria, however
15505 this can be worked round by performing a lookup first
15506 (which will fill the cache with candidate certificates)
15507 and then examining the cache for matches. This is probably
15508 the best we can do without throwing out X509_LOOKUP
15509 entirely (maybe later...).
15511 The X509_VERIFY_CTX structure has been enhanced considerably.
15513 All certificate lookup operations now go via a get_issuer()
15514 callback. Although this currently uses an X509_STORE it
15515 can be replaced by custom lookups. This is a simple way
15516 to bypass the X509_STORE hackery necessary to make this
15517 work and makes it possible to use more efficient techniques
15518 in future. A very simple version which uses a simple
15519 STACK for its trusted certificate store is also provided
15520 using X509_STORE_CTX_trusted_stack().
15522 The verify_cb() and verify() callbacks now have equivalents
15523 in the X509_STORE_CTX structure.
15525 X509_STORE_CTX also has a 'flags' field which can be used
15526 to customise the verify behaviour.
15530 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15531 excludes S/MIME capabilities.
15535 * When a certificate request is read in keep a copy of the
15536 original encoding of the signed data and use it when outputting
15537 again. Signatures then use the original encoding rather than
15538 a decoded, encoded version which may cause problems if the
15539 request is improperly encoded.
15543 * For consistency with other BIO_puts implementations, call
15544 buffer_write(b, ...) directly in buffer_puts instead of calling
15547 In BIO_puts, increment b->num_write as in BIO_write.
15549 *Peter.Sylvester@EdelWeb.fr*
15551 * Fix BN_mul_word for the case where the word is 0. (We have to use
15552 BN_zero, we may not return a BIGNUM with an array consisting of
15553 words set to zero.)
15557 * Avoid calling abort() from within the library when problems are
15558 detected, except if preprocessor symbols have been defined
15559 (such as REF_CHECK, BN_DEBUG etc.).
15563 * New openssl application 'rsautl'. This utility can be
15564 used for low-level RSA operations. DER public key
15565 BIO/fp routines also added.
15569 * New Configure entry and patches for compiling on QNX 4.
15571 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15573 * A demo state-machine implementation was sponsored by
15574 Nuron (<http://www.nuron.com/>) and is now available in
15575 demos/state_machine.
15579 * New options added to the 'dgst' utility for signature
15580 generation and verification.
15584 * Unrecognized PKCS#7 content types are now handled via a
15585 catch all ASN1_TYPE structure. This allows unsupported
15586 types to be stored as a "blob" and an application can
15587 encode and decode it manually.
15591 * Fix various signed/unsigned issues to make a_strex.c
15592 compile under VC++.
15594 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15596 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15597 length if passed a buffer. ASN1_INTEGER_to_BN failed
15598 if passed a NULL BN and its argument was negative.
15600 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15602 * Modification to PKCS#7 encoding routines to output definite
15603 length encoding. Since currently the whole structures are in
15604 memory there's not real point in using indefinite length
15605 constructed encoding. However if OpenSSL is compiled with
15606 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15610 * Added BIO_vprintf() and BIO_vsnprintf().
15614 * Added more prefixes to parse for in the strings written
15615 through a logging bio, to cover all the levels that are available
15616 through syslog. The prefixes are now:
15618 PANIC, EMERG, EMR => LOG_EMERG
15619 ALERT, ALR => LOG_ALERT
15620 CRIT, CRI => LOG_CRIT
15621 ERROR, ERR => LOG_ERR
15622 WARNING, WARN, WAR => LOG_WARNING
15623 NOTICE, NOTE, NOT => LOG_NOTICE
15624 INFO, INF => LOG_INFO
15625 DEBUG, DBG => LOG_DEBUG
15627 and as before, if none of those prefixes are present at the
15628 beginning of the string, LOG_ERR is chosen.
15630 On Win32, the `LOG_*` levels are mapped according to this:
15632 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15633 LOG_WARNING => EVENTLOG_WARNING_TYPE
15634 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15638 * Made it possible to reconfigure with just the configuration
15639 argument "reconf" or "reconfigure". The command line arguments
15640 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15641 and are retrieved from there when reconfiguring.
15647 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15649 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15653 * The obj_dat.pl script was messing up the sorting of object
15654 names. The reason was that it compared the quoted version
15655 of strings as a result "OCSP" > "OCSP Signing" because
15656 " > SPACE. Changed script to store unquoted versions of
15657 names and add quotes on output. It was also omitting some
15658 names from the lookup table if they were given a default
15659 value (that is if SN is missing it is given the same
15660 value as LN and vice versa), these are now added on the
15661 grounds that if an object has a name we should be able to
15662 look it up. Finally added warning output when duplicate
15663 short or long names are found.
15667 * Changes needed for Tandem NSK.
15669 *Scott Uroff <scott@xypro.com>*
15671 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15672 RSA_padding_check_SSLv23(), special padding was never detected
15673 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15674 version rollback attacks was not effective.
15676 In s23_clnt.c, don't use special rollback-attack detection padding
15677 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15678 client; similarly, in s23_srvr.c, don't do the rollback check if
15679 SSL 2.0 is the only protocol enabled in the server.
15683 * Make it possible to get hexdumps of unprintable data with 'openssl
15684 asn1parse'. By implication, the functions ASN1_parse_dump() and
15685 BIO_dump_indent() are added.
15689 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15690 these print out strings and name structures based on various
15691 flags including RFC2253 support and proper handling of
15692 multibyte characters. Added options to the 'x509' utility
15693 to allow the various flags to be set.
15697 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15698 Also change the functions X509_cmp_current_time() and
15699 X509_gmtime_adj() work with an ASN1_TIME structure,
15700 this will enable certificates using GeneralizedTime in validity
15701 dates to be checked.
15705 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15706 negative public key encodings) on by default,
15707 NO_NEG_PUBKEY_BUG can be set to disable it.
15711 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15712 content octets. An i2c_ASN1_OBJECT is unnecessary because
15713 the encoding can be trivially obtained from the structure.
15717 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15718 not read locks (`CRYPTO_r_[un]lock`).
15722 * A first attempt at creating official support for shared
15723 libraries through configuration. I've kept it so the
15724 default is static libraries only, and the OpenSSL programs
15725 are always statically linked for now, but there are
15726 preparations for dynamic linking in place.
15727 This has been tested on Linux and Tru64.
15731 * Randomness polling function for Win9x, as described in:
15732 Peter Gutmann, Software Generation of Practically Strong
15737 * Fix so PRNG is seeded in req if using an already existing
15742 * New options to smime application. -inform and -outform
15743 allow alternative formats for the S/MIME message including
15744 PEM and DER. The -content option allows the content to be
15745 specified separately. This should allow things like Netscape
15746 form signing output easier to verify.
15750 * Fix the ASN1 encoding of tags using the 'long form'.
15754 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15755 STRING types. These convert content octets to and from the
15756 underlying type. The actual tag and length octets are
15757 already assumed to have been read in and checked. These
15758 are needed because all other string types have virtually
15759 identical handling apart from the tag. By having versions
15760 of the ASN1 functions that just operate on content octets
15761 IMPLICIT tagging can be handled properly. It also allows
15762 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15763 and ASN1_INTEGER are identical apart from the tag.
15767 * Change the handling of OID objects as follows:
15769 - New object identifiers are inserted in objects.txt, following
15770 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15771 - objects.pl is used to process obj_mac.num and create a new
15773 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15776 This is currently kind of a hack, and the perl code in objects.pl
15777 isn't very elegant, but it works as I intended. The simplest way
15778 to check that it worked correctly is to look in obj_dat.h and
15779 check the array nid_objs and make sure the objects haven't moved
15780 around (this is important!). Additions are OK, as well as
15781 consistent name changes.
15785 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15789 * Addition of the command line parameter '-rand file' to 'openssl req'.
15790 The given file adds to whatever has already been seeded into the
15791 random pool through the RANDFILE configuration file option or
15792 environment variable, or the default random state file.
15796 * mkstack.pl now sorts each macro group into lexical order.
15797 Previously the output order depended on the order the files
15798 appeared in the directory, resulting in needless rewriting
15803 * Patches to make OpenSSL compile under Win32 again. Mostly
15804 work arounds for the VC++ problem that it treats func() as
15805 func(void). Also stripped out the parts of mkdef.pl that
15806 added extra typesafe functions: these no longer exist.
15810 * Reorganisation of the stack code. The macros are now all
15811 collected in safestack.h . Each macro is defined in terms of
15812 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15813 DEBUG_SAFESTACK is now handled in terms of function casts,
15814 this has the advantage of retaining type safety without the
15815 use of additional functions. If DEBUG_SAFESTACK is not defined
15816 then the non typesafe macros are used instead. Also modified the
15817 mkstack.pl script to handle the new form. Needs testing to see
15818 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15819 the default if no major problems. Similar behaviour for ASN1_SET_OF
15820 and PKCS12_STACK_OF.
15824 * When some versions of IIS use the 'NET' form of private key the
15825 key derivation algorithm is different. Normally MD5(password) is
15826 used as a 128 bit RC4 key. In the modified case
15827 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15828 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15829 as the old Netscape_RSA functions except they have an additional
15830 'sgckey' parameter which uses the modified algorithm. Also added
15831 an -sgckey command line option to the rsa utility. Thanks to
15832 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15833 algorithm to openssl-dev.
15837 * The evp_local.h macros were using 'c.##kname' which resulted in
15838 invalid expansion on some systems (SCO 5.0.5 for example).
15839 Corrected to 'c.kname'.
15841 *Phillip Porch <root@theporch.com>*
15843 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15844 a STACK of email addresses from a certificate or request, these look
15845 in the subject name and the subject alternative name extensions and
15846 omit any duplicate addresses.
15850 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15851 This makes DSA verification about 2 % faster.
15855 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15856 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15857 plus overhead for 1024 bit moduli).
15858 This makes exponentiations about 0.5 % faster for 1024 bit
15859 exponents (as measured by "openssl speed rsa2048").
15863 * Rename memory handling macros to avoid conflicts with other
15865 Malloc => OPENSSL_malloc
15866 Malloc_locked => OPENSSL_malloc_locked
15867 Realloc => OPENSSL_realloc
15868 Free => OPENSSL_free
15872 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15873 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15877 * CygWin32 support.
15879 *John Jarvie <jjarvie@newsguy.com>*
15881 * The type-safe stack code has been rejigged. It is now only compiled
15882 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15883 by default all type-specific stack functions are "#define"d back to
15884 standard stack functions. This results in more streamlined output
15885 but retains the type-safety checking possibilities of the original
15890 * The STACK code has been cleaned up, and certain type declarations
15891 that didn't make a lot of sense have been brought in line. This has
15892 also involved a cleanup of sorts in safestack.h to more correctly
15893 map type-safe stack functions onto their plain stack counterparts.
15894 This work has also resulted in a variety of "const"ifications of
15895 lots of the code, especially `_cmp` operations which should normally
15896 be prototyped with "const" parameters anyway.
15900 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15901 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15902 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15903 where all of 'md' is used each time the PRNG is used, but 'state'
15904 is used only indexed by a cyclic counter. As entropy may not be
15905 well distributed from the beginning, 'md' is important as a
15906 chaining variable. However, the output function chains only half
15907 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15908 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15909 in all of 'state' being rewritten, with the new values depending
15910 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15914 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15915 the handshake is continued after ssl_verify_cert_chain();
15916 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15917 can lead to 'unexplainable' connection aborts later.
15919 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15921 * Major EVP API cipher revision.
15922 Add hooks for extra EVP features. This allows various cipher
15923 parameters to be set in the EVP interface. Support added for variable
15924 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15925 setting of RC2 and RC5 parameters.
15927 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15930 Remove lots of duplicated code from the EVP library. For example *every*
15931 cipher init() function handles the 'iv' in the same way according to the
15932 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15933 for CFB and OFB modes they zero ctx->num.
15935 New functionality allows removal of S/MIME code RC2 hack.
15937 Most of the routines have the same form and so can be declared in terms
15940 By shifting this to the top level EVP_CipherInit() it can be removed from
15941 all individual ciphers. If the cipher wants to handle IVs or keys
15942 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15945 Change lots of functions like EVP_EncryptUpdate() to now return a
15946 value: although software versions of the algorithms cannot fail
15947 any installed hardware versions can.
15951 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15952 this option is set, tolerate broken clients that send the negotiated
15953 protocol version number instead of the requested protocol version
15958 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15959 i.e. non-zero for export ciphersuites, zero otherwise.
15960 Previous versions had this flag inverted, inconsistent with
15961 rsa_tmp_cb (..._TMP_RSA_CB).
15963 *Bodo Moeller; problem reported by Amit Chopra*
15965 * Add missing DSA library text string. Work around for some IIS
15966 key files with invalid SEQUENCE encoding.
15970 * Add a document (doc/standards.txt) that list all kinds of standards
15971 and so on that are implemented in OpenSSL.
15975 * Enhance c_rehash script. Old version would mishandle certificates
15976 with the same subject name hash and wouldn't handle CRLs at all.
15977 Added -fingerprint option to crl utility, to support new c_rehash
15982 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15986 * Fix for SSL server purpose checking. Server checking was
15987 rejecting certificates which had extended key usage present
15988 but no ssl client purpose.
15990 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15992 * Make PKCS#12 code work with no password. The PKCS#12 spec
15993 is a little unclear about how a blank password is handled.
15994 Since the password in encoded as a BMPString with terminating
15995 double NULL a zero length password would end up as just the
15996 double NULL. However no password at all is different and is
15997 handled differently in the PKCS#12 key generation code. NS
15998 treats a blank password as zero length. MSIE treats it as no
15999 password on export: but it will try both on import. We now do
16000 the same: PKCS12_parse() tries zero length and no password if
16001 the password is set to "" or NULL (NULL is now a valid password:
16002 it wasn't before) as does the pkcs12 application.
16006 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16007 perror when PEM_read_bio_X509_REQ fails, the error message must
16008 be obtained from the error queue.
16012 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16013 it in ERR_remove_state if appropriate, and change ERR_get_state
16014 accordingly to avoid race conditions (this is necessary because
16015 thread_hash is no longer constant once set).
16019 * Bugfix for linux-elf makefile.one.
16023 * RSA_get_default_method() will now cause a default
16024 RSA_METHOD to be chosen if one doesn't exist already.
16025 Previously this was only set during a call to RSA_new()
16026 or RSA_new_method(NULL) meaning it was possible for
16027 RSA_get_default_method() to return NULL.
16031 * Added native name translation to the existing DSO code
16032 that will convert (if the flag to do so is set) filenames
16033 that are sufficiently small and have no path information
16034 into a canonical native form. Eg. "blah" converted to
16035 "libblah.so" or "blah.dll" etc.
16039 * New function ERR_error_string_n(e, buf, len) which is like
16040 ERR_error_string(e, buf), but writes at most 'len' bytes
16041 including the 0 terminator. For ERR_error_string_n, 'buf'
16044 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16046 * CONF library reworked to become more general. A new CONF
16047 configuration file reader "class" is implemented as well as a
16048 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16049 old `CONF_*` functions are still there, but are reimplemented to
16050 work in terms of the new functions. Also, a set of functions
16051 to handle the internal storage of the configuration data is
16052 provided to make it easier to write new configuration file
16053 reader "classes" (I can definitely see something reading a
16054 configuration file in XML format, for example), called `_CONF_*`,
16055 or "the configuration storage API"...
16057 The new configuration file reading functions are:
16059 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16060 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16062 NCONF_default, NCONF_WIN32
16064 NCONF_dump_fp, NCONF_dump_bio
16066 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16067 NCONF_new creates a new CONF object. This works in the same way
16068 as other interfaces in OpenSSL, like the BIO interface.
16069 `NCONF_dump_*` dump the internal storage of the configuration file,
16070 which is useful for debugging. All other functions take the same
16071 arguments as the old `CONF_*` functions with the exception of the
16072 first that must be a `CONF *` instead of a `LHASH *`.
16074 To make it easier to use the new classes with the old `CONF_*` functions,
16075 the function CONF_set_default_method is provided.
16079 * Add '-tls1' option to 'openssl ciphers', which was already
16080 mentioned in the documentation but had not been implemented.
16081 (This option is not yet really useful because even the additional
16082 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16086 * Initial DSO code added into libcrypto for letting OpenSSL (and
16087 OpenSSL-based applications) load shared libraries and bind to
16088 them in a portable way.
16090 *Geoff Thorpe, with contributions from Richard Levitte*
16092 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16094 * Make sure _lrotl and _lrotr are only used with MSVC.
16096 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16097 (the default implementation of RAND_status).
16099 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16100 to '-clrext' (= clear extensions), as intended and documented.
16101 *Bodo Moeller; inconsistency pointed out by Michael Attili
16102 <attili@amaxo.com>*
16104 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16105 was larger than the MD block size.
16107 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16109 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16110 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16111 using the passed key: if the passed key was a private key the result
16112 of X509_print(), for example, would be to print out all the private key
16117 * des_quad_cksum() byte order bug fix.
16118 *Ulf Möller, using the problem description in krb4-0.9.7, where
16119 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16121 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16124 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16126 * For easily testing in shell scripts whether some command
16127 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16128 returns with exit code 0 iff no command of the given name is available.
16129 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16130 the output goes to stdout and nothing is printed to stderr.
16131 Additional arguments are always ignored.
16133 Since for each cipher there is a command of the same name,
16134 the 'no-cipher' compilation switches can be tested this way.
16136 ('openssl no-XXX' is not able to detect pseudo-commands such
16137 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16141 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16145 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16146 is set; it will be thrown away anyway because each handshake creates
16148 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16149 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16150 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16151 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16155 * New s_client option -ign_eof: EOF at stdin is ignored, and
16156 'Q' and 'R' lose their special meanings (quit/renegotiate).
16157 This is part of what -quiet does; unlike -quiet, -ign_eof
16158 does not suppress any output.
16162 * Add compatibility options to the purpose and trust code. The
16163 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16164 accepts a certificate or CA, this was the previous behaviour,
16165 with all the associated security issues.
16167 X509_TRUST_COMPAT is the old trust behaviour: only and
16168 automatically trust self signed roots in certificate store. A
16169 new trust setting X509_TRUST_DEFAULT is used to specify that
16170 a purpose has no associated trust setting and it should instead
16171 use the value in the default purpose.
16175 * Fix the PKCS#8 DSA private key code so it decodes keys again
16176 and fix a memory leak.
16180 * In util/mkerr.pl (which implements 'make errors'), preserve
16181 reason strings from the previous version of the .c file, as
16182 the default to have only downcase letters (and digits) in
16183 automatically generated reasons codes is not always appropriate.
16187 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16188 using strerror. Previously, ERR_reason_error_string() returned
16189 library names as reason strings for SYSerr; but SYSerr is a special
16190 case where small numbers are errno values, not library numbers.
16194 * Add '-dsaparam' option to 'openssl dhparam' application. This
16195 converts DSA parameters into DH parameters. (When creating parameters,
16196 DSA_generate_parameters is used.)
16200 * Include 'length' (recommended exponent length) in C code generated
16201 by 'openssl dhparam -C'.
16205 * The second argument to set_label in perlasm was already being used
16206 so couldn't be used as a "file scope" flag. Moved to third argument
16211 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16212 instead of RAND_bytes for encryption IVs and salts.
16216 * Include RAND_status() into RAND_METHOD instead of implementing
16217 it only for md_rand.c Otherwise replacing the PRNG by calling
16218 RAND_set_rand_method would be impossible.
16222 * Don't let DSA_generate_key() enter an infinite loop if the random
16223 number generation fails.
16227 * New 'rand' application for creating pseudo-random output.
16231 * Added configuration support for Linux/IA64
16233 *Rolf Haberrecker <rolf@suse.de>*
16235 * Assembler module support for Mingw32.
16239 * Shared library support for HPUX (in shlib/).
16241 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16243 * Shared library support for Solaris gcc.
16245 *Lutz Behnke <behnke@trustcenter.de>*
16247 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16249 * PKCS7_encrypt() was adding text MIME headers twice because they
16250 were added manually and by SMIME_crlf_copy().
16254 * In bntest.c don't call BN_rand with zero bits argument.
16256 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16258 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16259 case was implemented. This caused BN_div_recp() to fail occasionally.
16263 * Add an optional second argument to the set_label() in the perl
16264 assembly language builder. If this argument exists and is set
16265 to 1 it signals that the assembler should use a symbol whose
16266 scope is the entire file, not just the current function. This
16267 is needed with MASM which uses the format label:: for this scope.
16269 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16271 * Change the ASN1 types so they are typedefs by default. Before
16272 almost all types were #define'd to ASN1_STRING which was causing
16273 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16278 * Change names of new functions to the new get1/get0 naming
16279 convention: After 'get1', the caller owns a reference count
16280 and has to call `..._free`; 'get0' returns a pointer to some
16281 data structure without incrementing reference counters.
16282 (Some of the existing 'get' functions increment a reference
16283 counter, some don't.)
16284 Similarly, 'set1' and 'add1' functions increase reference
16285 counters or duplicate objects.
16289 * Allow for the possibility of temp RSA key generation failure:
16290 the code used to assume it always worked and crashed on failure.
16294 * Fix potential buffer overrun problem in BIO_printf().
16295 *Ulf Möller, using public domain code by Patrick Powell; problem
16296 pointed out by David Sacerdote <das33@cornell.edu>*
16298 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16299 RAND_egd() and RAND_status(). In the command line application,
16300 the EGD socket can be specified like a seed file using RANDFILE
16305 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16306 Some CAs (e.g. Verisign) distribute certificates in this form.
16310 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16311 list to exclude them. This means that no special compilation option
16312 is needed to use anonymous DH: it just needs to be included in the
16317 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16318 EVP_MD_type. The old functionality is available in a new macro called
16319 EVP_MD_md(). Change code that uses it and update docs.
16323 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16324 where the `void *` argument is replaced by a function pointer argument.
16325 Previously `void *` was abused to point to functions, which works on
16326 many platforms, but is not correct. As these functions are usually
16327 called by macros defined in OpenSSL header files, most source code
16328 should work without changes.
16332 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16333 sections with information on -D... compiler switches used for
16334 compiling the library so that applications can see them. To enable
16335 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16336 must be defined. E.g.,
16337 #define OPENSSL_ALGORITHM_DEFINES
16338 #include <openssl/opensslconf.h>
16339 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16341 *Richard Levitte, Ulf and Bodo Möller*
16343 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16348 * Change the 'other' type in certificate aux info to a STACK_OF
16349 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16350 the required ASN1 format: arbitrary types determined by an OID.
16354 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16355 argument to 'req'. This is not because the function is newer or
16356 better than others it just uses the work 'NEW' in the certificate
16357 request header lines. Some software needs this.
16361 * Reorganise password command line arguments: now passwords can be
16362 obtained from various sources. Delete the PEM_cb function and make
16363 it the default behaviour: i.e. if the callback is NULL and the
16364 usrdata argument is not NULL interpret it as a null terminated pass
16365 phrase. If usrdata and the callback are NULL then the pass phrase
16366 is prompted for as usual.
16370 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16371 the support is automatically enabled. The resulting binaries will
16372 autodetect the card and use it if present.
16374 *Ben Laurie and Compaq Inc.*
16376 * Work around for Netscape hang bug. This sends certificate request
16377 and server done in one record. Since this is perfectly legal in the
16378 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16379 the bugs/SSLv3 entry for more info.
16383 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16387 * Add -rand argument to smime and pkcs12 applications and read/write
16392 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16396 * Add command line password options to the remaining applications.
16400 * Bug fix for BN_div_recp() for numerators with an even number of
16405 * More tests in bntest.c, and changed test_bn output.
16409 * ./config recognizes MacOS X now.
16413 * Bug fix for BN_div() when the first words of num and divisor are
16414 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16418 * Add support for various broken PKCS#8 formats, and command line
16419 options to produce them.
16423 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16424 get temporary BIGNUMs from a BN_CTX.
16428 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16433 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16434 include a #define from the old name to the new. The original intent
16435 was that statically linked binaries could for example just call
16436 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16437 link with digests. This never worked because SSLeay_add_all_digests()
16438 and SSLeay_add_all_ciphers() were in the same source file so calling
16439 one would link with the other. They are now in separate source files.
16443 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16447 * Use a less unusual form of the Miller-Rabin primality test (it used
16448 a binary algorithm for exponentiation integrated into the Miller-Rabin
16449 loop, our standard modexp algorithms are faster).
16453 * Support for the EBCDIC character set completed.
16455 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16457 * Source code cleanups: use const where appropriate, eliminate casts,
16458 use `void *` instead of `char *` in lhash.
16462 * Bugfix: ssl3_send_server_key_exchange was not restartable
16463 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16464 this the server could overwrite ephemeral keys that the client
16469 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16470 using 50 iterations of the Rabin-Miller test.
16472 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16473 iterations of the Rabin-Miller test as required by the appendix
16474 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16475 As BN_is_prime_fasttest includes trial division, DSA parameter
16476 generation becomes much faster.
16478 This implies a change for the callback functions in DSA_is_prime
16479 and DSA_generate_parameters: The callback function is called once
16480 for each positive witness in the Rabin-Miller test, not just
16481 occasionally in the inner loop; and the parameters to the
16482 callback function now provide an iteration count for the outer
16483 loop rather than for the current invocation of the inner loop.
16484 DSA_generate_parameters additionally can call the callback
16485 function with an 'iteration count' of -1, meaning that a
16486 candidate has passed the trial division test (when q is generated
16487 from an application-provided seed, trial division is skipped).
16491 * New function BN_is_prime_fasttest that optionally does trial
16492 division before starting the Rabin-Miller test and has
16493 an additional BN_CTX * argument (whereas BN_is_prime always
16494 has to allocate at least one BN_CTX).
16495 'callback(1, -1, cb_arg)' is called when a number has passed the
16496 trial division stage.
16500 * Fix for bug in CRL encoding. The validity dates weren't being handled
16505 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16509 * New function BN_pseudo_rand().
16513 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16514 bignum version of BN_from_montgomery() with the working code from
16515 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16520 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16521 made it impossible to use the same SSL_SESSION data structure in
16522 SSL2 clients in multiple threads.
16526 * The return value of RAND_load_file() no longer counts bytes obtained
16527 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16528 to seed the PRNG (previously an explicit byte count was required).
16530 *Ulf Möller, Bodo Möller*
16532 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16533 used `char *` instead of `void *` and had casts all over the place.
16537 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16541 * Retain source code compatibility for BN_prime_checks macro:
16542 BN_is_prime(..., BN_prime_checks, ...) now uses
16543 BN_prime_checks_for_size to determine the appropriate number of
16544 Rabin-Miller iterations.
16548 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16549 DH_CHECK_P_NOT_SAFE_PRIME.
16550 (Check if this is true? OpenPGP calls them "strong".)
16554 * Merge the functionality of "dh" and "gendh" programs into a new program
16555 "dhparam". The old programs are retained for now but will handle DH keys
16556 (instead of parameters) in future.
16560 * Make the ciphers, s_server and s_client programs check the return values
16561 when a new cipher list is set.
16565 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16566 ciphers. Before when the 56bit ciphers were enabled the sorting was
16569 The syntax for the cipher sorting has been extended to support sorting by
16570 cipher-strength (using the strength_bits hard coded in the tables).
16571 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16573 Fix a bug in the cipher-command parser: when supplying a cipher command
16574 string with an "undefined" symbol (neither command nor alphanumeric
16575 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16576 an error is flagged.
16578 Due to the strength-sorting extension, the code of the
16579 ssl_create_cipher_list() function was completely rearranged. I hope that
16580 the readability was also increased :-)
16582 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16584 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16585 for the first serial number and places 2 in the serial number file. This
16586 avoids problems when the root CA is created with serial number zero and
16587 the first user certificate has the same issuer name and serial number
16592 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16593 the new code. Add documentation for this stuff.
16597 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16598 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16599 structures and behave in an analogous way to the X509v3 functions:
16600 they shouldn't be called directly but wrapper functions should be used
16603 So we also now have some wrapper functions that call the X509at functions
16604 when passed certificate requests. (TO DO: similar things can be done with
16605 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16606 things. Some of these need some d2i or i2d and print functionality
16607 because they handle more complex structures.)
16611 * Add missing #ifndefs that caused missing symbols when building libssl
16612 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16613 NO_RSA in `ssl/s2*.c`.
16615 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16617 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16618 has a return value which indicates the quality of the random data
16619 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16620 error queue. New function RAND_pseudo_bytes() generates output that is
16621 guaranteed to be unique but not unpredictable. RAND_add is like
16622 RAND_seed, but takes an extra argument for an entropy estimate
16623 (RAND_seed always assumes full entropy).
16627 * Do more iterations of Rabin-Miller probable prime test (specifically,
16628 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16629 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16630 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16631 false-positive rate of at most 2^-80 for random input.
16635 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16639 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16640 in the 0.9.5 release), this returns the chain
16641 from an X509_CTX structure with a dup of the stack and all
16642 the X509 reference counts upped: so the stack will exist
16643 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16646 Also make SSL_SESSION_print() print out the verify return
16651 * Add manpage for the pkcs12 command. Also change the default
16652 behaviour so MAC iteration counts are used unless the new
16653 -nomaciter option is used. This improves file security and
16654 only older versions of MSIE (4.0 for example) need it.
16658 * Honor the no-xxx Configure options when creating .DEF files.
16662 * Add PKCS#10 attributes to field table: challengePassword,
16663 unstructuredName and unstructuredAddress. These are taken from
16664 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16665 international characters are used.
16667 More changes to X509_ATTRIBUTE code: allow the setting of types
16668 based on strings. Remove the 'loc' parameter when adding
16669 attributes because these will be a SET OF encoding which is sorted
16674 * Initial changes to the 'req' utility to allow request generation
16675 automation. This will allow an application to just generate a template
16676 file containing all the field values and have req construct the
16679 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16680 used all over the place including certificate requests and PKCS#7
16681 structures. They are currently handled manually where necessary with
16682 some primitive wrappers for PKCS#7. The new functions behave in a
16683 manner analogous to the X509 extension functions: they allow
16684 attributes to be looked up by NID and added.
16686 Later something similar to the X509V3 code would be desirable to
16687 automatically handle the encoding, decoding and printing of the
16688 more complex types. The string types like challengePassword can
16689 be handled by the string table functions.
16691 Also modified the multi byte string table handling. Now there is
16692 a 'global mask' which masks out certain types. The table itself
16693 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16694 is useful when for example there is only one permissible type
16695 (as in countryName) and using the mask might result in no valid
16700 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16701 SSL_get_peer_finished to allow applications to obtain the latest
16702 Finished messages sent to the peer or expected from the peer,
16703 respectively. (SSL_get_peer_finished is usually the Finished message
16704 actually received from the peer, otherwise the protocol will be aborted.)
16706 As the Finished message are message digests of the complete handshake
16707 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16708 be used for external authentication procedures when the authentication
16709 provided by SSL/TLS is not desired or is not enough.
16713 * Enhanced support for Alpha Linux is added. Now ./config checks if
16714 the host supports BWX extension and if Compaq C is present on the
16715 $PATH. Just exploiting of the BWX extension results in 20-30%
16716 performance kick for some algorithms, e.g. DES and RC4 to mention
16717 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16722 * Add support for MS "fast SGC". This is arguably a violation of the
16723 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16724 weak crypto and after checking the certificate is SGC a second one
16725 with strong crypto. MS SGC stops the first handshake after receiving
16726 the server certificate message and sends a second client hello. Since
16727 a server will typically do all the time consuming operations before
16728 expecting any further messages from the client (server key exchange
16729 is the most expensive) there is little difference between the two.
16731 To get OpenSSL to support MS SGC we have to permit a second client
16732 hello message after we have sent server done. In addition we have to
16733 reset the MAC if we do get this second client hello.
16737 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16738 if a DER encoded private key is RSA or DSA traditional format. Changed
16739 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16740 format DER encoded private key. Newer code should use PKCS#8 format which
16741 has the key type encoded in the ASN1 structure. Added DER private key
16742 support to pkcs8 application.
16746 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16747 ciphersuites has been selected (as required by the SSL 3/TLS 1
16748 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16749 is set, we interpret this as a request to violate the specification
16750 (the worst that can happen is a handshake failure, and 'correct'
16751 behaviour would result in a handshake failure anyway).
16755 * In SSL_CTX_add_session, take into account that there might be multiple
16756 SSL_SESSION structures with the same session ID (e.g. when two threads
16757 concurrently obtain them from an external cache).
16758 The internal cache can handle only one SSL_SESSION with a given ID,
16759 so if there's a conflict, we now throw out the old one to achieve
16764 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16765 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16766 some routines that use cipher OIDs: some ciphers do not have OIDs
16767 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16772 * Simplify the trust setting structure and code. Now we just have
16773 two sequences of OIDs for trusted and rejected settings. These will
16774 typically have values the same as the extended key usage extension
16775 and any application specific purposes.
16777 The trust checking code now has a default behaviour: it will just
16778 check for an object with the same NID as the passed id. Functions can
16779 be provided to override either the default behaviour or the behaviour
16780 for a given id. SSL client, server and email already have functions
16781 in place for compatibility: they check the NID and also return "trusted"
16782 if the certificate is self signed.
16786 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16787 traditional format into an EVP_PKEY structure.
16791 * Add a password callback function PEM_cb() which either prompts for
16792 a password if usr_data is NULL or otherwise assumes it is a null
16793 terminated password. Allow passwords to be passed on command line
16794 environment or config files in a few more utilities.
16798 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16799 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16800 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16801 Update documentation.
16805 * Support for ASN1 "NULL" type. This could be handled before by using
16806 ASN1_TYPE but there wasn't any function that would try to read a NULL
16807 and produce an error if it couldn't. For compatibility we also have
16808 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16809 don't allocate anything because they don't need to.
16813 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16816 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16818 * Rebuild of the memory allocation routines used by OpenSSL code and
16819 possibly others as well. The purpose is to make an interface that
16820 provide hooks so anyone can build a separate set of allocation and
16821 deallocation routines to be used by OpenSSL, for example memory
16822 pool implementations, or something else, which was previously hard
16823 since Malloc(), Realloc() and Free() were defined as macros having
16824 the values malloc, realloc and free, respectively (except for Win32
16825 compilations). The same is provided for memory debugging code.
16826 OpenSSL already comes with functionality to find memory leaks, but
16827 this gives people a chance to debug other memory problems.
16829 With these changes, a new set of functions and macros have appeared:
16831 CRYPTO_set_mem_debug_functions() [F]
16832 CRYPTO_get_mem_debug_functions() [F]
16833 CRYPTO_dbg_set_options() [F]
16834 CRYPTO_dbg_get_options() [F]
16835 CRYPTO_malloc_debug_init() [M]
16837 The memory debug functions are NULL by default, unless the library
16838 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16839 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16840 gives the standard debugging functions that come with OpenSSL) or
16841 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16842 provided by the library user) must be used. When the standard
16843 debugging functions are used, CRYPTO_dbg_set_options can be used to
16844 request additional information:
16845 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16846 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16848 Also, things like CRYPTO_set_mem_functions will always give the
16849 expected result (the new set of functions is used for allocation
16850 and deallocation) at all times, regardless of platform and compiler
16853 To finish it up, some functions that were never use in any other
16854 way than through macros have a new API and new semantic:
16856 CRYPTO_dbg_malloc()
16857 CRYPTO_dbg_realloc()
16860 All macros of value have retained their old syntax.
16862 *Richard Levitte and Bodo Moeller*
16864 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16865 ordering of SMIMECapabilities wasn't in "strength order" and there
16866 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16871 * Some ASN1 types with illegal zero length encoding (INTEGER,
16872 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16874 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16876 * Merge in my S/MIME library for OpenSSL. This provides a simple
16877 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16878 functionality to handle multipart/signed properly) and a utility
16879 called 'smime' to call all this stuff. This is based on code I
16880 originally wrote for Celo who have kindly allowed it to be
16881 included in OpenSSL.
16885 * Add variants des_set_key_checked and des_set_key_unchecked of
16886 des_set_key (aka des_key_sched). Global variable des_check_key
16887 decides which of these is called by des_set_key; this way
16888 des_check_key behaves as it always did, but applications and
16889 the library itself, which was buggy for des_check_key == 1,
16890 have a cleaner way to pick the version they need.
16894 * New function PKCS12_newpass() which changes the password of a
16899 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16900 dynamic mix. In both cases the ids can be used as an index into the
16901 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16902 functions so they accept a list of the field values and the
16903 application doesn't need to directly manipulate the X509_TRUST
16908 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16913 * Modify the way the V3 extension code looks up extensions. This now
16914 works in a similar way to the object code: we have some "standard"
16915 extensions in a static table which is searched with OBJ_bsearch()
16916 and the application can add dynamic ones if needed. The file
16917 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16918 updated whenever a new extension is added to the core code and kept
16919 in ext_nid order. There is a simple program 'tabtest.c' which checks
16920 this. New extensions are not added too often so this file can readily
16921 be maintained manually.
16923 There are two big advantages in doing things this way. The extensions
16924 can be looked up immediately and no longer need to be "added" using
16925 X509V3_add_standard_extensions(): this function now does nothing.
16926 Side note: I get *lots* of email saying the extension code doesn't
16927 work because people forget to call this function.
16928 Also no dynamic allocation is done unless new extensions are added:
16929 so if we don't add custom extensions there is no need to call
16930 X509V3_EXT_cleanup().
16934 * Modify enc utility's salting as follows: make salting the default. Add a
16935 magic header, so unsalted files fail gracefully instead of just decrypting
16936 to garbage. This is because not salting is a big security hole, so people
16937 should be discouraged from doing it.
16941 * Fixes and enhancements to the 'x509' utility. It allowed a message
16942 digest to be passed on the command line but it only used this
16943 parameter when signing a certificate. Modified so all relevant
16944 operations are affected by the digest parameter including the
16945 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16946 DSA key was used because it didn't fix the digest.
16950 * Initial certificate chain verify code. Currently tests the untrusted
16951 certificates for consistency with the verify purpose (which is set
16952 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16954 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16955 this is because it will reject chains with invalid extensions whereas
16956 every previous version of OpenSSL and SSLeay made no checks at all.
16958 Trust code: checks the root CA for the relevant trust settings. Trust
16959 settings have an initial value consistent with the verify purpose: e.g.
16960 if the verify purpose is for SSL client use it expects the CA to be
16961 trusted for SSL client use. However the default value can be changed to
16962 permit custom trust settings: one example of this would be to only trust
16963 certificates from a specific "secure" set of CAs.
16965 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16966 which should be used for version portability: especially since the
16967 verify structure is likely to change more often now.
16969 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16970 to set them. If not set then assume SSL clients will verify SSL servers
16973 Two new options to the verify program: -untrusted allows a set of
16974 untrusted certificates to be passed in and -purpose which sets the
16975 intended purpose of the certificate. If a purpose is set then the
16976 new chain verify code is used to check extension consistency.
16980 * Support for the authority information access extension.
16984 * Modify RSA and DSA PEM read routines to transparently handle
16985 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16986 public keys in a format compatible with certificate
16987 SubjectPublicKeyInfo structures. Unfortunately there were already
16988 functions called *_PublicKey_* which used various odd formats so
16989 these are retained for compatibility: however the DSA variants were
16990 never in a public release so they have been deleted. Changed dsa/rsa
16991 utilities to handle the new format: note no releases ever handled public
16992 keys so we should be OK.
16994 The primary motivation for this change is to avoid the same fiasco
16995 that dogs private keys: there are several incompatible private key
16996 formats some of which are standard and some OpenSSL specific and
16997 require various evil hacks to allow partial transparent handling and
16998 even then it doesn't work with DER formats. Given the option anything
16999 other than PKCS#8 should be dumped: but the other formats have to
17000 stay in the name of compatibility.
17002 With public keys and the benefit of hindsight one standard format
17003 is used which works with EVP_PKEY, RSA or DSA structures: though
17004 it clearly returns an error if you try to read the wrong kind of key.
17006 Added a -pubkey option to the 'x509' utility to output the public key.
17007 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17008 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17009 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17010 that do the same as the `EVP_PKEY_assign_*()` except they up the
17011 reference count of the added key (they don't "swallow" the
17016 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17017 CRLs would fail if the file contained no certificates or no CRLs:
17018 added a new function to read in both types and return the number
17019 read: this means that if none are read it will be an error. The
17020 DER versions of the certificate and CRL reader would always fail
17021 because it isn't possible to mix certificates and CRLs in DER format
17022 without choking one or the other routine. Changed this to just read
17023 a certificate: this is the best we can do. Also modified the code
17024 in `apps/verify.c` to take notice of return codes: it was previously
17025 attempting to read in certificates from NULL pointers and ignoring
17026 any errors: this is one reason why the cert and CRL reader seemed
17027 to work. It doesn't check return codes from the default certificate
17028 routines: these may well fail if the certificates aren't installed.
17032 * Code to support otherName option in GeneralName.
17036 * First update to verify code. Change the verify utility
17037 so it warns if it is passed a self signed certificate:
17038 for consistency with the normal behaviour. X509_verify
17039 has been modified to it will now verify a self signed
17040 certificate if *exactly* the same certificate appears
17041 in the store: it was previously impossible to trust a
17042 single self signed certificate. This means that:
17043 openssl verify ss.pem
17044 now gives a warning about a self signed certificate but
17045 openssl verify -CAfile ss.pem ss.pem
17050 * For servers, store verify_result in SSL_SESSION data structure
17051 (and add it to external session representation).
17052 This is needed when client certificate verifications fails,
17053 but an application-provided verification callback (set by
17054 SSL_CTX_set_cert_verify_callback) allows accepting the session
17055 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17056 but returns 1): When the session is reused, we have to set
17057 ssl->verify_result to the appropriate error code to avoid
17060 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17062 * Fix a bug in the new PKCS#7 code: it didn't consider the
17063 case in PKCS7_dataInit() where the signed PKCS7 structure
17064 didn't contain any existing data because it was being created.
17066 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17068 * Add a salt to the key derivation routines in enc.c. This
17069 forms the first 8 bytes of the encrypted file. Also add a
17070 -S option to allow a salt to be input on the command line.
17074 * New function X509_cmp(). Oddly enough there wasn't a function
17075 to compare two certificates. We do this by working out the SHA1
17076 hash and comparing that. X509_cmp() will be needed by the trust
17081 * SSL_get1_session() is like SSL_get_session(), but increments
17082 the reference count in the SSL_SESSION returned.
17084 *Geoff Thorpe <geoff@eu.c2.net>*
17086 * Fix for 'req': it was adding a null to request attributes.
17087 Also change the X509_LOOKUP and X509_INFO code to handle
17088 certificate auxiliary information.
17092 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17097 * Add the possibility to add extra information to the memory leak
17098 detecting output, to form tracebacks, showing from where each
17099 allocation was originated: CRYPTO_push_info("constant string") adds
17100 the string plus current file name and line number to a per-thread
17101 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17102 is like calling CYRPTO_pop_info() until the stack is empty.
17103 Also updated memory leak detection code to be multi-thread-safe.
17107 * Add options -text and -noout to pkcs7 utility and delete the
17108 encryption options which never did anything. Update docs.
17112 * Add options to some of the utilities to allow the pass phrase
17113 to be included on either the command line (not recommended on
17114 OSes like Unix) or read from the environment. Update the
17115 manpages and fix a few bugs.
17119 * Add a few manpages for some of the openssl commands.
17123 * Fix the -revoke option in ca. It was freeing up memory twice,
17124 leaking and not finding already revoked certificates.
17128 * Extensive changes to support certificate auxiliary information.
17129 This involves the use of X509_CERT_AUX structure and X509_AUX
17130 functions. An X509_AUX function such as PEM_read_X509_AUX()
17131 can still read in a certificate file in the usual way but it
17132 will also read in any additional "auxiliary information". By
17133 doing things this way a fair degree of compatibility can be
17134 retained: existing certificates can have this information added
17135 using the new 'x509' options.
17137 Current auxiliary information includes an "alias" and some trust
17138 settings. The trust settings will ultimately be used in enhanced
17139 certificate chain verification routines: currently a certificate
17140 can only be trusted if it is self signed and then it is trusted
17145 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17146 The problem was that one of the replacement routines had not been working
17147 since SSLeay releases. For now the offending routine has been replaced
17148 with non-optimised assembler. Even so, this now gives around 95%
17149 performance improvement for 1024 bit RSA signs.
17153 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17154 handling. Most clients have the effective key size in bits equal to
17155 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17156 A few however don't do this and instead use the size of the decrypted key
17157 to determine the RC2 key length and the AlgorithmIdentifier to determine
17158 the effective key length. In this case the effective key length can still
17159 be 40 bits but the key length can be 168 bits for example. This is fixed
17160 by manually forcing an RC2 key into the EVP_PKEY structure because the
17161 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17162 the key length and effective key length are equal.
17166 * Add a bunch of functions that should simplify the creation of
17167 X509_NAME structures. Now you should be able to do:
17168 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17169 and have it automatically work out the correct field type and fill in
17170 the structures. The more adventurous can try:
17171 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17172 and it will (hopefully) work out the correct multibyte encoding.
17176 * Change the 'req' utility to use the new field handling and multibyte
17177 copy routines. Before the DN field creation was handled in an ad hoc
17178 way in req, ca, and x509 which was rather broken and didn't support
17179 BMPStrings or UTF8Strings. Since some software doesn't implement
17180 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17181 using the dirstring_type option. See the new comment in the default
17182 openssl.cnf for more info.
17186 * Make crypto/rand/md_rand.c more robust:
17187 - Assure unique random numbers after fork().
17188 - Make sure that concurrent threads access the global counter and
17189 md serializably so that we never lose entropy in them
17190 or use exactly the same state in multiple threads.
17191 Access to the large state is not always serializable because
17192 the additional locking could be a performance killer, and
17193 md should be large enough anyway.
17197 * New file `apps/app_rand.c` with commonly needed functionality
17198 for handling the random seed file.
17200 Use the random seed file in some applications that previously did not:
17202 dsaparam -genkey (which also ignored its '-rand' option),
17205 x509 (when signing).
17206 Except on systems with /dev/urandom, it is crucial to have a random
17207 seed file at least for key creation, DSA signing, and for DH exchanges;
17208 for RSA signatures we could do without one.
17210 gendh and gendsa (unlike genrsa) used to read only the first byte
17211 of each file listed in the '-rand' option. The function as previously
17212 found in genrsa is now in app_rand.c and is used by all programs
17213 that support '-rand'.
17217 * In RAND_write_file, use mode 0600 for creating files;
17218 don't just chmod when it may be too late.
17222 * Report an error from X509_STORE_load_locations
17223 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17227 * New function ASN1_mbstring_copy() this copies a string in either
17228 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17229 into an ASN1_STRING type. A mask of permissible types is passed
17230 and it chooses the "minimal" type to use or an error if not type
17235 * Add function equivalents to the various macros in asn1.h. The old
17236 macros are retained with an `M_` prefix. Code inside the library can
17237 use the `M_` macros. External code (including the openssl utility)
17238 should *NOT* in order to be "shared library friendly".
17242 * Add various functions that can check a certificate's extensions
17243 to see if it usable for various purposes such as SSL client,
17244 server or S/MIME and CAs of these types. This is currently
17245 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17246 verification. Also added a -purpose flag to x509 utility to
17247 print out all the purposes.
17251 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17256 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17257 for, obtain and decode and extension and obtain its critical flag.
17258 This allows all the necessary extension code to be handled in a
17259 single function call.
17263 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17264 platforms. See crypto/rc4/rc4_enc.c for further details.
17268 * New -noout option to asn1parse. This causes no output to be produced
17269 its main use is when combined with -strparse and -out to extract data
17270 from a file (which may not be in ASN.1 format).
17274 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17275 when producing the local key id.
17277 *Richard Levitte <levitte@stacken.kth.se>*
17279 * New option -dhparam in s_server. This allows a DH parameter file to be
17280 stated explicitly. If it is not stated then it tries the first server
17281 certificate file. The previous behaviour hard coded the filename
17286 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17287 a public key to be input or output. For example:
17288 openssl rsa -in key.pem -pubout -out pubkey.pem
17289 Also added necessary DSA public key functions to handle this.
17293 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17294 in the message. This was handled by allowing
17295 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17297 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17299 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17300 to the end of the strings whereas this didn't. This would cause problems
17301 if strings read with d2i_ASN1_bytes() were later modified.
17303 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17305 * Fix for base64 decode bug. When a base64 bio reads only one line of
17306 data and it contains EOF it will end up returning an error. This is
17307 caused by input 46 bytes long. The cause is due to the way base64
17308 BIOs find the start of base64 encoded data. They do this by trying a
17309 trial decode on each line until they find one that works. When they
17310 do a flag is set and it starts again knowing it can pass all the
17311 data directly through the decoder. Unfortunately it doesn't reset
17312 the context it uses. This means that if EOF is reached an attempt
17313 is made to pass two EOFs through the context and this causes the
17314 resulting error. This can also cause other problems as well. As is
17315 usual with these problems it takes *ages* to find and the fix is
17316 trivial: move one line.
17318 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17320 * Ugly workaround to get s_client and s_server working under Windows. The
17321 old code wouldn't work because it needed to select() on sockets and the
17322 tty (for keypresses and to see if data could be written). Win32 only
17323 supports select() on sockets so we select() with a 1s timeout on the
17324 sockets and then see if any characters are waiting to be read, if none
17325 are present then we retry, we also assume we can always write data to
17326 the tty. This isn't nice because the code then blocks until we've
17327 received a complete line of data and it is effectively polling the
17328 keyboard at 1s intervals: however it's quite a bit better than not
17329 working at all :-) A dedicated Windows application might handle this
17330 with an event loop for example.
17334 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17335 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17336 will be called when RSA_sign() and RSA_verify() are used. This is useful
17337 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17338 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17339 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17340 This necessitated the support of an extra signature type NID_md5_sha1
17341 for SSL signatures and modifications to the SSL library to use it instead
17342 of calling RSA_public_decrypt() and RSA_private_encrypt().
17346 * Add new -verify -CAfile and -CApath options to the crl program, these
17347 will lookup a CRL issuers certificate and verify the signature in a
17348 similar way to the verify program. Tidy up the crl program so it
17349 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17350 less strict. It will now permit CRL extensions even if it is not
17351 a V2 CRL: this will allow it to tolerate some broken CRLs.
17355 * Initialize all non-automatic variables each time one of the openssl
17356 sub-programs is started (this is necessary as they may be started
17357 multiple times from the "OpenSSL>" prompt).
17359 *Lennart Bang, Bodo Moeller*
17361 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17362 removing all other RSA functionality (this is what NO_RSA does). This
17363 is so (for example) those in the US can disable those operations covered
17364 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17369 * Non-copying interface to BIO pairs.
17370 (still largely untested)
17374 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17375 ASCII string. This was handled independently in various places before.
17379 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17380 UTF8 strings a character at a time.
17384 * Use client_version from client hello to select the protocol
17385 (s23_srvr.c) and for RSA client key exchange verification
17386 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17390 * Add various utility functions to handle SPKACs, these were previously
17391 handled by poking round in the structure internals. Added new function
17392 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17393 print, verify and generate SPKACs. Based on an original idea from
17394 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17398 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17402 * Allow the config file extension section to be overwritten on the
17403 command line. Based on an original idea from Massimiliano Pala
17404 <madwolf@comune.modena.it>. The new option is called -extensions
17405 and can be applied to ca, req and x509. Also -reqexts to override
17406 the request extensions in req and -crlexts to override the crl extensions
17411 * Add new feature to the SPKAC handling in ca. Now you can include
17412 the same field multiple times by preceding it by "XXXX." for example:
17415 this is the same syntax as used in the req config file.
17419 * Allow certificate extensions to be added to certificate requests. These
17420 are specified in a 'req_extensions' option of the req section of the
17421 config file. They can be printed out with the -text option to req but
17422 are otherwise ignored at present.
17426 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17427 data read consists of only the final block it would not decrypted because
17428 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17429 A misplaced 'break' also meant the decrypted final block might not be
17430 copied until the next read.
17434 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17435 a few extra parameters to the DH structure: these will be useful if
17436 for example we want the value of 'q' or implement X9.42 DH.
17440 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17441 provides hooks that allow the default DSA functions or functions on a
17442 "per key" basis to be replaced. This allows hardware acceleration and
17443 hardware key storage to be handled without major modification to the
17444 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17445 associated functions.
17449 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17450 as "read only": it can't be written to and the buffer it points to will
17451 not be freed. Reading from a read only BIO is much more efficient than
17452 a normal memory BIO. This was added because there are several times when
17453 an area of memory needs to be read from a BIO. The previous method was
17454 to create a memory BIO and write the data to it, this results in two
17455 copies of the data and an O(n^2) reading algorithm. There is a new
17456 function BIO_new_mem_buf() which creates a read only memory BIO from
17457 an area of memory. Also modified the PKCS#7 routines to use read only
17462 * Bugfix: ssl23_get_client_hello did not work properly when called in
17463 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17464 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17465 but a retry condition occurred while trying to read the rest.
17469 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17470 NID_pkcs7_encrypted by default: this was wrong since this should almost
17471 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17472 the encrypted data type: this is a more sensible place to put it and it
17473 allows the PKCS#12 code to be tidied up that duplicated this
17478 * Changed obj_dat.pl script so it takes its input and output files on
17479 the command line. This should avoid shell escape redirection problems
17484 * Initial support for certificate extension requests, these are included
17485 in things like Xenroll certificate requests. Included functions to allow
17486 extensions to be obtained and added.
17490 * -crlf option to s_client and s_server for sending newlines as
17491 CRLF (as required by many protocols).
17495 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17497 * Install libRSAglue.a when OpenSSL is built with RSAref.
17499 *Ralf S. Engelschall*
17501 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17503 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17505 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17510 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17511 DH parameters/keys (q is lost during that conversion, but the resulting
17512 DH parameters contain its length).
17514 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17515 much faster than DH_generate_parameters (which creates parameters
17516 where `p = 2*q + 1`), and also the smaller q makes DH computations
17517 much more efficient (160-bit exponentiation instead of 1024-bit
17518 exponentiation); so this provides a convenient way to support DHE
17519 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17520 utter importance to use
17521 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17523 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17524 when such DH parameters are used, because otherwise small subgroup
17525 attacks may become possible!
17529 * Avoid memory leak in i2d_DHparams.
17533 * Allow the -k option to be used more than once in the enc program:
17534 this allows the same encrypted message to be read by multiple recipients.
17538 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17539 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17540 it will always use the numerical form of the OID, even if it has a short
17545 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17546 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17547 otherwise bn_mod_exp was called. In the case of hardware keys for example
17548 no private key components need be present and it might store extra data
17549 in the RSA structure, which cannot be accessed from bn_mod_exp.
17550 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17551 private key operations.
17555 * Added support for SPARC Linux.
17559 * pem_password_cb function type incompatibly changed from
17560 typedef int pem_password_cb(char *buf, int size, int rwflag);
17562 ....(char *buf, int size, int rwflag, void *userdata);
17563 so that applications can pass data to their callbacks:
17564 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17565 additional void * argument, which is just handed through whenever
17566 the password callback is called.
17568 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17570 New function SSL_CTX_set_default_passwd_cb_userdata.
17572 Compatibility note: As many C implementations push function arguments
17573 onto the stack in reverse order, the new library version is likely to
17574 interoperate with programs that have been compiled with the old
17575 pem_password_cb definition (PEM_whatever takes some data that
17576 happens to be on the stack as its last argument, and the callback
17577 just ignores this garbage); but there is no guarantee whatsoever that
17580 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17581 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17582 problems not only on Windows, but also on some Unix platforms.
17583 To avoid problematic command lines, these definitions are now in an
17584 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17585 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17589 * MIPS III/IV assembler module is reimplemented.
17593 * More DES library cleanups: remove references to srand/rand and
17594 delete an unused file.
17598 * Add support for the free Netwide assembler (NASM) under Win32,
17599 since not many people have MASM (ml) and it can be hard to obtain.
17600 This is currently experimental but it seems to work OK and pass all
17601 the tests. Check out INSTALL.W32 for info.
17605 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17606 without temporary keys kept an extra copy of the server key,
17607 and connections with temporary keys did not free everything in case
17612 * New function RSA_check_key and new openssl rsa option -check
17613 for verifying the consistency of RSA keys.
17615 *Ulf Moeller, Bodo Moeller*
17617 * Various changes to make Win32 compile work:
17618 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17619 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17620 comparison" warnings.
17621 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17625 * Add a debugging option to PKCS#5 v2 key generation function: when
17626 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17627 derived keys are printed to stderr.
17631 * Copy the flags in ASN1_STRING_dup().
17633 *Roman E. Pavlov <pre@mo.msk.ru>*
17635 * The x509 application mishandled signing requests containing DSA
17636 keys when the signing key was also DSA and the parameters didn't match.
17638 It was supposed to omit the parameters when they matched the signing key:
17639 the verifying software was then supposed to automatically use the CA's
17640 parameters if they were absent from the end user certificate.
17642 Omitting parameters is no longer recommended. The test was also
17643 the wrong way round! This was probably due to unusual behaviour in
17644 EVP_cmp_parameters() which returns 1 if the parameters match.
17645 This meant that parameters were omitted when they *didn't* match and
17646 the certificate was useless. Certificates signed with 'ca' didn't have
17649 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17651 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17652 The interface is as follows:
17653 Applications can use
17654 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17655 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17656 "off" is now the default.
17657 The library internally uses
17658 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17659 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17660 to disable memory-checking temporarily.
17662 Some inconsistent states that previously were possible (and were
17663 even the default) are now avoided.
17665 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17666 with each memory chunk allocated; this is occasionally more helpful
17667 than just having a counter.
17669 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17671 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17676 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17677 which largely parallels "options", but is for changing API behaviour,
17678 whereas "options" are about protocol behaviour.
17679 Initial "mode" flags are:
17681 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17682 a single record has been written.
17683 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17684 retries use the same buffer location.
17685 (But all of the contents must be
17690 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17693 * Fix problems with no-hmac etc.
17695 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17697 * New functions RSA_get_default_method(), RSA_set_method() and
17698 RSA_get_method(). These allows replacement of RSA_METHODs without having
17699 to mess around with the internals of an RSA structure.
17703 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17704 Also really enable memory leak checks in openssl.c and in some
17707 *Chad C. Mulligan, Bodo Moeller*
17709 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17710 up the length of negative integers. This has now been simplified to just
17711 store the length when it is first determined and use it later, rather
17712 than trying to keep track of where data is copied and updating it to
17714 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17716 * Add a new function PKCS7_signatureVerify. This allows the verification
17717 of a PKCS#7 signature but with the signing certificate passed to the
17718 function itself. This contrasts with PKCS7_dataVerify which assumes the
17719 certificate is present in the PKCS#7 structure. This isn't always the
17720 case: certificates can be omitted from a PKCS#7 structure and be
17721 distributed by "out of band" means (such as a certificate database).
17725 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17726 function prototypes in pem.h, also change util/mkdef.pl to add the
17727 necessary function names.
17731 * mk1mf.pl (used by Windows builds) did not properly read the
17732 options set by Configure in the top level Makefile, and Configure
17733 was not even able to write more than one option correctly.
17734 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17738 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17739 file to be loaded from a BIO or FILE pointer. The BIO version will
17740 for example allow memory BIOs to contain config info.
17744 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17745 Whoever hopes to achieve shared-library compatibility across versions
17746 must use this, not the compile-time macro.
17747 (Exercise 0.9.4: Which is the minimum library version required by
17749 Note: All this applies only to multi-threaded programs, others don't
17754 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17755 through a BIO pair triggered the default case, i.e.
17756 SSLerr(...,SSL_R_UNKNOWN_STATE).
17760 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17761 can use the SSL library even if none of the specific BIOs is
17766 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17767 for the encoded length.
17769 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17771 * Add initial documentation of the X509V3 functions.
17775 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17776 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17777 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17778 secure PKCS#8 private key format with a high iteration count.
17782 * Fix determination of Perl interpreter: A perl or perl5
17783 *directory* in $PATH was also accepted as the interpreter.
17785 *Ralf S. Engelschall*
17787 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17788 wrong with it but it was very old and did things like calling
17789 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17790 unusual formatting.
17794 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17795 to use the new extension code.
17799 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17800 with macros. This should make it easier to change their form, add extra
17801 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17806 * Add to configuration table a new entry that can specify an alternative
17807 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17808 according to Mark Crispin <MRC@Panda.COM>.
17812 * DES CBC did not update the IV. Weird.
17816 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17817 Changing the behaviour of the former might break existing programs --
17818 where IV updating is needed, des_ncbc_encrypt can be used.
17821 * When bntest is run from "make test" it drives bc to check its
17822 calculations, as well as internally checking them. If an internal check
17823 fails, it needs to cause bc to give a non-zero result or make test carries
17824 on without noticing the failure. Fixed.
17828 * DES library cleanups.
17832 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17833 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17834 ciphers. NOTE: although the key derivation function has been verified
17835 against some published test vectors it has not been extensively tested
17836 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17841 * Instead of "mkdir -p", which is not fully portable, use new
17842 Perl script "util/mkdir-p.pl".
17846 * Rewrite the way password based encryption (PBE) is handled. It used to
17847 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17848 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17849 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17850 the 'parameter' field of the AlgorithmIdentifier is passed to the
17851 underlying key generation function so it must do its own ASN1 parsing.
17852 This has also changed the EVP_PBE_CipherInit() function which now has a
17853 'parameter' argument instead of literal salt and iteration count values
17854 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17858 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17859 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17860 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17861 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17862 value was just used as a "magic string" and not used directly its
17863 value doesn't matter.
17867 * Introduce some semblance of const correctness to BN. Shame C doesn't
17872 * "linux-sparc64" configuration (ultrapenguin).
17874 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17875 "linux-sparc" configuration.
17877 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17879 * config now generates no-xxx options for missing ciphers.
17883 * Support the EBCDIC character set (work in progress).
17884 File ebcdic.c not yet included because it has a different license.
17886 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17888 * Support BS2000/OSD-POSIX.
17890 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17892 * Make callbacks for key generation use `void *` instead of `char *`.
17896 * Make S/MIME samples compile (not yet tested).
17900 * Additional typesafe stacks.
17904 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17908 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17910 * New configuration variant "sco5-gcc".
17912 * Updated some demos.
17914 *Sean O Riordain, Wade Scholine*
17916 * Add missing BIO_free at exit of pkcs12 application.
17920 * Fix memory leak in conf.c.
17924 * Updates for Win32 to assembler version of MD5.
17928 * Set #! path to perl in `apps/der_chop` to where we found it
17929 instead of using a fixed path.
17933 * SHA library changes for irix64-mips4-cc.
17937 * Improvements for VMS support.
17941 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17943 * Bignum library bug fix. IRIX 6 passes "make test" now!
17944 This also avoids the problems with SC4.2 and unpatched SC5.
17946 *Andy Polyakov <appro@fy.chalmers.se>*
17948 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17949 These are required because of the typesafe stack would otherwise break
17950 existing code. If old code used a structure member which used to be STACK
17951 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17952 sk_num or sk_value it would produce an error because the num, data members
17953 are not present in STACK_OF. Now it just produces a warning. sk_set
17954 replaces the old method of assigning a value to sk_value
17955 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17956 that does this will no longer work (and should use sk_set instead) but
17957 this could be regarded as a "questionable" behaviour anyway.
17961 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17962 correctly handle encrypted S/MIME data.
17966 * Change type of various DES function arguments from des_cblock
17967 (which means, in function argument declarations, pointer to char)
17968 to des_cblock * (meaning pointer to array with 8 char elements),
17969 which allows the compiler to do more typechecking; it was like
17970 that back in SSLeay, but with lots of ugly casts.
17972 Introduce new type const_des_cblock.
17976 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17977 problems: find RecipientInfo structure that matches recipient certificate
17978 and initialise the ASN1 structures properly based on passed cipher.
17982 * Belatedly make the BN tests actually check the results.
17986 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17987 to and from BNs: it was completely broken. New compilation option
17988 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17989 key elements as negative integers.
17993 * Reorganize and speed up MD5.
17995 *Andy Polyakov <appro@fy.chalmers.se>*
17999 *Richard Levitte <richard@levitte.org>*
18001 * New option -out to asn1parse to allow the parsed structure to be
18002 output to a file. This is most useful when combined with the -strparse
18003 option to examine the output of things like OCTET STRINGS.
18007 * Make SSL library a little more fool-proof by not requiring any longer
18008 that `SSL_set_{accept,connect}_state` be called before
18009 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18010 in many applications because usually everything *appeared* to work as
18011 intended anyway -- now it really works as intended).
18015 * Move openssl.cnf out of lib/.
18019 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18020 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18021 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18023 *Ralf S. Engelschall*
18025 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18026 handle PKCS#7 enveloped data properly.
18028 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18030 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18031 copying pointers. The cert_st handling is changed by this in
18032 various ways (and thus what used to be known as ctx->default_cert
18033 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18034 any longer when s->cert does not give us what we need).
18035 ssl_cert_instantiate becomes obsolete by this change.
18036 As soon as we've got the new code right (possibly it already is?),
18037 we have solved a couple of bugs of the earlier code where s->cert
18038 was used as if it could not have been shared with other SSL structures.
18040 Note that using the SSL API in certain dirty ways now will result
18041 in different behaviour than observed with earlier library versions:
18042 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18043 does not influence s as it used to.
18045 In order to clean up things more thoroughly, inside SSL_SESSION
18046 we don't use CERT any longer, but a new structure SESS_CERT
18047 that holds per-session data (if available); currently, this is
18048 the peer's certificate chain and, for clients, the server's certificate
18049 and temporary key. CERT holds only those values that can have
18050 meaningful defaults in an SSL_CTX.
18054 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18055 from the internal representation. Various PKCS#7 fixes: remove some
18056 evil casts and set the enc_dig_alg field properly based on the signing
18061 * Allow PKCS#12 password to be set from the command line or the
18062 environment. Let 'ca' get its config file name from the environment
18063 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18068 * Allow certificate policies extension to use an IA5STRING for the
18069 organization field. This is contrary to the PKIX definition but
18070 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18075 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18076 without disallowing inline assembler and the like for non-pedantic builds.
18080 * Support Borland C++ builder.
18082 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18088 * SHA-1 cleanups and performance enhancements.
18090 *Andy Polyakov <appro@fy.chalmers.se>*
18092 * Sparc v8plus assembler for the bignum library.
18094 *Andy Polyakov <appro@fy.chalmers.se>*
18096 * Accept any -xxx and +xxx compiler options in Configure.
18100 * Update HPUX configuration.
18104 * Add missing `sk_<type>_unshift()` function to safestack.h
18106 *Ralf S. Engelschall*
18108 * New function SSL_CTX_use_certificate_chain_file that sets the
18109 "extra_cert"s in addition to the certificate. (This makes sense
18110 only for "PEM" format files, as chains as a whole are not
18115 * Support verify_depth from the SSL API.
18116 x509_vfy.c had what can be considered an off-by-one-error:
18117 Its depth (which was not part of the external interface)
18118 was actually counting the number of certificates in a chain;
18119 now it really counts the depth.
18123 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18124 instead of X509err, which often resulted in confusing error
18125 messages since the error codes are not globally unique
18126 (e.g. an alleged error in ssl3_accept when a certificate
18127 didn't match the private key).
18129 * New function SSL_CTX_set_session_id_context that allows to set a default
18130 value (so that you don't need SSL_set_session_id_context for each
18131 connection using the SSL_CTX).
18135 * OAEP decoding bug fix.
18139 * Support INSTALL_PREFIX for package builders, as proposed by
18144 * New Configure options "threads" and "no-threads". For systems
18145 where the proper compiler options are known (currently Solaris
18146 and Linux), "threads" is the default.
18150 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18154 * Install various scripts to $(OPENSSLDIR)/misc, not to
18155 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18156 such as /usr/local/bin.
18160 * "make linux-shared" to build shared libraries.
18162 *Niels Poppe <niels@netbox.org>*
18164 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18168 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18169 extension adding in x509 utility.
18173 * Remove NOPROTO sections and error code comments.
18177 * Partial rewrite of the DEF file generator to now parse the ANSI
18182 * New Configure options --prefix=DIR and --openssldir=DIR.
18186 * Complete rewrite of the error code script(s). It is all now handled
18187 by one script at the top level which handles error code gathering,
18188 header rewriting and C source file generation. It should be much better
18189 than the old method: it now uses a modified version of Ulf's parser to
18190 read the ANSI prototypes in all header files (thus the old K&R definitions
18191 aren't needed for error creation any more) and do a better job of
18192 translating function codes into names. The old 'ASN1 error code embedded
18193 in a comment' is no longer necessary and it doesn't use .err files which
18194 have now been deleted. Also the error code call doesn't have to appear all
18195 on one line (which resulted in some large lines...).
18199 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18203 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18204 0 (which usually indicates a closed connection), but continue reading.
18208 * Fix some race conditions.
18212 * Add support for CRL distribution points extension. Add Certificate
18213 Policies and CRL distribution points documentation.
18217 * Move the autogenerated header file parts to crypto/opensslconf.h.
18221 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18222 8 of keying material. Merlin has also confirmed interop with this fix
18223 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18225 *Merlin Hughes <merlin@baltimore.ie>*
18227 * Fix lots of warnings.
18229 *Richard Levitte <levitte@stacken.kth.se>*
18231 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18232 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18234 *Richard Levitte <levitte@stacken.kth.se>*
18236 * Fix problems with sizeof(long) == 8.
18238 *Andy Polyakov <appro@fy.chalmers.se>*
18240 * Change functions to ANSI C.
18244 * Fix typos in error codes.
18246 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18248 * Remove defunct assembler files from Configure.
18252 * SPARC v8 assembler BIGNUM implementation.
18254 *Andy Polyakov <appro@fy.chalmers.se>*
18256 * Support for Certificate Policies extension: both print and set.
18257 Various additions to support the r2i method this uses.
18261 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18262 return a const string when you are expecting an allocated buffer.
18266 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18267 types DirectoryString and DisplayText.
18271 * Add code to allow r2i extensions to access the configuration database,
18272 add an LHASH database driver and add several ctx helper functions.
18276 * Fix an evil bug in bn_expand2() which caused various BN functions to
18277 fail when they extended the size of a BIGNUM.
18281 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18282 support typesafe stack.
18286 * Fix typo in SSL_[gs]et_options().
18288 *Nils Frostberg <nils@medcom.se>*
18290 * Delete various functions and files that belonged to the (now obsolete)
18291 old X509V3 handling code.
18295 * New Configure option "rsaref".
18299 * Don't auto-generate pem.h.
18303 * Introduce type-safe ASN.1 SETs.
18307 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18309 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18311 * Introduce type-safe STACKs. This will almost certainly break lots of code
18312 that links with OpenSSL (well at least cause lots of warnings), but fear
18313 not: the conversion is trivial, and it eliminates loads of evil casts. A
18314 few STACKed things have been converted already. Feel free to convert more.
18315 In the fullness of time, I'll do away with the STACK type altogether.
18319 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18320 specified in `<certfile>` by updating the entry in the index.txt file.
18321 This way one no longer has to edit the index.txt file manually for
18322 revoking a certificate. The -revoke option does the gory details now.
18324 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18326 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18327 `-text` option at all and this way the `-noout -text` combination was
18328 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18330 *Ralf S. Engelschall*
18332 * Make sure a corresponding plain text error message exists for the
18333 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18334 verify callback function determined that a certificate was revoked.
18336 *Ralf S. Engelschall*
18338 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18339 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18340 all available ciphers including rc5, which was forgotten until now.
18341 In order to let the testing shell script know which algorithms
18342 are available, a new (up to now undocumented) command
18343 `openssl list-cipher-commands` is used.
18347 * Bugfix: s_client occasionally would sleep in select() when
18348 it should have checked SSL_pending() first.
18352 * New functions DSA_do_sign and DSA_do_verify to provide access to
18353 the raw DSA values prior to ASN.1 encoding.
18357 * Tweaks to Configure
18359 *Niels Poppe <niels@netbox.org>*
18361 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18366 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18370 * New config option to avoid instructions that are illegal on the 80386.
18371 The default code is faster, but requires at least a 486.
18375 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18376 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18377 same as SSL2_VERSION anyway.
18381 * New "-showcerts" option for s_client.
18385 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18386 application. Various cleanups and fixes.
18390 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18391 modify error routines to work internally. Add error codes and PBE init
18392 to library startup routines.
18396 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18397 packing functions to asn1 and evp. Changed function names and error
18398 codes along the way.
18402 * PKCS12 integration: and so it begins... First of several patches to
18403 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18404 objects to objects.h
18408 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18409 and display support for Thawte strong extranet extension.
18413 * Add LinuxPPC support.
18415 *Jeff Dubrule <igor@pobox.org>*
18417 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18418 bn_div_words in alpha.s.
18420 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18422 * Make sure the RSA OAEP test is skipped under -DRSAref because
18423 OAEP isn't supported when OpenSSL is built with RSAref.
18425 *Ulf Moeller <ulf@fitug.de>*
18427 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18428 so they no longer are missing under -DNOPROTO.
18430 *Soren S. Jorvang <soren@t.dk>*
18432 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18434 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18435 doesn't work when the session is reused. Coming soon!
18439 * Fix a security hole, that allows sessions to be reused in the wrong
18440 context thus bypassing client cert protection! All software that uses
18441 client certs and session caches in multiple contexts NEEDS PATCHING to
18442 allow session reuse! A fuller solution is in the works.
18444 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18446 * Some more source tree cleanups (removed obsolete files
18447 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18448 permission on "config" script to be executable) and a fix for the INSTALL
18451 *Ulf Moeller <ulf@fitug.de>*
18453 * Remove some legacy and erroneous uses of malloc, free instead of
18456 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18458 * Make rsa_oaep_test return non-zero on error.
18460 *Ulf Moeller <ulf@fitug.de>*
18462 * Add support for native Solaris shared libraries. Configure
18463 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18464 if someone would make that last step automatic.
18466 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18468 * ctx_size was not built with the right compiler during "make links". Fixed.
18472 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18473 except NULL ciphers". This means the default cipher list will no longer
18474 enable NULL ciphers. They need to be specifically enabled e.g. with
18475 the string "DEFAULT:eNULL".
18479 * Fix to RSA private encryption routines: if p < q then it would
18480 occasionally produce an invalid result. This will only happen with
18481 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18485 * Be less restrictive and allow also `perl util/perlpath.pl
18486 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18487 because this way one can also use an interpreter named `perl5` (which is
18488 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18489 installed as `perl`).
18491 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18493 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18495 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18497 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18498 advapi32.lib to Win32 build and change the pem test comparison
18499 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18500 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18501 and crypto/des/ede_cbcm_enc.c.
18505 * DES quad checksum was broken on big-endian architectures. Fixed.
18509 * Comment out two functions in bio.h that aren't implemented. Fix up the
18510 Win32 test batch file so it (might) work again. The Win32 test batch file
18511 is horrible: I feel ill....
18515 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18516 in e_os.h. Audit of header files to check ANSI and non ANSI
18517 sections: 10 functions were absent from non ANSI section and not exported
18518 from Windows DLLs. Fixed up libeay.num for new functions.
18522 * Make `openssl version` output lines consistent.
18524 *Ralf S. Engelschall*
18526 * Fix Win32 symbol export lists for BIO functions: Added
18527 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18528 to ms/libeay{16,32}.def.
18530 *Ralf S. Engelschall*
18532 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18533 fine under Unix and passes some trivial tests I've now added. But the
18534 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18535 added to make sure no one expects that this stuff really works in the
18536 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18537 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18540 *Ralf S. Engelschall*
18542 * Fix the generation of two part addresses in perl.
18544 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18546 * Add config entry for Linux on MIPS.
18548 *John Tobey <jtobey@channel1.com>*
18550 * Make links whenever Configure is run, unless we are on Windoze.
18554 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18555 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18560 * Add a useful kludge to allow package maintainers to specify compiler and
18561 other platforms details on the command line without having to patch the
18562 Configure script every time: One now can use
18563 `perl Configure <id>:<details>`,
18564 i.e. platform ids are allowed to have details appended
18565 to them (separated by colons). This is treated as there would be a static
18566 pre-configured entry in Configure's %table under key `<id>` with value
18567 `<details>` and `perl Configure <id>` is called. So, when you want to
18568 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18569 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18570 now, which overrides the FreeBSD-elf entry on-the-fly.
18572 *Ralf S. Engelschall*
18574 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18578 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18579 on the `perl Configure ...` command line. This way one can compile
18580 OpenSSL libraries with Position Independent Code (PIC) which is needed
18581 for linking it into DSOs.
18583 *Ralf S. Engelschall*
18585 * Remarkably, export ciphers were totally broken and no-one had noticed!
18590 * Cleaned up the LICENSE document: The official contact for any license
18591 questions now is the OpenSSL core team under openssl-core@openssl.org.
18592 And add a paragraph about the dual-license situation to make sure people
18593 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18594 to the OpenSSL toolkit.
18596 *Ralf S. Engelschall*
18598 * General source tree makefile cleanups: Made `making xxx in yyy...`
18599 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18600 Additionally cleaned up the `make links` target: Remove unnecessary
18601 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18602 to speed processing and no longer clutter the display with confusing
18603 stuff. Instead only the actually done links are displayed.
18605 *Ralf S. Engelschall*
18607 * Permit null encryption ciphersuites, used for authentication only. It used
18608 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18609 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18614 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18615 signed attributes when verifying signatures (this would break them),
18616 the detached data encoding was wrong and public keys obtained using
18617 X509_get_pubkey() weren't freed.
18621 * Add text documentation for the BUFFER functions. Also added a work around
18622 to a Win95 console bug. This was triggered by the password read stuff: the
18623 last character typed gets carried over to the next fread(). If you were
18624 generating a new cert request using 'req' for example then the last
18625 character of the passphrase would be CR which would then enter the first
18630 * Added the new 'Includes OpenSSL Cryptography Software' button as
18631 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18632 button and can be used by applications based on OpenSSL to show the
18633 relationship to the OpenSSL project.
18635 *Ralf S. Engelschall*
18637 * Remove confusing variables in function signatures in files
18638 ssl/ssl_lib.c and ssl/ssl.h.
18640 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18642 * Don't install bss_file.c under PREFIX/include/
18644 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18646 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18647 functions that return function pointers and has support for NT specific
18648 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18649 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18650 unsigned to signed types: this was killing the Win32 compile.
18654 * Add new certificate file to stack functions,
18655 SSL_add_dir_cert_subjects_to_stack() and
18656 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18657 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18658 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18659 This means that Apache-SSL and similar packages don't have to mess around
18660 to add as many CAs as they want to the preferred list.
18664 * Experiment with doxygen documentation. Currently only partially applied to
18666 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18667 openssl.doxy as the configuration file.
18671 * Get rid of remaining C++-style comments which strict C compilers hate.
18673 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18675 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18676 compiled in by default: it has problems with large keys.
18680 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18681 DH private keys and/or callback functions which directly correspond to
18682 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18683 is needed for applications which have to configure certificates on a
18684 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18686 For the RSA certificate situation is makes no difference, but
18687 for the DSA certificate situation this fixes the "no shared cipher"
18688 problem where the OpenSSL cipher selection procedure failed because the
18689 temporary keys were not overtaken from the context and the API provided
18690 no way to reconfigure them.
18691 The new functions now let applications reconfigure the stuff and they
18692 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18693 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18694 non-public-API function ssl_cert_instantiate() is used as a helper
18695 function and also to reduce code redundancy inside ssl_rsa.c.
18697 *Ralf S. Engelschall*
18699 * Move s_server -dcert and -dkey options out of the undocumented feature
18700 area because they are useful for the DSA situation and should be
18701 recognized by the users.
18703 *Ralf S. Engelschall*
18705 * Fix the cipher decision scheme for export ciphers: the export bits are
18706 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18707 SSL_EXP_MASK. So, the original variable has to be used instead of the
18708 already masked variable.
18710 *Richard Levitte <levitte@stacken.kth.se>*
18712 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18714 *Richard Levitte <levitte@stacken.kth.se>*
18716 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18717 from `int` to `unsigned int` because it is a length and initialized by
18718 EVP_DigestFinal() which expects an `unsigned int *`.
18720 *Richard Levitte <levitte@stacken.kth.se>*
18722 * Don't hard-code path to Perl interpreter on shebang line of Configure
18723 script. Instead use the usual Shell->Perl transition trick.
18725 *Ralf S. Engelschall*
18727 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18728 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18729 -noout -modulus` as it's already the case for `openssl rsa -noout
18730 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18731 currently the public key is printed (a decision which was already done by
18732 `openssl dsa -modulus` in the past) which serves a similar purpose.
18733 Additionally the NO_RSA no longer completely removes the whole -modulus
18734 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18737 *Ralf S. Engelschall*
18739 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18740 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18742 *Arne Ansper <arne@ats.cyber.ee>*
18744 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18745 to be added. Now both 'req' and 'ca' can use new objects defined in the
18750 * Add cool BIO that does syslog (or event log on NT).
18752 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18754 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18755 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18756 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18757 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18761 * Add preliminary config info for new extension code.
18765 * Make RSA_NO_PADDING really use no padding.
18767 *Ulf Moeller <ulf@fitug.de>*
18769 * Generate errors when private/public key check is done.
18773 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18774 for some CRL extensions and new objects added.
18778 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18779 key usage extension and fuller support for authority key id.
18783 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18784 padding method for RSA, which is recommended for new applications in PKCS
18785 #1 v2.0 (RFC 2437, October 1998).
18786 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18787 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18788 against Bleichbacher's attack on RSA.
18789 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18792 * Updates to the new SSL compression code
18794 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18796 * Fix so that the version number in the master secret, when passed
18797 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18798 (because the server will not accept higher), that the version number
18799 is 0x03,0x01, not 0x03,0x00
18801 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18803 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18804 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18805 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18809 * Support for RAW extensions where an arbitrary extension can be
18810 created by including its DER encoding. See `apps/openssl.cnf` for
18815 * Make sure latest Perl versions don't interpret some generated C array
18816 code as Perl array code in the crypto/err/err_genc.pl script.
18818 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18820 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18821 not many people have the assembler. Various Win32 compilation fixes and
18822 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18823 build instructions.
18827 * Modify configure script 'Configure' to automatically create crypto/date.h
18828 file under Win32 and also build pem.h from pem.org. New script
18829 util/mkfiles.pl to create the MINFO file on environments that can't do a
18830 'make files': perl util/mkfiles.pl >MINFO should work.
18834 * Major rework of DES function declarations, in the pursuit of correctness
18835 and purity. As a result, many evil casts evaporated, and some weirdness,
18836 too. You may find this causes warnings in your code. Zapping your evil
18837 casts will probably fix them. Mostly.
18841 * Fix for a typo in asn1.h. Bug fix to object creation script
18842 obj_dat.pl. It considered a zero in an object definition to mean
18843 "end of object": none of the objects in objects.h have any zeros
18844 so it wasn't spotted.
18846 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18848 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18849 Masking (CBCM). In the absence of test vectors, the best I have been able
18850 to do is check that the decrypt undoes the encrypt, so far. Send me test
18851 vectors if you have them.
18855 * Correct calculation of key length for export ciphers (too much space was
18856 allocated for null ciphers). This has not been tested!
18860 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18861 message is now correct (it understands "crypto" and "ssl" on its
18862 command line). There is also now an "update" option. This will update
18863 the util/ssleay.num and util/libeay.num files with any new functions.
18865 perl util/mkdef.pl crypto ssl update
18866 it will update them.
18870 * Overhauled the Perl interface:
18871 - ported BN stuff to OpenSSL's different BN library
18872 - made the perl/ source tree CVS-aware
18873 - renamed the package from SSLeay to OpenSSL (the files still contain
18874 their history because I've copied them in the repository)
18875 - removed obsolete files (the test scripts will be replaced
18876 by better Test::Harness variants in the future)
18878 *Ralf S. Engelschall*
18880 * First cut for a very conservative source tree cleanup:
18881 1. merge various obsolete readme texts into doc/ssleay.txt
18882 where we collect the old documents and readme texts.
18883 2. remove the first part of files where I'm already sure that we no
18884 longer need them because of three reasons: either they are just temporary
18885 files which were left by Eric or they are preserved original files where
18886 I've verified that the diff is also available in the CVS via "cvs diff
18887 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18888 the crypto/md/ stuff).
18890 *Ralf S. Engelschall*
18892 * More extension code. Incomplete support for subject and issuer alt
18893 name, issuer and authority key id. Change the i2v function parameters
18894 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18895 what that's for :-) Fix to ASN1 macro which messed up
18896 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18900 * Preliminary support for ENUMERATED type. This is largely copied from the
18905 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18907 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18909 * Make sure `make rehash` target really finds the `openssl` program.
18911 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18913 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18914 like to hear about it if this slows down other processors.
18918 * Add CygWin32 platform information to Configure script.
18920 *Alan Batie <batie@aahz.jf.intel.com>*
18922 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18924 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18926 * New program nseq to manipulate netscape certificate sequences
18930 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18935 * Fixes to BN code. Previously the default was to define BN_RECURSION
18936 but the BN code had some problems that would cause failures when
18937 doing certificate verification and some other functions.
18939 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18941 * Add ASN1 and PEM code to support netscape certificate sequences.
18945 * Add ASN1 and PEM code to support netscape certificate sequences.
18949 * Add several PKIX and private extended key usage OIDs.
18953 * Modify the 'ca' program to handle the new extension code. Modify
18954 openssl.cnf for new extension format, add comments.
18958 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18959 and add a sample to openssl.cnf so req -x509 now adds appropriate
18964 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18965 error code, add initial support to X509_print() and x509 application.
18969 * Takes a deep breath and start adding X509 V3 extension support code. Add
18970 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18971 stuff is currently isolated and isn't even compiled yet.
18975 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18976 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18977 Removed the versions check from X509 routines when loading extensions:
18978 this allows certain broken certificates that don't set the version
18979 properly to be processed.
18983 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18984 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18985 can still be regenerated with "make depend".
18989 * Spelling mistake in C version of CAST-128.
18991 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18993 * Changes to the error generation code. The perl script err-code.pl
18994 now reads in the old error codes and retains the old numbers, only
18995 adding new ones if necessary. It also only changes the .err files if new
18996 codes are added. The makefiles have been modified to only insert errors
18997 when needed (to avoid needlessly modifying header files). This is done
18998 by only inserting errors if the .err file is newer than the auto generated
18999 C file. To rebuild all the error codes from scratch (the old behaviour)
19000 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19001 or delete all the .err files.
19005 * CAST-128 was incorrectly implemented for short keys. The C version has
19006 been fixed, but is untested. The assembler versions are also fixed, but
19007 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19008 to regenerate it if needed.
19009 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19010 Hagino <itojun@kame.net>*
19012 * File was opened incorrectly in randfile.c.
19014 *Ulf Möller <ulf@fitug.de>*
19016 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19017 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19018 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19019 al: it's just almost always a UTCTime. Note this patch adds new error
19020 codes so do a "make errors" if there are problems.
19024 * Correct Linux 1 recognition in config.
19026 *Ulf Möller <ulf@fitug.de>*
19028 * Remove pointless MD5 hash when using DSA keys in ca.
19030 *Anonymous <nobody@replay.com>*
19032 * Generate an error if given an empty string as a cert directory. Also
19033 generate an error if handed NULL (previously returned 0 to indicate an
19034 error, but didn't set one).
19036 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19038 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19042 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19043 parameters. This was causing a warning which killed off the Win32 compile.
19047 * Remove C++ style comments from crypto/bn/bn_local.h.
19049 *Neil Costigan <neil.costigan@celocom.com>*
19051 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19052 based on a text string, looking up short and long names and finally
19053 "dot" format. The "dot" format stuff didn't work. Added new function
19054 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19055 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19056 OID is not part of the table.
19060 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19061 X509_LOOKUP_by_alias().
19065 * Sort openssl functions by name.
19069 * Get the `gendsa` command working and add it to the `list` command. Remove
19070 encryption from sample DSA keys (in case anyone is interested the password
19075 * Make *all* `*_free` functions accept a NULL pointer.
19077 *Frans Heymans <fheymans@isaserver.be>*
19079 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19082 *Anonymous <nobody@replay.com>*
19084 * s_server should send the CAfile as acceptable CAs, not its own cert.
19086 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19088 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19090 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19092 * Temp key "for export" tests were wrong in s3_srvr.c.
19094 *Anonymous <nobody@replay.com>*
19096 * Add prototype for temp key callback functions
19097 SSL_CTX_set_tmp_{rsa,dh}_callback().
19101 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19102 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19106 * X509_name_add_entry() freed the wrong thing after an error.
19108 *Arne Ansper <arne@ats.cyber.ee>*
19110 * rsa_eay.c would attempt to free a NULL context.
19112 *Arne Ansper <arne@ats.cyber.ee>*
19114 * BIO_s_socket() had a broken should_retry() on Windoze.
19116 *Arne Ansper <arne@ats.cyber.ee>*
19118 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19120 *Arne Ansper <arne@ats.cyber.ee>*
19122 * Make sure the already existing X509_STORE->depth variable is initialized
19123 in X509_STORE_new(), but document the fact that this variable is still
19124 unused in the certificate verification process.
19126 *Ralf S. Engelschall*
19128 * Fix the various library and `apps/` files to free up pkeys obtained from
19129 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19133 * Fix reference counting in X509_PUBKEY_get(). This makes
19134 demos/maurice/example2.c work, amongst others, probably.
19136 *Steve Henson and Ben Laurie*
19138 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19139 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19140 are no longer created. This way we have a single and consistent command
19141 line interface `openssl <command>`, similar to `cvs <command>`.
19143 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19145 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19146 BIT STRING wrapper always have zero unused bits.
19150 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19154 * Make the top-level INSTALL documentation easier to understand.
19158 * Makefiles updated to exit if an error occurs in a sub-directory
19159 make (including if user presses ^C) [Paul Sutton]
19161 * Make Montgomery context stuff explicit in RSA data structure.
19165 * Fix build order of pem and err to allow for generated pem.h.
19169 * Fix renumbering bug in X509_NAME_delete_entry().
19173 * Enhanced the err-ins.pl script so it makes the error library number
19174 global and can add a library name. This is needed for external ASN1 and
19175 other error libraries.
19179 * Fixed sk_insert which never worked properly.
19183 * Fix ASN1 macros so they can handle indefinite length constructed
19184 EXPLICIT tags. Some non standard certificates use these: they can now
19189 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19190 into a single doc/ssleay.txt bundle. This way the information is still
19191 preserved but no longer messes up this directory. Now it's new room for
19192 the new set of documentation files.
19194 *Ralf S. Engelschall*
19196 * SETs were incorrectly DER encoded. This was a major pain, because they
19197 shared code with SEQUENCEs, which aren't coded the same. This means that
19198 almost everything to do with SETs or SEQUENCEs has either changed name or
19199 number of arguments.
19201 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19203 * Fix test data to work with the above.
19207 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19208 was already fixed by Eric for 0.9.1 it seems.
19210 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19212 * Autodetect FreeBSD3.
19216 * Fix various bugs in Configure. This affects the following platforms:
19220 unixware-2.0-pentium
19225 * Eliminate generated files from CVS. Reorder tests to regenerate files
19226 before they are needed.
19230 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19234 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19236 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19237 changed SSLeay to OpenSSL in version strings.
19239 *Ralf S. Engelschall*
19241 * Some fixups to the top-level documents.
19245 * Fixed the nasty bug where rsaref.h was not found under compile-time
19246 because the symlink to include/ was missing.
19248 *Ralf S. Engelschall*
19250 * Incorporated the popular no-RSA/DSA-only patches
19251 which allow to compile a RSA-free SSLeay.
19253 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19255 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19256 when "ssleay" is still not found.
19258 *Ralf S. Engelschall*
19260 * Added more platforms to Configure: Cray T3E, HPUX 11,
19262 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19264 * Updated the README file.
19266 *Ralf S. Engelschall*
19268 * Added various .cvsignore files in the CVS repository subdirs
19269 to make a "cvs update" really silent.
19271 *Ralf S. Engelschall*
19273 * Recompiled the error-definition header files and added
19274 missing symbols to the Win32 linker tables.
19276 *Ralf S. Engelschall*
19278 * Cleaned up the top-level documents;
19279 o new files: CHANGES and LICENSE
19280 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19281 o merged COPYRIGHT into LICENSE
19282 o removed obsolete TODO file
19283 o renamed MICROSOFT to INSTALL.W32
19285 *Ralf S. Engelschall*
19287 * Removed dummy files from the 0.9.1b source tree:
19288 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19289 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19290 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19291 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19292 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19294 *Ralf S. Engelschall*
19296 * Added various platform portability fixes.
19300 * The Genesis of the OpenSSL rpject:
19301 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19302 Young and Tim J. Hudson created while they were working for C2Net until
19305 *The OpenSSL Project*
19307 ### Changes between 0.9.0b and 0.9.1b [not released]
19309 * Updated a few CA certificates under certs/
19313 * Changed some BIGNUM api stuff.
19317 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19318 DGUX x86, Linux Alpha, etc.
19322 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19323 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19328 * Add -strparse option to asn1pars program which parses nested
19331 *Dr Stephen Henson <shenson@bigfoot.com>*
19333 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19337 * DSA fix for "ca" program.
19341 * Added "-genkey" option to "dsaparam" program.
19345 * Added RIPE MD160 (rmd160) message digest.
19349 * Added -a (all) option to "ssleay version" command.
19353 * Added PLATFORM define which is the id given to Configure.
19357 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19361 * Extended the ASN.1 parser routines.
19365 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19369 * Added a BN_CTX to the BN library.
19373 * Fixed the weak key values in DES library
19377 * Changed API in EVP library for cipher aliases.
19381 * Added support for RC2/64bit cipher.
19385 * Converted the lhash library to the crypto/mem.c functions.
19389 * Added more recognized ASN.1 object ids.
19393 * Added more RSA padding checks for SSL/TLS.
19397 * Added BIO proxy/filter functionality.
19401 * Added extra_certs to SSL_CTX which can be used
19402 send extra CA certificates to the client in the CA cert chain sending
19403 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19407 * Now Fortezza is denied in the authentication phase because
19408 this is key exchange mechanism is not supported by SSLeay at all.
19412 * Additional PKCS1 checks.
19416 * Support the string "TLSv1" for all TLS v1 ciphers.
19420 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19421 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19425 * Fixed a few memory leaks.
19429 * Fixed various code and comment typos.
19433 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19434 bytes sent in the client random.
19436 *Edward Bishop <ebishop@spyglass.com>*
19440 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19441 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19442 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19443 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19444 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19445 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19446 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19447 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19448 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19449 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19450 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19451 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19452 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19453 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19454 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19455 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19456 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19457 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19458 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19459 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19460 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19461 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19462 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19463 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19464 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19465 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19466 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19467 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19468 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19469 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19470 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19471 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19472 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19473 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19474 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19475 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19476 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19477 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19478 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19479 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19480 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19481 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19482 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19483 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19484 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19485 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19486 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19487 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19488 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19489 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19490 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19491 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19492 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19493 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19494 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19495 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19496 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19497 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19498 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19499 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19500 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19501 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19502 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19503 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19504 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19505 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19506 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19507 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19508 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19509 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19510 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19511 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19512 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19513 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19514 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19515 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19516 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19517 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19518 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19519 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19520 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19521 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19522 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19523 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19524 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19525 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19526 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19527 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19528 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19529 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19530 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19531 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19532 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19533 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19534 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19535 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19536 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19537 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19538 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19539 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19540 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19541 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19542 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19543 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19544 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19545 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19546 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19547 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19548 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19549 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19550 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19551 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19552 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19553 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19554 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19555 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19556 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19557 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19558 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19559 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19560 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19561 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19562 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19563 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19564 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19565 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19566 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19567 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19568 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19569 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19570 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19571 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19572 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19573 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19574 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19575 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19576 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19577 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19578 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19579 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19580 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19581 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19582 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19583 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19584 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19585 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19586 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19587 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19588 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19589 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19590 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19591 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19592 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19593 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19594 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19595 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19596 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19597 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19598 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19599 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19600 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19601 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19602 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19603 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655