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]
32 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
33 listed here are only a brief description.
34 The migration guide contains more detailed information related to new features,
35 breaking changes, and mappings for the large list of deprecated functions.
37 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
39 ### Changes between 3.0.6 and 3.0.7 [xx XXX xxxx]
41 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT parameters
43 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
44 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
45 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
46 Using these invalid names may cause algorithms to use slower methods
47 that ignore the CRT parameters.
51 * Added RIPEMD160 to the default provider.
55 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
57 * OpenSSL supports creating a custom cipher via the legacy
58 EVP_CIPHER_meth_new() function and associated function calls. This function
59 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
60 to use the new provider mechanism in order to implement custom ciphers.
62 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
63 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
64 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
65 and decryption initialisation functions). Instead of using the custom cipher
66 directly it incorrectly tries to fetch an equivalent cipher from the
67 available providers. An equivalent cipher is found based on the NID passed to
68 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
69 given cipher. However it is possible for an application to incorrectly pass
70 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
71 is used in this way the OpenSSL encryption/decryption initialisation function
72 will match the NULL cipher as being equivalent and will fetch this from the
73 available providers. This will succeed if the default provider has been
74 loaded (or if a third party provider has been loaded that offers this
75 cipher). Using the NULL cipher means that the plaintext is emitted as the
78 Applications are only affected by this issue if they call
79 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
80 encryption/decryption initialisation function. Applications that only use
81 SSL/TLS are not impacted by this issue.
86 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
91 * Fixed the linux-mips64 Configure target which was missing the
92 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
97 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
102 * Correctly handle a retransmitted ClientHello in DTLS
106 * Fixed detection of ktls support in cross-compile environment on Linux
110 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
115 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
116 report correct results in some cases
120 * Fix UWP builds by defining VirtualLock
124 * For known safe primes use the minimum key length according to RFC 7919.
125 Longer private key sizes unnecessarily raise the cycles needed to compute the
126 shared secret without any increase of the real security. This fixes a
127 regression from 1.1.1 where these shorter keys were generated for the known
132 * Added the loongarch64 target
136 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
137 only passed to the FIPS provider and not to the default or legacy provider.
141 * Fixed reported performance degradation on aarch64. Restored the
142 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
143 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
144 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
145 The new algorithm is still used for 32 bit targets.
149 * Added a missing header for memcmp that caused compilation failure on some
154 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
156 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
157 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
158 This issue makes the RSA implementation with 2048 bit private keys
159 incorrect on such machines and memory corruption will happen during
160 the computation. As a consequence of the memory corruption an attacker
161 may be able to trigger a remote code execution on the machine performing
164 SSL/TLS servers or other servers using 2048 bit RSA private keys running
165 on machines supporting AVX512IFMA instructions of the X86_64 architecture
166 are affected by this issue.
171 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
172 implementation would not encrypt the entirety of the data under some
173 circumstances. This could reveal sixteen bytes of data that was
174 preexisting in the memory that wasn't written. In the special case of
175 "in place" encryption, sixteen bytes of the plaintext would be revealed.
177 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
178 they are both unaffected.
181 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
183 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
185 * In addition to the c_rehash shell command injection identified in
186 CVE-2022-1292, further bugs where the c_rehash script does not
187 properly sanitise shell metacharacters to prevent command injection have been
190 When the CVE-2022-1292 was fixed it was not discovered that there
191 are other places in the script where the file names of certificates
192 being hashed were possibly passed to a command executed through the shell.
194 This script is distributed by some operating systems in a manner where
195 it is automatically executed. On such operating systems, an attacker
196 could execute arbitrary commands with the privileges of the script.
198 Use of the c_rehash script is considered obsolete and should be replaced
199 by the OpenSSL rehash command line tool.
202 *Daniel Fiala, Tomáš Mráz*
204 * Case insensitive string comparison no longer uses locales. It has instead
205 been directly implemented.
209 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
211 * Case insensitive string comparison is reimplemented via new locale-agnostic
212 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
213 comparison. The previous implementation had problems when the Turkish locale
218 * Fixed a bug in the c_rehash script which was not properly sanitising shell
219 metacharacters to prevent command injection. This script is distributed by
220 some operating systems in a manner where it is automatically executed. On
221 such operating systems, an attacker could execute arbitrary commands with the
222 privileges of the script.
224 Use of the c_rehash script is considered obsolete and should be replaced
225 by the OpenSSL rehash command line tool.
230 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
231 certificate on an OCSP response. The bug caused the function in the case
232 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
233 response (meaning a successful verification) even in the case where the
234 response signing certificate fails to verify.
236 It is anticipated that most users of `OCSP_basic_verify` will not use the
237 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
238 a negative value (indicating a fatal error) in the case of a certificate
239 verification failure. The normal expected return value in this case would be
242 This issue also impacts the command line OpenSSL "ocsp" application. When
243 verifying an ocsp response with the "-no_cert_checks" option the command line
244 application will report that the verification is successful even though it
245 has in fact failed. In this case the incorrect successful response will also
246 be accompanied by error messages showing the failure and contradicting the
247 apparently successful result.
252 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
253 AAD data as the MAC key. This made the MAC key trivially predictable.
255 An attacker could exploit this issue by performing a man-in-the-middle attack
256 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
257 that the modified data would still pass the MAC integrity check.
259 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
260 endpoint will always be rejected by the recipient and the connection will
261 fail at that point. Many application protocols require data to be sent from
262 the client to the server first. Therefore, in such a case, only an OpenSSL
263 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
265 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
266 sent in both directions. In this case both clients and servers could be
267 affected, regardless of the application protocol.
269 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
270 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
271 the handshake when using this ciphersuite.
273 The confidentiality of data is not impacted by this issue, i.e. an attacker
274 cannot decrypt data that has been encrypted using this ciphersuite - they can
277 In order for this attack to work both endpoints must legitimately negotiate
278 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
279 OpenSSL 3.0, and is not available within the default provider or the default
280 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
281 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
282 following must have occurred:
284 1) OpenSSL must have been compiled with the (non-default) compile time option
285 enable-weak-ssl-ciphers
287 2) OpenSSL must have had the legacy provider explicitly loaded (either
288 through application code or via configuration)
290 3) The ciphersuite must have been explicitly added to the ciphersuite list
292 4) The libssl security level must have been set to 0 (default is 1)
294 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
296 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
297 others that both endpoints have in common
302 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
303 occuppied by the removed hash table entries.
305 This function is used when decoding certificates or keys. If a long lived
306 process periodically decodes certificates or keys its memory usage will
307 expand without bounds and the process might be terminated by the operating
308 system causing a denial of service. Also traversing the empty hash table
309 entries will take increasingly more time.
311 Typically such long lived processes might be TLS clients or TLS servers
312 configured to accept client certificate authentication.
315 *Hugo Landau, Aliaksei Levin*
317 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
318 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
319 statistics are no longer supported. For compatibility, these statistics are
320 still listed in the output but are now always reported as zero.
324 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
326 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
327 for non-prime moduli.
329 Internally this function is used when parsing certificates that contain
330 elliptic curve public keys in compressed form or explicit elliptic curve
331 parameters with a base point encoded in compressed form.
333 It is possible to trigger the infinite loop by crafting a certificate that
334 has invalid explicit curve parameters.
336 Since certificate parsing happens prior to verification of the certificate
337 signature, any process that parses an externally supplied certificate may thus
338 be subject to a denial of service attack. The infinite loop can also be
339 reached when parsing crafted private keys as they can contain explicit
340 elliptic curve parameters.
342 Thus vulnerable situations include:
344 - TLS clients consuming server certificates
345 - TLS servers consuming client certificates
346 - Hosting providers taking certificates or private keys from customers
347 - Certificate authorities parsing certification requests from subscribers
348 - Anything else which parses ASN.1 elliptic curve parameters
350 Also any other applications that use the BN_mod_sqrt() where the attacker
351 can control the parameter values are vulnerable to this DoS issue.
356 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
357 to the list of ciphersuites providing Perfect Forward Secrecy as
358 required by SECLEVEL >= 3.
360 *Dmitry Belyavskiy, Nicola Tuveri*
362 * Made the AES constant time code for no-asm configurations
363 optional due to the resulting 95% performance degradation.
364 The AES constant time code can be enabled, for no assembly
365 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
369 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
374 * The negative return value handling of the certificate verification callback
375 was reverted. The replacement is to set the verification retry state with
376 the SSL_set_retry_verify() function.
380 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
382 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
383 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
384 verify a certificate supplied by a server. That function may return a
385 negative return value to indicate an internal error (for example out of
386 memory). Such a negative return value is mishandled by OpenSSL and will cause
387 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
388 success and a subsequent call to SSL_get_error() to return the value
389 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
390 returned by OpenSSL if the application has previously called
391 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
392 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
393 totally unexpected and applications may not behave correctly as a result. The
394 exact behaviour will depend on the application but it could result in
395 crashes, infinite loops or other similar incorrect responses.
397 This issue is made more serious in combination with a separate bug in OpenSSL
398 3.0 that will cause X509_verify_cert() to indicate an internal error when
399 processing a certificate chain. This will occur where a certificate does not
400 include the Subject Alternative Name extension but where a Certificate
401 Authority has enforced name constraints. This issue can occur even with valid
407 * Corrected a few file name and file reference bugs in the build,
408 installation and setup scripts, which lead to installation verification
409 failures. Slightly enhanced the installation verification script.
413 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
418 * Fixed PVK encoder to properly query for the passphrase.
422 * Multiple fixes in the OSSL_HTTP API functions.
426 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
427 OSSL_PARAM_INTEGER data type and return error on negative numbers
428 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
429 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
433 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
437 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
441 * Multiple threading fixes.
445 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
449 * Allow fetching an operation from the provider that owns an unexportable key
450 as a fallback if that is still allowed by the property query.
454 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
456 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
461 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
462 S390X capability vector to zero. This simplifies testing of different code
463 paths on S390X architecture.
467 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
468 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
469 SP 800-38D". The communication will fail at this point.
473 * The EC_GROUP_clear_free() function is deprecated as there is nothing
474 confidential in EC_GROUP data.
478 * The byte order mark (BOM) character is ignored if encountered at the
479 beginning of a PEM-formatted file.
483 * Added CMS support for the Russian GOST algorithms.
487 * Due to move of the implementation of cryptographic operations
488 to the providers, validation of various operation parameters can
489 be postponed until the actual operation is executed where previously
490 it happened immediately when an operation parameter was set.
492 For example when setting an unsupported curve with
493 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
494 fail but later keygen operations with the EVP_PKEY_CTX will fail.
496 *OpenSSL team members and many third party contributors*
498 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
499 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
500 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
501 instead to retrieve these algorithms from a provider.
505 * On build targets where the multilib postfix is set in the build
506 configuration the libdir directory was changing based on whether
507 the lib directory with the multilib postfix exists on the system
508 or not. This unpredictable behavior was removed and eventual
509 multilib postfix is now always added to the default libdir. Use
510 `--libdir=lib` to override the libdir if adding the postfix is
515 * The triple DES key wrap functionality now conforms to RFC 3217 but is
516 no longer interoperable with OpenSSL 1.1.1.
520 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
521 function codes, this function can only cause problems for calling
526 * Add a configurable flag to output date formats as ISO 8601. Does not
527 change the default date format.
531 * Version of MSVC earlier than 1300 could get link warnings, which could
532 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
533 Support for this flag has been removed.
537 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
538 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
539 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
540 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
541 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
545 * The signatures of the functions to get and set options on SSL and
546 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
547 Some source code changes may be required.
551 * The public definitions of conf_method_st and conf_st have been
552 deprecated. They will be made opaque in a future release.
554 *Rich Salz and Tomáš Mráz*
556 * Client-initiated renegotiation is disabled by default. To allow it, use
557 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
558 flag, or the "ClientRenegotiation" config parameter as appropriate.
562 * Add "abspath" and "includedir" pragma's to config files, to prevent,
563 or modify relative pathname inclusion.
567 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
568 validated. Please consult the README-FIPS and
569 README-PROVIDERS files, as well as the migration guide.
571 *OpenSSL team members and many third party contributors*
573 * For the key types DH and DHX the allowed settable parameters are now different.
577 * The openssl commands that read keys, certificates, and CRLs now
578 automatically detect the PEM or DER format of the input files.
580 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
582 * Added enhanced PKCS#12 APIs which accept a library context.
586 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
590 * Added support for Kernel TLS (KTLS).
592 *Boris Pismenny, John Baldwin and Andrew Gallatin*
594 * Support for RFC 5746 secure renegotiation is now required by default for
595 SSL or TLS connections to succeed.
599 * The signature of the `copy` functional parameter of the
600 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
601 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
602 the signature of the `pub_decode` functional parameter of the
603 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
604 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
608 * The error return values from some control calls (ctrl) have changed.
612 * A public key check is now performed during EVP_PKEY_derive_set_peer().
616 * Many functions in the EVP_ namespace that are getters of values from
617 implementations or contexts were renamed to include get or get0 in their
618 names. Old names are provided as macro aliases for compatibility and
623 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
624 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
625 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
630 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
633 * The output from the command line applications may have minor
638 * The output from numerous "printing" may have minor changes.
642 * Windows thread synchronization uses read/write primitives (SRWLock) when
643 supported by the OS, otherwise CriticalSection continues to be used.
647 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
648 work on read only BIO source/sinks that do not support these functions.
649 This allows piping or redirection of a file BIO using stdin to be buffered
650 into memory. This is used internally in OSSL_DECODER_from_bio().
654 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
655 this function would return one of the values OSSL_STORE_INFO_NAME,
656 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
657 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
658 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
659 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
660 using this function should be amended to handle the changed return value.
664 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
665 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
666 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
667 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
668 contains more than one certificate identifier: This means that all
669 certificates referenced there MUST be part of the validation chain.
673 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
674 RC5, DESX and DES have been moved to the legacy provider.
678 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
679 RIPEMD-160 have been moved to the legacy provider.
683 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
688 * The deprecated functions EVP_PKEY_get0_RSA(),
689 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
690 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
691 well as the similarly named "get1" functions behave differently in
696 * A number of functions handling low-level keys or engines were deprecated
697 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
698 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
699 EVP_PKEY_get0_siphash().
703 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
704 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
705 will need to load the legacy crypto provider. This includes these PBE
706 algorithms which use this KDF:
707 - NID_pbeWithMD2AndDES_CBC
708 - NID_pbeWithMD5AndDES_CBC
709 - NID_pbeWithSHA1AndRC2_CBC
710 - NID_pbeWithMD2AndRC2_CBC
711 - NID_pbeWithMD5AndRC2_CBC
712 - NID_pbeWithSHA1AndDES_CBC
716 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
717 BIO_debug_callback() functions.
721 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
722 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
726 * The RAND_METHOD APIs have been deprecated.
730 * The SRP APIs have been deprecated.
734 * Add a compile time option to prevent the caching of provider fetched
735 algorithms. This is enabled by including the no-cached-fetch option
736 at configuration time.
740 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
741 count of PKCS12_DEFAULT_ITER.
743 *Tomáš Mráz and Sahana Prasad*
745 * The openssl speed command does not use low-level API calls anymore.
749 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
752 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
754 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
758 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
759 providers may supply their own group implementations (using either the "key
760 exchange" or the "key encapsulation" methods) which will automatically be
761 detected and used by libssl.
763 *Matt Caswell, Nicola Tuveri*
765 * The undocumented function X509_certificate_type() has been deprecated;
769 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
773 * Removed RSA padding mode for SSLv23 (which was only used for
774 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
775 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
780 * Deprecated the obsolete X9.31 RSA key generation related functions.
782 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
783 is not allowed to return a value > 1, this is no more taken as failure.
785 *Viktor Dukhovni and David von Oheimb*
787 * Deprecated the obsolete X9.31 RSA key generation related functions
788 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
789 BN_X931_generate_prime_ex().
793 * The default key generation method for the regular 2-prime RSA keys was
794 changed to the FIPS 186-4 B.3.6 method.
798 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
802 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
806 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
807 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
809 *Rich Salz, Richard Levitte, and David von Oheimb*
811 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
815 * Deprecated `OCSP_parse_url()`.
819 * Validation of SM2 keys has been separated from the validation of regular EC
824 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
825 switches: a validation failure triggers an early exit, returning a failure
826 exit status to the parent process.
830 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
831 to ignore unknown ciphers.
835 * The `-cipher-commands` and `-digest-commands` options
836 of the command line utility `list` have been deprecated.
837 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
841 * Added convenience functions for generating asymmetric key pairs:
842 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
843 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
847 * All of the low level EC_KEY functions have been deprecated.
849 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
851 * Deprecated all the libcrypto and libssl error string loading
856 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
857 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
862 * The `-crypt` option to the `passwd` command line tool has been removed.
866 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
871 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
875 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
876 EVP_PKEY_get1_tls_encodedpoint().
880 * The security callback, which can be customised by application code, supports
881 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
882 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
886 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
887 interface. Their functionality remains unchanged.
891 * Added new option for 'openssl list', '-providers', which will display the
892 list of loaded providers, their names, version and status. It optionally
893 displays their gettable parameters.
897 * Removed EVP_PKEY_set_alias_type().
901 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
902 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
906 * Changed all "STACK" functions to be macros instead of inline functions. Macro
907 parameters are still checked for type safety at compile time via helper
912 * Remove the RAND_DRBG API
914 *Paul Dale and Matthias St. Pierre*
916 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
917 as well as actual hostnames.
921 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
922 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
923 conversely, silently ignore DTLS protocol version bounds when configuring
924 TLS-based contexts. The commands can be repeated to set bounds of both
925 types. The same applies with the corresponding "min_protocol" and
926 "max_protocol" command-line switches, in case some application uses both TLS
929 SSL_CTX instances that are created for a fixed protocol version (e.g.
930 `TLSv1_server_method()`) also silently ignore version bounds. Previously
931 attempts to apply bounds to these protocol versions would result in an
932 error. Now only the "version-flexible" SSL_CTX instances are subject to
933 limits in configuration files in command-line options.
937 * Deprecated the `ENGINE` API. Engines should be replaced with providers
942 * Reworked the recorded ERR codes to make better space for system errors.
943 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
944 given code is a system error (true) or an OpenSSL error (false).
948 * Reworked the test perl framework to better allow parallel testing.
950 *Nicola Tuveri and David von Oheimb*
952 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
953 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
957 * 'Configure' has been changed to figure out the configuration target if
958 none is given on the command line. Consequently, the 'config' script is
959 now only a mere wrapper. All documentation is changed to only mention
962 *Rich Salz and Richard Levitte*
964 * Added a library context `OSSL_LIB_CTX` that applications as well as
965 other libraries can use to form a separate context within which
966 libcrypto operations are performed.
970 * Added various `_ex` functions to the OpenSSL API that support using
971 a non-default `OSSL_LIB_CTX`.
975 * Handshake now fails if Extended Master Secret extension is dropped
980 * Dropped interactive mode from the `openssl` program.
984 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
986 *David von Oheimb and Shane Lontis*
988 * Deprecated `EC_METHOD_get_field_type()`.
992 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
993 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
994 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
998 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1002 * Add CAdES-BES signature verification support, mostly derived
1003 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1005 *Filipe Raimundo da Silva*
1007 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1011 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1012 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1016 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1020 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1021 EC_KEY_precompute_mult().
1025 * Deprecated EC_POINTs_mul().
1029 * Removed FIPS_mode() and FIPS_mode_set().
1033 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1037 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1038 EC_POINT_get_Jprojective_coordinates_GFp().
1042 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1043 arrays to be more easily constructed via a series of utility functions.
1044 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1045 the various push functions and finally convert to a passable OSSL_PARAM
1046 array using OSSL_PARAM_BLD_to_param().
1050 * The security strength of SHA1 and MD5 based signatures in TLS has been
1055 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1056 contain a provider side internal key.
1060 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1064 * Project text documents not yet having a proper file name extension
1065 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1066 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1070 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1071 have been converted to Markdown with the goal to produce documents
1072 which not only look pretty when viewed online in the browser, but
1073 remain well readable inside a plain text editor.
1075 To achieve this goal, a 'minimalistic' Markdown style has been applied
1076 which avoids formatting elements that interfere too much with the
1077 reading flow in the text file. For example, it
1079 * avoids [ATX headings][] and uses [setext headings][] instead
1080 (which works for `<h1>` and `<h2>` headings only).
1081 * avoids [inline links][] and uses [reference links][] instead.
1082 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1084 [ATX headings]: https://github.github.com/gfm/#atx-headings
1085 [setext headings]: https://github.github.com/gfm/#setext-headings
1086 [inline links]: https://github.github.com/gfm/#inline-link
1087 [reference links]: https://github.github.com/gfm/#reference-link
1088 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1089 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1091 *Matthias St. Pierre*
1093 * The test suite is changed to preserve results of each test recipe.
1094 A new directory test-runs/ with subdirectories named like the
1095 test recipes are created in the build tree for this purpose.
1099 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1100 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1101 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1103 *David von Oheimb, Martin Peylo*
1105 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1106 It supports arbitrary request and response content types, GET redirection,
1107 TLS, connections via HTTP(S) proxies, connections and exchange via
1108 user-defined BIOs (allowing implicit connections), persistent connections,
1109 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1110 The legacy OCSP-focused (and only partly documented) API
1111 is retained for backward compatibility, while most of it is deprecated.
1115 * Added `util/check-format.pl`, a tool for checking adherence to the
1116 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1117 The checks performed are incomplete and yield some false positives.
1118 Still the tool should be useful for detecting most typical glitches.
1122 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1123 If domain name resolution yields multiple IP addresses all of them are tried
1124 after `connect()` failures.
1128 * All of the low level RSA functions have been deprecated.
1132 * X509 certificates signed using SHA1 are no longer allowed at security
1137 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1138 modified to use PKEY APIs. These commands are now in maintenance mode
1139 and no new features will be added to them.
1143 * The command line utility rsautl has been deprecated.
1147 * The command line utilities genrsa and rsa have been modified to use PKEY
1148 APIs. They now write PKCS#8 keys by default. These commands are now in
1149 maintenance mode and no new features will be added to them.
1153 * All of the low level DH functions have been deprecated.
1155 *Paul Dale and Matt Caswell*
1157 * All of the low level DSA functions have been deprecated.
1161 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1162 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1166 * Deprecated low level ECDH and ECDSA functions.
1170 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1174 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1175 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1176 a new formulation to include all the things it can be used for,
1177 as well as words of caution.
1181 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1185 * All of the low level HMAC functions have been deprecated.
1187 *Paul Dale and David von Oheimb*
1189 * Over two thousand fixes were made to the documentation, including:
1190 - Common options (such as -rand/-writerand, TLS version control, etc)
1191 were refactored and point to newly-enhanced descriptions in openssl.pod.
1192 - Added style conformance for all options (with help from Richard Levitte),
1193 documented all reported missing options, added a CI build to check
1194 that all options are documented and that no unimplemented options
1196 - Documented some internals, such as all use of environment variables.
1197 - Addressed all internal broken L<> references.
1201 * All of the low level CMAC functions have been deprecated.
1205 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1206 functions have been deprecated.
1208 *Paul Dale and David von Oheimb*
1210 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1211 set of functions. The documentation mentioned negative values for some
1212 errors, but this was never the case, so the mention of negative values
1215 Code that followed the documentation and thereby check with something
1216 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1220 * All of the low level cipher functions have been deprecated.
1222 *Matt Caswell and Paul Dale*
1224 * Removed include/openssl/opensslconf.h.in and replaced it with
1225 include/openssl/configuration.h.in, which differs in not including
1226 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1227 was added to include both.
1229 This allows internal hacks where one might need to modify the set
1230 of configured macros, for example this if deprecated symbols are
1231 still supposed to be available internally:
1233 #include <openssl/configuration.h>
1235 #undef OPENSSL_NO_DEPRECATED
1236 #define OPENSSL_SUPPRESS_DEPRECATED
1238 #include <openssl/macros.h>
1240 This should not be used by applications that use the exported
1241 symbols, as that will lead to linking errors.
1245 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1246 used in exponentiation with 512-bit moduli. No EC algorithms are
1247 affected. Analysis suggests that attacks against 2-prime RSA1024,
1248 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1249 difficult to perform and are not believed likely. Attacks against DH512
1250 are considered just feasible. However, for an attack the target would
1251 have to re-use the DH512 private key, which is not recommended anyway.
1252 Also applications directly using the low-level API BN_mod_exp may be
1253 affected if they use BN_FLG_CONSTTIME.
1258 * Most memory-debug features have been deprecated, and the functionality
1259 replaced with no-ops.
1263 * Added documentation for the STACK API.
1267 * Introduced a new method type and API, OSSL_ENCODER, to represent
1268 generic encoders. These do the same sort of job that PEM writers
1269 and d2i functions do, but with support for methods supplied by
1270 providers, and the possibility for providers to support other
1275 * Introduced a new method type and API, OSSL_DECODER, to represent
1276 generic decoders. These do the same sort of job that PEM readers
1277 and i2d functions do, but with support for methods supplied by
1278 providers, and the possibility for providers to support other
1283 * Added a .pragma directive to the syntax of configuration files, to
1284 allow varying behavior in a supported and predictable manner.
1285 Currently added pragma:
1289 This allows dollar signs to be a keyword character unless it's
1290 followed by a opening brace or parenthesis. This is useful for
1291 platforms where dollar signs are commonly used in names, such as
1292 volume names and system directory names on VMS.
1296 * Added functionality to create an EVP_PKEY from user data.
1300 * Change the interpretation of the '--api' configuration option to
1301 mean that this is a desired API compatibility level with no
1302 further meaning. The previous interpretation, that this would
1303 also mean to remove all deprecated symbols up to and including
1304 the given version, no requires that 'no-deprecated' is also used
1305 in the configuration.
1307 When building applications, the desired API compatibility level
1308 can be set with the OPENSSL_API_COMPAT macro like before. For
1309 API compatibility version below 3.0, the old style numerical
1310 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1311 For version 3.0 and on, the value is expected to be the decimal
1312 value calculated from the major and minor version like this:
1314 MAJOR * 10000 + MINOR * 100
1318 -DOPENSSL_API_COMPAT=30000 For 3.0
1319 -DOPENSSL_API_COMPAT=30200 For 3.2
1321 To hide declarations that are deprecated up to and including the
1322 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1323 given when building the application as well.
1327 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1328 access to certificate and CRL stores via URIs and OSSL_STORE
1331 This adds the following functions:
1333 - X509_LOOKUP_store()
1334 - X509_STORE_load_file()
1335 - X509_STORE_load_path()
1336 - X509_STORE_load_store()
1337 - SSL_add_store_cert_subjects_to_stack()
1338 - SSL_CTX_set_default_verify_store()
1339 - SSL_CTX_load_verify_file()
1340 - SSL_CTX_load_verify_dir()
1341 - SSL_CTX_load_verify_store()
1345 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1346 The presence of this system service is determined at run-time.
1350 * Added functionality to create an EVP_PKEY context based on data
1351 for methods from providers. This takes an algorithm name and a
1352 property query string and simply stores them, with the intent
1353 that any operation that uses this context will use those strings
1354 to fetch the needed methods implicitly, thereby making the port
1355 of application written for pre-3.0 OpenSSL easier.
1359 * The undocumented function NCONF_WIN32() has been deprecated; for
1360 conversion details see the HISTORY section of doc/man5/config.pod
1364 * Introduced the new functions EVP_DigestSignInit_ex() and
1365 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1366 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1367 pages for further details.
1371 * Over two thousand fixes were made to the documentation, including:
1372 adding missing command flags, better style conformance, documentation
1375 *Rich Salz, Richard Levitte*
1377 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1378 X25519, X448, Ed25519 and Ed448.
1382 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1387 * Deprecated the public definition of `ERR_STATE` as well as the function
1388 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1393 * Added ERR functionality to give callers access to the stored function
1394 names that have replaced the older function code based functions.
1396 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1397 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1398 ERR_peek_error_all() and ERR_peek_last_error_all().
1400 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1401 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1402 ERR_func_error_string().
1406 * Extended testing to be verbose for failing tests only. The make variables
1407 VERBOSE_FAILURE or VF can be used to enable this:
1409 $ make VF=1 test # Unix
1410 $ mms /macro=(VF=1) test ! OpenVMS
1411 $ nmake VF=1 test # Windows
1415 * Added the `-copy_extensions` option to the `x509` command for use with
1416 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1417 all extensions in the request are copied to the certificate or vice versa.
1419 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1421 * Added the `-copy_extensions` option to the `req` command for use with
1422 `-x509`. When given with the `copy` or `copyall` argument,
1423 all extensions in the certification request are copied to the certificate.
1427 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1428 they generate are by default RFC 5280 compliant in the following sense:
1429 There is a subjectKeyIdentifier extension with a hash value of the public key
1430 and for not self-signed certs there is an authorityKeyIdentifier extension
1431 with a keyIdentifier field or issuer information identifying the signing key.
1432 This is done unless some configuration overrides the new default behavior,
1433 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1437 * Added several checks to `X509_verify_cert()` according to requirements in
1438 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1439 (which may be done by using the CLI option `-x509_strict`):
1440 * The basicConstraints of CA certificates must be marked critical.
1441 * CA certificates must explicitly include the keyUsage extension.
1442 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1443 * The issuer name of any certificate must not be empty.
1444 * The subject name of CA certs, certs with keyUsage crlSign,
1445 and certs without subjectAlternativeName must not be empty.
1446 * If a subjectAlternativeName extension is given it must not be empty.
1447 * The signatureAlgorithm field and the cert signature must be consistent.
1448 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1449 must not be marked critical.
1450 * The authorityKeyIdentifier must be given for X.509v3 certs
1451 unless they are self-signed.
1452 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1456 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1457 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1461 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1462 used even when parsing explicit parameters, when loading a encoded key
1463 or calling `EC_GROUP_new_from_ecpkparameters()`/
1464 `EC_GROUP_new_from_ecparameters()`.
1465 This prevents bypass of security hardening and performance gains,
1466 especially for curves with specialized EC_METHODs.
1467 By default, if a key encoded with explicit parameters is loaded and later
1468 encoded, the output is still encoded with explicit parameters, even if
1469 internally a "named" EC_GROUP is used for computation.
1473 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1474 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1475 NULL. After this change, only the cofactor parameter can be NULL. It also
1476 does some minimal sanity checks on the passed order.
1481 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1482 An attack is simple, if the first CMS_recipientInfo is valid but the
1483 second CMS_recipientInfo is chosen ciphertext. If the second
1484 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1485 encryption key will be replaced by garbage, and the message cannot be
1486 decoded, but if the RSA decryption fails, the correct encryption key is
1487 used and the recipient will not notice the attack.
1488 As a work around for this potential attack the length of the decrypted
1489 key must be equal to the cipher default key length, in case the
1490 certifiate is not given and all recipientInfo are tried out.
1491 The old behaviour can be re-enabled in the CMS code by setting the
1492 CMS_DEBUG_DECRYPT flag.
1496 * Early start up entropy quality from the DEVRANDOM seed source has been
1497 improved for older Linux systems. The RAND subsystem will wait for
1498 /dev/random to be producing output before seeding from /dev/urandom.
1499 The seeded state is stored for future library initialisations using
1500 a system global shared memory segment. The shared memory identifier
1501 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1502 the desired value. The default identifier is 114.
1506 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1507 when primes for RSA keys are computed.
1508 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1509 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1510 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1511 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1512 This avoids possible fingerprinting of newly generated RSA modules.
1516 * Correct the extended master secret constant on EBCDIC systems. Without this
1517 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1518 negotiate EMS will fail. Unfortunately this also means that TLS connections
1519 between EBCDIC systems with this fix, and EBCDIC systems without this
1520 fix will fail if they negotiate EMS.
1524 * Changed the library initialisation so that the config file is now loaded
1525 by default. This was already the case for libssl. It now occurs for both
1526 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1527 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1531 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1532 where the former acts as a replacement for `ERR_put_error()`, and the
1533 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1534 `ERR_raise_data()` adds more flexibility by taking a format string and
1535 an arbitrary number of arguments following it, to be processed with
1540 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1541 to check if a named provider is loaded and available. When called, it
1542 will also activate all fallback providers if such are still present.
1546 * Enforce a minimum DH modulus size of 512 bits.
1550 * Changed DH parameters to generate the order q subgroup instead of 2q.
1551 Previously generated DH parameters are still accepted by DH_check
1552 but DH_generate_key works around that by clearing bit 0 of the
1553 private key for those. This avoids leaking bit 0 of the private key.
1557 * Significantly reduce secure memory usage by the randomness pools.
1561 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1566 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1567 algorithms. An implementation of a key exchange algorithm can be obtained
1568 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1569 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1570 the older EVP_PKEY_derive_init() function. See the man pages for the new
1571 functions for further details.
1575 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1579 * Removed the function names from error messages and deprecated the
1584 * Removed NextStep support and the macro OPENSSL_UNISTD
1588 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1589 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1590 Also removed "export var as function" capability; we do not export
1591 variables, only functions.
1595 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1596 an error and 1 indicating success. In previous versions of OpenSSL this
1597 was a void type. If a key was set longer than the maximum possible this
1602 * Support SM2 signing and verification schemes with X509 certificate.
1606 * Use SHA256 as the default digest for TS query in the `ts` app.
1610 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1614 * Default cipher lists/suites are now available via a function, the
1615 #defines are deprecated.
1619 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1620 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1621 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1625 * Join the directories crypto/x509 and crypto/x509v3
1629 * Added command 'openssl kdf' that uses the EVP_KDF API.
1633 * Added command 'openssl mac' that uses the EVP_MAC API.
1637 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1638 as default directories. Also added the command 'openssl info'
1639 for scripting purposes.
1643 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1648 * Add prediction resistance to the DRBG reseeding process.
1652 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1653 mandated by IEEE Std 1619-2018.
1657 * Added newline escaping functionality to a filename when using openssl dgst.
1658 This output format is to replicate the output format found in the `*sum`
1659 checksum programs. This aims to preserve backward compatibility.
1661 *Matt Eaton, Richard Levitte, and Paul Dale*
1663 * Removed the heartbeat message in DTLS feature, as it has very
1664 little usage and doesn't seem to fulfill a valuable purpose.
1665 The configuration option is now deprecated.
1669 * Changed the output of 'openssl {digestname} < file' to display the
1670 digest name in its output.
1674 * Added a new generic trace API which provides support for enabling
1675 instrumentation through trace output.
1677 *Richard Levitte & Matthias St. Pierre*
1679 * Added build tests for C++. These are generated files that only do one
1680 thing, to include one public OpenSSL head file each. This tests that
1681 the public header files can be usefully included in a C++ application.
1683 This test isn't enabled by default. It can be enabled with the option
1684 'enable-buildtest-c++'.
1688 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1692 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1696 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1700 * Added KMAC to EVP_MAC.
1704 * Added property based algorithm implementation selection framework to
1709 * Added SCA hardening for modular field inversion in EC_GROUP through
1710 a new dedicated field_inv() pointer in EC_METHOD.
1711 This also addresses a leakage affecting conversions from projective
1712 to affine coordinates.
1714 *Billy Bob Brumley, Nicola Tuveri*
1716 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1717 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1718 those algorithms that were already supported through the EVP_PKEY API
1719 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1720 and scrypt are now wrappers that call EVP_KDF.
1724 * Build devcrypto engine as a dynamic engine.
1726 *Eneas U de Queiroz*
1728 * Add keyed BLAKE2 to EVP_MAC.
1732 * Fix a bug in the computation of the endpoint-pair shared secret used
1733 by DTLS over SCTP. This breaks interoperability with older versions
1734 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1735 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1736 interoperability with such broken implementations. However, enabling
1737 this switch breaks interoperability with correct implementations.
1739 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1740 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1744 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1748 * Changed the license to the Apache License v2.0.
1752 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1754 - Major releases (indicated by incrementing the MAJOR release number)
1755 may introduce incompatible API/ABI changes.
1756 - Minor releases (indicated by incrementing the MINOR release number)
1757 may introduce new features but retain API/ABI compatibility.
1758 - Patch releases (indicated by incrementing the PATCH number)
1759 are intended for bug fixes and other improvements of existing
1760 features only (like improving performance or adding documentation)
1761 and retain API/ABI compatibility.
1765 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1769 * Remove the 'dist' target and add a tarball building script. The
1770 'dist' target has fallen out of use, and it shouldn't be
1771 necessary to configure just to create a source distribution.
1775 * Recreate the OS390-Unix config target. It no longer relies on a
1776 special script like it did for OpenSSL pre-1.1.0.
1780 * Instead of having the source directories listed in Configure, add
1781 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1786 * Add GMAC to EVP_MAC.
1790 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1794 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1795 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1796 to facilitate the continued use of MACs through raw private keys in
1797 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1801 * Deprecate ECDH_KDF_X9_62().
1805 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1806 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1807 are retained for backwards compatibility.
1811 * AES-XTS mode now enforces that its two keys are different to mitigate
1812 the attacked described in "Efficient Instantiations of Tweakable
1813 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1814 Details of this attack can be obtained from:
1815 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1819 * Rename the object files, i.e. give them other names than in previous
1820 versions. Their names now include the name of the final product, as
1821 well as its type mnemonic (bin, lib, shlib).
1825 * Added new option for 'openssl list', '-objects', which will display the
1826 list of built in objects, i.e. OIDs with names.
1830 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1831 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1836 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1837 improves application performance by removing data copies and providing
1838 applications with zero-copy system calls such as sendfile and splice.
1842 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1846 * `PKCS12_parse` now maintains the order of the parsed certificates
1847 when outputting them via `*ca` (rather than reversing it).
1851 * Deprecated pthread fork support methods.
1855 * Added support for FFDHE key exchange in TLS 1.3.
1859 * Added a new concept for OpenSSL plugability: providers. This
1860 functionality is designed to replace the ENGINE API and ENGINE
1861 implementations, and to be much more dynamic, allowing provider
1862 authors to introduce new algorithms among other things, as long as
1863 there's an API that supports the algorithm type.
1865 With this concept comes a new core API for interaction between
1866 libcrypto and provider implementations. Public libcrypto functions
1867 that want to use providers do so through this core API.
1869 The main documentation for this core API is found in
1870 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1871 refer to other manuals describing the API specific for supported
1872 algorithm types (also called operations).
1879 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1881 * Avoid loading of a dynamic engine twice.
1885 * Prioritise DANE TLSA issuer certs over peer certs
1889 * Fixed random API for MacOS prior to 10.12
1891 These MacOS versions don't support the CommonCrypto APIs
1895 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1897 * Fixed an SM2 Decryption Buffer Overflow.
1899 In order to decrypt SM2 encrypted data an application is expected to
1900 call the API function EVP_PKEY_decrypt(). Typically an application will
1901 call this function twice. The first time, on entry, the "out" parameter
1902 can be NULL and, on exit, the "outlen" parameter is populated with the
1903 buffer size required to hold the decrypted plaintext. The application
1904 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1905 again, but this time passing a non-NULL value for the "out" parameter.
1907 A bug in the implementation of the SM2 decryption code means that the
1908 calculation of the buffer size required to hold the plaintext returned
1909 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1910 size required by the second call. This can lead to a buffer overflow
1911 when EVP_PKEY_decrypt() is called by the application a second time with
1912 a buffer that is too small.
1914 A malicious attacker who is able present SM2 content for decryption to
1915 an application could cause attacker chosen data to overflow the buffer
1916 by up to a maximum of 62 bytes altering the contents of other data held
1917 after the buffer, possibly changing application behaviour or causing
1918 the application to crash. The location of the buffer is application
1919 dependent but is typically heap allocated.
1924 * Fixed various read buffer overruns processing ASN.1 strings
1926 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1927 structure which contains a buffer holding the string data and a field
1928 holding the buffer length. This contrasts with normal C strings which
1929 are repesented as a buffer for the string data which is terminated
1930 with a NUL (0) byte.
1932 Although not a strict requirement, ASN.1 strings that are parsed using
1933 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1934 well as any string whose value has been set with the ASN1_STRING_set()
1935 function will additionally NUL terminate the byte array in the
1936 ASN1_STRING structure.
1938 However, it is possible for applications to directly construct valid
1939 ASN1_STRING structures which do not NUL terminate the byte array by
1940 directly setting the "data" and "length" fields in the ASN1_STRING
1941 array. This can also happen by using the ASN1_STRING_set0() function.
1943 Numerous OpenSSL functions that print ASN.1 data have been found to
1944 assume that the ASN1_STRING byte array will be NUL terminated, even
1945 though this is not guaranteed for strings that have been directly
1946 constructed. Where an application requests an ASN.1 structure to be
1947 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1948 been directly constructed by the application without NUL terminating
1949 the "data" field, then a read buffer overrun can occur.
1951 The same thing can also occur during name constraints processing
1952 of certificates (for example if a certificate has been directly
1953 constructed by the application instead of loading it via the OpenSSL
1954 parsing functions, and the certificate contains non NUL terminated
1955 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1956 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1958 If a malicious actor can cause an application to directly construct an
1959 ASN1_STRING and then process it through one of the affected OpenSSL
1960 functions then this issue could be hit. This might result in a crash
1961 (causing a Denial of Service attack). It could also result in the
1962 disclosure of private memory contents (such as private keys, or
1963 sensitive plaintext).
1968 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1970 * Fixed a problem with verifying a certificate chain when using the
1971 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1972 the certificates present in a certificate chain. It is not set by default.
1974 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1975 the chain that have explicitly encoded elliptic curve parameters was added
1976 as an additional strict check.
1978 An error in the implementation of this check meant that the result of a
1979 previous check to confirm that certificates in the chain are valid CA
1980 certificates was overwritten. This effectively bypasses the check
1981 that non-CA certificates must not be able to issue other certificates.
1983 If a "purpose" has been configured then there is a subsequent opportunity
1984 for checks that the certificate is a valid CA. All of the named "purpose"
1985 values implemented in libcrypto perform this check. Therefore, where
1986 a purpose is set the certificate chain will still be rejected even when the
1987 strict flag has been used. A purpose is set by default in libssl client and
1988 server certificate verification routines, but it can be overridden or
1989 removed by an application.
1991 In order to be affected, an application must explicitly set the
1992 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1993 for the certificate verification or, in the case of TLS client or server
1994 applications, override the default purpose.
1999 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2000 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2001 renegotiation ClientHello omits the signature_algorithms extension (where it
2002 was present in the initial ClientHello), but includes a
2003 signature_algorithms_cert extension then a NULL pointer dereference will
2004 result, leading to a crash and a denial of service attack.
2006 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2007 (which is the default configuration). OpenSSL TLS clients are not impacted by
2011 *Peter Kästle and Samuel Sapalski*
2013 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2015 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2016 create a unique hash value based on the issuer and serial number data
2017 contained within an X509 certificate. However it was failing to correctly
2018 handle any errors that may occur while parsing the issuer field (which might
2019 occur if the issuer field is maliciously constructed). This may subsequently
2020 result in a NULL pointer deref and a crash leading to a potential denial of
2026 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2027 padding mode to correctly check for rollback attacks. This is considered a
2028 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2033 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2034 functions. Previously they could overflow the output length argument in some
2035 cases where the input length is close to the maximum permissable length for
2036 an integer on the platform. In such cases the return value from the function
2037 call would be 1 (indicating success), but the output length value would be
2038 negative. This could cause applications to behave incorrectly or crash.
2043 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2044 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2045 could be exploited in a side channel attack to recover the password. Since
2046 the attack is local host only this is outside of the current OpenSSL
2047 threat model and therefore no CVE is assigned.
2049 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2054 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2056 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2057 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2058 If an attacker can control both items being compared then this could lead
2059 to a possible denial of service attack. OpenSSL itself uses the
2060 GENERAL_NAME_cmp function for two purposes:
2061 1) Comparing CRL distribution point names between an available CRL and a
2062 CRL distribution point embedded in an X509 certificate
2063 2) When verifying that a timestamp response token signer matches the
2064 timestamp authority name (exposed via the API functions
2065 TS_RESP_verify_response and TS_RESP_verify_token)
2070 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2072 * Certificates with explicit curve parameters are now disallowed in
2073 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2077 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2078 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2079 conversely, silently ignore DTLS protocol version bounds when configuring
2080 TLS-based contexts. The commands can be repeated to set bounds of both
2081 types. The same applies with the corresponding "min_protocol" and
2082 "max_protocol" command-line switches, in case some application uses both TLS
2085 SSL_CTX instances that are created for a fixed protocol version (e.g.
2086 TLSv1_server_method()) also silently ignore version bounds. Previously
2087 attempts to apply bounds to these protocol versions would result in an
2088 error. Now only the "version-flexible" SSL_CTX instances are subject to
2089 limits in configuration files in command-line options.
2093 * Handshake now fails if Extended Master Secret extension is dropped
2098 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2100 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2102 * Fixed segmentation fault in SSL_check_chain()
2103 Server or client applications that call the SSL_check_chain() function
2104 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2105 dereference as a result of incorrect handling of the
2106 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2107 or unrecognised signature algorithm is received from the peer. This could
2108 be exploited by a malicious peer in a Denial of Service attack.
2113 * Added AES consttime code for no-asm configurations
2114 an optional constant time support for AES was added
2115 when building openssl for no-asm.
2116 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2117 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2118 At this time this feature is by default disabled.
2119 It will be enabled by default in 3.0.
2123 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2125 * Revert the change of EOF detection while reading in libssl to avoid
2126 regressions in applications depending on the current way of reporting
2127 the EOF. As the existing method is not fully accurate the change to
2128 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2129 branch and will be present in the 3.0 release.
2133 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2134 when primes for RSA keys are computed.
2135 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2136 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2137 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2138 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2139 This avoids possible fingerprinting of newly generated RSA modules.
2143 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2145 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2146 while reading in libssl then we would report an error back to the
2147 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2148 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2149 therefore give a hint as to what went wrong.
2153 * Check that ed25519 and ed448 are allowed by the security level. Previously
2154 signature algorithms not using an MD were not being checked that they were
2155 allowed by the security level.
2159 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2160 was not quite right. The behaviour was not consistent between resumption
2161 and normal handshakes, and also not quite consistent with historical
2162 behaviour. The behaviour in various scenarios has been clarified and
2163 it has been updated to make it match historical behaviour as closely as
2168 * *[VMS only]* The header files that the VMS compilers include automatically,
2169 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2170 that the C++ compiler doesn't understand. This is a shortcoming in the
2171 compiler, but can be worked around with `__cplusplus` guards.
2173 C++ applications that use OpenSSL libraries must be compiled using the
2174 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2175 functions. Otherwise, only functions with symbols of less than 31
2176 characters can be used, as the linker will not be able to successfully
2177 resolve symbols with longer names.
2181 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2182 The presence of this system service is determined at run-time.
2186 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2191 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2193 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2194 number generator (RNG). This was intended to include protection in the
2195 event of a fork() system call in order to ensure that the parent and child
2196 processes did not share the same RNG state. However this protection was not
2197 being used in the default case.
2199 A partial mitigation for this issue is that the output from a high
2200 precision timer is mixed into the RNG state so the likelihood of a parent
2201 and child process sharing state is significantly reduced.
2203 If an application already calls OPENSSL_init_crypto() explicitly using
2204 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2207 *Matthias St. Pierre*
2209 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2210 used even when parsing explicit parameters, when loading a encoded key
2211 or calling `EC_GROUP_new_from_ecpkparameters()`/
2212 `EC_GROUP_new_from_ecparameters()`.
2213 This prevents bypass of security hardening and performance gains,
2214 especially for curves with specialized EC_METHODs.
2215 By default, if a key encoded with explicit parameters is loaded and later
2216 encoded, the output is still encoded with explicit parameters, even if
2217 internally a "named" EC_GROUP is used for computation.
2221 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2222 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2223 NULL. After this change, only the cofactor parameter can be NULL. It also
2224 does some minimal sanity checks on the passed order.
2229 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2230 An attack is simple, if the first CMS_recipientInfo is valid but the
2231 second CMS_recipientInfo is chosen ciphertext. If the second
2232 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2233 encryption key will be replaced by garbage, and the message cannot be
2234 decoded, but if the RSA decryption fails, the correct encryption key is
2235 used and the recipient will not notice the attack.
2236 As a work around for this potential attack the length of the decrypted
2237 key must be equal to the cipher default key length, in case the
2238 certifiate is not given and all recipientInfo are tried out.
2239 The old behaviour can be re-enabled in the CMS code by setting the
2240 CMS_DEBUG_DECRYPT flag.
2245 * Early start up entropy quality from the DEVRANDOM seed source has been
2246 improved for older Linux systems. The RAND subsystem will wait for
2247 /dev/random to be producing output before seeding from /dev/urandom.
2248 The seeded state is stored for future library initialisations using
2249 a system global shared memory segment. The shared memory identifier
2250 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2251 the desired value. The default identifier is 114.
2255 * Correct the extended master secret constant on EBCDIC systems. Without this
2256 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2257 negotiate EMS will fail. Unfortunately this also means that TLS connections
2258 between EBCDIC systems with this fix, and EBCDIC systems without this
2259 fix will fail if they negotiate EMS.
2263 * Use Windows installation paths in the mingw builds
2265 Mingw isn't a POSIX environment per se, which means that Windows
2266 paths should be used for installation.
2271 * Changed DH_check to accept parameters with order q and 2q subgroups.
2272 With order 2q subgroups the bit 0 of the private key is not secret
2273 but DH_generate_key works around that by clearing bit 0 of the
2274 private key for those. This avoids leaking bit 0 of the private key.
2278 * Significantly reduce secure memory usage by the randomness pools.
2282 * Revert the DEVRANDOM_WAIT feature for Linux systems
2284 The DEVRANDOM_WAIT feature added a select() call to wait for the
2285 /dev/random device to become readable before reading from the
2286 /dev/urandom device.
2288 It turned out that this change had negative side effects on
2289 performance which were not acceptable. After some discussion it
2290 was decided to revert this feature and leave it up to the OS
2291 resp. the platform maintainer to ensure a proper initialization
2292 during early boot time.
2294 *Matthias St. Pierre*
2296 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2298 * Add build tests for C++. These are generated files that only do one
2299 thing, to include one public OpenSSL head file each. This tests that
2300 the public header files can be usefully included in a C++ application.
2302 This test isn't enabled by default. It can be enabled with the option
2303 'enable-buildtest-c++'.
2307 * Enable SHA3 pre-hashing for ECDSA and DSA.
2311 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2312 This changes the size when using the `genpkey` command when no size is given.
2313 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2314 generation commands to use 2048 bits by default.
2318 * Reorganize the manual pages to consistently have RETURN VALUES,
2319 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2320 util/fix-doc-nits accordingly.
2322 *Paul Yang, Joshua Lock*
2324 * Add the missing accessor EVP_PKEY_get0_engine()
2328 * Have commands like `s_client` and `s_server` output the signature scheme
2329 along with other cipher suite parameters when debugging.
2333 * Make OPENSSL_config() error agnostic again.
2337 * Do the error handling in RSA decryption constant time.
2341 * Prevent over long nonces in ChaCha20-Poly1305.
2343 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2344 for every encryption operation. RFC 7539 specifies that the nonce value
2345 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2346 and front pads the nonce with 0 bytes if it is less than 12
2347 bytes. However it also incorrectly allows a nonce to be set of up to 16
2348 bytes. In this case only the last 12 bytes are significant and any
2349 additional leading bytes are ignored.
2351 It is a requirement of using this cipher that nonce values are
2352 unique. Messages encrypted using a reused nonce value are susceptible to
2353 serious confidentiality and integrity attacks. If an application changes
2354 the default nonce length to be longer than 12 bytes and then makes a
2355 change to the leading bytes of the nonce expecting the new value to be a
2356 new unique nonce then such an application could inadvertently encrypt
2357 messages with a reused nonce.
2359 Additionally the ignored bytes in a long nonce are not covered by the
2360 integrity guarantee of this cipher. Any application that relies on the
2361 integrity of these ignored leading bytes of a long nonce may be further
2362 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2363 is safe because no such use sets such a long nonce value. However user
2364 applications that use this cipher directly and set a non-default nonce
2365 length to be longer than 12 bytes may be vulnerable.
2367 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2373 * Add DEVRANDOM_WAIT feature for Linux systems
2375 On older Linux systems where the getrandom() system call is not available,
2376 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2377 Contrary to getrandom(), the /dev/urandom device will not block during
2378 early boot when the kernel CSPRNG has not been seeded yet.
2380 To mitigate this known weakness, use select() to wait for /dev/random to
2381 become readable before reading from /dev/urandom.
2383 * Ensure that SM2 only uses SM3 as digest algorithm
2387 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2389 * Change the info callback signals for the start and end of a post-handshake
2390 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2391 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2392 confused by this and assume that a TLSv1.2 renegotiation has started. This
2393 can break KeyUpdate handling. Instead we no longer signal the start and end
2394 of a post handshake message exchange (although the messages themselves are
2395 still signalled). This could break some applications that were expecting
2396 the old signals. However without this KeyUpdate is not usable for many
2401 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2403 * Timing vulnerability in DSA signature generation
2405 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2406 timing side channel attack. An attacker could use variations in the signing
2407 algorithm to recover the private key.
2409 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2414 * Timing vulnerability in ECDSA signature generation
2416 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2417 timing side channel attack. An attacker could use variations in the signing
2418 algorithm to recover the private key.
2420 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2425 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2426 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2427 of two gigabytes and the error handling improved.
2429 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2430 categorized as a normal bug, not a security issue, because the DRBG reseeds
2431 automatically and is fully functional even without additional randomness
2432 provided by the application.
2434 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2436 * Add a new ClientHello callback. Provides a callback interface that gives
2437 the application the ability to adjust the nascent SSL object at the
2438 earliest stage of ClientHello processing, immediately after extensions have
2439 been collected but before they have been processed. In particular, this
2440 callback can adjust the supported TLS versions in response to the contents
2445 * Add SM2 base algorithm support.
2449 * s390x assembly pack: add (improved) hardware-support for the following
2450 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2451 aes-cfb/cfb8, aes-ecb.
2455 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2456 parameter is no longer accepted, as it leads to a corrupt table. NULL
2457 pem_str is reserved for alias entries only.
2461 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2462 step for prime curves. The new implementation is based on formulae from
2463 differential addition-and-doubling in homogeneous projective coordinates
2464 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2465 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2466 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2467 to work in projective coordinates.
2469 *Billy Bob Brumley, Nicola Tuveri*
2471 * Change generating and checking of primes so that the error rate of not
2472 being prime depends on the intended use based on the size of the input.
2473 For larger primes this will result in more rounds of Miller-Rabin.
2474 The maximal error rate for primes with more than 1080 bits is lowered
2477 *Kurt Roeckx, Annie Yousar*
2479 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2483 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2484 moving between systems, and to avoid confusion when a Windows build is
2485 done with mingw vs with MSVC. For POSIX installs, there's still a
2486 symlink or copy named 'tsget' to avoid that confusion as well.
2490 * Revert blinding in ECDSA sign and instead make problematic addition
2491 length-invariant. Switch even to fixed-length Montgomery multiplication.
2495 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2496 step for binary curves. The new implementation is based on formulae from
2497 differential addition-and-doubling in mixed Lopez-Dahab projective
2498 coordinates, modified to independently blind the operands.
2500 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2502 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2503 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2504 EC_METHODs to implement their own specialized "ladder step", to take
2505 advantage of more favorable coordinate systems or more efficient
2506 differential addition-and-doubling algorithms.
2508 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2510 * Modified the random device based seed sources to keep the relevant
2511 file descriptors open rather than reopening them on each access.
2512 This allows such sources to operate in a chroot() jail without
2513 the associated device nodes being available. This behaviour can be
2514 controlled using RAND_keep_random_devices_open().
2518 * Numerous side-channel attack mitigations have been applied. This may have
2519 performance impacts for some algorithms for the benefit of improved
2520 security. Specific changes are noted in this change log by their respective
2525 * AIX shared library support overhaul. Switch to AIX "natural" way of
2526 handling shared libraries, which means collecting shared objects of
2527 different versions and bitnesses in one common archive. This allows to
2528 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2529 doesn't affect the way 3rd party applications are linked, only how
2530 multi-version installation is managed.
2534 * Make ec_group_do_inverse_ord() more robust and available to other
2535 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2536 mitigations are applied to the fallback BN_mod_inverse().
2537 When using this function rather than BN_mod_inverse() directly, new
2538 EC cryptosystem implementations are then safer-by-default.
2542 * Add coordinate blinding for EC_POINT and implement projective
2543 coordinate blinding for generic prime curves as a countermeasure to
2544 chosen point SCA attacks.
2546 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2548 * Add blinding to ECDSA and DSA signatures to protect against side channel
2549 attacks discovered by Keegan Ryan (NCC Group).
2553 * Enforce checking in the `pkeyutl` command to ensure that the input
2554 length does not exceed the maximum supported digest length when performing
2555 a sign, verify or verifyrecover operation.
2559 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2560 I/O in combination with something like select() or poll() will hang. This
2561 can be turned off again using SSL_CTX_clear_mode().
2562 Many applications do not properly handle non-application data records, and
2563 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2564 around the problems in those applications, but can also break some.
2565 It's recommended to read the manpages about SSL_read(), SSL_write(),
2566 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2567 SSL_CTX_set_read_ahead() again.
2571 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2572 now allow empty (zero character) pass phrases.
2576 * Apply blinding to binary field modular inversion and remove patent
2577 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2581 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2582 binary and prime elliptic curves.
2586 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2587 constant time fixed point multiplication.
2591 * Revise elliptic curve scalar multiplication with timing attack
2592 defenses: ec_wNAF_mul redirects to a constant time implementation
2593 when computing fixed point and variable point multiplication (which
2594 in OpenSSL are mostly used with secret scalars in keygen, sign,
2595 ECDH derive operations).
2596 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2599 * Updated CONTRIBUTING
2603 * Updated DRBG / RAND to request nonce and additional low entropy
2604 randomness from the system.
2606 *Matthias St. Pierre*
2608 * Updated 'openssl rehash' to use OpenSSL consistent default.
2612 * Moved the load of the ssl_conf module to libcrypto, which helps
2613 loading engines that libssl uses before libssl is initialised.
2617 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2621 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2623 *Ingo Schwarze, Rich Salz*
2625 * Added output of accepting IP address and port for 'openssl s_server'
2629 * Added a new API for TLSv1.3 ciphersuites:
2630 SSL_CTX_set_ciphersuites()
2631 SSL_set_ciphersuites()
2635 * Memory allocation failures consistently add an error to the error
2640 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2641 in libcrypto when run as setuid/setgid.
2645 * Load any config file by default when libssl is used.
2649 * Added new public header file <openssl/rand_drbg.h> and documentation
2650 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2652 *Matthias St. Pierre*
2654 * QNX support removed (cannot find contributors to get their approval
2655 for the license change).
2659 * TLSv1.3 replay protection for early data has been implemented. See the
2660 SSL_read_early_data() man page for further details.
2664 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2665 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2666 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2667 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2668 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2669 configuration has been separated out. See the ciphers man page or the
2670 SSL_CTX_set_ciphersuites() man page for more information.
2674 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2675 in responder mode now supports the new "-multi" option, which
2676 spawns the specified number of child processes to handle OCSP
2677 requests. The "-timeout" option now also limits the OCSP
2678 responder's patience to wait to receive the full client request
2679 on a newly accepted connection. Child processes are respawned
2680 as needed, and the CA index file is automatically reloaded
2681 when changed. This makes it possible to run the "ocsp" responder
2682 as a long-running service, making the OpenSSL CA somewhat more
2683 feature-complete. In this mode, most diagnostic messages logged
2684 after entering the event loop are logged via syslog(3) rather than
2689 * Added support for X448 and Ed448. Heavily based on original work by
2694 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2695 objects loaded. This adds the functions OSSL_STORE_expect() and
2696 OSSL_STORE_find() as well as needed tools to construct searches and
2697 get the search data out of them.
2701 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2702 version of OpenSSL should review their configuration settings to ensure
2703 that they are still appropriate for TLSv1.3. For further information see:
2704 <https://wiki.openssl.org/index.php/TLS1.3>
2708 * Grand redesign of the OpenSSL random generator
2710 The default RAND method now utilizes an AES-CTR DRBG according to
2711 NIST standard SP 800-90Ar1. The new random generator is essentially
2712 a port of the default random generator from the OpenSSL FIPS 2.0
2713 object module. It is a hybrid deterministic random bit generator
2714 using an AES-CTR bit stream and which seeds and reseeds itself
2715 automatically using trusted system entropy sources.
2717 Some of its new features are:
2718 - Support for multiple DRBG instances with seed chaining.
2719 - The default RAND method makes use of a DRBG.
2720 - There is a public and private DRBG instance.
2721 - The DRBG instances are fork-safe.
2722 - Keep all global DRBG instances on the secure heap if it is enabled.
2723 - The public and private DRBG instance are per thread for lock free
2726 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2728 * Changed Configure so it only says what it does and doesn't dump
2729 so much data. Instead, ./configdata.pm should be used as a script
2730 to display all sorts of configuration data.
2734 * Added processing of "make variables" to Configure.
2738 * Added SHA512/224 and SHA512/256 algorithm support.
2742 * The last traces of Netware support, first removed in 1.1.0, have
2747 * Get rid of Makefile.shared, and in the process, make the processing
2748 of certain files (rc.obj, or the .def/.map/.opt files produced from
2749 the ordinal files) more visible and hopefully easier to trace and
2750 debug (or make silent).
2754 * Make it possible to have environment variable assignments as
2755 arguments to config / Configure.
2759 * Add multi-prime RSA (RFC 8017) support.
2763 * Add SM3 implemented according to GB/T 32905-2016
2764 *Jack Lloyd <jack.lloyd@ribose.com>,*
2765 *Ronald Tse <ronald.tse@ribose.com>,*
2766 *Erick Borsboom <erick.borsboom@ribose.com>*
2768 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2769 as documented in RFC6066.
2770 Based on a patch from Tomasz Moń
2772 *Filipe Raimundo da Silva*
2774 * Add SM4 implemented according to GB/T 32907-2016.
2775 *Jack Lloyd <jack.lloyd@ribose.com>,*
2776 *Ronald Tse <ronald.tse@ribose.com>,*
2777 *Erick Borsboom <erick.borsboom@ribose.com>*
2779 * Reimplement -newreq-nodes and ERR_error_string_n; the
2780 original author does not agree with the license change.
2784 * Add ARIA AEAD TLS support.
2788 * Some macro definitions to support VS6 have been removed. Visual
2789 Studio 6 has not worked since 1.1.0
2793 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2794 without clearing the errors.
2798 * Add "atfork" functions. If building on a system that without
2799 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2800 requirements. The RAND facility now uses/requires this.
2808 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2809 not possible to disable entirely. However, it's still possible to
2810 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2813 To disable, configure with 'no-ui-console'. 'no-ui' is still
2814 possible to use as an alias. Check at compile time with the
2815 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2816 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2820 * Add a STORE module, which implements a uniform and URI based reader of
2821 stores that can contain keys, certificates, CRLs and numerous other
2822 objects. The main API is loosely based on a few stdio functions,
2823 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2824 OSSL_STORE_error and OSSL_STORE_close.
2825 The implementation uses backends called "loaders" to implement arbitrary
2826 URI schemes. There is one built in "loader" for the 'file' scheme.
2830 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2831 then adjusted to work on FreeBSD 8.4 as well.
2832 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2833 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2837 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2838 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2839 error code calls like this:
2841 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2843 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2844 that can be encoded in C. For the foreseeable future, this will only
2847 *Richard Levitte and Tim Hudson*
2849 * Removed BSD cryptodev engine.
2853 * Add a build target 'build_all_generated', to build all generated files
2854 and only that. This can be used to prepare everything that requires
2855 things like perl for a system that lacks perl and then move everything
2856 to that system and do the rest of the build there.
2860 * In the UI interface, make it possible to duplicate the user data. This
2861 can be used by engines that need to retain the data for a longer time
2862 than just the call where this user data is passed.
2866 * Ignore the '-named_curve auto' value for compatibility of applications
2869 *Tomáš Mráz <tmraz@fedoraproject.org>*
2871 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2872 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2873 alerts across multiple records (some of which could be empty). In practice
2874 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2875 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2876 support this at all. Supporting it adds significant complexity to the
2877 record layer, and its removal is unlikely to cause interoperability
2882 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2883 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2884 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2889 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2890 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2892 *Richard Levitte, Andy Polyakov*
2894 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2899 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2900 platform rather than 'mingw'.
2904 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2905 success if they are asked to add an object which already exists
2906 in the store. This change cascades to other functions which load
2907 certificates and CRLs.
2911 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2912 facilitate stack unwinding even from assembly subroutines.
2916 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2917 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2921 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2922 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2923 which is the minimum version we support.
2927 * Certificate time validation (X509_cmp_time) enforces stricter
2928 compliance with RFC 5280. Fractional seconds and timezone offsets
2929 are no longer allowed.
2933 * Add support for ARIA
2937 * s_client will now send the Server Name Indication (SNI) extension by
2938 default unless the new "-noservername" option is used. The server name is
2939 based on the host provided to the "-connect" option unless overridden by
2940 using "-servername".
2944 * Add support for SipHash
2948 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2949 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2950 prevent issues where no progress is being made and the peer continually
2951 sends unrecognised record types, using up resources processing them.
2955 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2956 using the algorithm defined in
2957 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2961 * Heartbeat support has been removed; the ABI is changed for now.
2963 *Richard Levitte, Rich Salz*
2965 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2969 * The RSA "null" method, which was partially supported to avoid patent
2970 issues, has been replaced to always returns NULL.
2977 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2979 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2980 used even when parsing explicit parameters, when loading a encoded key
2981 or calling `EC_GROUP_new_from_ecpkparameters()`/
2982 `EC_GROUP_new_from_ecparameters()`.
2983 This prevents bypass of security hardening and performance gains,
2984 especially for curves with specialized EC_METHODs.
2985 By default, if a key encoded with explicit parameters is loaded and later
2986 encoded, the output is still encoded with explicit parameters, even if
2987 internally a "named" EC_GROUP is used for computation.
2991 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2992 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2993 NULL. After this change, only the cofactor parameter can be NULL. It also
2994 does some minimal sanity checks on the passed order.
2999 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3000 An attack is simple, if the first CMS_recipientInfo is valid but the
3001 second CMS_recipientInfo is chosen ciphertext. If the second
3002 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3003 encryption key will be replaced by garbage, and the message cannot be
3004 decoded, but if the RSA decryption fails, the correct encryption key is
3005 used and the recipient will not notice the attack.
3006 As a work around for this potential attack the length of the decrypted
3007 key must be equal to the cipher default key length, in case the
3008 certifiate is not given and all recipientInfo are tried out.
3009 The old behaviour can be re-enabled in the CMS code by setting the
3010 CMS_DEBUG_DECRYPT flag.
3015 * Use Windows installation paths in the mingw builds
3017 Mingw isn't a POSIX environment per se, which means that Windows
3018 paths should be used for installation.
3023 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3025 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3026 This changes the size when using the `genpkey` command when no size is given.
3027 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3028 generation commands to use 2048 bits by default.
3032 * Prevent over long nonces in ChaCha20-Poly1305.
3034 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3035 for every encryption operation. RFC 7539 specifies that the nonce value
3036 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3037 and front pads the nonce with 0 bytes if it is less than 12
3038 bytes. However it also incorrectly allows a nonce to be set of up to 16
3039 bytes. In this case only the last 12 bytes are significant and any
3040 additional leading bytes are ignored.
3042 It is a requirement of using this cipher that nonce values are
3043 unique. Messages encrypted using a reused nonce value are susceptible to
3044 serious confidentiality and integrity attacks. If an application changes
3045 the default nonce length to be longer than 12 bytes and then makes a
3046 change to the leading bytes of the nonce expecting the new value to be a
3047 new unique nonce then such an application could inadvertently encrypt
3048 messages with a reused nonce.
3050 Additionally the ignored bytes in a long nonce are not covered by the
3051 integrity guarantee of this cipher. Any application that relies on the
3052 integrity of these ignored leading bytes of a long nonce may be further
3053 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3054 is safe because no such use sets such a long nonce value. However user
3055 applications that use this cipher directly and set a non-default nonce
3056 length to be longer than 12 bytes may be vulnerable.
3058 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3064 * Added SCA hardening for modular field inversion in EC_GROUP through
3065 a new dedicated field_inv() pointer in EC_METHOD.
3066 This also addresses a leakage affecting conversions from projective
3067 to affine coordinates.
3069 *Billy Bob Brumley, Nicola Tuveri*
3071 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3072 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3076 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3080 * Remove the 'dist' target and add a tarball building script. The
3081 'dist' target has fallen out of use, and it shouldn't be
3082 necessary to configure just to create a source distribution.
3086 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3088 * Timing vulnerability in DSA signature generation
3090 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3091 timing side channel attack. An attacker could use variations in the signing
3092 algorithm to recover the private key.
3094 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3099 * Timing vulnerability in ECDSA signature generation
3101 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3102 timing side channel attack. An attacker could use variations in the signing
3103 algorithm to recover the private key.
3105 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3110 * Add coordinate blinding for EC_POINT and implement projective
3111 coordinate blinding for generic prime curves as a countermeasure to
3112 chosen point SCA attacks.
3114 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3116 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3118 * Client DoS due to large DH parameter
3120 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3121 malicious server can send a very large prime value to the client. This will
3122 cause the client to spend an unreasonably long period of time generating a
3123 key for this prime resulting in a hang until the client has finished. This
3124 could be exploited in a Denial Of Service attack.
3126 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3131 * Cache timing vulnerability in RSA Key Generation
3133 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3134 a cache timing side channel attack. An attacker with sufficient access to
3135 mount cache timing attacks during the RSA key generation process could
3136 recover the private key.
3138 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3139 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3144 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3145 parameter is no longer accepted, as it leads to a corrupt table. NULL
3146 pem_str is reserved for alias entries only.
3150 * Revert blinding in ECDSA sign and instead make problematic addition
3151 length-invariant. Switch even to fixed-length Montgomery multiplication.
3155 * Change generating and checking of primes so that the error rate of not
3156 being prime depends on the intended use based on the size of the input.
3157 For larger primes this will result in more rounds of Miller-Rabin.
3158 The maximal error rate for primes with more than 1080 bits is lowered
3161 *Kurt Roeckx, Annie Yousar*
3163 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3167 * Add blinding to ECDSA and DSA signatures to protect against side channel
3168 attacks discovered by Keegan Ryan (NCC Group).
3172 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3173 now allow empty (zero character) pass phrases.
3177 * Certificate time validation (X509_cmp_time) enforces stricter
3178 compliance with RFC 5280. Fractional seconds and timezone offsets
3179 are no longer allowed.
3183 * Fixed a text canonicalisation bug in CMS
3185 Where a CMS detached signature is used with text content the text goes
3186 through a canonicalisation process first prior to signing or verifying a
3187 signature. This process strips trailing space at the end of lines, converts
3188 line terminators to CRLF and removes additional trailing line terminators
3189 at the end of a file. A bug in the canonicalisation process meant that
3190 some characters, such as form-feed, were incorrectly treated as whitespace
3191 and removed. This is contrary to the specification (RFC5485). This fix
3192 could mean that detached text data signed with an earlier version of
3193 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3194 signed with a fixed OpenSSL may fail to verify with an earlier version of
3195 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3196 and use the "-binary" flag (for the "cms" command line application) or set
3197 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3201 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3203 * Constructed ASN.1 types with a recursive definition could exceed the stack
3205 Constructed ASN.1 types with a recursive definition (such as can be found
3206 in PKCS7) could eventually exceed the stack given malicious input with
3207 excessive recursion. This could result in a Denial Of Service attack. There
3208 are no such structures used within SSL/TLS that come from untrusted sources
3209 so this is considered safe.
3211 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3217 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3219 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3220 effectively reduced to only comparing the least significant bit of each
3221 byte. This allows an attacker to forge messages that would be considered as
3222 authenticated in an amount of tries lower than that guaranteed by the
3223 security claims of the scheme. The module can only be compiled by the
3224 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3226 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3232 * Add a build target 'build_all_generated', to build all generated files
3233 and only that. This can be used to prepare everything that requires
3234 things like perl for a system that lacks perl and then move everything
3235 to that system and do the rest of the build there.
3239 * Backport SSL_OP_NO_RENGOTIATION
3241 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3242 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3243 changes this is no longer possible in 1.1.0. Therefore the new
3244 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3245 1.1.0 to provide equivalent functionality.
3247 Note that if an application built against 1.1.0h headers (or above) is run
3248 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3249 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3253 * Removed the OS390-Unix config target. It relied on a script that doesn't
3258 * rsaz_1024_mul_avx2 overflow bug on x86_64
3260 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3261 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3262 Analysis suggests that attacks against RSA and DSA as a result of this
3263 defect would be very difficult to perform and are not believed likely.
3264 Attacks against DH1024 are considered just feasible, because most of the
3265 work necessary to deduce information about a private key may be performed
3266 offline. The amount of resources required for such an attack would be
3267 significant. However, for an attack on TLS to be meaningful, the server
3268 would have to share the DH1024 private key among multiple clients, which is
3269 no longer an option since CVE-2016-0701.
3271 This only affects processors that support the AVX2 but not ADX extensions
3272 like Intel Haswell (4th generation).
3274 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3275 was originally found via the OSS-Fuzz project.
3280 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3282 * bn_sqrx8x_internal carry bug on x86_64
3284 There is a carry propagating bug in the x86_64 Montgomery squaring
3285 procedure. No EC algorithms are affected. Analysis suggests that attacks
3286 against RSA and DSA as a result of this defect would be very difficult to
3287 perform and are not believed likely. Attacks against DH are considered just
3288 feasible (although very difficult) because most of the work necessary to
3289 deduce information about a private key may be performed offline. The amount
3290 of resources required for such an attack would be very significant and
3291 likely only accessible to a limited number of attackers. An attacker would
3292 additionally need online access to an unpatched system using the target
3293 private key in a scenario with persistent DH parameters and a private
3294 key that is shared between multiple clients.
3296 This only affects processors that support the BMI1, BMI2 and ADX extensions
3297 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3299 This issue was reported to OpenSSL by the OSS-Fuzz project.
3304 * Malformed X.509 IPAddressFamily could cause OOB read
3306 If an X.509 certificate has a malformed IPAddressFamily extension,
3307 OpenSSL could do a one-byte buffer overread. The most likely result
3308 would be an erroneous display of the certificate in text format.
3310 This issue was reported to OpenSSL by the OSS-Fuzz project.
3315 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3317 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3318 platform rather than 'mingw'.
3322 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3323 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3324 which is the minimum version we support.
3328 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3330 * Encrypt-Then-Mac renegotiation crash
3332 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3333 negotiated where it was not in the original handshake (or vice-versa) then
3334 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3335 and servers are affected.
3337 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3342 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3344 * Truncated packet could crash via OOB read
3346 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3347 cipher is being used, then a truncated packet can cause that host to
3348 perform an out-of-bounds read, usually resulting in a crash.
3350 This issue was reported to OpenSSL by Robert Święcki of Google.
3355 * Bad (EC)DHE parameters cause a client crash
3357 If a malicious server supplies bad parameters for a DHE or ECDHE key
3358 exchange then this can result in the client attempting to dereference a
3359 NULL pointer leading to a client crash. This could be exploited in a Denial
3362 This issue was reported to OpenSSL by Guido Vranken.
3367 * BN_mod_exp may produce incorrect results on x86_64
3369 There is a carry propagating bug in the x86_64 Montgomery squaring
3370 procedure. No EC algorithms are affected. Analysis suggests that attacks
3371 against RSA and DSA as a result of this defect would be very difficult to
3372 perform and are not believed likely. Attacks against DH are considered just
3373 feasible (although very difficult) because most of the work necessary to
3374 deduce information about a private key may be performed offline. The amount
3375 of resources required for such an attack would be very significant and
3376 likely only accessible to a limited number of attackers. An attacker would
3377 additionally need online access to an unpatched system using the target
3378 private key in a scenario with persistent DH parameters and a private
3379 key that is shared between multiple clients. For example this can occur by
3380 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3381 similar to CVE-2015-3193 but must be treated as a separate problem.
3383 This issue was reported to OpenSSL by the OSS-Fuzz project.
3388 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3390 * ChaCha20/Poly1305 heap-buffer-overflow
3392 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3393 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3394 crash. This issue is not considered to be exploitable beyond a DoS.
3396 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3401 * CMS Null dereference
3403 Applications parsing invalid CMS structures can crash with a NULL pointer
3404 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3405 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3406 structure callback if an attempt is made to free certain invalid encodings.
3407 Only CHOICE structures using a callback which do not handle NULL value are
3410 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3415 * Montgomery multiplication may produce incorrect results
3417 There is a carry propagating bug in the Broadwell-specific Montgomery
3418 multiplication procedure that handles input lengths divisible by, but
3419 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3420 and DH private keys are impossible. This is because the subroutine in
3421 question is not used in operations with the private key itself and an input
3422 of the attacker's direct choice. Otherwise the bug can manifest itself as
3423 transient authentication and key negotiation failures or reproducible
3424 erroneous outcome of public-key operations with specially crafted input.
3425 Among EC algorithms only Brainpool P-512 curves are affected and one
3426 presumably can attack ECDH key negotiation. Impact was not analyzed in
3427 detail, because pre-requisites for attack are considered unlikely. Namely
3428 multiple clients have to choose the curve in question and the server has to
3429 share the private key among them, neither of which is default behaviour.
3430 Even then only clients that chose the curve will be affected.
3432 This issue was publicly reported as transient failures and was not
3433 initially recognized as a security issue. Thanks to Richard Morgan for
3434 providing reproducible case.
3439 * Removed automatic addition of RPATH in shared libraries and executables,
3440 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3444 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3446 * Fix Use After Free for large message sizes
3448 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3449 message larger than approx 16k is received then the underlying buffer to
3450 store the incoming message is reallocated and moved. Unfortunately a
3451 dangling pointer to the old location is left which results in an attempt to
3452 write to the previously freed location. This is likely to result in a
3453 crash, however it could potentially lead to execution of arbitrary code.
3455 This issue only affects OpenSSL 1.1.0a.
3457 This issue was reported to OpenSSL by Robert Święcki.
3462 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3464 * OCSP Status Request extension unbounded memory growth
3466 A malicious client can send an excessively large OCSP Status Request
3467 extension. If that client continually requests renegotiation, sending a
3468 large OCSP Status Request extension each time, then there will be unbounded
3469 memory growth on the server. This will eventually lead to a Denial Of
3470 Service attack through memory exhaustion. Servers with a default
3471 configuration are vulnerable even if they do not support OCSP. Builds using
3472 the "no-ocsp" build time option are not affected.
3474 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3479 * SSL_peek() hang on empty record
3481 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3482 sends an empty record. This could be exploited by a malicious peer in a
3483 Denial Of Service attack.
3485 This issue was reported to OpenSSL by Alex Gaynor.
3490 * Excessive allocation of memory in tls_get_message_header() and
3491 dtls1_preprocess_fragment()
3493 A (D)TLS message includes 3 bytes for its length in the header for the
3494 message. This would allow for messages up to 16Mb in length. Messages of
3495 this length are excessive and OpenSSL includes a check to ensure that a
3496 peer is sending reasonably sized messages in order to avoid too much memory
3497 being consumed to service a connection. A flaw in the logic of version
3498 1.1.0 means that memory for the message is allocated too early, prior to
3499 the excessive message length check. Due to way memory is allocated in
3500 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3501 to service a connection. This could lead to a Denial of Service through
3502 memory exhaustion. However, the excessive message length check still takes
3503 place, and this would cause the connection to immediately fail. Assuming
3504 that the application calls SSL_free() on the failed connection in a timely
3505 manner then the 21Mb of allocated memory will then be immediately freed
3506 again. Therefore the excessive memory allocation will be transitory in
3507 nature. This then means that there is only a security impact if:
3509 1) The application does not call SSL_free() in a timely manner in the event
3510 that the connection fails
3512 2) The application is working in a constrained environment where there is
3513 very little free memory
3515 3) The attacker initiates multiple connection attempts such that there are
3516 multiple connections in a state where memory has been allocated for the
3517 connection; SSL_free() has not yet been called; and there is insufficient
3518 memory to service the multiple requests.
3520 Except in the instance of (1) above any Denial Of Service is likely to be
3521 transitory because as soon as the connection fails the memory is
3522 subsequently freed again in the SSL_free() call. However there is an
3523 increased risk during this period of application crashes due to the lack of
3524 memory - which would then mean a more serious Denial of Service.
3526 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3527 (CVE-2016-6307 and CVE-2016-6308)
3531 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3532 had to be removed. Primary reason is that vendor assembler can't
3533 assemble our modules with -KPIC flag. As result it, assembly
3534 support, was not even available as option. But its lack means
3535 lack of side-channel resistant code, which is incompatible with
3536 security by todays standards. Fortunately gcc is readily available
3537 prepackaged option, which we firmly point at...
3541 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3543 * Windows command-line tool supports UTF-8 opt-in option for arguments
3544 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3545 (to any value) allows Windows user to access PKCS#12 file generated
3546 with Windows CryptoAPI and protected with non-ASCII password, as well
3547 as files generated under UTF-8 locale on Linux also protected with
3552 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3553 have been disabled by default and removed from DEFAULT, just like RC4.
3554 See the RC4 item below to re-enable both.
3558 * The method for finding the storage location for the Windows RAND seed file
3559 has changed. First we check %RANDFILE%. If that is not set then we check
3560 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3561 all else fails we fall back to C:\.
3565 * The EVP_EncryptUpdate() function has had its return type changed from void
3566 to int. A return of 0 indicates and error while a return of 1 indicates
3571 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3572 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3573 off the constant time implementation for RSA, DSA and DH have been made
3574 no-ops and deprecated.
3578 * Windows RAND implementation was simplified to only get entropy by
3579 calling CryptGenRandom(). Various other RAND-related tickets
3582 *Joseph Wylie Yandle, Rich Salz*
3584 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3585 and `OPENSSL_LH_`, respectively. The old names are available
3586 with API compatibility. They new names are now completely documented.
3590 * Unify TYPE_up_ref(obj) methods signature.
3591 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3592 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3593 int (instead of void) like all others TYPE_up_ref() methods.
3594 So now these methods also check the return value of CRYPTO_atomic_add(),
3595 and the validity of object reference counter.
3597 *fdasilvayy@gmail.com*
3599 * With Windows Visual Studio builds, the .pdb files are installed
3600 alongside the installed libraries and executables. For a static
3601 library installation, ossl_static.pdb is the associate compiler
3602 generated .pdb file to be used when linking programs.
3606 * Remove openssl.spec. Packaging files belong with the packagers.
3610 * Automatic Darwin/OSX configuration has had a refresh, it will now
3611 recognise x86_64 architectures automatically. You can still decide
3612 to build for a different bitness with the environment variable
3613 KERNEL_BITS (can be 32 or 64), for example:
3615 KERNEL_BITS=32 ./config
3619 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3620 256 bit AES and HMAC with SHA256.
3624 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3628 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3632 * To enable users to have their own config files and build file templates,
3633 Configure looks in the directory indicated by the environment variable
3634 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3635 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3636 name and is used as is.
3640 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3641 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3642 X509_CERT_FILE_CTX was removed.
3646 * "shared" builds are now the default. To create only static libraries use
3647 the "no-shared" Configure option.
3651 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3652 All of these option have not worked for some while and are fundamental
3657 * Make various cleanup routines no-ops and mark them as deprecated. Most
3658 global cleanup functions are no longer required because they are handled
3659 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3660 Explicitly de-initing can cause problems (e.g. where a library that uses
3661 OpenSSL de-inits, but an application is still using it). The affected
3662 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3663 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3664 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3665 COMP_zlib_cleanup().
3669 * --strict-warnings no longer enables runtime debugging options
3670 such as REF_DEBUG. Instead, debug options are automatically
3671 enabled with '--debug' builds.
3673 *Andy Polyakov, Emilia Käsper*
3675 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3676 have been moved out of the public header files. New functions for managing
3677 these have been added.
3681 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3682 objects have been moved out of the public header files. New
3683 functions for managing these have been added.
3687 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3688 have been moved out of the public header files. New functions for managing
3689 these have been added.
3693 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3694 moved out of the public header files. New functions for managing these
3699 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3703 * Removed the mk1mf build scripts.
3707 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3708 it is always safe to #include a header now.
3712 * Removed the aged BC-32 config and all its supporting scripts
3716 * Removed support for Ultrix, Netware, and OS/2.
3720 * Add support for HKDF.
3722 *Alessandro Ghedini*
3724 * Add support for blake2b and blake2s
3728 * Added support for "pipelining". Ciphers that have the
3729 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3730 encryptions/decryptions simultaneously. There are currently no built-in
3731 ciphers with this property but the expectation is that engines will be able
3732 to offer it to significantly improve throughput. Support has been extended
3733 into libssl so that multiple records for a single connection can be
3734 processed in one go (for >=TLS 1.1).
3738 * Added the AFALG engine. This is an async capable engine which is able to
3739 offload work to the Linux kernel. In this initial version it only supports
3740 AES128-CBC. The kernel must be version 4.1.0 or greater.
3744 * OpenSSL now uses a new threading API. It is no longer necessary to
3745 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3746 are two supported threading models: pthreads and windows threads. It is
3747 also possible to configure OpenSSL at compile time for "no-threads". The
3748 old threading API should no longer be used. The functions have been
3749 replaced with "no-op" compatibility macros.
3751 *Alessandro Ghedini, Matt Caswell*
3753 * Modify behavior of ALPN to invoke callback after SNI/servername
3754 callback, such that updates to the SSL_CTX affect ALPN.
3758 * Add SSL_CIPHER queries for authentication and key-exchange.
3762 * Changes to the DEFAULT cipherlist:
3763 - Prefer (EC)DHE handshakes over plain RSA.
3764 - Prefer AEAD ciphers over legacy ciphers.
3765 - Prefer ECDSA over RSA when both certificates are available.
3766 - Prefer TLSv1.2 ciphers/PRF.
3767 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3772 * Change the ECC default curve list to be this, in order: x25519,
3773 secp256r1, secp521r1, secp384r1.
3777 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3778 disabled by default. They can be re-enabled using the
3779 enable-weak-ssl-ciphers option to Configure.
3783 * If the server has ALPN configured, but supports no protocols that the
3784 client advertises, send a fatal "no_application_protocol" alert.
3785 This behaviour is SHALL in RFC 7301, though it isn't universally
3786 implemented by other servers.
3790 * Add X25519 support.
3791 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3792 for public and private key encoding using the format documented in
3793 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3794 key generation and key derivation.
3796 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3801 * Deprecate SRP_VBASE_get_by_user.
3802 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3803 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3804 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3805 seed, even if the seed is configured.
3807 Users should use SRP_VBASE_get1_by_user instead. Note that in
3808 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3809 also that even though configuring the SRP seed attempts to hide
3810 invalid usernames by continuing the handshake with fake
3811 credentials, this behaviour is not constant time and no strong
3812 guarantees are made that the handshake is indistinguishable from
3813 that of a valid user.
3817 * Configuration change; it's now possible to build dynamic engines
3818 without having to build shared libraries and vice versa. This
3819 only applies to the engines in `engines/`, those in `crypto/engine/`
3820 will always be built into libcrypto (i.e. "static").
3822 Building dynamic engines is enabled by default; to disable, use
3823 the configuration option "disable-dynamic-engine".
3825 The only requirements for building dynamic engines are the
3826 presence of the DSO module and building with position independent
3827 code, so they will also automatically be disabled if configuring
3828 with "disable-dso" or "disable-pic".
3830 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3831 are also taken away from openssl/opensslconf.h, as they are
3836 * Configuration change; if there is a known flag to compile
3837 position independent code, it will always be applied on the
3838 libcrypto and libssl object files, and never on the application
3839 object files. This means other libraries that use routines from
3840 libcrypto / libssl can be made into shared libraries regardless
3841 of how OpenSSL was configured.
3843 If this isn't desirable, the configuration options "disable-pic"
3844 or "no-pic" can be used to disable the use of PIC. This will
3845 also disable building shared libraries and dynamic engines.
3849 * Removed JPAKE code. It was experimental and has no wide use.
3853 * The INSTALL_PREFIX Makefile variable has been renamed to
3854 DESTDIR. That makes for less confusion on what this variable
3855 is for. Also, the configuration option --install_prefix is
3860 * Heartbeat for TLS has been removed and is disabled by default
3861 for DTLS; configure with enable-heartbeats. Code that uses the
3862 old #define's might need to be updated.
3864 *Emilia Käsper, Rich Salz*
3866 * Rename REF_CHECK to REF_DEBUG.
3870 * New "unified" build system
3872 The "unified" build system is aimed to be a common system for all
3873 platforms we support. With it comes new support for VMS.
3875 This system builds supports building in a different directory tree
3876 than the source tree. It produces one Makefile (for unix family
3877 or lookalikes), or one descrip.mms (for VMS).
3879 The source of information to make the Makefile / descrip.mms is
3880 small files called 'build.info', holding the necessary
3881 information for each directory with source to compile, and a
3882 template in Configurations, like unix-Makefile.tmpl or
3885 With this change, the library names were also renamed on Windows
3886 and on VMS. They now have names that are closer to the standard
3887 on Unix, and include the major version number, and in certain
3888 cases, the architecture they are built for. See "Notes on shared
3889 libraries" in INSTALL.
3891 We rely heavily on the perl module Text::Template.
3895 * Added support for auto-initialisation and de-initialisation of the library.
3896 OpenSSL no longer requires explicit init or deinit routines to be called,
3897 except in certain circumstances. See the OPENSSL_init_crypto() and
3898 OPENSSL_init_ssl() man pages for further information.
3902 * The arguments to the DTLSv1_listen function have changed. Specifically the
3903 "peer" argument is now expected to be a BIO_ADDR object.
3905 * Rewrite of BIO networking library. The BIO library lacked consistent
3906 support of IPv6, and adding it required some more extensive
3907 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3908 which hold all types of addresses and chains of address information.
3909 It also introduces a new API, with functions like BIO_socket,
3910 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3911 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3912 have been adapted accordingly.
3916 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3921 * CRIME protection: disable compression by default, even if OpenSSL is
3922 compiled with zlib enabled. Applications can still enable compression
3923 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3924 using the SSL_CONF library to configure compression.
3928 * The signature of the session callback configured with
3929 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3930 was explicitly marked as `const unsigned char*` instead of
3935 * Always DPURIFY. Remove the use of uninitialized memory in the
3936 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3940 * Removed many obsolete configuration items, including
3941 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3942 MD2_CHAR, MD2_INT, MD2_LONG
3944 IDEA_SHORT, IDEA_LONG
3945 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3947 *Rich Salz, with advice from Andy Polyakov*
3949 * Many BN internals have been moved to an internal header file.
3951 *Rich Salz with help from Andy Polyakov*
3953 * Configuration and writing out the results from it has changed.
3954 Files such as Makefile include/openssl/opensslconf.h and are now
3955 produced through general templates, such as Makefile.in and
3956 crypto/opensslconf.h.in and some help from the perl module
3959 Also, the center of configuration information is no longer
3960 Makefile. Instead, Configure produces a perl module in
3961 configdata.pm which holds most of the config data (in the hash
3962 table %config), the target data that comes from the target
3963 configuration in one of the `Configurations/*.conf` files (in
3968 * To clarify their intended purposes, the Configure options
3969 --prefix and --openssldir change their semantics, and become more
3970 straightforward and less interdependent.
3972 --prefix shall be used exclusively to give the location INSTALLTOP
3973 where programs, scripts, libraries, include files and manuals are
3974 going to be installed. The default is now /usr/local.
3976 --openssldir shall be used exclusively to give the default
3977 location OPENSSLDIR where certificates, private keys, CRLs are
3978 managed. This is also where the default openssl.cnf gets
3980 If the directory given with this option is a relative path, the
3981 values of both the --prefix value and the --openssldir value will
3982 be combined to become OPENSSLDIR.
3983 The default for --openssldir is INSTALLTOP/ssl.
3985 Anyone who uses --openssldir to specify where OpenSSL is to be
3986 installed MUST change to use --prefix instead.
3990 * The GOST engine was out of date and therefore it has been removed. An up
3991 to date GOST engine is now being maintained in an external repository.
3992 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3993 support for GOST ciphersuites (these are only activated if a GOST engine
3998 * EGD is no longer supported by default; use enable-egd when
4001 *Ben Kaduk and Rich Salz*
4003 * The distribution now has Makefile.in files, which are used to
4004 create Makefile's when Configure is run. *Configure must be run
4005 before trying to build now.*
4009 * The return value for SSL_CIPHER_description() for error conditions
4014 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4016 Obtaining and performing DNSSEC validation of TLSA records is
4017 the application's responsibility. The application provides
4018 the TLSA records of its choice to OpenSSL, and these are then
4019 used to authenticate the peer.
4021 The TLSA records need not even come from DNS. They can, for
4022 example, be used to implement local end-entity certificate or
4023 trust-anchor "pinning", where the "pin" data takes the form
4024 of TLSA records, which can augment or replace verification
4025 based on the usual WebPKI public certification authorities.
4029 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4030 continues to support deprecated interfaces in default builds.
4031 However, applications are strongly advised to compile their
4032 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4033 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4034 or the 1.1.0 releases.
4036 In environments in which all applications have been ported to
4037 not use any deprecated interfaces OpenSSL's Configure script
4038 should be used with the --api=1.1.0 option to entirely remove
4039 support for the deprecated features from the library and
4040 unconditionally disable them in the installed headers.
4041 Essentially the same effect can be achieved with the "no-deprecated"
4042 argument to Configure, except that this will always restrict
4043 the build to just the latest API, rather than a fixed API
4046 As applications are ported to future revisions of the API,
4047 they should update their compile-time OPENSSL_API_COMPAT define
4048 accordingly, but in most cases should be able to continue to
4049 compile with later releases.
4051 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4052 0x10000000L and 0x00908000L, respectively. However those
4053 versions did not support the OPENSSL_API_COMPAT feature, and
4054 so applications are not typically tested for explicit support
4055 of just the undeprecated features of either release.
4059 * Add support for setting the minimum and maximum supported protocol.
4060 It can bet set via the SSL_set_min_proto_version() and
4061 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4062 MaxProtocol. It's recommended to use the new APIs to disable
4063 protocols instead of disabling individual protocols using
4064 SSL_set_options() or SSL_CONF's Protocol. This change also
4065 removes support for disabling TLS 1.2 in the OpenSSL TLS
4066 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4070 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4074 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4075 and integrates ECDSA and ECDH functionality into EC. Implementations can
4076 now redirect key generation and no longer need to convert to or from
4079 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4080 include the ec.h header file instead.
4084 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4085 ciphers who are no longer supported and drops support the ephemeral RSA key
4086 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4090 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4091 opaque. For HMAC_CTX, the following constructors and destructors
4094 HMAC_CTX *HMAC_CTX_new(void);
4095 void HMAC_CTX_free(HMAC_CTX *ctx);
4097 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4098 destroy such methods has been added. See EVP_MD_meth_new(3) and
4099 EVP_CIPHER_meth_new(3) for documentation.
4102 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4103 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4104 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4105 an already created structure.
4106 2) For consistency with the majority of our object creators and
4107 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4108 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4109 for deprecated builds.
4113 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4114 cryptographic operations to be performed asynchronously as long as an
4115 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4116 further details. Libssl has also had this capability integrated with the
4117 introduction of the new mode SSL_MODE_ASYNC and associated error
4118 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4119 pages. This work was developed in partnership with Intel Corp.
4123 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4124 always enabled now. If you want to disable the support you should
4125 exclude it using the list of supported ciphers. This also means that the
4126 "-no_ecdhe" option has been removed from s_server.
4130 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4131 SSL_{CTX_}set1_curves() which can set a list.
4135 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4136 curve you want to support using SSL_{CTX_}set1_curves().
4140 * State machine rewrite. The state machine code has been significantly
4141 refactored in order to remove much duplication of code and solve issues
4142 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4143 further details). This change does have some associated API changes.
4144 Notably the SSL_state() function has been removed and replaced by
4145 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4146 SSL_set_state() has been removed altogether. The previous handshake states
4147 defined in ssl.h and ssl3.h have also been removed.
4151 * All instances of the string "ssleay" in the public API were replaced
4152 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4153 Some error codes related to internal RSA_eay API's were renamed.
4157 * The demo files in crypto/threads were moved to demo/threads.
4161 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4164 *Matt Caswell, Rich Salz*
4166 * New ASN.1 embed macro.
4168 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4169 structure is not allocated: it is part of the parent. That is instead of
4177 This reduces memory fragmentation and make it impossible to accidentally
4178 set a mandatory field to NULL.
4180 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4181 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4182 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4187 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4191 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4192 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4193 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4194 DES and RC4 ciphersuites.
4198 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4199 This changes the decoding behaviour for some invalid messages,
4200 though the change is mostly in the more lenient direction, and
4201 legacy behaviour is preserved as much as possible.
4205 * Fix no-stdio build.
4206 *David Woodhouse <David.Woodhouse@intel.com> and also*
4207 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4209 * New testing framework
4210 The testing framework has been largely rewritten and is now using
4211 perl and the perl modules Test::Harness and an extended variant of
4212 Test::More called OpenSSL::Test to do its work. All test scripts in
4213 test/ have been rewritten into test recipes, and all direct calls to
4214 executables in test/Makefile have become individual recipes using the
4215 simplified testing OpenSSL::Test::Simple.
4217 For documentation on our testing modules, do:
4219 perldoc test/testlib/OpenSSL/Test/Simple.pm
4220 perldoc test/testlib/OpenSSL/Test.pm
4224 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4225 are used; the latter aborts on memory leaks (usually checked on exit).
4226 Some undocumented "set malloc, etc., hooks" functions were removed
4227 and others were changed. All are now documented.
4231 * In DSA_generate_parameters_ex, if the provided seed is too short,
4234 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4236 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4237 from RFC4279, RFC4785, RFC5487, RFC5489.
4239 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4240 original RSA_PSK patch.
4244 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4245 era flag was never set throughout the codebase (only read). Also removed
4246 SSL3_FLAGS_POP_BUFFER which was only used if
4247 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4251 * Changed the default name options in the "ca", "crl", "req" and "x509"
4252 to be "oneline" instead of "compat".
4256 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4257 not aware of clients that still exhibit this bug, and the workaround
4258 hasn't been working properly for a while.
4262 * The return type of BIO_number_read() and BIO_number_written() as well as
4263 the corresponding num_read and num_write members in the BIO structure has
4264 changed from unsigned long to uint64_t. On platforms where an unsigned
4265 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4270 * Given the pervasive nature of TLS extensions it is inadvisable to run
4271 OpenSSL without support for them. It also means that maintaining
4272 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4273 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4277 * Removed support for the two export grade static DH ciphersuites
4278 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4279 were newly added (along with a number of other static DH ciphersuites) to
4280 1.0.2. However the two export ones have *never* worked since they were
4281 introduced. It seems strange in any case to be adding new export
4282 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4286 * Version negotiation has been rewritten. In particular SSLv23_method(),
4287 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4288 and turned into macros which simply call the new preferred function names
4289 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4290 should use the new names instead. Also as part of this change the ssl23.h
4291 header file has been removed.
4295 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4296 code and the associated standard is no longer considered fit-for-purpose.
4300 * RT2547 was closed. When generating a private key, try to make the
4301 output file readable only by the owner. This behavior change might
4302 be noticeable when interacting with other software.
4304 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4309 * Added HTTP GET support to the ocsp command.
4313 * Changed default digest for the dgst and enc commands from MD5 to
4318 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4322 * Added support for TLS extended master secret from
4323 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4324 initial patch which was a great help during development.
4328 * All libssl internal structures have been removed from the public header
4329 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4330 now redundant). Users should not attempt to access internal structures
4331 directly. Instead they should use the provided API functions.
4335 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4336 Access to deprecated functions can be re-enabled by running config with
4337 "enable-deprecated". In addition applications wishing to use deprecated
4338 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4339 will, by default, disable some transitive includes that previously existed
4340 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4344 * Added support for OCB mode. OpenSSL has been granted a patent license
4345 compatible with the OpenSSL license for use of OCB. Details are available
4346 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4347 for OCB can be removed by calling config with no-ocb.
4351 * SSLv2 support has been removed. It still supports receiving a SSLv2
4352 compatible client hello.
4356 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4357 done while fixing the error code for the key-too-small case.
4359 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4361 * CA.sh has been removed; use CA.pl instead.
4365 * Removed old DES API.
4369 * Remove various unsupported platforms:
4375 Sinix/ReliantUNIX RM400
4380 16-bit platforms such as WIN16
4384 * Clean up OPENSSL_NO_xxx #define's
4385 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4386 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4387 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4388 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4389 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4390 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4391 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4392 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4393 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4394 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4398 * Cleaned up dead code
4399 Remove all but one '#ifdef undef' which is to be looked at.
4403 * Clean up calling of xxx_free routines.
4404 Just like free(), fix most of the xxx_free routines to accept
4405 NULL. Remove the non-null checks from callers. Save much code.
4409 * Add secure heap for storage of private keys (when possible).
4410 Add BIO_s_secmem(), CBIGNUM, etc.
4411 Contributed by Akamai Technologies under our Corporate CLA.
4415 * Experimental support for a new, fast, unbiased prime candidate generator,
4416 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4418 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4420 * New output format NSS in the sess_id command line tool. This allows
4421 exporting the session id and the master key in NSS keylog format.
4423 *Martin Kaiser <martin@kaiser.cx>*
4425 * Harmonize version and its documentation. -f flag is used to display
4428 *mancha <mancha1@zoho.com>*
4430 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4431 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4433 *mancha <mancha1@zoho.com>*
4435 * Fix some double frees. These are not thought to be exploitable.
4437 *mancha <mancha1@zoho.com>*
4439 * A missing bounds check in the handling of the TLS heartbeat extension
4440 can be used to reveal up to 64k of memory to a connected client or
4443 Thanks for Neel Mehta of Google Security for discovering this bug and to
4444 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4445 preparing the fix ([CVE-2014-0160])
4447 *Adam Langley, Bodo Moeller*
4449 * Fix for the attack described in the paper "Recovering OpenSSL
4450 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4451 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4452 <http://eprint.iacr.org/2014/140>
4454 Thanks to Yuval Yarom and Naomi Benger for discovering this
4455 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4457 *Yuval Yarom and Naomi Benger*
4459 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4460 this fixes a limitation in previous versions of OpenSSL.
4464 * Experimental encrypt-then-mac support.
4466 Experimental support for encrypt then mac from
4467 draft-gutmann-tls-encrypt-then-mac-02.txt
4469 To enable it set the appropriate extension number (0x42 for the test
4470 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4472 For non-compliant peers (i.e. just about everything) this should have no
4475 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4479 * Add EVP support for key wrapping algorithms, to avoid problems with
4480 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4481 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4482 algorithms and include tests cases.
4486 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4491 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4492 MGF1 digest and OAEP label.
4496 * Make openssl verify return errors.
4498 *Chris Palmer <palmer@google.com> and Ben Laurie*
4500 * New function ASN1_TIME_diff to calculate the difference between two
4501 ASN1_TIME structures or one structure and the current time.
4505 * Update fips_test_suite to support multiple command line options. New
4506 test to induce all self test errors in sequence and check expected
4511 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4512 sign or verify all in one operation.
4516 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4517 test programs and fips_test_suite. Includes functionality to parse
4518 the minimal script output of fipsalgest.pl directly.
4522 * Add authorisation parameter to FIPS_module_mode_set().
4526 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4530 * Use separate DRBG fields for internal and external flags. New function
4531 FIPS_drbg_health_check() to perform on demand health checking. Add
4532 generation tests to fips_test_suite with reduced health check interval to
4533 demonstrate periodic health checking. Add "nodh" option to
4534 fips_test_suite to skip very slow DH test.
4538 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4543 * More extensive health check for DRBG checking many more failure modes.
4544 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4545 combination: call this in fips_test_suite.
4549 * Add support for canonical generation of DSA parameter 'g'. See
4552 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4553 POST to handle HMAC cases.
4557 * Add functions FIPS_module_version() and FIPS_module_version_text()
4558 to return numerical and string versions of the FIPS module number.
4562 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4563 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4564 outside the validated module in the FIPS capable OpenSSL.
4568 * Minor change to DRBG entropy callback semantics. In some cases
4569 there is no multiple of the block length between min_len and
4570 max_len. Allow the callback to return more than max_len bytes
4571 of entropy but discard any extra: it is the callback's responsibility
4572 to ensure that the extra data discarded does not impact the
4573 requested amount of entropy.
4577 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4578 information in FIPS186-3, SP800-57 and SP800-131A.
4582 * CCM support via EVP. Interface is very similar to GCM case except we
4583 must supply all data in one chunk (i.e. no update, final) and the
4584 message length must be supplied if AAD is used. Add algorithm test
4589 * Initial version of POST overhaul. Add POST callback to allow the status
4590 of POST to be monitored and/or failures induced. Modify fips_test_suite
4591 to use callback. Always run all selftests even if one fails.
4595 * XTS support including algorithm test driver in the fips_gcmtest program.
4596 Note: this does increase the maximum key length from 32 to 64 bytes but
4597 there should be no binary compatibility issues as existing applications
4598 will never use XTS mode.
4602 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4603 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4604 performs algorithm blocking for unapproved PRNG types. Also do not
4605 set PRNG type in FIPS_mode_set(): leave this to the application.
4606 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4607 the standard OpenSSL PRNG: set additional data to a date time vector.
4611 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4612 This shouldn't present any incompatibility problems because applications
4613 shouldn't be using these directly and any that are will need to rethink
4614 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4618 * Extensive self tests and health checking required by SP800-90 DRBG.
4619 Remove strength parameter from FIPS_drbg_instantiate and always
4620 instantiate at maximum supported strength.
4624 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4628 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4632 * New function DH_compute_key_padded() to compute a DH key and pad with
4633 leading zeroes if needed: this complies with SP800-56A et al.
4637 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4638 anything, incomplete, subject to change and largely untested at present.
4642 * Modify fipscanisteronly build option to only build the necessary object
4643 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4647 * Add experimental option FIPSSYMS to give all symbols in
4648 fipscanister.o and FIPS or fips prefix. This will avoid
4649 conflicts with future versions of OpenSSL. Add perl script
4650 util/fipsas.pl to preprocess assembly language source files
4651 and rename any affected symbols.
4655 * Add selftest checks and algorithm block of non-fips algorithms in
4656 FIPS mode. Remove DES2 from selftests.
4660 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4661 return internal method without any ENGINE dependencies. Add new
4662 tiny fips sign and verify functions.
4666 * New build option no-ec2m to disable characteristic 2 code.
4670 * New build option "fipscanisteronly". This only builds fipscanister.o
4671 and (currently) associated fips utilities. Uses the file Makefile.fips
4672 instead of Makefile.org as the prototype.
4676 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4677 Update fips_gcmtest to use IV generator.
4681 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4682 setting output buffer to NULL. The `*Final` function must be
4683 called although it will not retrieve any additional data. The tag
4684 can be set or retrieved with a ctrl. The IV length is by default 12
4685 bytes (96 bits) but can be set to an alternative value. If the IV
4686 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4691 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4692 underlying do_cipher function handles all cipher semantics itself
4693 including padding and finalisation. This is useful if (for example)
4694 an ENGINE cipher handles block padding itself. The behaviour of
4695 do_cipher is subtly changed if this flag is set: the return value
4696 is the number of characters written to the output buffer (zero is
4697 no longer an error code) or a negative error code. Also if the
4698 input buffer is NULL and length 0 finalisation should be performed.
4702 * If a candidate issuer certificate is already part of the constructed
4703 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4707 * Improve forward-security support: add functions
4709 void SSL_CTX_set_not_resumable_session_callback(
4710 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4711 void SSL_set_not_resumable_session_callback(
4712 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4714 for use by SSL/TLS servers; the callback function will be called whenever a
4715 new session is created, and gets to decide whether the session may be
4716 cached to make it resumable (return 0) or not (return 1). (As by the
4717 SSL/TLS protocol specifications, the session_id sent by the server will be
4718 empty to indicate that the session is not resumable; also, the server will
4719 not generate RFC 4507 (RFC 5077) session tickets.)
4721 A simple reasonable callback implementation is to return is_forward_secure.
4722 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4723 by the SSL/TLS server library, indicating whether it can provide forward
4726 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4728 * New -verify_name option in command line utilities to set verification
4733 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4734 Add CMAC pkey methods.
4738 * Experimental renegotiation in s_server -www mode. If the client
4739 browses /reneg connection is renegotiated. If /renegcert it is
4740 renegotiated requesting a certificate.
4744 * Add an "external" session cache for debugging purposes to s_server. This
4745 should help trace issues which normally are only apparent in deployed
4746 multi-process servers.
4750 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4751 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4752 BIO_set_cipher() and some obscure PEM functions were changed so they
4753 can now return an error. The RAND changes required a change to the
4754 RAND_METHOD structure.
4758 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4759 a gcc attribute to warn if the result of a function is ignored. This
4760 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4761 whose return value is often ignored.
4765 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4766 These allow SCTs (signed certificate timestamps) to be requested and
4767 validated when establishing a connection.
4769 *Rob Percival <robpercival@google.com>*
4774 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4776 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4777 used even when parsing explicit parameters, when loading a encoded key
4778 or calling `EC_GROUP_new_from_ecpkparameters()`/
4779 `EC_GROUP_new_from_ecparameters()`.
4780 This prevents bypass of security hardening and performance gains,
4781 especially for curves with specialized EC_METHODs.
4782 By default, if a key encoded with explicit parameters is loaded and later
4783 encoded, the output is still encoded with explicit parameters, even if
4784 internally a "named" EC_GROUP is used for computation.
4788 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4789 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4790 NULL. After this change, only the cofactor parameter can be NULL. It also
4791 does some minimal sanity checks on the passed order.
4796 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4797 An attack is simple, if the first CMS_recipientInfo is valid but the
4798 second CMS_recipientInfo is chosen ciphertext. If the second
4799 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4800 encryption key will be replaced by garbage, and the message cannot be
4801 decoded, but if the RSA decryption fails, the correct encryption key is
4802 used and the recipient will not notice the attack.
4803 As a work around for this potential attack the length of the decrypted
4804 key must be equal to the cipher default key length, in case the
4805 certifiate is not given and all recipientInfo are tried out.
4806 The old behaviour can be re-enabled in the CMS code by setting the
4807 CMS_DEBUG_DECRYPT flag.
4812 * Document issue with installation paths in diverse Windows builds
4814 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4815 binaries and run-time config file.
4820 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4822 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4823 This changes the size when using the `genpkey` command when no size is given.
4824 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4825 generation commands to use 2048 bits by default.
4829 * Add FIPS support for Android Arm 64-bit
4831 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4832 Module in Version 2.0.10. For some reason, the corresponding target
4833 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4834 built with FIPS support on Android Arm 64-bit. This omission has been
4837 *Matthias St. Pierre*
4839 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4841 * 0-byte record padding oracle
4843 If an application encounters a fatal protocol error and then calls
4844 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4845 then OpenSSL can respond differently to the calling application if a 0 byte
4846 record is received with invalid padding compared to if a 0 byte record is
4847 received with an invalid MAC. If the application then behaves differently
4848 based on that in a way that is detectable to the remote peer, then this
4849 amounts to a padding oracle that could be used to decrypt data.
4851 In order for this to be exploitable "non-stitched" ciphersuites must be in
4852 use. Stitched ciphersuites are optimised implementations of certain
4853 commonly used ciphersuites. Also the application must call SSL_shutdown()
4854 twice even if a protocol error has occurred (applications should not do
4855 this but some do anyway).
4857 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4858 Aviram, with additional investigation by Steven Collison and Andrew
4859 Hourselt. It was reported to OpenSSL on 10th December 2018.
4864 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4868 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4870 * Microarchitecture timing vulnerability in ECC scalar multiplication
4872 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4873 shown to be vulnerable to a microarchitecture timing side channel attack.
4874 An attacker with sufficient access to mount local timing attacks during
4875 ECDSA signature generation could recover the private key.
4877 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4878 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4884 * Timing vulnerability in DSA signature generation
4886 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4887 timing side channel attack. An attacker could use variations in the signing
4888 algorithm to recover the private key.
4890 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4895 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4896 Module, accidentally introduced while backporting security fixes from the
4897 development branch and hindering the use of ECC in FIPS mode.
4901 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4903 * Client DoS due to large DH parameter
4905 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4906 malicious server can send a very large prime value to the client. This will
4907 cause the client to spend an unreasonably long period of time generating a
4908 key for this prime resulting in a hang until the client has finished. This
4909 could be exploited in a Denial Of Service attack.
4911 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4916 * Cache timing vulnerability in RSA Key Generation
4918 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4919 a cache timing side channel attack. An attacker with sufficient access to
4920 mount cache timing attacks during the RSA key generation process could
4921 recover the private key.
4923 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4924 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4929 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4930 parameter is no longer accepted, as it leads to a corrupt table. NULL
4931 pem_str is reserved for alias entries only.
4935 * Revert blinding in ECDSA sign and instead make problematic addition
4936 length-invariant. Switch even to fixed-length Montgomery multiplication.
4940 * Change generating and checking of primes so that the error rate of not
4941 being prime depends on the intended use based on the size of the input.
4942 For larger primes this will result in more rounds of Miller-Rabin.
4943 The maximal error rate for primes with more than 1080 bits is lowered
4946 *Kurt Roeckx, Annie Yousar*
4948 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4952 * Add blinding to ECDSA and DSA signatures to protect against side channel
4953 attacks discovered by Keegan Ryan (NCC Group).
4957 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4958 now allow empty (zero character) pass phrases.
4962 * Certificate time validation (X509_cmp_time) enforces stricter
4963 compliance with RFC 5280. Fractional seconds and timezone offsets
4964 are no longer allowed.
4968 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4970 * Constructed ASN.1 types with a recursive definition could exceed the stack
4972 Constructed ASN.1 types with a recursive definition (such as can be found
4973 in PKCS7) could eventually exceed the stack given malicious input with
4974 excessive recursion. This could result in a Denial Of Service attack. There
4975 are no such structures used within SSL/TLS that come from untrusted sources
4976 so this is considered safe.
4978 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4984 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4986 * Read/write after SSL object in error state
4988 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4989 mechanism. The intent was that if a fatal error occurred during a handshake
4990 then OpenSSL would move into the error state and would immediately fail if
4991 you attempted to continue the handshake. This works as designed for the
4992 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4993 SSL_connect()), however due to a bug it does not work correctly if
4994 SSL_read() or SSL_write() is called directly. In that scenario, if the
4995 handshake fails then a fatal error will be returned in the initial function
4996 call. If SSL_read()/SSL_write() is subsequently called by the application
4997 for the same SSL object then it will succeed and the data is passed without
4998 being decrypted/encrypted directly from the SSL/TLS record layer.
5000 In order to exploit this issue an application bug would have to be present
5001 that resulted in a call to SSL_read()/SSL_write() being issued after having
5002 already received a fatal error.
5004 This issue was reported to OpenSSL by David Benjamin (Google).
5009 * rsaz_1024_mul_avx2 overflow bug on x86_64
5011 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5012 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5013 Analysis suggests that attacks against RSA and DSA as a result of this
5014 defect would be very difficult to perform and are not believed likely.
5015 Attacks against DH1024 are considered just feasible, because most of the
5016 work necessary to deduce information about a private key may be performed
5017 offline. The amount of resources required for such an attack would be
5018 significant. However, for an attack on TLS to be meaningful, the server
5019 would have to share the DH1024 private key among multiple clients, which is
5020 no longer an option since CVE-2016-0701.
5022 This only affects processors that support the AVX2 but not ADX extensions
5023 like Intel Haswell (4th generation).
5025 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5026 was originally found via the OSS-Fuzz project.
5031 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5033 * bn_sqrx8x_internal carry bug on x86_64
5035 There is a carry propagating bug in the x86_64 Montgomery squaring
5036 procedure. No EC algorithms are affected. Analysis suggests that attacks
5037 against RSA and DSA as a result of this defect would be very difficult to
5038 perform and are not believed likely. Attacks against DH are considered just
5039 feasible (although very difficult) because most of the work necessary to
5040 deduce information about a private key may be performed offline. The amount
5041 of resources required for such an attack would be very significant and
5042 likely only accessible to a limited number of attackers. An attacker would
5043 additionally need online access to an unpatched system using the target
5044 private key in a scenario with persistent DH parameters and a private
5045 key that is shared between multiple clients.
5047 This only affects processors that support the BMI1, BMI2 and ADX extensions
5048 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5050 This issue was reported to OpenSSL by the OSS-Fuzz project.
5055 * Malformed X.509 IPAddressFamily could cause OOB read
5057 If an X.509 certificate has a malformed IPAddressFamily extension,
5058 OpenSSL could do a one-byte buffer overread. The most likely result
5059 would be an erroneous display of the certificate in text format.
5061 This issue was reported to OpenSSL by the OSS-Fuzz project.
5065 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5067 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5068 platform rather than 'mingw'.
5072 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5074 * Truncated packet could crash via OOB read
5076 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5077 cipher is being used, then a truncated packet can cause that host to
5078 perform an out-of-bounds read, usually resulting in a crash.
5080 This issue was reported to OpenSSL by Robert Święcki of Google.
5085 * BN_mod_exp may produce incorrect results on x86_64
5087 There is a carry propagating bug in the x86_64 Montgomery squaring
5088 procedure. No EC algorithms are affected. Analysis suggests that attacks
5089 against RSA and DSA as a result of this defect would be very difficult to
5090 perform and are not believed likely. Attacks against DH are considered just
5091 feasible (although very difficult) because most of the work necessary to
5092 deduce information about a private key may be performed offline. The amount
5093 of resources required for such an attack would be very significant and
5094 likely only accessible to a limited number of attackers. An attacker would
5095 additionally need online access to an unpatched system using the target
5096 private key in a scenario with persistent DH parameters and a private
5097 key that is shared between multiple clients. For example this can occur by
5098 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5099 similar to CVE-2015-3193 but must be treated as a separate problem.
5101 This issue was reported to OpenSSL by the OSS-Fuzz project.
5106 * Montgomery multiplication may produce incorrect results
5108 There is a carry propagating bug in the Broadwell-specific Montgomery
5109 multiplication procedure that handles input lengths divisible by, but
5110 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5111 and DH private keys are impossible. This is because the subroutine in
5112 question is not used in operations with the private key itself and an input
5113 of the attacker's direct choice. Otherwise the bug can manifest itself as
5114 transient authentication and key negotiation failures or reproducible
5115 erroneous outcome of public-key operations with specially crafted input.
5116 Among EC algorithms only Brainpool P-512 curves are affected and one
5117 presumably can attack ECDH key negotiation. Impact was not analyzed in
5118 detail, because pre-requisites for attack are considered unlikely. Namely
5119 multiple clients have to choose the curve in question and the server has to
5120 share the private key among them, neither of which is default behaviour.
5121 Even then only clients that chose the curve will be affected.
5123 This issue was publicly reported as transient failures and was not
5124 initially recognized as a security issue. Thanks to Richard Morgan for
5125 providing reproducible case.
5130 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5131 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5132 prevent issues where no progress is being made and the peer continually
5133 sends unrecognised record types, using up resources processing them.
5137 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5139 * Missing CRL sanity check
5141 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5142 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5143 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5145 This issue only affects the OpenSSL 1.0.2i
5150 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5152 * OCSP Status Request extension unbounded memory growth
5154 A malicious client can send an excessively large OCSP Status Request
5155 extension. If that client continually requests renegotiation, sending a
5156 large OCSP Status Request extension each time, then there will be unbounded
5157 memory growth on the server. This will eventually lead to a Denial Of
5158 Service attack through memory exhaustion. Servers with a default
5159 configuration are vulnerable even if they do not support OCSP. Builds using
5160 the "no-ocsp" build time option are not affected.
5162 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5167 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5170 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5176 * OOB write in MDC2_Update()
5178 An overflow can occur in MDC2_Update() either if called directly or
5179 through the EVP_DigestUpdate() function using MDC2. If an attacker
5180 is able to supply very large amounts of input data after a previous
5181 call to EVP_EncryptUpdate() with a partial block then a length check
5182 can overflow resulting in a heap corruption.
5184 The amount of data needed is comparable to SIZE_MAX which is impractical
5187 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5192 * Malformed SHA512 ticket DoS
5194 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5195 DoS attack where a malformed ticket will result in an OOB read which will
5198 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5199 a custom server callback and ticket lookup mechanism.
5201 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5206 * OOB write in BN_bn2dec()
5208 The function BN_bn2dec() does not check the return value of BN_div_word().
5209 This can cause an OOB write if an application uses this function with an
5210 overly large BIGNUM. This could be a problem if an overly large certificate
5211 or CRL is printed out from an untrusted source. TLS is not affected because
5212 record limits will reject an oversized certificate before it is parsed.
5214 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5219 * OOB read in TS_OBJ_print_bio()
5221 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5222 the total length the OID text representation would use and not the amount
5223 of data written. This will result in OOB reads when large OIDs are
5226 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5231 * Pointer arithmetic undefined behaviour
5233 Avoid some undefined pointer arithmetic
5235 A common idiom in the codebase is to check limits in the following manner:
5238 Where "p" points to some malloc'd data of SIZE bytes and
5241 "len" here could be from some externally supplied data (e.g. from a TLS
5244 The rules of C pointer arithmetic are such that "p + len" is only well
5245 defined where len <= SIZE. Therefore the above idiom is actually
5246 undefined behaviour.
5248 For example this could cause problems if some malloc implementation
5249 provides an address for "p" such that "p + len" actually overflows for
5250 values of len that are too big and therefore p + len < limit.
5252 This issue was reported to OpenSSL by Guido Vranken
5257 * Constant time flag not preserved in DSA signing
5259 Operations in the DSA signing algorithm should run in constant time in
5260 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5261 implementation means that a non-constant time codepath is followed for
5262 certain operations. This has been demonstrated through a cache-timing
5263 attack to be sufficient for an attacker to recover the private DSA key.
5265 This issue was reported by César Pereida (Aalto University), Billy Brumley
5266 (Tampere University of Technology), and Yuval Yarom (The University of
5267 Adelaide and NICTA).
5272 * DTLS buffered message DoS
5274 In a DTLS connection where handshake messages are delivered out-of-order
5275 those messages that OpenSSL is not yet ready to process will be buffered
5276 for later use. Under certain circumstances, a flaw in the logic means that
5277 those messages do not get removed from the buffer even though the handshake
5278 has been completed. An attacker could force up to approx. 15 messages to
5279 remain in the buffer when they are no longer required. These messages will
5280 be cleared when the DTLS connection is closed. The default maximum size for
5281 a message is 100k. Therefore the attacker could force an additional 1500k
5282 to be consumed per connection. By opening many simulataneous connections an
5283 attacker could cause a DoS attack through memory exhaustion.
5285 This issue was reported to OpenSSL by Quan Luo.
5290 * DTLS replay protection DoS
5292 A flaw in the DTLS replay attack protection mechanism means that records
5293 that arrive for future epochs update the replay protection "window" before
5294 the MAC for the record has been validated. This could be exploited by an
5295 attacker by sending a record for the next epoch (which does not have to
5296 decrypt or have a valid MAC), with a very large sequence number. This means
5297 that all subsequent legitimate packets are dropped causing a denial of
5298 service for a specific DTLS connection.
5300 This issue was reported to OpenSSL by the OCAP audit team.
5305 * Certificate message OOB reads
5307 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5308 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5309 theoretical DoS risk but this has not been observed in practice on common
5312 The messages affected are client certificate, client certificate request
5313 and server certificate. As a result the attack can only be performed
5314 against a client or a server which enables client authentication.
5316 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5321 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5323 * Prevent padding oracle in AES-NI CBC MAC check
5325 A MITM attacker can use a padding oracle attack to decrypt traffic
5326 when the connection uses an AES CBC cipher and the server support
5329 This issue was introduced as part of the fix for Lucky 13 padding
5330 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5331 constant time by making sure that always the same bytes are read and
5332 compared against either the MAC or padding bytes. But it no longer
5333 checked that there was enough data to have both the MAC and padding
5336 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5340 * Fix EVP_EncodeUpdate overflow
5342 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5343 Base64 encoding of binary data. If an attacker is able to supply very large
5344 amounts of input data then a length check can overflow resulting in a heap
5347 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5348 the `PEM_write_bio*` family of functions. These are mainly used within the
5349 OpenSSL command line applications, so any application which processes data
5350 from an untrusted source and outputs it as a PEM file should be considered
5351 vulnerable to this issue. User applications that call these APIs directly
5352 with large amounts of untrusted data may also be vulnerable.
5354 This issue was reported by Guido Vranken.
5359 * Fix EVP_EncryptUpdate overflow
5361 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5362 is able to supply very large amounts of input data after a previous call to
5363 EVP_EncryptUpdate() with a partial block then a length check can overflow
5364 resulting in a heap corruption. Following an analysis of all OpenSSL
5365 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5366 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5367 the first called function after an EVP_EncryptInit(), and therefore that
5368 specific call must be safe. The second form is where the length passed to
5369 EVP_EncryptUpdate() can be seen from the code to be some small value and
5370 therefore there is no possibility of an overflow. Since all instances are
5371 one of these two forms, it is believed that there can be no overflows in
5372 internal code due to this problem. It should be noted that
5373 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5374 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5375 of these calls have also been analysed too and it is believed there are no
5376 instances in internal usage where an overflow could occur.
5378 This issue was reported by Guido Vranken.
5383 * Prevent ASN.1 BIO excessive memory allocation
5385 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5386 a short invalid encoding can cause allocation of large amounts of memory
5387 potentially consuming excessive resources or exhausting memory.
5389 Any application parsing untrusted data through d2i BIO functions is
5390 affected. The memory based functions such as d2i_X509() are *not* affected.
5391 Since the memory based functions are used by the TLS library, TLS
5392 applications are not affected.
5394 This issue was reported by Brian Carpenter.
5401 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5402 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5403 in arbitrary stack data being returned in the buffer.
5405 This issue was reported by Guido Vranken.
5410 * Modify behavior of ALPN to invoke callback after SNI/servername
5411 callback, such that updates to the SSL_CTX affect ALPN.
5415 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5420 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5421 methods are enabled and ssl2 is disabled the methods return NULL.
5425 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5427 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5428 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5429 provide any "EXPORT" or "LOW" strength ciphers.
5433 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5434 is by default disabled at build-time. Builds that are not configured with
5435 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5436 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5437 will need to explicitly call either of:
5439 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5441 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5443 as appropriate. Even if either of those is used, or the application
5444 explicitly uses the version-specific SSLv2_method() or its client and
5445 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5446 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5447 ciphers, and SSLv2 56-bit DES are no longer available.
5452 * Fix a double-free in DSA code
5454 A double free bug was discovered when OpenSSL parses malformed DSA private
5455 keys and could lead to a DoS attack or memory corruption for applications
5456 that receive DSA private keys from untrusted sources. This scenario is
5459 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5465 * Disable SRP fake user seed to address a server memory leak.
5467 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5469 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5470 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5471 was changed to ignore the "fake user" SRP seed, even if the seed
5474 Users should use SRP_VBASE_get1_by_user instead. Note that in
5475 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5476 also that even though configuring the SRP seed attempts to hide
5477 invalid usernames by continuing the handshake with fake
5478 credentials, this behaviour is not constant time and no strong
5479 guarantees are made that the handshake is indistinguishable from
5480 that of a valid user.
5485 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5487 In the BN_hex2bn function the number of hex digits is calculated using an
5488 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5489 large values of `i` this can result in `bn_expand` not allocating any
5490 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5491 field as NULL leading to a subsequent NULL ptr deref. For very large values
5492 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5493 In this case memory is allocated to the internal BIGNUM data field, but it
5494 is insufficiently sized leading to heap corruption. A similar issue exists
5495 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5496 is ever called by user applications with very large untrusted hex/dec data.
5497 This is anticipated to be a rare occurrence.
5499 All OpenSSL internal usage of these functions use data that is not expected
5500 to be untrusted, e.g. config file data or application command line
5501 arguments. If user developed applications generate config file data based
5502 on untrusted data then it is possible that this could also lead to security
5503 consequences. This is also anticipated to be rare.
5505 This issue was reported to OpenSSL by Guido Vranken.
5510 * Fix memory issues in `BIO_*printf` functions
5512 The internal `fmtstr` function used in processing a "%s" format string in
5513 the `BIO_*printf` functions could overflow while calculating the length of a
5514 string and cause an OOB read when printing very long strings.
5516 Additionally the internal `doapr_outch` function can attempt to write to an
5517 OOB memory location (at an offset from the NULL pointer) in the event of a
5518 memory allocation failure. In 1.0.2 and below this could be caused where
5519 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5520 could be in processing a very long "%s" format string. Memory leaks can
5523 The first issue may mask the second issue dependent on compiler behaviour.
5524 These problems could enable attacks where large amounts of untrusted data
5525 is passed to the `BIO_*printf` functions. If applications use these functions
5526 in this way then they could be vulnerable. OpenSSL itself uses these
5527 functions when printing out human-readable dumps of ASN.1 data. Therefore
5528 applications that print this data could be vulnerable if the data is from
5529 untrusted sources. OpenSSL command line applications could also be
5530 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5531 as command line arguments.
5533 Libssl is not considered directly vulnerable. Additionally certificates etc
5534 received via remote connections via libssl are also unlikely to be able to
5535 trigger these issues because of message size limits enforced within libssl.
5537 This issue was reported to OpenSSL Guido Vranken.
5542 * Side channel attack on modular exponentiation
5544 A side-channel attack was found which makes use of cache-bank conflicts on
5545 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5546 of RSA keys. The ability to exploit this issue is limited as it relies on
5547 an attacker who has control of code in a thread running on the same
5548 hyper-threaded core as the victim thread which is performing decryptions.
5550 This issue was reported to OpenSSL by Yuval Yarom, The University of
5551 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5552 Nadia Heninger, University of Pennsylvania with more information at
5553 <http://cachebleed.info>.
5558 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5559 if no keysize is specified with default_bits. This fixes an
5560 omission in an earlier change that changed all RSA/DSA key generation
5561 commands to use 2048 bits by default.
5565 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5567 * DH small subgroups
5569 Historically OpenSSL only ever generated DH parameters based on "safe"
5570 primes. More recently (in version 1.0.2) support was provided for
5571 generating X9.42 style parameter files such as those required for RFC 5114
5572 support. The primes used in such files may not be "safe". Where an
5573 application is using DH configured with parameters based on primes that are
5574 not "safe" then an attacker could use this fact to find a peer's private
5575 DH exponent. This attack requires that the attacker complete multiple
5576 handshakes in which the peer uses the same private DH exponent. For example
5577 this could be used to discover a TLS server's private DH exponent if it's
5578 reusing the private DH exponent or it's using a static DH ciphersuite.
5580 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5581 TLS. It is not on by default. If the option is not set then the server
5582 reuses the same private DH exponent for the life of the server process and
5583 would be vulnerable to this attack. It is believed that many popular
5584 applications do set this option and would therefore not be at risk.
5586 The fix for this issue adds an additional check where a "q" parameter is
5587 available (as is the case in X9.42 based parameters). This detects the
5588 only known attack, and is the only possible defense for static DH
5589 ciphersuites. This could have some performance impact.
5591 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5592 default and cannot be disabled. This could have some performance impact.
5594 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5599 * SSLv2 doesn't block disabled ciphers
5601 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5602 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5603 been disabled, provided that the SSLv2 protocol was not also disabled via
5606 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5607 and Sebastian Schinzel.
5612 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5614 * BN_mod_exp may produce incorrect results on x86_64
5616 There is a carry propagating bug in the x86_64 Montgomery squaring
5617 procedure. No EC algorithms are affected. Analysis suggests that attacks
5618 against RSA and DSA as a result of this defect would be very difficult to
5619 perform and are not believed likely. Attacks against DH are considered just
5620 feasible (although very difficult) because most of the work necessary to
5621 deduce information about a private key may be performed offline. The amount
5622 of resources required for such an attack would be very significant and
5623 likely only accessible to a limited number of attackers. An attacker would
5624 additionally need online access to an unpatched system using the target
5625 private key in a scenario with persistent DH parameters and a private
5626 key that is shared between multiple clients. For example this can occur by
5627 default in OpenSSL DHE based SSL/TLS ciphersuites.
5629 This issue was reported to OpenSSL by Hanno Böck.
5634 * Certificate verify crash with missing PSS parameter
5636 The signature verification routines will crash with a NULL pointer
5637 dereference if presented with an ASN.1 signature using the RSA PSS
5638 algorithm and absent mask generation function parameter. Since these
5639 routines are used to verify certificate signature algorithms this can be
5640 used to crash any certificate verification operation and exploited in a
5641 DoS attack. Any application which performs certificate verification is
5642 vulnerable including OpenSSL clients and servers which enable client
5645 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5650 * X509_ATTRIBUTE memory leak
5652 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5653 memory. This structure is used by the PKCS#7 and CMS routines so any
5654 application which reads PKCS#7 or CMS data from untrusted sources is
5655 affected. SSL/TLS is not affected.
5657 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5663 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5664 This changes the decoding behaviour for some invalid messages,
5665 though the change is mostly in the more lenient direction, and
5666 legacy behaviour is preserved as much as possible.
5670 * In DSA_generate_parameters_ex, if the provided seed is too short,
5673 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5675 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5677 * Alternate chains certificate forgery
5679 During certificate verification, OpenSSL will attempt to find an
5680 alternative certificate chain if the first attempt to build such a chain
5681 fails. An error in the implementation of this logic can mean that an
5682 attacker could cause certain checks on untrusted certificates to be
5683 bypassed, such as the CA flag, enabling them to use a valid leaf
5684 certificate to act as a CA and "issue" an invalid certificate.
5686 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5691 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5693 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5694 incompatibility in the handling of HMAC. The previous ABI has now been
5699 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5701 * Malformed ECParameters causes infinite loop
5703 When processing an ECParameters structure OpenSSL enters an infinite loop
5704 if the curve specified is over a specially malformed binary polynomial
5707 This can be used to perform denial of service against any
5708 system which processes public keys, certificate requests or
5709 certificates. This includes TLS clients and TLS servers with
5710 client authentication enabled.
5712 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5717 * Exploitable out-of-bounds read in X509_cmp_time
5719 X509_cmp_time does not properly check the length of the ASN1_TIME
5720 string and can read a few bytes out of bounds. In addition,
5721 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5724 An attacker can use this to craft malformed certificates and CRLs of
5725 various sizes and potentially cause a segmentation fault, resulting in
5726 a DoS on applications that verify certificates or CRLs. TLS clients
5727 that verify CRLs are affected. TLS clients and servers with client
5728 authentication enabled may be affected if they use custom verification
5731 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5732 independently by Hanno Böck.
5737 * PKCS7 crash with missing EnvelopedContent
5739 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5740 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5741 with missing content and trigger a NULL pointer dereference on parsing.
5743 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5744 structures from untrusted sources are affected. OpenSSL clients and
5745 servers are not affected.
5747 This issue was reported to OpenSSL by Michal Zalewski (Google).
5752 * CMS verify infinite loop with unknown hash function
5754 When verifying a signedData message the CMS code can enter an infinite loop
5755 if presented with an unknown hash function OID. This can be used to perform
5756 denial of service against any system which verifies signedData messages using
5758 This issue was reported to OpenSSL by Johannes Bauer.
5763 * Race condition handling NewSessionTicket
5765 If a NewSessionTicket is received by a multi-threaded client when attempting to
5766 reuse a previous ticket then a race condition can occur potentially leading to
5767 a double free of the ticket data.
5772 * Only support 256-bit or stronger elliptic curves with the
5773 'ecdh_auto' setting (server) or by default (client). Of supported
5774 curves, prefer P-256 (both).
5778 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5780 * ClientHello sigalgs DoS fix
5782 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5783 invalid signature algorithms extension a NULL pointer dereference will
5784 occur. This can be exploited in a DoS attack against the server.
5786 This issue was was reported to OpenSSL by David Ramos of Stanford
5790 *Stephen Henson and Matt Caswell*
5792 * Multiblock corrupted pointer fix
5794 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5795 feature only applies on 64 bit x86 architecture platforms that support AES
5796 NI instructions. A defect in the implementation of "multiblock" can cause
5797 OpenSSL's internal write buffer to become incorrectly set to NULL when
5798 using non-blocking IO. Typically, when the user application is using a
5799 socket BIO for writing, this will only result in a failed connection.
5800 However if some other BIO is used then it is likely that a segmentation
5801 fault will be triggered, thus enabling a potential DoS attack.
5803 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5808 * Segmentation fault in DTLSv1_listen fix
5810 The DTLSv1_listen function is intended to be stateless and processes the
5811 initial ClientHello from many peers. It is common for user code to loop
5812 over the call to DTLSv1_listen until a valid ClientHello is received with
5813 an associated cookie. A defect in the implementation of DTLSv1_listen means
5814 that state is preserved in the SSL object from one invocation to the next
5815 that can lead to a segmentation fault. Errors processing the initial
5816 ClientHello can trigger this scenario. An example of such an error could be
5817 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5820 This issue was reported to OpenSSL by Per Allansson.
5825 * Segmentation fault in ASN1_TYPE_cmp fix
5827 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5828 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5829 certificate signature algorithm consistency this can be used to crash any
5830 certificate verification operation and exploited in a DoS attack. Any
5831 application which performs certificate verification is vulnerable including
5832 OpenSSL clients and servers which enable client authentication.
5837 * Segmentation fault for invalid PSS parameters fix
5839 The signature verification routines will crash with a NULL pointer
5840 dereference if presented with an ASN.1 signature using the RSA PSS
5841 algorithm and invalid parameters. Since these routines are used to verify
5842 certificate signature algorithms this can be used to crash any
5843 certificate verification operation and exploited in a DoS attack. Any
5844 application which performs certificate verification is vulnerable including
5845 OpenSSL clients and servers which enable client authentication.
5847 This issue was was reported to OpenSSL by Brian Carpenter.
5852 * ASN.1 structure reuse memory corruption fix
5854 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5855 memory corruption via an invalid write. Such reuse is and has been
5856 strongly discouraged and is believed to be rare.
5858 Applications that parse structures containing CHOICE or ANY DEFINED BY
5859 components may be affected. Certificate parsing (d2i_X509 and related
5860 functions) are however not affected. OpenSSL clients and servers are
5866 * PKCS7 NULL pointer dereferences fix
5868 The PKCS#7 parsing code does not handle missing outer ContentInfo
5869 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5870 missing content and trigger a NULL pointer dereference on parsing.
5872 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5873 otherwise parse PKCS#7 structures from untrusted sources are
5874 affected. OpenSSL clients and servers are not affected.
5876 This issue was reported to OpenSSL by Michal Zalewski (Google).
5881 * DoS via reachable assert in SSLv2 servers fix
5883 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5884 servers that both support SSLv2 and enable export cipher suites by sending
5885 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5887 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5888 (OpenSSL development team).
5893 * Empty CKE with client auth and DHE fix
5895 If client auth is used then a server can seg fault in the event of a DHE
5896 ciphersuite being selected and a zero length ClientKeyExchange message
5897 being sent by the client. This could be exploited in a DoS attack.
5902 * Handshake with unseeded PRNG fix
5904 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5905 with an unseeded PRNG. The conditions are:
5906 - The client is on a platform where the PRNG has not been seeded
5907 automatically, and the user has not seeded manually
5908 - A protocol specific client method version has been used (i.e. not
5909 SSL_client_methodv23)
5910 - A ciphersuite is used that does not require additional random data from
5911 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5913 If the handshake succeeds then the client random that has been used will
5914 have been generated from a PRNG with insufficient entropy and therefore the
5915 output may be predictable.
5917 For example using the following command with an unseeded openssl will
5918 succeed on an unpatched platform:
5920 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5925 * Use After Free following d2i_ECPrivatekey error fix
5927 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5928 could cause a use after free condition. This, in turn, could cause a double
5929 free in several private key parsing functions (such as d2i_PrivateKey
5930 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5931 for applications that receive EC private keys from untrusted
5932 sources. This scenario is considered rare.
5934 This issue was discovered by the BoringSSL project and fixed in their
5940 * X509_to_X509_REQ NULL pointer deref fix
5942 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5943 the certificate key is invalid. This function is rarely used in practice.
5945 This issue was discovered by Brian Carpenter.
5950 * Removed the export ciphers from the DEFAULT ciphers
5954 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5956 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5957 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5958 So far those who have to target multiple platforms would compromise
5959 and argue that binary targeting say ARMv5 would still execute on
5960 ARMv8. "Universal" build resolves this compromise by providing
5961 near-optimal performance even on newer platforms.
5965 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5966 (other platforms pending).
5968 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5970 * Add support for the SignedCertificateTimestampList certificate and
5971 OCSP response extensions from RFC6962.
5975 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5976 for corner cases. (Certain input points at infinity could lead to
5977 bogus results, with non-infinity inputs mapped to infinity too.)
5981 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5982 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5983 common cases are optimized and there still is room for further
5984 improvements. Vector Permutation AES for Altivec is also added.
5988 * Add support for little-endian ppc64 Linux target.
5990 *Marcelo Cerri (IBM)*
5992 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5993 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5994 are optimized and there still is room for further improvements.
5995 Both 32- and 64-bit modes are supported.
5997 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5999 * Improved ARMv7 NEON support.
6003 * Support for SPARC Architecture 2011 crypto extensions, first
6004 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6005 SHA256/512, MD5, GHASH and modular exponentiation.
6007 *Andy Polyakov, David Miller*
6009 * Accelerated modular exponentiation for Intel processors, a.k.a.
6012 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6014 * Support for new and upcoming Intel processors, including AVX2,
6015 BMI and SHA ISA extensions. This includes additional "stitched"
6016 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6019 This work was sponsored by Intel Corp.
6023 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6024 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6025 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6029 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6030 this fixes a limitation in previous versions of OpenSSL.
6034 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6035 MGF1 digest and OAEP label.
6039 * Add EVP support for key wrapping algorithms, to avoid problems with
6040 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6041 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6042 algorithms and include tests cases.
6046 * Add functions to allocate and set the fields of an ECDSA_METHOD
6049 *Douglas E. Engert, Steve Henson*
6051 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6052 difference in days and seconds between two tm or ASN1_TIME structures.
6056 * Add -rev test option to s_server to just reverse order of characters
6057 received by client and send back to server. Also prints an abbreviated
6058 summary of the connection parameters.
6062 * New option -brief for s_client and s_server to print out a brief summary
6063 of connection parameters.
6067 * Add callbacks for arbitrary TLS extensions.
6069 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6071 * New option -crl_download in several openssl utilities to download CRLs
6072 from CRLDP extension in certificates.
6076 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6080 * New function X509_CRL_diff to generate a delta CRL from the difference
6081 of two full CRLs. Add support to "crl" utility.
6085 * New functions to set lookup_crls function and to retrieve
6086 X509_STORE from X509_STORE_CTX.
6090 * Print out deprecated issuer and subject unique ID fields in
6095 * Extend OCSP I/O functions so they can be used for simple general purpose
6096 HTTP as well as OCSP. New wrapper function which can be used to download
6097 CRLs using the OCSP API.
6101 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6105 * `SSL_CONF*` functions. These provide a common framework for application
6106 configuration using configuration files or command lines.
6110 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6111 message callback and prints the results. Needs compile time option
6112 "enable-ssl-trace". New options to s_client and s_server to enable
6117 * New ctrl and macro to retrieve supported points extensions.
6118 Print out extension in s_server and s_client.
6122 * New functions to retrieve certificate signature and signature
6127 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6132 * New Suite B modes for TLS code. These use and enforce the requirements
6133 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6134 only use Suite B curves. The Suite B modes can be set by using the
6135 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6139 * New chain verification flags for Suite B levels of security. Check
6140 algorithms are acceptable when flags are set in X509_verify_cert.
6144 * Make tls1_check_chain return a set of flags indicating checks passed
6145 by a certificate chain. Add additional tests to handle client
6146 certificates: checks for matching certificate type and issuer name
6151 * If an attempt is made to use a signature algorithm not in the peer
6152 preference list abort the handshake. If client has no suitable
6153 signature algorithms in response to a certificate request do not
6154 use the certificate.
6158 * If server EC tmp key is not in client preference list abort handshake.
6162 * Add support for certificate stores in CERT structure. This makes it
6163 possible to have different stores per SSL structure or one store in
6164 the parent SSL_CTX. Include distinct stores for certificate chain
6165 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6166 to build and store a certificate chain in CERT structure: returning
6167 an error if the chain cannot be built: this will allow applications
6168 to test if a chain is correctly configured.
6170 Note: if the CERT based stores are not set then the parent SSL_CTX
6171 store is used to retain compatibility with existing behaviour.
6175 * New function ssl_set_client_disabled to set a ciphersuite disabled
6176 mask based on the current session, check mask when sending client
6177 hello and checking the requested ciphersuite.
6181 * New ctrls to retrieve and set certificate types in a certificate
6182 request message. Print out received values in s_client. If certificate
6183 types is not set with custom values set sensible values based on
6184 supported signature algorithms.
6188 * Support for distinct client and server supported signature algorithms.
6192 * Add certificate callback. If set this is called whenever a certificate
6193 is required by client or server. An application can decide which
6194 certificate chain to present based on arbitrary criteria: for example
6195 supported signature algorithms. Add very simple example to s_server.
6196 This fixes many of the problems and restrictions of the existing client
6197 certificate callback: for example you can now clear an existing
6198 certificate and specify the whole chain.
6202 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6203 the certificate can be used for (if anything). Set valid_flags field
6204 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6205 to have similar checks in it.
6207 Add new "cert_flags" field to CERT structure and include a "strict mode".
6208 This enforces some TLS certificate requirements (such as only permitting
6209 certificate signature algorithms contained in the supported algorithms
6210 extension) which some implementations ignore: this option should be used
6211 with caution as it could cause interoperability issues.
6215 * Update and tidy signature algorithm extension processing. Work out
6216 shared signature algorithms based on preferences and peer algorithms
6217 and print them out in s_client and s_server. Abort handshake if no
6218 shared signature algorithms.
6222 * Add new functions to allow customised supported signature algorithms
6223 for SSL and SSL_CTX structures. Add options to s_client and s_server
6228 * New function SSL_certs_clear() to delete all references to certificates
6229 from an SSL structure. Before this once a certificate had been added
6230 it couldn't be removed.
6234 * Integrate hostname, email address and IP address checking with certificate
6235 verification. New verify options supporting checking in openssl utility.
6239 * Fixes and wildcard matching support to hostname and email checking
6240 functions. Add manual page.
6242 *Florian Weimer (Red Hat Product Security Team)*
6244 * New functions to check a hostname email or IP address against a
6245 certificate. Add options x509 utility to print results of checks against
6250 * Fix OCSP checking.
6252 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6254 * Initial experimental support for explicitly trusted non-root CAs.
6255 OpenSSL still tries to build a complete chain to a root but if an
6256 intermediate CA has a trust setting included that is used. The first
6257 setting is used: whether to trust (e.g., -addtrust option to the x509
6262 * Add -trusted_first option which attempts to find certificates in the
6263 trusted store even if an untrusted chain is also supplied.
6267 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6268 platform support for Linux and Android.
6272 * Support for linux-x32, ILP32 environment in x86_64 framework.
6276 * Experimental multi-implementation support for FIPS capable OpenSSL.
6277 When in FIPS mode the approved implementations are used as normal,
6278 when not in FIPS mode the internal unapproved versions are used instead.
6279 This means that the FIPS capable OpenSSL isn't forced to use the
6280 (often lower performance) FIPS implementations outside FIPS mode.
6284 * Transparently support X9.42 DH parameters when calling
6285 PEM_read_bio_DHparameters. This means existing applications can handle
6286 the new parameter format automatically.
6290 * Initial experimental support for X9.42 DH parameter format: mainly
6291 to support use of 'q' parameter for RFC5114 parameters.
6295 * Add DH parameters from RFC5114 including test data to dhtest.
6299 * Support for automatic EC temporary key parameter selection. If enabled
6300 the most preferred EC parameters are automatically used instead of
6301 hardcoded fixed parameters. Now a server just has to call:
6302 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6303 support ECDH and use the most appropriate parameters.
6307 * Enhance and tidy EC curve and point format TLS extension code. Use
6308 static structures instead of allocation if default values are used.
6309 New ctrls to set curves we wish to support and to retrieve shared curves.
6310 Print out shared curves in s_server. New options to s_server and s_client
6311 to set list of supported curves.
6315 * New ctrls to retrieve supported signature algorithms and
6316 supported curve values as an array of NIDs. Extend openssl utility
6317 to print out received values.
6321 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6322 between NIDs and the more common NIST names such as "P-256". Enhance
6323 ecparam utility and ECC method to recognise the NIST names for curves.
6327 * Enhance SSL/TLS certificate chain handling to support different
6328 chains for each certificate instead of one chain in the parent SSL_CTX.
6332 * Support for fixed DH ciphersuite client authentication: where both
6333 server and client use DH certificates with common parameters.
6337 * Support for fixed DH ciphersuites: those requiring DH server
6342 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6344 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6345 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6346 X509_CINF_get_signature were reverted post internal team review.
6351 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6353 * OCSP Status Request extension unbounded memory growth
6355 A malicious client can send an excessively large OCSP Status Request
6356 extension. If that client continually requests renegotiation, sending a
6357 large OCSP Status Request extension each time, then there will be unbounded
6358 memory growth on the server. This will eventually lead to a Denial Of
6359 Service attack through memory exhaustion. Servers with a default
6360 configuration are vulnerable even if they do not support OCSP. Builds using
6361 the "no-ocsp" build time option are not affected.
6363 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6368 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6371 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6377 * OOB write in MDC2_Update()
6379 An overflow can occur in MDC2_Update() either if called directly or
6380 through the EVP_DigestUpdate() function using MDC2. If an attacker
6381 is able to supply very large amounts of input data after a previous
6382 call to EVP_EncryptUpdate() with a partial block then a length check
6383 can overflow resulting in a heap corruption.
6385 The amount of data needed is comparable to SIZE_MAX which is impractical
6388 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6393 * Malformed SHA512 ticket DoS
6395 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6396 DoS attack where a malformed ticket will result in an OOB read which will
6399 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6400 a custom server callback and ticket lookup mechanism.
6402 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6407 * OOB write in BN_bn2dec()
6409 The function BN_bn2dec() does not check the return value of BN_div_word().
6410 This can cause an OOB write if an application uses this function with an
6411 overly large BIGNUM. This could be a problem if an overly large certificate
6412 or CRL is printed out from an untrusted source. TLS is not affected because
6413 record limits will reject an oversized certificate before it is parsed.
6415 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6420 * OOB read in TS_OBJ_print_bio()
6422 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6423 the total length the OID text representation would use and not the amount
6424 of data written. This will result in OOB reads when large OIDs are
6427 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6432 * Pointer arithmetic undefined behaviour
6434 Avoid some undefined pointer arithmetic
6436 A common idiom in the codebase is to check limits in the following manner:
6439 Where "p" points to some malloc'd data of SIZE bytes and
6442 "len" here could be from some externally supplied data (e.g. from a TLS
6445 The rules of C pointer arithmetic are such that "p + len" is only well
6446 defined where len <= SIZE. Therefore the above idiom is actually
6447 undefined behaviour.
6449 For example this could cause problems if some malloc implementation
6450 provides an address for "p" such that "p + len" actually overflows for
6451 values of len that are too big and therefore p + len < limit.
6453 This issue was reported to OpenSSL by Guido Vranken
6458 * Constant time flag not preserved in DSA signing
6460 Operations in the DSA signing algorithm should run in constant time in
6461 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6462 implementation means that a non-constant time codepath is followed for
6463 certain operations. This has been demonstrated through a cache-timing
6464 attack to be sufficient for an attacker to recover the private DSA key.
6466 This issue was reported by César Pereida (Aalto University), Billy Brumley
6467 (Tampere University of Technology), and Yuval Yarom (The University of
6468 Adelaide and NICTA).
6473 * DTLS buffered message DoS
6475 In a DTLS connection where handshake messages are delivered out-of-order
6476 those messages that OpenSSL is not yet ready to process will be buffered
6477 for later use. Under certain circumstances, a flaw in the logic means that
6478 those messages do not get removed from the buffer even though the handshake
6479 has been completed. An attacker could force up to approx. 15 messages to
6480 remain in the buffer when they are no longer required. These messages will
6481 be cleared when the DTLS connection is closed. The default maximum size for
6482 a message is 100k. Therefore the attacker could force an additional 1500k
6483 to be consumed per connection. By opening many simulataneous connections an
6484 attacker could cause a DoS attack through memory exhaustion.
6486 This issue was reported to OpenSSL by Quan Luo.
6491 * DTLS replay protection DoS
6493 A flaw in the DTLS replay attack protection mechanism means that records
6494 that arrive for future epochs update the replay protection "window" before
6495 the MAC for the record has been validated. This could be exploited by an
6496 attacker by sending a record for the next epoch (which does not have to
6497 decrypt or have a valid MAC), with a very large sequence number. This means
6498 that all subsequent legitimate packets are dropped causing a denial of
6499 service for a specific DTLS connection.
6501 This issue was reported to OpenSSL by the OCAP audit team.
6506 * Certificate message OOB reads
6508 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6509 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6510 theoretical DoS risk but this has not been observed in practice on common
6513 The messages affected are client certificate, client certificate request
6514 and server certificate. As a result the attack can only be performed
6515 against a client or a server which enables client authentication.
6517 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6522 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6524 * Prevent padding oracle in AES-NI CBC MAC check
6526 A MITM attacker can use a padding oracle attack to decrypt traffic
6527 when the connection uses an AES CBC cipher and the server support
6530 This issue was introduced as part of the fix for Lucky 13 padding
6531 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6532 constant time by making sure that always the same bytes are read and
6533 compared against either the MAC or padding bytes. But it no longer
6534 checked that there was enough data to have both the MAC and padding
6537 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6542 * Fix EVP_EncodeUpdate overflow
6544 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6545 Base64 encoding of binary data. If an attacker is able to supply very large
6546 amounts of input data then a length check can overflow resulting in a heap
6549 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6550 the `PEM_write_bio*` family of functions. These are mainly used within the
6551 OpenSSL command line applications, so any application which processes data
6552 from an untrusted source and outputs it as a PEM file should be considered
6553 vulnerable to this issue. User applications that call these APIs directly
6554 with large amounts of untrusted data may also be vulnerable.
6556 This issue was reported by Guido Vranken.
6561 * Fix EVP_EncryptUpdate overflow
6563 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6564 is able to supply very large amounts of input data after a previous call to
6565 EVP_EncryptUpdate() with a partial block then a length check can overflow
6566 resulting in a heap corruption. Following an analysis of all OpenSSL
6567 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6568 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6569 the first called function after an EVP_EncryptInit(), and therefore that
6570 specific call must be safe. The second form is where the length passed to
6571 EVP_EncryptUpdate() can be seen from the code to be some small value and
6572 therefore there is no possibility of an overflow. Since all instances are
6573 one of these two forms, it is believed that there can be no overflows in
6574 internal code due to this problem. It should be noted that
6575 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6576 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6577 of these calls have also been analysed too and it is believed there are no
6578 instances in internal usage where an overflow could occur.
6580 This issue was reported by Guido Vranken.
6585 * Prevent ASN.1 BIO excessive memory allocation
6587 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6588 a short invalid encoding can casuse allocation of large amounts of memory
6589 potentially consuming excessive resources or exhausting memory.
6591 Any application parsing untrusted data through d2i BIO functions is
6592 affected. The memory based functions such as d2i_X509() are *not* affected.
6593 Since the memory based functions are used by the TLS library, TLS
6594 applications are not affected.
6596 This issue was reported by Brian Carpenter.
6603 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6604 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6605 in arbitrary stack data being returned in the buffer.
6607 This issue was reported by Guido Vranken.
6612 * Modify behavior of ALPN to invoke callback after SNI/servername
6613 callback, such that updates to the SSL_CTX affect ALPN.
6617 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6622 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6623 methods are enabled and ssl2 is disabled the methods return NULL.
6627 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6629 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6630 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6631 provide any "EXPORT" or "LOW" strength ciphers.
6635 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6636 is by default disabled at build-time. Builds that are not configured with
6637 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6638 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6639 will need to explicitly call either of:
6641 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6643 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6645 as appropriate. Even if either of those is used, or the application
6646 explicitly uses the version-specific SSLv2_method() or its client and
6647 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6648 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6649 ciphers, and SSLv2 56-bit DES are no longer available.
6654 * Fix a double-free in DSA code
6656 A double free bug was discovered when OpenSSL parses malformed DSA private
6657 keys and could lead to a DoS attack or memory corruption for applications
6658 that receive DSA private keys from untrusted sources. This scenario is
6661 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6667 * Disable SRP fake user seed to address a server memory leak.
6669 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6671 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6672 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6673 was changed to ignore the "fake user" SRP seed, even if the seed
6676 Users should use SRP_VBASE_get1_by_user instead. Note that in
6677 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6678 also that even though configuring the SRP seed attempts to hide
6679 invalid usernames by continuing the handshake with fake
6680 credentials, this behaviour is not constant time and no strong
6681 guarantees are made that the handshake is indistinguishable from
6682 that of a valid user.
6687 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6689 In the BN_hex2bn function the number of hex digits is calculated using an
6690 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6691 large values of `i` this can result in `bn_expand` not allocating any
6692 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6693 field as NULL leading to a subsequent NULL ptr deref. For very large values
6694 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6695 In this case memory is allocated to the internal BIGNUM data field, but it
6696 is insufficiently sized leading to heap corruption. A similar issue exists
6697 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6698 is ever called by user applications with very large untrusted hex/dec data.
6699 This is anticipated to be a rare occurrence.
6701 All OpenSSL internal usage of these functions use data that is not expected
6702 to be untrusted, e.g. config file data or application command line
6703 arguments. If user developed applications generate config file data based
6704 on untrusted data then it is possible that this could also lead to security
6705 consequences. This is also anticipated to be rare.
6707 This issue was reported to OpenSSL by Guido Vranken.
6712 * Fix memory issues in `BIO_*printf` functions
6714 The internal `fmtstr` function used in processing a "%s" format string in
6715 the `BIO_*printf` functions could overflow while calculating the length of a
6716 string and cause an OOB read when printing very long strings.
6718 Additionally the internal `doapr_outch` function can attempt to write to an
6719 OOB memory location (at an offset from the NULL pointer) in the event of a
6720 memory allocation failure. In 1.0.2 and below this could be caused where
6721 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6722 could be in processing a very long "%s" format string. Memory leaks can
6725 The first issue may mask the second issue dependent on compiler behaviour.
6726 These problems could enable attacks where large amounts of untrusted data
6727 is passed to the `BIO_*printf` functions. If applications use these functions
6728 in this way then they could be vulnerable. OpenSSL itself uses these
6729 functions when printing out human-readable dumps of ASN.1 data. Therefore
6730 applications that print this data could be vulnerable if the data is from
6731 untrusted sources. OpenSSL command line applications could also be
6732 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6733 as command line arguments.
6735 Libssl is not considered directly vulnerable. Additionally certificates etc
6736 received via remote connections via libssl are also unlikely to be able to
6737 trigger these issues because of message size limits enforced within libssl.
6739 This issue was reported to OpenSSL Guido Vranken.
6744 * Side channel attack on modular exponentiation
6746 A side-channel attack was found which makes use of cache-bank conflicts on
6747 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6748 of RSA keys. The ability to exploit this issue is limited as it relies on
6749 an attacker who has control of code in a thread running on the same
6750 hyper-threaded core as the victim thread which is performing decryptions.
6752 This issue was reported to OpenSSL by Yuval Yarom, The University of
6753 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6754 Nadia Heninger, University of Pennsylvania with more information at
6755 <http://cachebleed.info>.
6760 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6761 if no keysize is specified with default_bits. This fixes an
6762 omission in an earlier change that changed all RSA/DSA key generation
6763 commands to use 2048 bits by default.
6767 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6769 * Protection for DH small subgroup attacks
6771 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6772 switched on by default and cannot be disabled. This could have some
6777 * SSLv2 doesn't block disabled ciphers
6779 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6780 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6781 been disabled, provided that the SSLv2 protocol was not also disabled via
6784 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6785 and Sebastian Schinzel.
6790 * Reject DH handshakes with parameters shorter than 1024 bits.
6794 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6796 * Certificate verify crash with missing PSS parameter
6798 The signature verification routines will crash with a NULL pointer
6799 dereference if presented with an ASN.1 signature using the RSA PSS
6800 algorithm and absent mask generation function parameter. Since these
6801 routines are used to verify certificate signature algorithms this can be
6802 used to crash any certificate verification operation and exploited in a
6803 DoS attack. Any application which performs certificate verification is
6804 vulnerable including OpenSSL clients and servers which enable client
6807 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6812 * X509_ATTRIBUTE memory leak
6814 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6815 memory. This structure is used by the PKCS#7 and CMS routines so any
6816 application which reads PKCS#7 or CMS data from untrusted sources is
6817 affected. SSL/TLS is not affected.
6819 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6825 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6826 This changes the decoding behaviour for some invalid messages,
6827 though the change is mostly in the more lenient direction, and
6828 legacy behaviour is preserved as much as possible.
6832 * In DSA_generate_parameters_ex, if the provided seed is too short,
6833 use a random seed, as already documented.
6835 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6837 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6839 * Alternate chains certificate forgery
6841 During certificate verfification, OpenSSL will attempt to find an
6842 alternative certificate chain if the first attempt to build such a chain
6843 fails. An error in the implementation of this logic can mean that an
6844 attacker could cause certain checks on untrusted certificates to be
6845 bypassed, such as the CA flag, enabling them to use a valid leaf
6846 certificate to act as a CA and "issue" an invalid certificate.
6848 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6854 * Race condition handling PSK identify hint
6856 If PSK identity hints are received by a multi-threaded client then
6857 the values are wrongly updated in the parent SSL_CTX structure. This can
6858 result in a race condition potentially leading to a double free of the
6864 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6866 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6867 incompatibility in the handling of HMAC. The previous ABI has now been
6870 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6872 * Malformed ECParameters causes infinite loop
6874 When processing an ECParameters structure OpenSSL enters an infinite loop
6875 if the curve specified is over a specially malformed binary polynomial
6878 This can be used to perform denial of service against any
6879 system which processes public keys, certificate requests or
6880 certificates. This includes TLS clients and TLS servers with
6881 client authentication enabled.
6883 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6888 * Exploitable out-of-bounds read in X509_cmp_time
6890 X509_cmp_time does not properly check the length of the ASN1_TIME
6891 string and can read a few bytes out of bounds. In addition,
6892 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6895 An attacker can use this to craft malformed certificates and CRLs of
6896 various sizes and potentially cause a segmentation fault, resulting in
6897 a DoS on applications that verify certificates or CRLs. TLS clients
6898 that verify CRLs are affected. TLS clients and servers with client
6899 authentication enabled may be affected if they use custom verification
6902 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6903 independently by Hanno Böck.
6908 * PKCS7 crash with missing EnvelopedContent
6910 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6911 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6912 with missing content and trigger a NULL pointer dereference on parsing.
6914 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6915 structures from untrusted sources are affected. OpenSSL clients and
6916 servers are not affected.
6918 This issue was reported to OpenSSL by Michal Zalewski (Google).
6923 * CMS verify infinite loop with unknown hash function
6925 When verifying a signedData message the CMS code can enter an infinite loop
6926 if presented with an unknown hash function OID. This can be used to perform
6927 denial of service against any system which verifies signedData messages using
6929 This issue was reported to OpenSSL by Johannes Bauer.
6934 * Race condition handling NewSessionTicket
6936 If a NewSessionTicket is received by a multi-threaded client when attempting to
6937 reuse a previous ticket then a race condition can occur potentially leading to
6938 a double free of the ticket data.
6943 * Reject DH handshakes with parameters shorter than 768 bits.
6945 *Kurt Roeckx and Emilia Kasper*
6947 * dhparam: generate 2048-bit parameters by default.
6949 *Kurt Roeckx and Emilia Kasper*
6951 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6953 * Segmentation fault in ASN1_TYPE_cmp fix
6955 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6956 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6957 certificate signature algorithm consistency this can be used to crash any
6958 certificate verification operation and exploited in a DoS attack. Any
6959 application which performs certificate verification is vulnerable including
6960 OpenSSL clients and servers which enable client authentication.
6965 * ASN.1 structure reuse memory corruption fix
6967 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6968 memory corruption via an invalid write. Such reuse is and has been
6969 strongly discouraged and is believed to be rare.
6971 Applications that parse structures containing CHOICE or ANY DEFINED BY
6972 components may be affected. Certificate parsing (d2i_X509 and related
6973 functions) are however not affected. OpenSSL clients and servers are
6979 * PKCS7 NULL pointer dereferences fix
6981 The PKCS#7 parsing code does not handle missing outer ContentInfo
6982 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6983 missing content and trigger a NULL pointer dereference on parsing.
6985 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6986 otherwise parse PKCS#7 structures from untrusted sources are
6987 affected. OpenSSL clients and servers are not affected.
6989 This issue was reported to OpenSSL by Michal Zalewski (Google).
6994 * DoS via reachable assert in SSLv2 servers fix
6996 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6997 servers that both support SSLv2 and enable export cipher suites by sending
6998 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7000 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7001 (OpenSSL development team).
7006 * Use After Free following d2i_ECPrivatekey error fix
7008 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7009 could cause a use after free condition. This, in turn, could cause a double
7010 free in several private key parsing functions (such as d2i_PrivateKey
7011 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7012 for applications that receive EC private keys from untrusted
7013 sources. This scenario is considered rare.
7015 This issue was discovered by the BoringSSL project and fixed in their
7021 * X509_to_X509_REQ NULL pointer deref fix
7023 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7024 the certificate key is invalid. This function is rarely used in practice.
7026 This issue was discovered by Brian Carpenter.
7031 * Removed the export ciphers from the DEFAULT ciphers
7035 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7037 * Build fixes for the Windows and OpenVMS platforms
7039 *Matt Caswell and Richard Levitte*
7041 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7043 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7044 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7045 dereference. This could lead to a Denial Of Service attack. Thanks to
7046 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7051 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7052 dtls1_buffer_record function under certain conditions. In particular this
7053 could occur if an attacker sent repeated DTLS records with the same
7054 sequence number but for the next epoch. The memory leak could be exploited
7055 by an attacker in a Denial of Service attack through memory exhaustion.
7056 Thanks to Chris Mueller for reporting this issue.
7061 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7062 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7063 method would be set to NULL which could later result in a NULL pointer
7064 dereference. Thanks to Frank Schmirler for reporting this issue.
7069 * Abort handshake if server key exchange message is omitted for ephemeral
7072 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7073 reporting this issue.
7078 * Remove non-export ephemeral RSA code on client and server. This code
7079 violated the TLS standard by allowing the use of temporary RSA keys in
7080 non-export ciphersuites and could be used by a server to effectively
7081 downgrade the RSA key length used to a value smaller than the server
7082 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7083 INRIA or reporting this issue.
7088 * Fixed issue where DH client certificates are accepted without verification.
7089 An OpenSSL server will accept a DH certificate for client authentication
7090 without the certificate verify message. This effectively allows a client to
7091 authenticate without the use of a private key. This only affects servers
7092 which trust a client certificate authority which issues certificates
7093 containing DH keys: these are extremely rare and hardly ever encountered.
7094 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7100 * Ensure that the session ID context of an SSL is updated when its
7101 SSL_CTX is updated via SSL_set_SSL_CTX.
7103 The session ID context is typically set from the parent SSL_CTX,
7104 and can vary with the CTX.
7108 * Fix various certificate fingerprint issues.
7110 By using non-DER or invalid encodings outside the signed portion of a
7111 certificate the fingerprint can be changed without breaking the signature.
7112 Although no details of the signed portion of the certificate can be changed
7113 this can cause problems with some applications: e.g. those using the
7114 certificate fingerprint for blacklists.
7116 1. Reject signatures with non zero unused bits.
7118 If the BIT STRING containing the signature has non zero unused bits reject
7119 the signature. All current signature algorithms require zero unused bits.
7121 2. Check certificate algorithm consistency.
7123 Check the AlgorithmIdentifier inside TBS matches the one in the
7124 certificate signature. NB: this will result in signature failure
7125 errors for some broken certificates.
7127 Thanks to Konrad Kraszewski from Google for reporting this issue.
7129 3. Check DSA/ECDSA signatures use DER.
7131 Re-encode DSA/ECDSA signatures and compare with the original received
7132 signature. Return an error if there is a mismatch.
7134 This will reject various cases including garbage after signature
7135 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7136 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7137 (negative or with leading zeroes).
7139 Further analysis was conducted and fixes were developed by Stephen Henson
7140 of the OpenSSL core team.
7146 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7147 results on some platforms, including x86_64. This bug occurs at random
7148 with a very low probability, and is not known to be exploitable in any
7149 way, though its exact impact is difficult to determine. Thanks to Pieter
7150 Wuille (Blockstream) who reported this issue and also suggested an initial
7151 fix. Further analysis was conducted by the OpenSSL development team and
7152 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7153 the OpenSSL core team.
7158 * Do not resume sessions on the server if the negotiated protocol
7159 version does not match the session's version. Resuming with a different
7160 version, while not strictly forbidden by the RFC, is of questionable
7161 sanity and breaks all known clients.
7163 *David Benjamin, Emilia Käsper*
7165 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7166 early CCS messages during renegotiation. (Note that because
7167 renegotiation is encrypted, this early CCS was not exploitable.)
7171 * Tighten client-side session ticket handling during renegotiation:
7172 ensure that the client only accepts a session ticket if the server sends
7173 the extension anew in the ServerHello. Previously, a TLS client would
7174 reuse the old extension state and thus accept a session ticket if one was
7175 announced in the initial ServerHello.
7177 Similarly, ensure that the client requires a session ticket if one
7178 was advertised in the ServerHello. Previously, a TLS client would
7179 ignore a missing NewSessionTicket message.
7183 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7187 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7188 sends a carefully crafted handshake message, to cause OpenSSL to fail
7189 to free up to 64k of memory causing a memory leak. This could be
7190 exploited in a Denial Of Service attack. This issue affects OpenSSL
7191 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7192 whether SRTP is used or configured. Implementations of OpenSSL that
7193 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7195 The fix was developed by the OpenSSL team.
7200 * Session Ticket Memory Leak.
7202 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7203 integrity of that ticket is first verified. In the event of a session
7204 ticket integrity check failing, OpenSSL will fail to free memory
7205 causing a memory leak. By sending a large number of invalid session
7206 tickets an attacker could exploit this issue in a Denial Of Service
7212 * Build option no-ssl3 is incomplete.
7214 When OpenSSL is configured with "no-ssl3" as a build option, servers
7215 could accept and complete a SSL 3.0 handshake, and clients could be
7216 configured to send them.
7219 *Akamai and the OpenSSL team*
7221 * Add support for TLS_FALLBACK_SCSV.
7222 Client applications doing fallback retries should call
7223 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7226 *Adam Langley, Bodo Moeller*
7228 * Add additional DigestInfo checks.
7230 Re-encode DigestInto in DER and check against the original when
7231 verifying RSA signature: this will reject any improperly encoded
7232 DigestInfo structures.
7234 Note: this is a precautionary measure and no attacks are currently known.
7238 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7240 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7241 SRP code can be overrun an internal buffer. Add sanity check that
7242 g, A, B < N to SRP code.
7244 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7245 Group for discovering this issue.
7250 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7251 TLS 1.0 instead of higher protocol versions when the ClientHello message
7252 is badly fragmented. This allows a man-in-the-middle attacker to force a
7253 downgrade to TLS 1.0 even if both the server and the client support a
7254 higher protocol version, by modifying the client's TLS records.
7256 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7257 researching this issue.
7262 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7263 to a denial of service attack. A malicious server can crash the client
7264 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7265 ciphersuite and sending carefully crafted handshake messages.
7267 Thanks to Felix Gröbert (Google) for discovering and researching this
7273 * By sending carefully crafted DTLS packets an attacker could cause openssl
7274 to leak memory. This can be exploited through a Denial of Service attack.
7275 Thanks to Adam Langley for discovering and researching this issue.
7280 * An attacker can force openssl to consume large amounts of memory whilst
7281 processing DTLS handshake messages. This can be exploited through a
7282 Denial of Service attack.
7283 Thanks to Adam Langley for discovering and researching this issue.
7288 * An attacker can force an error condition which causes openssl to crash
7289 whilst processing DTLS packets due to memory being freed twice. This
7290 can be exploited through a Denial of Service attack.
7291 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7297 * If a multithreaded client connects to a malicious server using a resumed
7298 session and the server sends an ec point format extension it could write
7299 up to 255 bytes to freed memory.
7301 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7307 * A malicious server can crash an OpenSSL client with a null pointer
7308 dereference (read) by specifying an SRP ciphersuite even though it was not
7309 properly negotiated with the client. This can be exploited through a
7310 Denial of Service attack.
7312 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7313 discovering and researching this issue.
7318 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7319 X509_name_oneline, X509_name_print_ex et al. to leak some information
7320 from the stack. Applications may be affected if they echo pretty printing
7321 output to the attacker.
7323 Thanks to Ivan Fratric (Google) for discovering this issue.
7326 *Emilia Käsper, and Steve Henson*
7328 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7329 for corner cases. (Certain input points at infinity could lead to
7330 bogus results, with non-infinity inputs mapped to infinity too.)
7334 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7336 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7337 handshake can force the use of weak keying material in OpenSSL
7338 SSL/TLS clients and servers.
7340 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7341 researching this issue. ([CVE-2014-0224])
7343 *KIKUCHI Masashi, Steve Henson*
7345 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7346 OpenSSL DTLS client the code can be made to recurse eventually crashing
7349 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7352 *Imre Rad, Steve Henson*
7354 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7355 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7356 client or server. This is potentially exploitable to run arbitrary
7357 code on a vulnerable client or server.
7359 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7361 *Jüri Aedla, Steve Henson*
7363 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7364 are subject to a denial of service attack.
7366 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7367 this issue. ([CVE-2014-3470])
7369 *Felix Gröbert, Ivan Fratric, Steve Henson*
7371 * Harmonize version and its documentation. -f flag is used to display
7374 *mancha <mancha1@zoho.com>*
7376 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7377 in i2d_ECPrivateKey.
7379 *mancha <mancha1@zoho.com>*
7381 * Fix some double frees. These are not thought to be exploitable.
7383 *mancha <mancha1@zoho.com>*
7385 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7387 * A missing bounds check in the handling of the TLS heartbeat extension
7388 can be used to reveal up to 64k of memory to a connected client or
7391 Thanks for Neel Mehta of Google Security for discovering this bug and to
7392 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7393 preparing the fix ([CVE-2014-0160])
7395 *Adam Langley, Bodo Moeller*
7397 * Fix for the attack described in the paper "Recovering OpenSSL
7398 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7399 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7400 <http://eprint.iacr.org/2014/140>
7402 Thanks to Yuval Yarom and Naomi Benger for discovering this
7403 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7405 *Yuval Yarom and Naomi Benger*
7407 * TLS pad extension: draft-agl-tls-padding-03
7409 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7410 TLS client Hello record length value would otherwise be > 255 and
7411 less that 512 pad with a dummy extension containing zeroes so it
7412 is at least 512 bytes long.
7414 *Adam Langley, Steve Henson*
7416 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7418 * Fix for TLS record tampering bug. A carefully crafted invalid
7419 handshake could crash OpenSSL with a NULL pointer exception.
7420 Thanks to Anton Johansson for reporting this issues.
7423 * Keep original DTLS digest and encryption contexts in retransmission
7424 structures so we can use the previous session parameters if they need
7425 to be resent. ([CVE-2013-6450])
7429 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7430 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7431 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7432 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7433 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7434 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7436 *Rob Stradling, Adam Langley*
7438 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7440 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7441 supporting platforms or when small records were transferred.
7443 *Andy Polyakov, Steve Henson*
7445 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7447 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7449 This addresses the flaw in CBC record processing discovered by
7450 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7451 at: <http://www.isg.rhul.ac.uk/tls/>
7453 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7454 Security Group at Royal Holloway, University of London
7455 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7456 Emilia Käsper for the initial patch.
7459 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7461 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7462 ciphersuites which can be exploited in a denial of service attack.
7463 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7464 and detecting this bug and to Wolfgang Ettlinger
7465 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7470 * Return an error when checking OCSP signatures when key is NULL.
7471 This fixes a DoS attack. ([CVE-2013-0166])
7475 * Make openssl verify return errors.
7477 *Chris Palmer <palmer@google.com> and Ben Laurie*
7479 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7480 the right response is stapled. Also change SSL_get_certificate()
7481 so it returns the certificate actually sent.
7482 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7484 *Rob Stradling <rob.stradling@comodo.com>*
7486 * Fix possible deadlock when decoding public keys.
7490 * Don't use TLS 1.0 record version number in initial client hello
7495 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7497 * Sanity check record length before skipping explicit IV in TLS
7498 1.2, 1.1 and DTLS to fix DoS attack.
7500 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7501 fuzzing as a service testing platform.
7506 * Initialise tkeylen properly when encrypting CMS messages.
7507 Thanks to Solar Designer of Openwall for reporting this issue.
7511 * In FIPS mode don't try to use composite ciphers as they are not
7516 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7518 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7519 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7520 mean any application compiled against OpenSSL 1.0.0 headers setting
7521 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7522 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7523 0x10000000L Any application which was previously compiled against
7524 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7525 will need to be recompiled as a result. Letting be results in
7526 inability to disable specifically TLS 1.1 and in client context,
7527 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7531 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7532 disable just protocol X, but all protocols above X *if* there are
7533 protocols *below* X still enabled. In more practical terms it means
7534 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7535 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7536 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7541 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7543 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7544 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7545 in CRYPTO_realloc_clean.
7547 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7548 issue and to Adam Langley <agl@chromium.org> for fixing it.
7551 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7553 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7557 * Workarounds for some broken servers that "hang" if a client hello
7558 record length exceeds 255 bytes.
7560 1. Do not use record version number > TLS 1.0 in initial client
7561 hello: some (but not all) hanging servers will now work.
7562 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7563 the number of ciphers sent in the client hello. This should be
7564 set to an even number, such as 50, for example by passing:
7565 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7566 Most broken servers should now work.
7567 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7568 TLS 1.2 client support entirely.
7572 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7576 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7578 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7579 STRING form instead of a DigestInfo.
7583 * The format used for MDC2 RSA signatures is inconsistent between EVP
7584 and the RSA_sign/RSA_verify functions. This was made more apparent when
7585 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7586 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7587 the correct format in RSA_verify so both forms transparently work.
7591 * Some servers which support TLS 1.0 can choke if we initially indicate
7592 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7593 encrypted premaster secret. As a workaround use the maximum permitted
7594 client version in client hello, this should keep such servers happy
7595 and still work with previous versions of OpenSSL.
7599 * Add support for TLS/DTLS heartbeats.
7601 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7603 * Add support for SCTP.
7605 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7607 * Improved PRNG seeding for VOS.
7609 *Paul Green <Paul.Green@stratus.com>*
7611 * Extensive assembler packs updates, most notably:
7613 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7614 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7615 - x86_64: bit-sliced AES implementation;
7616 - ARM: NEON support, contemporary platforms optimizations;
7617 - s390x: z196 support;
7618 - `*`: GHASH and GF(2^m) multiplication implementations;
7622 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7623 (removal of unnecessary code)
7625 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7627 * Add TLS key material exporter from RFC 5705.
7631 * Add DTLS-SRTP negotiation from RFC 5764.
7635 * Add Next Protocol Negotiation,
7636 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7637 disabled with a no-npn flag to config or Configure. Code donated
7640 *Adam Langley <agl@google.com> and Ben Laurie*
7642 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7643 NIST-P256, NIST-P521, with constant-time single point multiplication on
7644 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7645 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7646 Code made available under Apache License version 2.0.
7648 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7649 line to include this in your build of OpenSSL, and run "make depend" (or
7650 "make update"). This enables the following EC_METHODs:
7652 EC_GFp_nistp224_method()
7653 EC_GFp_nistp256_method()
7654 EC_GFp_nistp521_method()
7656 EC_GROUP_new_by_curve_name() will automatically use these (while
7657 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7660 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7662 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7663 all platforms. Move ssize_t definition from e_os.h to the public
7664 header file e_os2.h as it now appears in public header file cms.h
7668 * New -sigopt option to the ca, req and x509 utilities. Additional
7669 signature parameters can be passed using this option and in
7674 * Add RSA PSS signing function. This will generate and set the
7675 appropriate AlgorithmIdentifiers for PSS based on those in the
7676 corresponding EVP_MD_CTX structure. No application support yet.
7680 * Support for companion algorithm specific ASN1 signing routines.
7681 New function ASN1_item_sign_ctx() signs a pre-initialised
7682 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7683 the appropriate parameters.
7687 * Add new algorithm specific ASN1 verification initialisation function
7688 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7689 handling will be the same no matter what EVP_PKEY_METHOD is used.
7690 Add a PSS handler to support verification of PSS signatures: checked
7691 against a number of sample certificates.
7695 * Add signature printing for PSS. Add PSS OIDs.
7697 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7699 * Add algorithm specific signature printing. An individual ASN1 method
7700 can now print out signatures instead of the standard hex dump.
7702 More complex signatures (e.g. PSS) can print out more meaningful
7703 information. Include DSA version that prints out the signature
7708 * Password based recipient info support for CMS library: implementing
7713 * Split password based encryption into PBES2 and PBKDF2 functions. This
7714 neatly separates the code into cipher and PBE sections and is required
7715 for some algorithms that split PBES2 into separate pieces (such as
7716 password based CMS).
7720 * Session-handling fixes:
7721 - Fix handling of connections that are resuming with a session ID,
7722 but also support Session Tickets.
7723 - Fix a bug that suppressed issuing of a new ticket if the client
7724 presented a ticket with an expired session.
7725 - Try to set the ticket lifetime hint to something reasonable.
7726 - Make tickets shorter by excluding irrelevant information.
7727 - On the client side, don't ignore renewed tickets.
7729 *Adam Langley, Bodo Moeller (Google)*
7731 * Fix PSK session representation.
7735 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7737 This work was sponsored by Intel.
7741 * Add GCM support to TLS library. Some custom code is needed to split
7742 the IV between the fixed (from PRF) and explicit (from TLS record)
7743 portions. This adds all GCM ciphersuites supported by RFC5288 and
7744 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7745 add a special AESGCM string for GCM only.
7749 * Expand range of ctrls for AES GCM. Permit setting invocation
7750 field on decrypt and retrieval of invocation field only on encrypt.
7754 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7755 As required by RFC5289 these ciphersuites cannot be used if for
7756 versions of TLS earlier than 1.2.
7760 * For FIPS capable OpenSSL interpret a NULL default public key method
7761 as unset and return the appropriate default but do *not* set the default.
7762 This means we can return the appropriate method in applications that
7763 switch between FIPS and non-FIPS modes.
7767 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7768 ENGINE is used then we cannot handle that in the FIPS module so we
7769 keep original code iff non-FIPS operations are allowed.
7773 * Add -attime option to openssl utilities.
7775 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7777 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7781 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7782 FIPS EC methods unconditionally for now.
7786 * New build option no-ec2m to disable characteristic 2 code.
7790 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7791 all cases can be covered as some introduce binary incompatibilities.
7795 * Redirect RSA operations to FIPS module including keygen,
7796 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7800 * Add similar low-level API blocking to ciphers.
7804 * low-level digest APIs are not approved in FIPS mode: any attempt
7805 to use these will cause a fatal error. Applications that *really* want
7806 to use them can use the `private_*` version instead.
7810 * Redirect cipher operations to FIPS module for FIPS builds.
7814 * Redirect digest operations to FIPS module for FIPS builds.
7818 * Update build system to add "fips" flag which will link in fipscanister.o
7819 for static and shared library builds embedding a signature if needed.
7823 * Output TLS supported curves in preference order instead of numerical
7824 order. This is currently hardcoded for the highest order curves first.
7825 This should be configurable so applications can judge speed vs strength.
7829 * Add TLS v1.2 server support for client authentication.
7833 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7838 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7839 FIPS modules versions.
7843 * Add TLS v1.2 client side support for client authentication. Keep cache
7844 of handshake records longer as we don't know the hash algorithm to use
7845 until after the certificate request message is received.
7849 * Initial TLS v1.2 client support. Add a default signature algorithms
7850 extension including all the algorithms we support. Parse new signature
7851 format in client key exchange. Relax some ECC signing restrictions for
7852 TLS v1.2 as indicated in RFC5246.
7856 * Add server support for TLS v1.2 signature algorithms extension. Switch
7857 to new signature format when needed using client digest preference.
7858 All server ciphersuites should now work correctly in TLS v1.2. No client
7859 support yet and no support for client certificates.
7863 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7864 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7865 ciphersuites. At present only RSA key exchange ciphersuites work with
7866 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7867 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7868 and version checking.
7872 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7873 with this defined it will not be affected by any changes to ssl internal
7874 structures. Add several utility functions to allow openssl application
7875 to work with OPENSSL_NO_SSL_INTERN defined.
7879 * A long standing patch to add support for SRP from EdelWeb (Peter
7880 Sylvester and Christophe Renou) was integrated.
7881 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7882 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7885 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7889 * Permit abbreviated handshakes when renegotiating using the function
7890 SSL_renegotiate_abbreviated().
7892 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7894 * Add call to ENGINE_register_all_complete() to
7895 ENGINE_load_builtin_engines(), so some implementations get used
7896 automatically instead of needing explicit application support.
7900 * Add support for TLS key exporter as described in RFC5705.
7902 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7904 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7905 a few changes are required:
7907 Add SSL_OP_NO_TLSv1_1 flag.
7908 Add TLSv1_1 methods.
7909 Update version checking logic to handle version 1.1.
7910 Add explicit IV handling (ported from DTLS code).
7911 Add command line options to s_client/s_server.
7918 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7920 * X509_ATTRIBUTE memory leak
7922 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7923 memory. This structure is used by the PKCS#7 and CMS routines so any
7924 application which reads PKCS#7 or CMS data from untrusted sources is
7925 affected. SSL/TLS is not affected.
7927 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7933 * Race condition handling PSK identify hint
7935 If PSK identity hints are received by a multi-threaded client then
7936 the values are wrongly updated in the parent SSL_CTX structure. This can
7937 result in a race condition potentially leading to a double free of the
7943 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7945 * Malformed ECParameters causes infinite loop
7947 When processing an ECParameters structure OpenSSL enters an infinite loop
7948 if the curve specified is over a specially malformed binary polynomial
7951 This can be used to perform denial of service against any
7952 system which processes public keys, certificate requests or
7953 certificates. This includes TLS clients and TLS servers with
7954 client authentication enabled.
7956 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7961 * Exploitable out-of-bounds read in X509_cmp_time
7963 X509_cmp_time does not properly check the length of the ASN1_TIME
7964 string and can read a few bytes out of bounds. In addition,
7965 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7968 An attacker can use this to craft malformed certificates and CRLs of
7969 various sizes and potentially cause a segmentation fault, resulting in
7970 a DoS on applications that verify certificates or CRLs. TLS clients
7971 that verify CRLs are affected. TLS clients and servers with client
7972 authentication enabled may be affected if they use custom verification
7975 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7976 independently by Hanno Böck.
7981 * PKCS7 crash with missing EnvelopedContent
7983 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7984 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7985 with missing content and trigger a NULL pointer dereference on parsing.
7987 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7988 structures from untrusted sources are affected. OpenSSL clients and
7989 servers are not affected.
7991 This issue was reported to OpenSSL by Michal Zalewski (Google).
7996 * CMS verify infinite loop with unknown hash function
7998 When verifying a signedData message the CMS code can enter an infinite loop
7999 if presented with an unknown hash function OID. This can be used to perform
8000 denial of service against any system which verifies signedData messages using
8002 This issue was reported to OpenSSL by Johannes Bauer.
8007 * Race condition handling NewSessionTicket
8009 If a NewSessionTicket is received by a multi-threaded client when attempting to
8010 reuse a previous ticket then a race condition can occur potentially leading to
8011 a double free of the ticket data.
8016 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8018 * Segmentation fault in ASN1_TYPE_cmp fix
8020 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8021 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8022 certificate signature algorithm consistency this can be used to crash any
8023 certificate verification operation and exploited in a DoS attack. Any
8024 application which performs certificate verification is vulnerable including
8025 OpenSSL clients and servers which enable client authentication.
8030 * ASN.1 structure reuse memory corruption fix
8032 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8033 memory corruption via an invalid write. Such reuse is and has been
8034 strongly discouraged and is believed to be rare.
8036 Applications that parse structures containing CHOICE or ANY DEFINED BY
8037 components may be affected. Certificate parsing (d2i_X509 and related
8038 functions) are however not affected. OpenSSL clients and servers are
8044 * PKCS7 NULL pointer dereferences fix
8046 The PKCS#7 parsing code does not handle missing outer ContentInfo
8047 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8048 missing content and trigger a NULL pointer dereference on parsing.
8050 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8051 otherwise parse PKCS#7 structures from untrusted sources are
8052 affected. OpenSSL clients and servers are not affected.
8054 This issue was reported to OpenSSL by Michal Zalewski (Google).
8059 * DoS via reachable assert in SSLv2 servers fix
8061 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8062 servers that both support SSLv2 and enable export cipher suites by sending
8063 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8065 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8066 (OpenSSL development team).
8071 * Use After Free following d2i_ECPrivatekey error fix
8073 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8074 could cause a use after free condition. This, in turn, could cause a double
8075 free in several private key parsing functions (such as d2i_PrivateKey
8076 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8077 for applications that receive EC private keys from untrusted
8078 sources. This scenario is considered rare.
8080 This issue was discovered by the BoringSSL project and fixed in their
8086 * X509_to_X509_REQ NULL pointer deref fix
8088 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8089 the certificate key is invalid. This function is rarely used in practice.
8091 This issue was discovered by Brian Carpenter.
8096 * Removed the export ciphers from the DEFAULT ciphers
8100 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8102 * Build fixes for the Windows and OpenVMS platforms
8104 *Matt Caswell and Richard Levitte*
8106 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8108 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8109 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8110 dereference. This could lead to a Denial Of Service attack. Thanks to
8111 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8116 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8117 dtls1_buffer_record function under certain conditions. In particular this
8118 could occur if an attacker sent repeated DTLS records with the same
8119 sequence number but for the next epoch. The memory leak could be exploited
8120 by an attacker in a Denial of Service attack through memory exhaustion.
8121 Thanks to Chris Mueller for reporting this issue.
8126 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8127 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8128 method would be set to NULL which could later result in a NULL pointer
8129 dereference. Thanks to Frank Schmirler for reporting this issue.
8134 * Abort handshake if server key exchange message is omitted for ephemeral
8137 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8138 reporting this issue.
8143 * Remove non-export ephemeral RSA code on client and server. This code
8144 violated the TLS standard by allowing the use of temporary RSA keys in
8145 non-export ciphersuites and could be used by a server to effectively
8146 downgrade the RSA key length used to a value smaller than the server
8147 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8148 INRIA or reporting this issue.
8153 * Fixed issue where DH client certificates are accepted without verification.
8154 An OpenSSL server will accept a DH certificate for client authentication
8155 without the certificate verify message. This effectively allows a client to
8156 authenticate without the use of a private key. This only affects servers
8157 which trust a client certificate authority which issues certificates
8158 containing DH keys: these are extremely rare and hardly ever encountered.
8159 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8165 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8166 results on some platforms, including x86_64. This bug occurs at random
8167 with a very low probability, and is not known to be exploitable in any
8168 way, though its exact impact is difficult to determine. Thanks to Pieter
8169 Wuille (Blockstream) who reported this issue and also suggested an initial
8170 fix. Further analysis was conducted by the OpenSSL development team and
8171 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8172 the OpenSSL core team.
8177 * Fix various certificate fingerprint issues.
8179 By using non-DER or invalid encodings outside the signed portion of a
8180 certificate the fingerprint can be changed without breaking the signature.
8181 Although no details of the signed portion of the certificate can be changed
8182 this can cause problems with some applications: e.g. those using the
8183 certificate fingerprint for blacklists.
8185 1. Reject signatures with non zero unused bits.
8187 If the BIT STRING containing the signature has non zero unused bits reject
8188 the signature. All current signature algorithms require zero unused bits.
8190 2. Check certificate algorithm consistency.
8192 Check the AlgorithmIdentifier inside TBS matches the one in the
8193 certificate signature. NB: this will result in signature failure
8194 errors for some broken certificates.
8196 Thanks to Konrad Kraszewski from Google for reporting this issue.
8198 3. Check DSA/ECDSA signatures use DER.
8200 Reencode DSA/ECDSA signatures and compare with the original received
8201 signature. Return an error if there is a mismatch.
8203 This will reject various cases including garbage after signature
8204 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8205 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8206 (negative or with leading zeroes).
8208 Further analysis was conducted and fixes were developed by Stephen Henson
8209 of the OpenSSL core team.
8215 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8217 * Session Ticket Memory Leak.
8219 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8220 integrity of that ticket is first verified. In the event of a session
8221 ticket integrity check failing, OpenSSL will fail to free memory
8222 causing a memory leak. By sending a large number of invalid session
8223 tickets an attacker could exploit this issue in a Denial Of Service
8229 * Build option no-ssl3 is incomplete.
8231 When OpenSSL is configured with "no-ssl3" as a build option, servers
8232 could accept and complete a SSL 3.0 handshake, and clients could be
8233 configured to send them.
8236 *Akamai and the OpenSSL team*
8238 * Add support for TLS_FALLBACK_SCSV.
8239 Client applications doing fallback retries should call
8240 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8243 *Adam Langley, Bodo Moeller*
8245 * Add additional DigestInfo checks.
8247 Reencode DigestInto in DER and check against the original when
8248 verifying RSA signature: this will reject any improperly encoded
8249 DigestInfo structures.
8251 Note: this is a precautionary measure and no attacks are currently known.
8255 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8257 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8258 to a denial of service attack. A malicious server can crash the client
8259 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8260 ciphersuite and sending carefully crafted handshake messages.
8262 Thanks to Felix Gröbert (Google) for discovering and researching this
8268 * By sending carefully crafted DTLS packets an attacker could cause openssl
8269 to leak memory. This can be exploited through a Denial of Service attack.
8270 Thanks to Adam Langley for discovering and researching this issue.
8275 * An attacker can force openssl to consume large amounts of memory whilst
8276 processing DTLS handshake messages. This can be exploited through a
8277 Denial of Service attack.
8278 Thanks to Adam Langley for discovering and researching this issue.
8283 * An attacker can force an error condition which causes openssl to crash
8284 whilst processing DTLS packets due to memory being freed twice. This
8285 can be exploited through a Denial of Service attack.
8286 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8292 * If a multithreaded client connects to a malicious server using a resumed
8293 session and the server sends an ec point format extension it could write
8294 up to 255 bytes to freed memory.
8296 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8302 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8303 X509_name_oneline, X509_name_print_ex et al. to leak some information
8304 from the stack. Applications may be affected if they echo pretty printing
8305 output to the attacker.
8307 Thanks to Ivan Fratric (Google) for discovering this issue.
8310 *Emilia Käsper, and Steve Henson*
8312 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8313 for corner cases. (Certain input points at infinity could lead to
8314 bogus results, with non-infinity inputs mapped to infinity too.)
8318 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8320 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8321 handshake can force the use of weak keying material in OpenSSL
8322 SSL/TLS clients and servers.
8324 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8325 researching this issue. ([CVE-2014-0224])
8327 *KIKUCHI Masashi, Steve Henson*
8329 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8330 OpenSSL DTLS client the code can be made to recurse eventually crashing
8333 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8336 *Imre Rad, Steve Henson*
8338 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8339 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8340 client or server. This is potentially exploitable to run arbitrary
8341 code on a vulnerable client or server.
8343 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8345 *Jüri Aedla, Steve Henson*
8347 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8348 are subject to a denial of service attack.
8350 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8351 this issue. ([CVE-2014-3470])
8353 *Felix Gröbert, Ivan Fratric, Steve Henson*
8355 * Harmonize version and its documentation. -f flag is used to display
8358 *mancha <mancha1@zoho.com>*
8360 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8361 in i2d_ECPrivateKey.
8363 *mancha <mancha1@zoho.com>*
8365 * Fix some double frees. These are not thought to be exploitable.
8367 *mancha <mancha1@zoho.com>*
8369 * Fix for the attack described in the paper "Recovering OpenSSL
8370 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8371 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8372 <http://eprint.iacr.org/2014/140>
8374 Thanks to Yuval Yarom and Naomi Benger for discovering this
8375 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8377 *Yuval Yarom and Naomi Benger*
8379 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8381 * Keep original DTLS digest and encryption contexts in retransmission
8382 structures so we can use the previous session parameters if they need
8383 to be resent. ([CVE-2013-6450])
8387 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8388 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8389 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8390 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8391 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8392 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8394 *Rob Stradling, Adam Langley*
8396 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8398 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8400 This addresses the flaw in CBC record processing discovered by
8401 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8402 at: <http://www.isg.rhul.ac.uk/tls/>
8404 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8405 Security Group at Royal Holloway, University of London
8406 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8407 Emilia Käsper for the initial patch.
8410 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8412 * Return an error when checking OCSP signatures when key is NULL.
8413 This fixes a DoS attack. ([CVE-2013-0166])
8417 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8418 the right response is stapled. Also change SSL_get_certificate()
8419 so it returns the certificate actually sent.
8420 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8421 (This is a backport)
8423 *Rob Stradling <rob.stradling@comodo.com>*
8425 * Fix possible deadlock when decoding public keys.
8429 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8431 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8434 * Sanity check record length before skipping explicit IV in DTLS
8437 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8438 fuzzing as a service testing platform.
8443 * Initialise tkeylen properly when encrypting CMS messages.
8444 Thanks to Solar Designer of Openwall for reporting this issue.
8448 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8450 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8451 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8452 in CRYPTO_realloc_clean.
8454 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8455 issue and to Adam Langley <agl@chromium.org> for fixing it.
8458 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8460 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8462 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8463 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8464 content decryption and always return the same error. Note: this attack
8465 needs on average 2^20 messages so it only affects automated senders. The
8466 old behaviour can be re-enabled in the CMS code by setting the
8467 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8468 an MMA defence is not necessary.
8469 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8470 this issue. ([CVE-2012-0884])
8474 * Fix CVE-2011-4619: make sure we really are receiving a
8475 client hello before rejecting multiple SGC restarts. Thanks to
8476 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8480 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8482 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8483 Thanks to Antonio Martin, Enterprise Secure Access Research and
8484 Development, Cisco Systems, Inc. for discovering this bug and
8485 preparing a fix. ([CVE-2012-0050])
8489 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8491 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8492 of the Vaudenay padding oracle attack on CBC mode encryption
8493 which enables an efficient plaintext recovery attack against
8494 the OpenSSL implementation of DTLS. Their attack exploits timing
8495 differences arising during decryption processing. A research
8496 paper describing this attack can be found at:
8497 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8498 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8499 Security Group at Royal Holloway, University of London
8500 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8501 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8502 for preparing the fix. ([CVE-2011-4108])
8504 *Robin Seggelmann, Michael Tuexen*
8506 * Clear bytes used for block padding of SSL 3.0 records.
8509 *Adam Langley (Google)*
8511 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8512 Kadianakis <desnacked@gmail.com> for discovering this issue and
8513 Adam Langley for preparing the fix. ([CVE-2011-4619])
8515 *Adam Langley (Google)*
8517 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8519 *Andrey Kulikov <amdeich@gmail.com>*
8521 * Prevent malformed RFC3779 data triggering an assertion failure.
8522 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8523 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8525 *Rob Austein <sra@hactrn.net>*
8527 * Improved PRNG seeding for VOS.
8529 *Paul Green <Paul.Green@stratus.com>*
8531 * Fix ssl_ciph.c set-up race.
8533 *Adam Langley (Google)*
8535 * Fix spurious failures in ecdsatest.c.
8537 *Emilia Käsper (Google)*
8539 * Fix the BIO_f_buffer() implementation (which was mixing different
8540 interpretations of the `..._len` fields).
8542 *Adam Langley (Google)*
8544 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8545 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8546 threads won't reuse the same blinding coefficients.
8548 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8549 lock to call BN_BLINDING_invert_ex, and avoids one use of
8550 BN_BLINDING_update for each BN_BLINDING structure (previously,
8551 the last update always remained unused).
8553 *Emilia Käsper (Google)*
8555 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8557 *Bob Buckholz (Google)*
8559 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8561 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8562 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8564 *Kaspar Brand <ossl@velox.ch>*
8566 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8567 for multi-threaded use of ECDH. ([CVE-2011-3210])
8569 *Adam Langley (Google)*
8571 * Fix x509_name_ex_d2i memory leak on bad inputs.
8575 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8576 signature public key algorithm by using OID xref utilities instead.
8577 Before this you could only use some ECC ciphersuites with SHA1 only.
8581 * Add protection against ECDSA timing attacks as mentioned in the paper
8582 by Billy Bob Brumley and Nicola Tuveri, see:
8583 <http://eprint.iacr.org/2011/232.pdf>
8585 *Billy Bob Brumley and Nicola Tuveri*
8587 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8589 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8591 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8593 * Fix bug in string printing code: if *any* escaping is enabled we must
8594 escape the escape character (backslash) or the resulting string is
8599 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8601 * Disable code workaround for ancient and obsolete Netscape browsers
8602 and servers: an attacker can use it in a ciphersuite downgrade attack.
8603 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8607 * Fixed J-PAKE implementation error, originally discovered by
8608 Sebastien Martini, further info and confirmation from Stefan
8609 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8613 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8615 * Fix extension code to avoid race conditions which can result in a buffer
8616 overrun vulnerability: resumed sessions must not be modified as they can
8617 be shared by multiple threads. CVE-2010-3864
8621 * Fix WIN32 build system to correctly link an ENGINE directory into
8626 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8628 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8631 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8633 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8635 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8636 context. The operation can be customised via the ctrl mechanism in
8637 case ENGINEs want to include additional functionality.
8641 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8645 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8646 output hashes compatible with older versions of OpenSSL.
8648 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8650 * Fix compression algorithm handling: if resuming a session use the
8651 compression algorithm of the resumed session instead of determining
8652 it from client hello again. Don't allow server to change algorithm.
8656 * Add load_crls() function to commands tidying load_certs() too. Add option
8657 to verify utility to allow additional CRLs to be included.
8661 * Update OCSP request code to permit adding custom headers to the request:
8662 some responders need this.
8666 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8669 *Julia Lawall <julia@diku.dk>*
8671 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8672 needlessly dereferenced structures, used obsolete functions and
8673 didn't handle all updated verify codes correctly.
8677 * Disable MD2 in the default configuration.
8681 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8682 indicate the initial BIO being pushed or popped. This makes it possible
8683 to determine whether the BIO is the one explicitly called or as a result
8684 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8685 it handles reference counts correctly and doesn't zero out the I/O bio
8686 when it is not being explicitly popped. WARNING: applications which
8687 included workarounds for the old buggy behaviour will need to be modified
8688 or they could free up already freed BIOs.
8692 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8693 renaming to all platforms (within the 0.9.8 branch, this was
8694 done conditionally on Netware platforms to avoid a name clash).
8696 *Guenter <lists@gknw.net>*
8698 * Add ECDHE and PSK support to DTLS.
8700 *Michael Tuexen <tuexen@fh-muenster.de>*
8702 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8707 * Add "missing" function EVP_MD_flags() (without this the only way to
8708 retrieve a digest flags is by accessing the structure directly. Update
8709 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8710 or cipher is registered as in the "from" argument. Print out all
8711 registered digests in the dgst usage message instead of manually
8712 attempting to work them out.
8716 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8717 this allows the use of compression and extensions. Change default cipher
8718 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8719 by default unless an application cipher string requests it.
8723 * Alter match criteria in PKCS12_parse(). It used to try to use local
8724 key ids to find matching certificates and keys but some PKCS#12 files
8725 don't follow the (somewhat unwritten) rules and this strategy fails.
8726 Now just gather all certificates together and the first private key
8727 then look for the first certificate that matches the key.
8731 * Support use of registered digest and cipher names for dgst and cipher
8732 commands instead of having to add each one as a special case. So now
8739 openssl dgst -sha256 foo
8741 and this works for ENGINE based algorithms too.
8745 * Update Gost ENGINE to support parameter files.
8747 *Victor B. Wagner <vitus@cryptocom.ru>*
8749 * Support GeneralizedTime in ca utility.
8751 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8753 * Enhance the hash format used for certificate directory links. The new
8754 form uses the canonical encoding (meaning equivalent names will work
8755 even if they aren't identical) and uses SHA1 instead of MD5. This form
8756 is incompatible with the older format and as a result c_rehash should
8757 be used to rebuild symbolic links.
8761 * Make PKCS#8 the default write format for private keys, replacing the
8762 traditional format. This form is standardised, more secure and doesn't
8763 include an implicit MD5 dependency.
8767 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8768 committed to OpenSSL should pass this lot as a minimum.
8772 * Add session ticket override functionality for use by EAP-FAST.
8774 *Jouni Malinen <j@w1.fi>*
8776 * Modify HMAC functions to return a value. Since these can be implemented
8777 in an ENGINE errors can occur.
8781 * Type-checked OBJ_bsearch_ex.
8785 * Type-checked OBJ_bsearch. Also some constification necessitated
8786 by type-checking. Still to come: TXT_DB, bsearch(?),
8787 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8792 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8793 seconds to a tm structure directly, instead of going through OS
8794 specific date routines. This avoids any issues with OS routines such
8795 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8796 and X509_time_adj_ex() to cover the extended range. The existing
8797 X509_time_adj() is still usable and will no longer have any date issues.
8801 * Delta CRL support. New use deltas option which will attempt to locate
8802 and search any appropriate delta CRLs available.
8804 This work was sponsored by Google.
8808 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8809 code and add additional score elements. Validate alternate CRL paths
8810 as part of the CRL checking and indicate a new error "CRL path validation
8811 error" in this case. Applications wanting additional details can use
8812 the verify callback and check the new "parent" field. If this is not
8813 NULL CRL path validation is taking place. Existing applications won't
8814 see this because it requires extended CRL support which is off by
8817 This work was sponsored by Google.
8821 * Support for freshest CRL extension.
8823 This work was sponsored by Google.
8827 * Initial indirect CRL support. Currently only supported in the CRLs
8828 passed directly and not via lookup. Process certificate issuer
8829 CRL entry extension and lookup CRL entries by bother issuer name
8830 and serial number. Check and process CRL issuer entry in IDP extension.
8832 This work was sponsored by Google.
8836 * Add support for distinct certificate and CRL paths. The CRL issuer
8837 certificate is validated separately in this case. Only enabled if
8838 an extended CRL support flag is set: this flag will enable additional
8839 CRL functionality in future.
8841 This work was sponsored by Google.
8845 * Add support for policy mappings extension.
8847 This work was sponsored by Google.
8851 * Fixes to pathlength constraint, self issued certificate handling,
8852 policy processing to align with RFC3280 and PKITS tests.
8854 This work was sponsored by Google.
8858 * Support for name constraints certificate extension. DN, email, DNS
8859 and URI types are currently supported.
8861 This work was sponsored by Google.
8865 * To cater for systems that provide a pointer-based thread ID rather
8866 than numeric, deprecate the current numeric thread ID mechanism and
8867 replace it with a structure and associated callback type. This
8868 mechanism allows a numeric "hash" to be extracted from a thread ID in
8869 either case, and on platforms where pointers are larger than 'long',
8870 mixing is done to help ensure the numeric 'hash' is usable even if it
8871 can't be guaranteed unique. The default mechanism is to use "&errno"
8872 as a pointer-based thread ID to distinguish between threads.
8874 Applications that want to provide their own thread IDs should now use
8875 CRYPTO_THREADID_set_callback() to register a callback that will call
8876 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8878 Note that ERR_remove_state() is now deprecated, because it is tied
8879 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8880 to free the current thread's error state should be replaced by
8881 ERR_remove_thread_state(NULL).
8883 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8884 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8885 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8886 application was previously providing a numeric thread callback that
8887 was inappropriate for distinguishing threads, then uniqueness might
8888 have been obtained with &errno that happened immediately in the
8889 intermediate development versions of OpenSSL; this is no longer the
8890 case, the numeric thread callback will now override the automatic use
8893 *Geoff Thorpe, with help from Bodo Moeller*
8895 * Initial support for different CRL issuing certificates. This covers a
8896 simple case where the self issued certificates in the chain exist and
8897 the real CRL issuer is higher in the existing chain.
8899 This work was sponsored by Google.
8903 * Removed effectively defunct crypto/store from the build.
8907 * Revamp of STACK to provide stronger type-checking. Still to come:
8908 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8909 ASN1_STRING, CONF_VALUE.
8913 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8914 RAM on SSL connections. This option can save about 34k per idle SSL.
8918 * Revamp of LHASH to provide stronger type-checking. Still to come:
8919 STACK, TXT_DB, bsearch, qsort.
8923 * Initial support for Cryptographic Message Syntax (aka CMS) based
8924 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8925 support for data, signedData, compressedData, digestedData and
8926 encryptedData, envelopedData types included. Scripts to check against
8927 RFC4134 examples draft and interop and consistency checks of many
8928 content types and variants.
8932 * Add options to enc utility to support use of zlib compression BIO.
8936 * Extend mk1mf to support importing of options and assembly language
8937 files from Configure script, currently only included in VC-WIN32.
8938 The assembly language rules can now optionally generate the source
8939 files from the associated perl scripts.
8943 * Implement remaining functionality needed to support GOST ciphersuites.
8944 Interop testing has been performed using CryptoPro implementations.
8946 *Victor B. Wagner <vitus@cryptocom.ru>*
8948 * s390x assembler pack.
8952 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8957 * Implement Opaque PRF Input TLS extension as specified in
8958 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8959 official specification yet and no extension type assignment by
8960 IANA exists, this extension (for now) will have to be explicitly
8961 enabled when building OpenSSL by providing the extension number
8962 to use. For example, specify an option
8964 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8966 to the "config" or "Configure" script to enable the extension,
8967 assuming extension number 0x9527 (which is a completely arbitrary
8968 and unofficial assignment based on the MD5 hash of the Internet
8969 Draft). Note that by doing so, you potentially lose
8970 interoperability with other TLS implementations since these might
8971 be using the same extension number for other purposes.
8973 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8974 opaque PRF input value to use in the handshake. This will create
8975 an internal copy of the length-'len' string at 'src', and will
8976 return non-zero for success.
8978 To get more control and flexibility, provide a callback function
8981 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8982 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8986 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8989 Callback function 'cb' will be called in handshakes, and is
8990 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8991 Argument 'arg' is for application purposes (the value as given to
8992 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8993 be provided to the callback function). The callback function
8994 has to return non-zero to report success: usually 1 to use opaque
8995 PRF input just if possible, or 2 to enforce use of the opaque PRF
8996 input. In the latter case, the library will abort the handshake
8997 if opaque PRF input is not successfully negotiated.
8999 Arguments 'peerinput' and 'len' given to the callback function
9000 will always be NULL and 0 in the case of a client. A server will
9001 see the client's opaque PRF input through these variables if
9002 available (NULL and 0 otherwise). Note that if the server
9003 provides an opaque PRF input, the length must be the same as the
9004 length of the client's opaque PRF input.
9006 Note that the callback function will only be called when creating
9007 a new session (session resumption can resume whatever was
9008 previously negotiated), and will not be called in SSL 2.0
9009 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9010 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9011 for applications that need to enforce opaque PRF input.
9015 * Update ssl code to support digests other than SHA1+MD5 for handshake
9018 *Victor B. Wagner <vitus@cryptocom.ru>*
9020 * Add RFC4507 support to OpenSSL. This includes the corrections in
9021 RFC4507bis. The encrypted ticket format is an encrypted encoded
9022 SSL_SESSION structure, that way new session features are automatically
9025 If a client application caches session in an SSL_SESSION structure
9026 support is transparent because tickets are now stored in the encoded
9029 The SSL_CTX structure automatically generates keys for ticket
9030 protection in servers so again support should be possible
9031 with no application modification.
9033 If a client or server wishes to disable RFC4507 support then the option
9034 SSL_OP_NO_TICKET can be set.
9036 Add a TLS extension debugging callback to allow the contents of any client
9037 or server extensions to be examined.
9039 This work was sponsored by Google.
9043 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9044 OpenSSL should now compile cleanly on gcc 4.2
9046 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9048 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9049 support including streaming MAC support: this is required for GOST
9050 ciphersuite support.
9052 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9054 * Add option -stream to use PKCS#7 streaming in smime utility. New
9055 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9056 to output in BER and PEM format.
9060 * Experimental support for use of HMAC via EVP_PKEY interface. This
9061 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9062 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9063 ENGINE support for HMAC keys which are unextractable. New -mac and
9064 -macopt options to dgst utility.
9068 * New option -sigopt to dgst utility. Update dgst to use
9069 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9070 alternative signing parameters such as X9.31 or PSS in the dgst
9075 * Change ssl_cipher_apply_rule(), the internal function that does
9076 the work each time a ciphersuite string requests enabling
9077 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9078 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9079 the order of disabled ciphersuites such that those ciphersuites
9080 that most recently went from enabled to disabled not only stay
9081 in order with respect to each other, but also have higher priority
9082 than other disabled ciphersuites the next time ciphersuites are
9085 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9086 the same ciphersuites as with "HIGH" alone, but in a specific
9087 order where the PSK ciphersuites come first (since they are the
9088 most recently disabled ciphersuites when "HIGH" is parsed).
9090 Also, change ssl_create_cipher_list() (using this new
9091 functionality) such that between otherwise identical
9092 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9097 * Change ssl_create_cipher_list() so that it automatically
9098 arranges the ciphersuites in reasonable order before starting
9099 to process the rule string. Thus, the definition for "DEFAULT"
9100 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9101 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9102 This makes it much easier to arrive at a reasonable default order
9103 in applications for which anonymous ciphers are OK (meaning
9104 that you can't actually use DEFAULT).
9106 *Bodo Moeller; suggested by Victor Duchovni*
9108 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9109 processing) into multiple integers instead of setting
9110 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9111 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9112 (These masks as well as the individual bit definitions are hidden
9113 away into the non-exported interface ssl/ssl_locl.h, so this
9114 change to the definition of the SSL_CIPHER structure shouldn't
9115 affect applications.) This give us more bits for each of these
9116 categories, so there is no longer a need to coagulate AES128 and
9117 AES256 into a single algorithm bit, and to coagulate Camellia128
9118 and Camellia256 into a single algorithm bit, which has led to all
9121 Thus, among other things, the kludge introduced in 0.9.7m and
9122 0.9.8e for masking out AES256 independently of AES128 or masking
9123 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9125 With the change, we also introduce new ciphersuite aliases that
9126 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9131 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9132 Use the leftmost N bytes of the signature input if the input is
9133 larger than the prime q (with N being the size in bytes of q).
9137 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9138 it yet and it is largely untested.
9142 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9146 * Initial incomplete changes to avoid need for function casts in OpenSSL
9147 some compilers (gcc 4.2 and later) reject their use. Safestack is
9148 reimplemented. Update ASN1 to avoid use of legacy functions.
9152 * Win32/64 targets are linked with Winsock2.
9156 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9157 to external functions. This can be used to increase CRL handling
9158 efficiency especially when CRLs are very large by (for example) storing
9159 the CRL revoked certificates in a database.
9163 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9164 new CRLs added to a directory can be used. New command line option
9165 -verify_return_error to s_client and s_server. This causes real errors
9166 to be returned by the verify callback instead of carrying on no matter
9167 what. This reflects the way a "real world" verify callback would behave.
9171 * GOST engine, supporting several GOST algorithms and public key formats.
9172 Kindly donated by Cryptocom.
9176 * Partial support for Issuing Distribution Point CRL extension. CRLs
9177 partitioned by DP are handled but no indirect CRL or reason partitioning
9178 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9179 selected via a scoring technique which handles IDP and AKID in CRLs.
9183 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9184 will ultimately be used for all verify operations: this will remove the
9185 X509_STORE dependency on certificate verification and allow alternative
9186 lookup methods. X509_STORE based implementations of these two callbacks.
9190 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9191 Modify get_crl() to find a valid (unexpired) CRL if possible.
9195 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9196 this would be called X509_CRL_cmp() but that name is already used by
9197 a function that just compares CRL issuer names. Cache several CRL
9198 extensions in X509_CRL structure and cache CRLDP in X509.
9202 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9203 this maps equivalent X509_NAME structures into a consistent structure.
9204 Name comparison can then be performed rapidly using memcmp().
9208 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9213 * Allow digests to supply their own micalg string for S/MIME type using
9214 the ctrl EVP_MD_CTRL_MICALG.
9218 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9219 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9220 ctrl. It can then customise the structure before and/or after signing
9225 * New function OBJ_add_sigid() to allow application defined signature OIDs
9226 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9227 to free up any added signature OIDs.
9231 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9232 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9233 digest and cipher tables. New options added to openssl utility:
9234 list-message-digest-algorithms and list-cipher-algorithms.
9238 * Change the array representation of binary polynomials: the list
9239 of degrees of non-zero coefficients is now terminated with -1.
9240 Previously it was terminated with 0, which was also part of the
9241 value; thus, the array representation was not applicable to
9242 polynomials where t^0 has coefficient zero. This change makes
9243 the array representation useful in a more general context.
9247 * Various modifications and fixes to SSL/TLS cipher string
9248 handling. For ECC, the code now distinguishes between fixed ECDH
9249 with RSA certificates on the one hand and with ECDSA certificates
9250 on the other hand, since these are separate ciphersuites. The
9251 unused code for Fortezza ciphersuites has been removed.
9253 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9254 (not "ECDHE"). For consistency with the code for DH
9255 certificates, use of ECDH certificates is now considered ECDH
9256 authentication, not RSA or ECDSA authentication (the latter is
9257 merely the CA's signing algorithm and not actively used in the
9260 The temporary ciphersuite alias "ECCdraft" is no longer
9261 available, and ECC ciphersuites are no longer excluded from "ALL"
9262 and "DEFAULT". The following aliases now exist for RFC 4492
9263 ciphersuites, most of these by analogy with the DH case:
9265 kECDHr - ECDH cert, signed with RSA
9266 kECDHe - ECDH cert, signed with ECDSA
9267 kECDH - ECDH cert (signed with either RSA or ECDSA)
9268 kEECDH - ephemeral ECDH
9269 ECDH - ECDH cert or ephemeral ECDH
9275 AECDH - anonymous ECDH
9276 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9280 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9281 Use correct micalg parameters depending on digest(s) in signed message.
9285 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9286 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9290 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9291 an engine to register a method. Add ENGINE lookups for methods and
9292 functional reference processing.
9296 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9297 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9302 * New -resign option to smime utility. This adds one or more signers
9303 to an existing PKCS#7 signedData structure. Also -md option to use an
9304 alternative message digest algorithm for signing.
9308 * Tidy up PKCS#7 routines and add new functions to make it easier to
9309 create PKCS7 structures containing multiple signers. Update smime
9310 application to support multiple signers.
9314 * New -macalg option to pkcs12 utility to allow setting of an alternative
9319 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9320 Reorganize PBE internals to lookup from a static table using NIDs,
9321 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9322 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9323 PRF which will be automatically used with PBES2.
9327 * Replace the algorithm specific calls to generate keys in "req" with the
9332 * Update PKCS#7 enveloped data routines to use new API. This is now
9333 supported by any public key method supporting the encrypt operation. A
9334 ctrl is added to allow the public key algorithm to examine or modify
9335 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9340 * Add a ctrl to asn1 method to allow a public key algorithm to express
9341 a default digest type to use. In most cases this will be SHA1 but some
9342 algorithms (such as GOST) need to specify an alternative digest. The
9343 return value indicates how strong the preference is 1 means optional and
9344 2 is mandatory (that is it is the only supported type). Modify
9345 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9346 use the default md. Update openssl utilities to use the default digest
9347 type for signing if it is not explicitly indicated.
9351 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9352 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9353 signing method from the key type. This effectively removes the link
9354 between digests and public key types.
9358 * Add an OID cross reference table and utility functions. Its purpose is to
9359 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9360 rsaEncryption. This will allow some of the algorithm specific hackery
9361 needed to use the correct OID to be removed.
9365 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9366 structures for PKCS7_sign(). They are now set up by the relevant public
9371 * Add provisional EC pkey method with support for ECDSA and ECDH.
9375 * Add support for key derivation (agreement) in the API, DH method and
9380 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9381 public and private key formats. As a side effect these add additional
9382 command line functionality not previously available: DSA signatures can be
9383 generated and verified using pkeyutl and DH key support and generation in
9390 *Oliver Tappe <zooey@hirschkaefer.de>*
9392 * New make target "install_html_docs" installs HTML renditions of the
9395 *Oliver Tappe <zooey@hirschkaefer.de>*
9397 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9398 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9399 support key and parameter generation and add initial key generation
9400 functionality for RSA.
9404 * Add functions for main EVP_PKEY_method operations. The undocumented
9405 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9406 `EVP_PKEY_{encrypt,decrypt}_old`.
9410 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9411 key API, doesn't do much yet.
9415 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9416 public key algorithms. New option to openssl utility:
9417 "list-public-key-algorithms" to print out info.
9421 * Implement the Supported Elliptic Curves Extension for
9422 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9426 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9427 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9431 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9432 utilities such as rsa, dsa, dsaparam etc except they process any key
9437 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9438 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9439 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9444 * Initial support for pluggable public key ASN1.
9445 De-spaghettify the public key ASN1 handling. Move public and private
9446 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9447 algorithm specific handling to a single module within the relevant
9448 algorithm directory. Add functions to allow (near) opaque processing
9449 of public and private key structures.
9453 * Implement the Supported Point Formats Extension for
9454 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9458 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9459 for the psk identity [hint] and the psk callback functions to the
9460 SSL_SESSION, SSL and SSL_CTX structure.
9463 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9467 SSL_CTX_use_psk_identity_hint
9468 SSL_get_psk_identity_hint
9469 SSL_get_psk_identity
9470 SSL_use_psk_identity_hint
9472 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9474 * Add RFC 3161 compliant time stamp request creation, response generation
9475 and response verification functionality.
9477 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9479 * Add initial support for TLS extensions, specifically for the server_name
9480 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9481 have new members for a host name. The SSL data structure has an
9482 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9483 stored in that context to allow for session resumption, even after the
9484 SSL has been switched to a new SSL_CTX in reaction to a client's
9485 server_name extension.
9487 New functions (subject to change):
9489 SSL_get_servername()
9490 SSL_get_servername_type()
9493 New CTRL codes and macros (subject to change):
9495 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9496 - SSL_CTX_set_tlsext_servername_callback()
9497 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9498 - SSL_CTX_set_tlsext_servername_arg()
9499 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9501 openssl s_client has a new '-servername ...' option.
9503 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9504 '-key2 ...', '-servername_fatal' (subject to change). This allows
9505 testing the HostName extension for a specific single host name ('-cert'
9506 and '-key' remain fallbacks for handshakes without HostName
9507 negotiation). If the unrecognized_name alert has to be sent, this by
9508 default is a warning; it becomes fatal with the '-servername_fatal'
9511 *Peter Sylvester, Remy Allais, Christophe Renou*
9513 * Whirlpool hash implementation is added.
9517 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9518 bn(64,32). Because of instruction set limitations it doesn't have
9519 any negative impact on performance. This was done mostly in order
9520 to make it possible to share assembler modules, such as bn_mul_mont
9521 implementations, between 32- and 64-bit builds without hassle.
9525 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9526 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9531 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9532 dedicated Montgomery multiplication procedure, is introduced.
9533 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9534 "64-bit" performance on certain 32-bit targets.
9538 * New option SSL_OP_NO_COMP to disable use of compression selectively
9539 in SSL structures. New SSL ctrl to set maximum send fragment size.
9540 Save memory by setting the I/O buffer sizes dynamically instead of
9541 using the maximum available value.
9545 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9546 in addition to the text details.
9550 * Very, very preliminary EXPERIMENTAL support for printing of general
9551 ASN1 structures. This currently produces rather ugly output and doesn't
9552 handle several customised structures at all.
9556 * Integrated support for PVK file format and some related formats such
9557 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9558 these in the 'rsa' and 'dsa' utilities.
9562 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9566 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9567 place for the (very old) "NETSCAPE" format certificates which are now
9568 handled using new ASN1 code equivalents.
9572 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9573 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9574 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9578 * Modify CRL distribution points extension code to print out previously
9579 unsupported fields. Enhance extension setting code to allow setting of
9584 * Add print and set support for Issuing Distribution Point CRL extension.
9588 * Change 'Configure' script to enable Camellia by default.
9595 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9597 * When rejecting SSL/TLS records due to an incorrect version number, never
9598 update s->server with a new major version number. As of
9599 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9600 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9601 the previous behavior could result in a read attempt at NULL when
9602 receiving specific incorrect SSL/TLS records once record payload
9603 protection is active. ([CVE-2010-0740])
9605 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9607 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9608 could be crashed if the relevant tables were not present (e.g. chrooted).
9610 *Tomas Hoger <thoger@redhat.com>*
9612 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9614 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9616 *Martin Olsson, Neel Mehta*
9618 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9619 accommodate for stack sorting, always a write lock!).
9623 * On some versions of WIN32 Heap32Next is very slow. This can cause
9624 excessive delays in the RAND_poll(): over a minute. As a workaround
9625 include a time check in the inner Heap32Next loop too.
9629 * The code that handled flushing of data in SSL/TLS originally used the
9630 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9631 the problem outlined in PR#1949. The fix suggested there however can
9632 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9633 of Apache). So instead simplify the code to flush unconditionally.
9634 This should be fine since flushing with no data to flush is a no op.
9638 * Handle TLS versions 2.0 and later properly and correctly use the
9639 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9640 off ancient servers have a habit of sticking around for a while...
9644 * Modify compression code so it frees up structures without using the
9645 ex_data callbacks. This works around a problem where some applications
9646 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9647 restarting) then use compression (e.g. SSL with compression) later.
9648 This results in significant per-connection memory leaks and
9649 has caused some security issues including CVE-2008-1678 and
9654 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9655 change when encrypting or decrypting.
9659 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9660 connect and renegotiate with servers which do not support RI.
9661 Until RI is more widely deployed this option is enabled by default.
9665 * Add "missing" ssl ctrls to clear options and mode.
9669 * If client attempts to renegotiate and doesn't support RI respond with
9670 a no_renegotiation alert as required by RFC5746. Some renegotiating
9671 TLS clients will continue a connection gracefully when they receive
9672 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9673 waiting for a server hello which it will never receive. Now we treat a
9674 received no_renegotiation alert as a fatal error. This is because
9675 applications requesting a renegotiation might well expect it to succeed
9676 and would have no code in place to handle the server denying it so the
9677 only safe thing to do is to terminate the connection.
9681 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9682 peer supports secure renegotiation and 0 otherwise. Print out peer
9683 renegotiation support in s_client/s_server.
9687 * Replace the highly broken and deprecated SPKAC certification method with
9688 the updated NID creation version. This should correctly handle UTF8.
9692 * Implement RFC5746. Re-enable renegotiation but require the extension
9693 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9694 turns out to be a bad idea. It has been replaced by
9695 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9696 SSL_CTX_set_options(). This is really not recommended unless you
9697 know what you are doing.
9699 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9701 * Fixes to stateless session resumption handling. Use initial_ctx when
9702 issuing and attempting to decrypt tickets in case it has changed during
9703 servername handling. Use a non-zero length session ID when attempting
9704 stateless session resumption: this makes it possible to determine if
9705 a resumption has occurred immediately after receiving server hello
9706 (several places in OpenSSL subtly assume this) instead of later in
9711 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9712 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9713 fixes for a few places where the return code is not checked
9716 *Julia Lawall <julia@diku.dk>*
9718 * Add --strict-warnings option to Configure script to include devteam
9719 warnings in other configurations.
9723 * Add support for --libdir option and LIBDIR variable in makefiles. This
9724 makes it possible to install openssl libraries in locations which
9725 have names other than "lib", for example "/usr/lib64" which some
9728 *Steve Henson, based on patch from Jeremy Utley*
9730 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9731 X690 8.9.12 and can produce some misleading textual output of OIDs.
9733 *Steve Henson, reported by Dan Kaminsky*
9735 * Delete MD2 from algorithm tables. This follows the recommendation in
9736 several standards that it is not used in new applications due to
9737 several cryptographic weaknesses. For binary compatibility reasons
9738 the MD2 API is still compiled in by default.
9742 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9747 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9748 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9751 *Guenter <lists@gknw.net>*
9753 * Fix the server certificate chain building code to use X509_verify_cert(),
9754 it used to have an ad-hoc builder which was unable to cope with anything
9755 other than a simple chain.
9757 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9759 * Don't check self signed certificate signatures in X509_verify_cert()
9760 by default (a flag can override this): it just wastes time without
9761 adding any security. As a useful side effect self signed root CAs
9762 with non-FIPS digests are now usable in FIPS mode.
9766 * In dtls1_process_out_of_seq_message() the check if the current message
9767 is already buffered was missing. For every new message was memory
9768 allocated, allowing an attacker to perform an denial of service attack
9769 with sending out of seq handshake messages until there is no memory
9770 left. Additionally every future message was buffered, even if the
9771 sequence number made no sense and would be part of another handshake.
9772 So only messages with sequence numbers less than 10 in advance will be
9773 buffered. ([CVE-2009-1378])
9775 *Robin Seggelmann, discovered by Daniel Mentz*
9777 * Records are buffered if they arrive with a future epoch to be
9778 processed after finishing the corresponding handshake. There is
9779 currently no limitation to this buffer allowing an attacker to perform
9780 a DOS attack with sending records with future epochs until there is no
9781 memory left. This patch adds the pqueue_size() function to determine
9782 the size of a buffer and limits the record buffer to 100 entries.
9785 *Robin Seggelmann, discovered by Daniel Mentz*
9787 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9788 parent structure is freed. ([CVE-2009-1379])
9792 * Handle non-blocking I/O properly in SSL_shutdown() call.
9794 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9796 * Add `2.5.4.*` OIDs
9798 *Ilya O. <vrghost@gmail.com>*
9800 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9802 * Disable renegotiation completely - this fixes a severe security
9803 problem ([CVE-2009-3555]) at the cost of breaking all
9804 renegotiation. Renegotiation can be re-enabled by setting
9805 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9806 run-time. This is really not recommended unless you know what
9811 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9813 * Don't set val to NULL when freeing up structures, it is freed up by
9814 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9815 zeroing past the valid field. ([CVE-2009-0789])
9817 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9819 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9820 checked correctly. This would allow some invalid signed attributes to
9821 appear to verify correctly. ([CVE-2009-0591])
9823 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9825 * Reject UniversalString and BMPString types with invalid lengths. This
9826 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9827 a legal length. ([CVE-2009-0590])
9831 * Set S/MIME signing as the default purpose rather than setting it
9832 unconditionally. This allows applications to override it at the store
9837 * Permit restricted recursion of ASN1 strings. This is needed in practice
9838 to handle some structures.
9842 * Improve efficiency of mem_gets: don't search whole buffer each time
9845 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9847 * New -hex option for openssl rand.
9851 * Print out UTF8String and NumericString when parsing ASN1.
9855 * Support NumericString type for name components.
9859 * Allow CC in the environment to override the automatically chosen
9860 compiler. Note that nothing is done to ensure flags work with the
9865 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9867 * Properly check EVP_VerifyFinal() and similar return values
9870 *Ben Laurie, Bodo Moeller, Google Security Team*
9872 * Enable TLS extensions by default.
9876 * Allow the CHIL engine to be loaded, whether the application is
9877 multithreaded or not. (This does not release the developer from the
9878 obligation to set up the dynamic locking callbacks.)
9880 *Sander Temme <sander@temme.net>*
9882 * Use correct exit code if there is an error in dgst command.
9884 *Steve Henson; problem pointed out by Roland Dirlewanger*
9886 * Tweak Configure so that you need to say "experimental-jpake" to enable
9887 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9891 * Add experimental JPAKE support, including demo authentication in
9892 s_client and s_server.
9896 * Set the comparison function in v3_addr_canonize().
9898 *Rob Austein <sra@hactrn.net>*
9900 * Add support for XMPP STARTTLS in s_client.
9902 *Philip Paeps <philip@freebsd.org>*
9904 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9905 to ensure that even with this option, only ciphersuites in the
9906 server's preference list will be accepted. (Note that the option
9907 applies only when resuming a session, so the earlier behavior was
9908 just about the algorithm choice for symmetric cryptography.)
9912 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9914 * Fix NULL pointer dereference if a DTLS server received
9915 ChangeCipherSpec as first record ([CVE-2009-1386]).
9919 * Fix a state transition in s3_srvr.c and d1_srvr.c
9920 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9924 * The fix in 0.9.8c that supposedly got rid of unsafe
9925 double-checked locking was incomplete for RSA blinding,
9926 addressing just one layer of what turns out to have been
9927 doubly unsafe triple-checked locking.
9929 So now fix this for real by retiring the MONT_HELPER macro
9930 in crypto/rsa/rsa_eay.c.
9932 *Bodo Moeller; problem pointed out by Marius Schilder*
9934 * Various precautionary measures:
9936 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9938 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9939 (NB: This would require knowledge of the secret session ticket key
9940 to exploit, in which case you'd be SOL either way.)
9942 - Change bn_nist.c so that it will properly handle input BIGNUMs
9943 outside the expected range.
9945 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9948 *Neel Mehta, Bodo Moeller*
9950 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9951 the load fails. Useful for distros.
9953 *Ben Laurie and the FreeBSD team*
9955 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9959 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9963 * Expand ENGINE to support engine supplied SSL client certificate functions.
9965 This work was sponsored by Logica.
9969 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9970 keystores. Support for SSL/TLS client authentication too.
9971 Not compiled unless enable-capieng specified to Configure.
9973 This work was sponsored by Logica.
9977 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9978 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9979 attribute creation routines such as certificate requests and PKCS#12
9984 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9986 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9987 handshake which could lead to a client crash as found using the
9988 Codenomicon TLS test suite ([CVE-2008-1672])
9990 *Steve Henson, Mark Cox*
9992 * Fix double free in TLS server name extensions which could lead to
9993 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9997 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9999 Clear the error queue to ensure that error entries left from
10000 older function calls do not interfere with the correct operation.
10002 *Lutz Jaenicke, Erik de Castro Lopo*
10004 * Remove root CA certificates of commercial CAs:
10006 The OpenSSL project does not recommend any specific CA and does not
10007 have any policy with respect to including or excluding any CA.
10008 Therefore it does not make any sense to ship an arbitrary selection
10009 of root CA certificates with the OpenSSL software.
10013 * RSA OAEP patches to fix two separate invalid memory reads.
10014 The first one involves inputs when 'lzero' is greater than
10015 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10016 before the beginning of from). The second one involves inputs where
10017 the 'db' section contains nothing but zeroes (there is a one-byte
10018 invalid read after the end of 'db').
10020 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10022 * Partial backport from 0.9.9-dev:
10024 Introduce bn_mul_mont (dedicated Montgomery multiplication
10025 procedure) as a candidate for BIGNUM assembler implementation.
10026 While 0.9.9-dev uses assembler for various architectures, only
10027 x86_64 is available by default here in the 0.9.8 branch, and
10028 32-bit x86 is available through a compile-time setting.
10030 To try the 32-bit x86 assembler implementation, use Configure
10031 option "enable-montasm" (which exists only for this backport).
10033 As "enable-montasm" for 32-bit x86 disclaims code stability
10034 anyway, in this constellation we activate additional code
10035 backported from 0.9.9-dev for further performance improvements,
10036 namely BN_from_montgomery_word. (To enable this otherwise,
10037 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10039 *Andy Polyakov (backport partially by Bodo Moeller)*
10041 * Add TLS session ticket callback. This allows an application to set
10042 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10043 values. This is useful for key rollover for example where several key
10044 sets may exist with different names.
10048 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10049 This was broken until now in 0.9.8 releases, such that the only way
10050 a registered ENGINE could be used (assuming it initialises
10051 successfully on the host) was to explicitly set it as the default
10052 for the relevant algorithms. This is in contradiction with 0.9.7
10053 behaviour and the documentation. With this fix, when an ENGINE is
10054 registered into a given algorithm's table of implementations, the
10055 'uptodate' flag is reset so that auto-discovery will be used next
10056 time a new context for that algorithm attempts to select an
10059 *Ian Lister (tweaked by Geoff Thorpe)*
10061 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10062 implementation in the following ways:
10064 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10067 Lack of BER streaming support means one pass streaming processing is
10068 only supported if data is detached: setting the streaming flag is
10069 ignored for embedded content.
10071 CMS support is disabled by default and must be explicitly enabled
10072 with the enable-cms configuration option.
10076 * Update the GMP engine glue to do direct copies between BIGNUM and
10077 mpz_t when openssl and GMP use the same limb size. Otherwise the
10078 existing "conversion via a text string export" trick is still used.
10080 *Paul Sheer <paulsheer@gmail.com>*
10082 * Zlib compression BIO. This is a filter BIO which compressed and
10083 uncompresses any data passed through it.
10087 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10088 RFC3394 compatible AES key wrapping.
10092 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10093 sets string data without copying. X509_ALGOR_set0() and
10094 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10095 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10096 from an X509_ATTRIBUTE structure optionally checking it occurs only
10097 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10102 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10103 to get the expected BN_FLG_CONSTTIME behavior.
10105 *Bodo Moeller (Google)*
10109 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10110 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10111 - added some more tests to do_tests.pl
10112 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10113 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10114 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10115 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10116 - various changes to netware.pl to enable gcc-cross builds on Win32
10118 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10119 - various changes to fix missing prototype warnings
10120 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10121 - added AES, WHIRLPOOL and CPUID assembler code to build files
10122 - added missing AES assembler make rules to mk1mf.pl
10123 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10125 *Guenter Knauf <eflash@gmx.net>*
10127 * Implement certificate status request TLS extension defined in RFC3546.
10128 A client can set the appropriate parameters and receive the encoded
10129 OCSP response via a callback. A server can query the supplied parameters
10130 and set the encoded OCSP response in the callback. Add simplified examples
10131 to s_client and s_server.
10135 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10137 * Fix various bugs:
10138 + Binary incompatibility of ssl_ctx_st structure
10139 + DTLS interoperation with non-compliant servers
10140 + Don't call get_session_cb() without proposed session
10141 + Fix ia64 assembler code
10143 *Andy Polyakov, Steve Henson*
10145 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10147 * DTLS Handshake overhaul. There were longstanding issues with
10148 OpenSSL DTLS implementation, which were making it impossible for
10149 RFC 4347 compliant client to communicate with OpenSSL server.
10150 Unfortunately just fixing these incompatibilities would "cut off"
10151 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10152 server keeps tolerating non RFC compliant syntax. The opposite is
10153 not true, 0.9.8f client can not communicate with earlier server.
10154 This update even addresses CVE-2007-4995.
10158 * Changes to avoid need for function casts in OpenSSL: some compilers
10159 (gcc 4.2 and later) reject their use.
10160 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10163 * Add RFC4507 support to OpenSSL. This includes the corrections in
10164 RFC4507bis. The encrypted ticket format is an encrypted encoded
10165 SSL_SESSION structure, that way new session features are automatically
10168 If a client application caches session in an SSL_SESSION structure
10169 support is transparent because tickets are now stored in the encoded
10172 The SSL_CTX structure automatically generates keys for ticket
10173 protection in servers so again support should be possible
10174 with no application modification.
10176 If a client or server wishes to disable RFC4507 support then the option
10177 SSL_OP_NO_TICKET can be set.
10179 Add a TLS extension debugging callback to allow the contents of any client
10180 or server extensions to be examined.
10182 This work was sponsored by Google.
10186 * Add initial support for TLS extensions, specifically for the server_name
10187 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10188 have new members for a host name. The SSL data structure has an
10189 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10190 stored in that context to allow for session resumption, even after the
10191 SSL has been switched to a new SSL_CTX in reaction to a client's
10192 server_name extension.
10194 New functions (subject to change):
10196 SSL_get_servername()
10197 SSL_get_servername_type()
10200 New CTRL codes and macros (subject to change):
10202 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10203 - SSL_CTX_set_tlsext_servername_callback()
10204 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10205 - SSL_CTX_set_tlsext_servername_arg()
10206 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10208 openssl s_client has a new '-servername ...' option.
10210 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10211 '-key2 ...', '-servername_fatal' (subject to change). This allows
10212 testing the HostName extension for a specific single host name ('-cert'
10213 and '-key' remain fallbacks for handshakes without HostName
10214 negotiation). If the unrecognized_name alert has to be sent, this by
10215 default is a warning; it becomes fatal with the '-servername_fatal'
10218 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10220 * Add AES and SSE2 assembly language support to VC++ build.
10224 * Mitigate attack on final subtraction in Montgomery reduction.
10228 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10229 (which previously caused an internal error).
10233 * Squeeze another 10% out of IGE mode when in != out.
10237 * AES IGE mode speedup.
10239 *Dean Gaudet (Google)*
10241 * Add the Korean symmetric 128-bit cipher SEED (see
10242 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10243 add SEED ciphersuites from RFC 4162:
10245 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10246 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10247 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10248 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10250 To minimize changes between patchlevels in the OpenSSL 0.9.8
10251 series, SEED remains excluded from compilation unless OpenSSL
10252 is configured with 'enable-seed'.
10254 *KISA, Bodo Moeller*
10256 * Mitigate branch prediction attacks, which can be practical if a
10257 single processor is shared, allowing a spy process to extract
10258 information. For detailed background information, see
10259 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10260 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10261 and Necessary Software Countermeasures"). The core of the change
10262 are new versions BN_div_no_branch() and
10263 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10264 respectively, which are slower, but avoid the security-relevant
10265 conditional branches. These are automatically called by BN_div()
10266 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10267 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10268 remove a conditional branch.
10270 BN_FLG_CONSTTIME is the new name for the previous
10271 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10272 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10273 in the exponent causes BN_mod_exp_mont() to use the alternative
10274 implementation in BN_mod_exp_mont_consttime().) The old name
10275 remains as a deprecated alias.
10277 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10278 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10279 constant-time implementations for more than just exponentiation.
10280 Here too the old name is kept as a deprecated alias.
10282 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10283 the BN_BLINDING structure gets an independent copy of the
10284 modulus. This means that the previous `BIGNUM *m` argument to
10285 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10286 essentially becomes `const BIGNUM *m`, although we can't actually
10287 change this in the header file before 0.9.9. It allows
10288 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10289 enable BN_FLG_CONSTTIME.
10291 *Matthew D Wood (Intel Corp)*
10293 * In the SSL/TLS server implementation, be strict about session ID
10294 context matching (which matters if an application uses a single
10295 external cache for different purposes). Previously,
10296 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10297 set. This did ensure strict client verification, but meant that,
10298 with applications using a single external cache for quite
10299 different requirements, clients could circumvent ciphersuite
10300 restrictions for a given session ID context by starting a session
10301 in a different context.
10305 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10306 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10307 authentication-only ciphersuites.
10311 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10312 not complete and could lead to a possible single byte overflow
10313 ([CVE-2007-5135]) [Ben Laurie]
10315 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10317 * Since AES128 and AES256 (and similarly Camellia128 and
10318 Camellia256) share a single mask bit in the logic of
10319 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10320 kludge to work properly if AES128 is available and AES256 isn't
10321 (or if Camellia128 is available and Camellia256 isn't).
10325 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10326 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10327 When a point or a seed is encoded in a BIT STRING, we need to
10328 prevent the removal of trailing zero bits to get the proper DER
10329 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10330 of a NamedBitList, for which trailing 0 bits need to be removed.)
10334 * Have SSL/TLS server implementation tolerate "mismatched" record
10335 protocol version while receiving ClientHello even if the
10336 ClientHello is fragmented. (The server can't insist on the
10337 particular protocol version it has chosen before the ServerHello
10338 message has informed the client about his choice.)
10342 * Add RFC 3779 support.
10344 *Rob Austein for ARIN, Ben Laurie*
10346 * Load error codes if they are not already present instead of using a
10347 static variable. This allows them to be cleanly unloaded and reloaded.
10348 Improve header file function name parsing.
10352 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10353 or CAPABILITY handshake as required by RFCs.
10355 *Goetz Babin-Ebell*
10357 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10359 * Introduce limits to prevent malicious keys being able to
10360 cause a denial of service. ([CVE-2006-2940])
10362 *Steve Henson, Bodo Moeller*
10364 * Fix ASN.1 parsing of certain invalid structures that can result
10365 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10367 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10368 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10370 * Fix SSL client code which could crash if connecting to a
10371 malicious SSLv2 server. ([CVE-2006-4343])
10373 *Tavis Ormandy and Will Drewry, Google Security Team*
10375 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10376 match only those. Before that, "AES256-SHA" would be interpreted
10377 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10378 the same strength classification in 0.9.7h) as we currently only
10379 have a single AES bit in the ciphersuite description bitmap.
10380 That change, however, also applied to ciphersuite strings such as
10381 "RC4-MD5" that intentionally matched multiple ciphersuites --
10382 namely, SSL 2.0 ciphersuites in addition to the more common ones
10383 from SSL 3.0/TLS 1.0.
10385 So we change the selection algorithm again: Naming an explicit
10386 ciphersuite selects this one ciphersuite, and any other similar
10387 ciphersuite (same bitmap) from *other* protocol versions.
10388 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10389 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10391 Since SSL 2.0 does not have any ciphersuites for which the
10392 128/256 bit distinction would be relevant, this works for now.
10393 The proper fix will be to use different bits for AES128 and
10394 AES256, which would have avoided the problems from the beginning;
10395 however, bits are scarce, so we can only do this in a new release
10396 (not just a patchlevel) when we can change the SSL_CIPHER
10397 definition to split the single 'unsigned long mask' bitmap into
10398 multiple values to extend the available space.
10402 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10404 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10405 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10407 * Add AES IGE and biIGE modes.
10411 * Change the Unix randomness entropy gathering to use poll() when
10412 possible instead of select(), since the latter has some
10413 undesirable limitations.
10415 *Darryl Miles via Richard Levitte and Bodo Moeller*
10417 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10418 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10419 cannot be implicitly activated as part of, e.g., the "AES" alias.
10420 However, please upgrade to OpenSSL 0.9.9[-dev] for
10421 non-experimental use of the ECC ciphersuites to get TLS extension
10422 support, which is required for curve and point format negotiation
10423 to avoid potential handshake problems.
10427 * Disable rogue ciphersuites:
10429 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10430 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10431 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10433 The latter two were purportedly from
10434 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10437 Also deactivate the remaining ciphersuites from
10438 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10439 unofficial, and the ID has long expired.
10443 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10444 dual-core machines) and other potential thread-safety issues.
10448 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10449 versions), which is now available for royalty-free use
10450 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10451 Also, add Camellia TLS ciphersuites from RFC 4132.
10453 To minimize changes between patchlevels in the OpenSSL 0.9.8
10454 series, Camellia remains excluded from compilation unless OpenSSL
10455 is configured with 'enable-camellia'.
10459 * Disable the padding bug check when compression is in use. The padding
10460 bug check assumes the first packet is of even length, this is not
10461 necessarily true if compression is enabled and can result in false
10462 positives causing handshake failure. The actual bug test is ancient
10463 code so it is hoped that implementations will either have fixed it by
10464 now or any which still have the bug do not support compression.
10468 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10470 * When applying a cipher rule check to see if string match is an explicit
10471 cipher suite and only match that one cipher suite if it is.
10475 * Link in manifests for VC++ if needed.
10477 *Austin Ziegler <halostatue@gmail.com>*
10479 * Update support for ECC-based TLS ciphersuites according to
10480 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10481 TLS extensions, which are supported starting with the 0.9.9
10482 branch, not in the OpenSSL 0.9.8 branch).
10486 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10487 opaque EVP_CIPHER_CTX handling.
10491 * Fixes and enhancements to zlib compression code. We now only use
10492 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10493 to conform with the standards mentioned here:
10494 <http://www.zlib.net/DLL_FAQ.txt>
10495 Static zlib linking now works on Windows and the new --with-zlib-include
10496 --with-zlib-lib options to Configure can be used to supply the location
10497 of the headers and library. Gracefully handle case where zlib library
10502 * Several fixes and enhancements to the OID generation code. The old code
10503 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10504 handle numbers larger than ULONG_MAX, truncated printing and had a
10505 non standard OBJ_obj2txt() behaviour.
10509 * Add support for building of engines under engine/ as shared libraries
10510 under VC++ build system.
10514 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10515 Hopefully, we will not see any false combination of paths any more.
10519 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10521 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10522 (part of SSL_OP_ALL). This option used to disable the
10523 countermeasure against man-in-the-middle protocol-version
10524 rollback in the SSL 2.0 server implementation, which is a bad
10525 idea. ([CVE-2005-2969])
10527 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10528 for Information Security, National Institute of Advanced Industrial
10529 Science and Technology [AIST], Japan)*
10531 * Add two function to clear and return the verify parameter flags.
10535 * Keep cipherlists sorted in the source instead of sorting them at
10536 runtime, thus removing the need for a lock.
10540 * Avoid some small subgroup attacks in Diffie-Hellman.
10542 *Nick Mathewson and Ben Laurie*
10544 * Add functions for well-known primes.
10548 * Extended Windows CE support.
10550 *Satoshi Nakamura and Andy Polyakov*
10552 * Initialize SSL_METHOD structures at compile time instead of during
10553 runtime, thus removing the need for a lock.
10557 * Make PKCS7_decrypt() work even if no certificate is supplied by
10558 attempting to decrypt each encrypted key in turn. Add support to
10563 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10565 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10568 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10572 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10573 key into the same file any more.
10577 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10581 * Add -utf8 command line and config file option to 'ca'.
10583 *Stefan <stf@udoma.org*
10585 * Removed the macro des_crypt(), as it seems to conflict with some
10586 libraries. Use DES_crypt().
10590 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10591 involves renaming the source and generated shared-libs for
10592 both. The engines will accept the corrected or legacy ids
10593 ('ncipher' and '4758_cca' respectively) when binding. NB,
10594 this only applies when building 'shared'.
10596 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10598 * Add attribute functions to EVP_PKEY structure. Modify
10599 PKCS12_create() to recognize a CSP name attribute and
10600 use it. Make -CSP option work again in pkcs12 utility.
10604 * Add new functionality to the bn blinding code:
10605 - automatic re-creation of the BN_BLINDING parameters after
10606 a fixed number of uses (currently 32)
10607 - add new function for parameter creation
10608 - introduce flags to control the update behaviour of the
10609 BN_BLINDING parameters
10610 - hide BN_BLINDING structure
10611 Add a second BN_BLINDING slot to the RSA structure to improve
10612 performance when a single RSA object is shared among several
10617 * Add support for DTLS.
10619 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10621 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10622 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10626 * Remove buggy and incomplete DH cert support from
10627 ssl/ssl_rsa.c and ssl/s3_both.c
10631 * Use SHA-1 instead of MD5 as the default digest algorithm for
10632 the `apps/openssl` commands.
10636 * Compile clean with "-Wall -Wmissing-prototypes
10637 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10638 DEBUG_SAFESTACK must also be set.
10642 * Change ./Configure so that certain algorithms can be disabled by default.
10643 The new counterpiece to "no-xxx" is "enable-xxx".
10645 The patented RC5 and MDC2 algorithms will now be disabled unless
10646 "enable-rc5" and "enable-mdc2", respectively, are specified.
10648 (IDEA remains enabled despite being patented. This is because IDEA
10649 is frequently required for interoperability, and there is no license
10650 fee for non-commercial use. As before, "no-idea" can be used to
10651 avoid this algorithm.)
10655 * Add processing of proxy certificates (see RFC 3820). This work was
10656 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10657 EGEE (Enabling Grids for E-science in Europe).
10661 * RC4 performance overhaul on modern architectures/implementations, such
10662 as Intel P4, IA-64 and AMD64.
10666 * New utility extract-section.pl. This can be used specify an alternative
10667 section number in a pod file instead of having to treat each file as
10668 a separate case in Makefile. This can be done by adding two lines to the
10671 =for comment openssl_section:XXX
10673 The blank line is mandatory.
10677 * New arguments -certform, -keyform and -pass for s_client and s_server
10678 to allow alternative format key and certificate files and passphrase
10683 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10684 update associated structures and add various utility functions.
10686 Add new policy related verify parameters, include policy checking in
10687 standard verify code. Enhance 'smime' application with extra parameters
10688 to support policy checking and print out.
10692 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10693 Nehemiah processors. These extensions support AES encryption in hardware
10694 as well as RNG (though RNG support is currently disabled).
10696 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10698 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10702 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10704 *Andy Polyakov and a number of other people*
10706 * Improved PowerPC platform support. Most notably BIGNUM assembler
10707 implementation contributed by IBM.
10709 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10711 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10712 exponent rather than 'unsigned long'. There is a corresponding change to
10713 the new 'rsa_keygen' element of the RSA_METHOD structure.
10715 *Jelte Jansen, Geoff Thorpe*
10717 * Functionality for creating the initial serial number file is now
10718 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10720 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10721 number file to 1, which is bound to cause problems. To avoid
10722 the problems while respecting compatibility between different 0.9.7
10723 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10724 CA.pl for serial number initialization. With the new release 0.9.8,
10725 we can fix the problem directly in the 'ca' utility.)
10729 * Reduced header interdependencies by declaring more opaque objects in
10730 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10731 give fewer recursive includes, which could break lazy source code - so
10732 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10733 developers should define this symbol when building and using openssl to
10734 ensure they track the recommended behaviour, interfaces, [etc], but
10735 backwards-compatible behaviour prevails when this isn't defined.
10739 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10743 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10744 This will generate a random key of the appropriate length based on the
10745 cipher context. The EVP_CIPHER can provide its own random key generation
10746 routine to support keys of a specific form. This is used in the des and
10747 3des routines to generate a key of the correct parity. Update S/MIME
10748 code to use new functions and hence generate correct parity DES keys.
10749 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10750 valid (weak or incorrect parity).
10754 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10755 as looking them up. This is useful when the verified structure may contain
10756 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10757 present unless the new PKCS7_NO_CRL flag is asserted.
10761 * Extend ASN1 oid configuration module. It now additionally accepts the
10764 shortName = some long name, 1.2.3.4
10768 * Reimplemented the BN_CTX implementation. There is now no more static
10769 limitation on the number of variables it can handle nor the depth of the
10770 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10771 information can now expand as required, and rather than having a single
10772 static array of bignums, BN_CTX now uses a linked-list of such arrays
10773 allowing it to expand on demand whilst maintaining the usefulness of
10774 BN_CTX's "bundling".
10778 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10779 to allow all RSA operations to function using a single BN_CTX.
10783 * Preliminary support for certificate policy evaluation and checking. This
10784 is initially intended to pass the tests outlined in "Conformance Testing
10785 of Relying Party Client Certificate Path Processing Logic" v1.07.
10789 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10790 remained unused and not that useful. A variety of other little bignum
10791 tweaks and fixes have also been made continuing on from the audit (see
10796 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10797 associated ASN1, EVP and SSL functions and old ASN1 macros.
10801 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10802 and this should never fail. So the return value from the use of
10803 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10804 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10808 * BN_CTX_get() should return zero-valued bignums, providing the same
10809 initialised value as BN_new().
10811 *Geoff Thorpe, suggested by Ulf Möller*
10813 * Support for inhibitAnyPolicy certificate extension.
10817 * An audit of the BIGNUM code is underway, for which debugging code is
10818 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10819 is considered valid when processing BIGNUMs, and causes execution to
10820 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10821 further steps are taken to deliberately pollute unused data in BIGNUM
10822 structures to try and expose faulty code further on. For now, openssl will
10823 (in its default mode of operation) continue to tolerate the inconsistent
10824 forms that it has tolerated in the past, but authors and packagers should
10825 consider trying openssl and their own applications when compiled with
10826 these debugging symbols defined. It will help highlight potential bugs in
10827 their own code, and will improve the test coverage for OpenSSL itself. At
10828 some point, these tighter rules will become openssl's default to improve
10829 maintainability, though the assert()s and other overheads will remain only
10830 in debugging configurations. See bn.h for more details.
10832 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10834 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10835 that can only be obtained through BN_CTX_new() (which implicitly
10836 initialises it). The presence of this function only made it possible
10837 to overwrite an existing structure (and cause memory leaks).
10841 * Because of the callback-based approach for implementing LHASH as a
10842 template type, lh_insert() adds opaque objects to hash-tables and
10843 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10844 to clean up those corresponding objects before destroying the hash table
10845 (and losing the object pointers). So some over-zealous constifications in
10846 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10847 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10848 prototyped to have "const" restrictions on the object pointers they are
10849 given (and so aren't required to cast them away any more).
10853 * The tmdiff.h API was so ugly and minimal that our own timing utility
10854 (speed) prefers to use its own implementation. The two implementations
10855 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10856 its object type properly exposed (MS_TM) instead of casting to/from
10857 `char *`. This may still change yet if someone realises MS_TM and
10859 aren't necessarily the greatest nomenclatures - but this is what was used
10860 internally to the implementation so I've used that for now.
10864 * Ensure that deprecated functions do not get compiled when
10865 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10866 the self-tests were still using deprecated key-generation functions so
10867 these have been updated also.
10871 * Reorganise PKCS#7 code to separate the digest location functionality
10872 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10873 New function PKCS7_set_digest() to set the digest type for PKCS#7
10874 digestedData type. Add additional code to correctly generate the
10875 digestedData type and add support for this type in PKCS7 initialization
10880 * New function PKCS7_set0_type_other() this initializes a PKCS7
10881 structure of type "other".
10885 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10886 sure the loop does correctly stop and breaking ("division by zero")
10887 modulus operations are not performed. The (pre-generated) prime
10888 table crypto/bn/bn_prime.h was already correct, but it could not be
10889 re-generated on some platforms because of the "division by zero"
10890 situation in the script.
10892 *Ralf S. Engelschall*
10894 * Update support for ECC-based TLS ciphersuites according to
10895 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10896 SHA-1 now is only used for "small" curves (where the
10897 representation of a field element takes up to 24 bytes); for
10898 larger curves, the field element resulting from ECDH is directly
10899 used as premaster secret.
10901 *Douglas Stebila (Sun Microsystems Laboratories)*
10903 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10904 curve secp160r1 to the tests.
10906 *Douglas Stebila (Sun Microsystems Laboratories)*
10908 * Add the possibility to load symbols globally with DSO.
10910 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10912 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10913 control of the error stack.
10917 * Add support for STORE in ENGINE.
10921 * Add the STORE type. The intention is to provide a common interface
10922 to certificate and key stores, be they simple file-based stores, or
10923 HSM-type store, or LDAP stores, or...
10924 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10928 * Add a generic structure called OPENSSL_ITEM. This can be used to
10929 pass a list of arguments to any function as well as provide a way
10930 for a function to pass data back to the caller.
10934 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10935 works like BUF_strdup() but can be used to duplicate a portion of
10936 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10941 * Add the function sk_find_ex() which works like sk_find(), but will
10942 return an index to an element even if an exact match couldn't be
10943 found. The index is guaranteed to point at the element where the
10944 searched-for key would be inserted to preserve sorting order.
10948 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10949 takes an extra flags argument for optional functionality. Currently,
10950 the following flags are defined:
10952 OBJ_BSEARCH_VALUE_ON_NOMATCH
10953 This one gets OBJ_bsearch_ex() to return a pointer to the first
10954 element where the comparing function returns a negative or zero
10957 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10958 This one gets OBJ_bsearch_ex() to return a pointer to the first
10959 element where the comparing function returns zero. This is useful
10960 if there are more than one element where the comparing function
10965 * Make it possible to create self-signed certificates with 'openssl ca'
10966 in such a way that the self-signed certificate becomes part of the
10967 CA database and uses the same mechanisms for serial number generation
10968 as all other certificate signing. The new flag '-selfsign' enables
10969 this functionality. Adapt CA.sh and CA.pl.in.
10973 * Add functionality to check the public key of a certificate request
10974 against a given private. This is useful to check that a certificate
10975 request can be signed by that key (self-signing).
10979 * Make it possible to have multiple active certificates with the same
10980 subject in the CA index file. This is done only if the keyword
10981 'unique_subject' is set to 'no' in the main CA section (default
10982 if 'CA_default') of the configuration file. The value is saved
10983 with the database itself in a separate index attribute file,
10984 named like the index file with '.attr' appended to the name.
10988 * Generate multi-valued AVAs using '+' notation in config files for
10993 * Support for nameConstraints certificate extension.
10997 * Support for policyConstraints certificate extension.
11001 * Support for policyMappings certificate extension.
11005 * Make sure the default DSA_METHOD implementation only uses its
11006 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11007 and change its own handlers to be NULL so as to remove unnecessary
11008 indirection. This lets alternative implementations fallback to the
11009 default implementation more easily.
11013 * Support for directoryName in GeneralName related extensions
11018 * Make it possible to link applications using Makefile.shared.
11019 Make that possible even when linking against static libraries!
11023 * Support for single pass processing for S/MIME signing. This now
11024 means that S/MIME signing can be done from a pipe, in addition
11025 cleartext signing (multipart/signed type) is effectively streaming
11026 and the signed data does not need to be all held in memory.
11028 This is done with a new flag PKCS7_STREAM. When this flag is set
11029 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11030 is done after the data is output (and digests calculated) in
11031 SMIME_write_PKCS7().
11035 * Add full support for -rpath/-R, both in shared libraries and
11036 applications, at least on the platforms where it's known how
11041 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11042 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11043 will now compute a table of multiples of the generator that
11044 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11045 faster (notably in the case of a single point multiplication,
11046 scalar * generator).
11048 *Nils Larsch, Bodo Moeller*
11050 * IPv6 support for certificate extensions. The various extensions
11051 which use the IP:a.b.c.d can now take IPv6 addresses using the
11052 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11057 * Added an ENGINE that implements RSA by performing private key
11058 exponentiations with the GMP library. The conversions to and from
11059 GMP's mpz_t format aren't optimised nor are any montgomery forms
11060 cached, and on x86 it appears OpenSSL's own performance has caught up.
11061 However there are likely to be other architectures where GMP could
11062 provide a boost. This ENGINE is not built in by default, but it can be
11063 specified at Configure time and should be accompanied by the necessary
11064 linker additions, eg;
11065 ./config -DOPENSSL_USE_GMP -lgmp
11069 * "openssl engine" will not display ENGINE/DSO load failure errors when
11070 testing availability of engines with "-t" - the old behaviour is
11071 produced by increasing the feature's verbosity with "-tt".
11075 * ECDSA routines: under certain error conditions uninitialized BN objects
11076 could be freed. Solution: make sure initialization is performed early
11077 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11082 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11083 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11084 software implementations. For DSA and DH, parameter generation can
11085 also be overridden by providing the appropriate method callbacks.
11089 * Change the "progress" mechanism used in key-generation and
11090 primality testing to functions that take a new BN_GENCB pointer in
11091 place of callback/argument pairs. The new API functions have `_ex`
11092 postfixes and the older functions are reimplemented as wrappers for
11093 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11094 declarations of the old functions to help (graceful) attempts to
11095 migrate to the new functions. Also, the new key-generation API
11096 functions operate on a caller-supplied key-structure and return
11097 success/failure rather than returning a key or NULL - this is to
11098 help make "keygen" another member function of RSA_METHOD etc.
11100 Example for using the new callback interface:
11102 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11103 void *my_arg = ...;
11106 BN_GENCB_set(&my_cb, my_callback, my_arg);
11108 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11109 /* For the meaning of a, b in calls to my_callback(), see the
11110 * documentation of the function that calls the callback.
11111 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11112 * my_callback should return 1 if it wants BN_is_prime_ex()
11113 * to continue, or 0 to stop.
11118 * Change the ZLIB compression method to be stateful, and make it
11119 available to TLS with the number defined in
11120 draft-ietf-tls-compression-04.txt.
11124 * Add the ASN.1 structures and functions for CertificatePair, which
11125 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11127 CertificatePair ::= SEQUENCE {
11128 forward [0] Certificate OPTIONAL,
11129 reverse [1] Certificate OPTIONAL,
11130 -- at least one of the pair shall be present -- }
11132 Also implement the PEM functions to read and write certificate
11133 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11135 This needed to be defined, mostly for the sake of the LDAP
11136 attribute crossCertificatePair, but may prove useful elsewhere as
11141 * Make it possible to inhibit symlinking of shared libraries in
11142 Makefile.shared, for Cygwin's sake.
11146 * Extend the BIGNUM API by creating a function
11147 void BN_set_negative(BIGNUM *a, int neg);
11148 and a macro that behave like
11149 int BN_is_negative(const BIGNUM *a);
11151 to avoid the need to access 'a->neg' directly in applications.
11155 * Implement fast modular reduction for pseudo-Mersenne primes
11156 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11157 EC_GROUP_new_curve_GFp() will now automatically use this
11160 *Nils Larsch <nla@trustcenter.de>*
11162 * Add new lock type (CRYPTO_LOCK_BN).
11166 * Change the ENGINE framework to automatically load engines
11167 dynamically from specific directories unless they could be
11168 found to already be built in or loaded. Move all the
11169 current engines except for the cryptodev one to a new
11170 directory engines/.
11171 The engines in engines/ are built as shared libraries if
11172 the "shared" options was given to ./Configure or ./config.
11173 Otherwise, they are inserted in libcrypto.a.
11174 /usr/local/ssl/engines is the default directory for dynamic
11175 engines, but that can be overridden at configure time through
11176 the usual use of --prefix and/or --openssldir, and at run
11177 time with the environment variable OPENSSL_ENGINES.
11179 *Geoff Thorpe and Richard Levitte*
11181 * Add Makefile.shared, a helper makefile to build shared
11182 libraries. Adapt Makefile.org.
11186 * Add version info to Win32 DLLs.
11188 *Peter 'Luna' Runestig" <peter@runestig.com>*
11190 * Add new 'medium level' PKCS#12 API. Certificates and keys
11191 can be added using this API to created arbitrary PKCS#12
11192 files while avoiding the low-level API.
11194 New options to PKCS12_create(), key or cert can be NULL and
11195 will then be omitted from the output file. The encryption
11196 algorithm NIDs can be set to -1 for no encryption, the mac
11197 iteration count can be set to 0 to omit the mac.
11199 Enhance pkcs12 utility by making the -nokeys and -nocerts
11200 options work when creating a PKCS#12 file. New option -nomac
11201 to omit the mac, NONE can be set for an encryption algorithm.
11202 New code is modified to use the enhanced PKCS12_create()
11203 instead of the low-level API.
11207 * Extend ASN1 encoder to support indefinite length constructed
11208 encoding. This can output sequences tags and octet strings in
11209 this form. Modify pk7_asn1.c to support indefinite length
11210 encoding. This is experimental and needs additional code to
11211 be useful, such as an ASN1 bio and some enhanced streaming
11214 Extend template encode functionality so that tagging is passed
11215 down to the template encoder.
11219 * Let 'openssl req' fail if an argument to '-newkey' is not
11220 recognized instead of using RSA as a default.
11224 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11225 As these are not official, they are not included in "ALL";
11226 the "ECCdraft" ciphersuite group alias can be used to select them.
11228 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11230 * Add ECDH engine support.
11232 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11234 * Add ECDH in new directory crypto/ecdh/.
11236 *Douglas Stebila (Sun Microsystems Laboratories)*
11238 * Let BN_rand_range() abort with an error after 100 iterations
11239 without success (which indicates a broken PRNG).
11243 * Change BN_mod_sqrt() so that it verifies that the input value
11244 is really the square of the return value. (Previously,
11245 BN_mod_sqrt would show GIGO behaviour.)
11249 * Add named elliptic curves over binary fields from X9.62, SECG,
11250 and WAP/WTLS; add OIDs that were still missing.
11252 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11254 * Extend the EC library for elliptic curves over binary fields
11255 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11258 EC_GF2m_simple_method
11262 EC_GROUP_new_curve_GF2m
11263 EC_GROUP_set_curve_GF2m
11264 EC_GROUP_get_curve_GF2m
11265 EC_POINT_set_affine_coordinates_GF2m
11266 EC_POINT_get_affine_coordinates_GF2m
11267 EC_POINT_set_compressed_coordinates_GF2m
11269 Point compression for binary fields is disabled by default for
11270 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11273 As binary polynomials are represented as BIGNUMs, various members
11274 of the EC_GROUP and EC_POINT data structures can be shared
11275 between the implementations for prime fields and binary fields;
11276 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11277 are essentially identical to their `..._GFp` counterparts.
11278 (For simplicity, the `..._GFp` prefix has been dropped from
11279 various internal method names.)
11281 An internal 'field_div' method (similar to 'field_mul' and
11282 'field_sqr') has been added; this is used only for binary fields.
11284 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11286 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11287 through methods ('mul', 'precompute_mult').
11289 The generic implementations (now internally called 'ec_wNAF_mul'
11290 and 'ec_wNAF_precomputed_mult') remain the default if these
11291 methods are undefined.
11293 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11295 * New function EC_GROUP_get_degree, which is defined through
11296 EC_METHOD. For curves over prime fields, this returns the bit
11297 length of the modulus.
11299 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11301 * New functions EC_GROUP_dup, EC_POINT_dup.
11302 (These simply call ..._new and ..._copy).
11304 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11306 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11307 Polynomials are represented as BIGNUMs (where the sign bit is not
11308 used) in the following functions [macros]:
11311 BN_GF2m_sub [= BN_GF2m_add]
11312 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11313 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11314 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11316 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11317 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11318 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11319 BN_GF2m_cmp [= BN_ucmp]
11321 (Note that only the 'mod' functions are actually for fields GF(2^m).
11322 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11324 For some functions, an the irreducible polynomial defining a
11325 field can be given as an 'unsigned int[]' with strictly
11326 decreasing elements giving the indices of those bits that are set;
11327 i.e., p[] represents the polynomial
11328 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11330 p[0] > p[1] > ... > p[k] = 0.
11331 This applies to the following functions:
11334 BN_GF2m_mod_mul_arr
11335 BN_GF2m_mod_sqr_arr
11336 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11337 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11338 BN_GF2m_mod_exp_arr
11339 BN_GF2m_mod_sqrt_arr
11340 BN_GF2m_mod_solve_quad_arr
11344 Conversion can be performed by the following functions:
11349 bntest.c has additional tests for binary polynomial arithmetic.
11351 Two implementations for BN_GF2m_mod_div() are available.
11352 The default algorithm simply uses BN_GF2m_mod_inv() and
11353 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11354 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11355 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11357 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11359 * Add new error code 'ERR_R_DISABLED' that can be used when some
11360 functionality is disabled at compile-time.
11362 *Douglas Stebila <douglas.stebila@sun.com>*
11364 * Change default behaviour of 'openssl asn1parse' so that more
11365 information is visible when viewing, e.g., a certificate:
11367 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11368 mode the content of non-printable OCTET STRINGs is output in a
11369 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11370 avoid the appearance of a printable string.
11372 *Nils Larsch <nla@trustcenter.de>*
11374 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11376 EC_GROUP_set_asn1_flag()
11377 EC_GROUP_get_asn1_flag()
11378 EC_GROUP_set_point_conversion_form()
11379 EC_GROUP_get_point_conversion_form()
11380 These control ASN1 encoding details:
11381 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11382 has been set to OPENSSL_EC_NAMED_CURVE.
11383 - Points are encoded in uncompressed form by default; options for
11384 asn1_for are as for point2oct, namely
11385 POINT_CONVERSION_COMPRESSED
11386 POINT_CONVERSION_UNCOMPRESSED
11387 POINT_CONVERSION_HYBRID
11389 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11391 EC_GROUP_set_seed()
11392 EC_GROUP_get0_seed()
11393 EC_GROUP_get_seed_len()
11394 This is used only for ASN1 purposes (so far).
11396 *Nils Larsch <nla@trustcenter.de>*
11398 * Add 'field_type' member to EC_METHOD, which holds the NID
11399 of the appropriate field type OID. The new function
11400 EC_METHOD_get_field_type() returns this value.
11402 *Nils Larsch <nla@trustcenter.de>*
11405 EC_POINT_point2bn()
11406 EC_POINT_bn2point()
11407 EC_POINT_point2hex()
11408 EC_POINT_hex2point()
11409 providing useful interfaces to EC_POINT_point2oct() and
11410 EC_POINT_oct2point().
11412 *Nils Larsch <nla@trustcenter.de>*
11414 * Change internals of the EC library so that the functions
11415 EC_GROUP_set_generator()
11416 EC_GROUP_get_generator()
11417 EC_GROUP_get_order()
11418 EC_GROUP_get_cofactor()
11419 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11420 to methods, which would lead to unnecessary code duplication when
11421 adding different types of curves.
11423 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11425 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11426 arithmetic, and such that modified wNAFs are generated
11427 (which avoid length expansion in many cases).
11431 * Add a function EC_GROUP_check_discriminant() (defined via
11432 EC_METHOD) that verifies that the curve discriminant is non-zero.
11434 Add a function EC_GROUP_check() that makes some sanity tests
11435 on a EC_GROUP, its generator and order. This includes
11436 EC_GROUP_check_discriminant().
11438 *Nils Larsch <nla@trustcenter.de>*
11440 * Add ECDSA in new directory crypto/ecdsa/.
11442 Add applications 'openssl ecparam' and 'openssl ecdsa'
11443 (these are based on 'openssl dsaparam' and 'openssl dsa').
11445 ECDSA support is also included in various other files across the
11446 library. Most notably,
11447 - 'openssl req' now has a '-newkey ecdsa:file' option;
11448 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11449 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11450 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11451 them suitable for ECDSA where domain parameters must be
11452 extracted before the specific public key;
11453 - ECDSA engine support has been added.
11455 *Nils Larsch <nla@trustcenter.de>*
11457 * Include some named elliptic curves, and add OIDs from X9.62,
11458 SECG, and WAP/WTLS. Each curve can be obtained from the new
11460 EC_GROUP_new_by_curve_name(),
11461 and the list of available named curves can be obtained with
11462 EC_get_builtin_curves().
11463 Also add a 'curve_name' member to EC_GROUP objects, which can be
11465 EC_GROUP_set_curve_name()
11466 EC_GROUP_get_curve_name()
11468 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11470 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11471 was actually never needed) and in BN_mul(). The removal in BN_mul()
11472 required a small change in bn_mul_part_recursive() and the addition
11473 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11474 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11475 bn_sub_words() and bn_add_words() except they take arrays with
11480 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11482 * Cleanse PEM buffers before freeing them since they may contain
11485 *Benjamin Bennett <ben@psc.edu>*
11487 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11488 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11489 authentication-only ciphersuites.
11493 * Since AES128 and AES256 share a single mask bit in the logic of
11494 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11495 kludge to work properly if AES128 is available and AES256 isn't.
11499 * Expand security boundary to match 1.1.1 module.
11503 * Remove redundant features: hash file source, editing of test vectors
11504 modify fipsld to use external fips_premain.c signature.
11508 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11509 run algorithm test programs.
11513 * Make algorithm test programs more tolerant of whitespace.
11517 * Have SSL/TLS server implementation tolerate "mismatched" record
11518 protocol version while receiving ClientHello even if the
11519 ClientHello is fragmented. (The server can't insist on the
11520 particular protocol version it has chosen before the ServerHello
11521 message has informed the client about his choice.)
11525 * Load error codes if they are not already present instead of using a
11526 static variable. This allows them to be cleanly unloaded and reloaded.
11530 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11532 * Introduce limits to prevent malicious keys being able to
11533 cause a denial of service. ([CVE-2006-2940])
11535 *Steve Henson, Bodo Moeller*
11537 * Fix ASN.1 parsing of certain invalid structures that can result
11538 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11540 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11541 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11543 * Fix SSL client code which could crash if connecting to a
11544 malicious SSLv2 server. ([CVE-2006-4343])
11546 *Tavis Ormandy and Will Drewry, Google Security Team*
11548 * Change ciphersuite string processing so that an explicit
11549 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11550 will no longer include "AES128-SHA"), and any other similar
11551 ciphersuite (same bitmap) from *other* protocol versions (so that
11552 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11553 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11554 changes from 0.9.8b and 0.9.8d.
11558 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11560 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11561 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11563 * Change the Unix randomness entropy gathering to use poll() when
11564 possible instead of select(), since the latter has some
11565 undesirable limitations.
11567 *Darryl Miles via Richard Levitte and Bodo Moeller*
11569 * Disable rogue ciphersuites:
11571 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11572 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11573 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11575 The latter two were purportedly from
11576 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11579 Also deactivate the remaining ciphersuites from
11580 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11581 unofficial, and the ID has long expired.
11585 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11586 dual-core machines) and other potential thread-safety issues.
11590 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11592 * Adapt fipsld and the build system to link against the validated FIPS
11593 module in FIPS mode.
11597 * Fixes for VC++ 2005 build under Windows.
11601 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11602 from a Windows bash shell such as MSYS. It is autodetected from the
11603 "config" script when run from a VC++ environment. Modify standard VC++
11604 build to use fipscanister.o from the GNU make build.
11608 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11610 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11611 The value now differs depending on if you build for FIPS or not.
11612 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11613 safely run with a non-FIPSed libcrypto, as it may crash because of
11614 the difference induced by this change.
11618 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11620 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11621 (part of SSL_OP_ALL). This option used to disable the
11622 countermeasure against man-in-the-middle protocol-version
11623 rollback in the SSL 2.0 server implementation, which is a bad
11624 idea. ([CVE-2005-2969])
11626 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11627 for Information Security, National Institute of Advanced Industrial
11628 Science and Technology [AIST, Japan)]*
11630 * Minimal support for X9.31 signatures and PSS padding modes. This is
11631 mainly for FIPS compliance and not fully integrated at this stage.
11635 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11636 the exponentiation using a fixed-length exponent. (Otherwise,
11637 the information leaked through timing could expose the secret key
11638 after many signatures; cf. Bleichenbacher's attack on DSA with
11643 * Make a new fixed-window mod_exp implementation the default for
11644 RSA, DSA, and DH private-key operations so that the sequence of
11645 squares and multiplies and the memory access pattern are
11646 independent of the particular secret key. This will mitigate
11647 cache-timing and potential related attacks.
11649 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11650 and this is automatically used by BN_mod_exp_mont() if the new flag
11651 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11652 will use this BN flag for private exponents unless the flag
11653 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11654 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11656 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11658 * Change the client implementation for SSLv23_method() and
11659 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11660 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11661 (Previously, the SSL 2.0 backwards compatible Client Hello
11662 message format would be used even with SSL_OP_NO_SSLv2.)
11666 * Add support for smime-type MIME parameter in S/MIME messages which some
11671 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11672 a threadsafe manner. Modify rsa code to use new function and add calls
11673 to dsa and dh code (which had race conditions before).
11677 * Include the fixed error library code in the C error file definitions
11678 instead of fixing them up at runtime. This keeps the error code
11679 structures constant.
11683 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11685 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11688 * Fixes for newer kerberos headers. NB: the casts are needed because
11689 the 'length' field is signed on one version and unsigned on another
11690 with no (?) obvious way to tell the difference, without these VC++
11691 complains. Also the "definition" of FAR (blank) is no longer included
11692 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11693 some needed definitions.
11697 * Undo Cygwin change.
11701 * Added support for proxy certificates according to RFC 3820.
11702 Because they may be a security thread to unaware applications,
11703 they must be explicitly allowed in run-time. See
11704 docs/HOWTO/proxy_certificates.txt for further information.
11708 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11710 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11711 server and client random values. Previously
11712 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11713 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11715 This change has negligible security impact because:
11717 1. Server and client random values still have 24 bytes of pseudo random
11720 2. Server and client random values are sent in the clear in the initial
11723 3. The master secret is derived using the premaster secret (48 bytes in
11724 size for static RSA ciphersuites) as well as client server and random
11727 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11730 *Stephen Henson, reported by UK NISCC*
11732 * Use Windows randomness collection on Cygwin.
11736 * Fix hang in EGD/PRNGD query when communication socket is closed
11737 prematurely by EGD/PRNGD.
11739 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11741 * Prompt for pass phrases when appropriate for PKCS12 input format.
11745 * Back-port of selected performance improvements from development
11746 branch, as well as improved support for PowerPC platforms.
11750 * Add lots of checks for memory allocation failure, error codes to indicate
11751 failure and freeing up memory if a failure occurs.
11753 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11755 * Add new -passin argument to dgst.
11759 * Perform some character comparisons of different types in X509_NAME_cmp:
11760 this is needed for some certificates that re-encode DNs into UTF8Strings
11761 (in violation of RFC3280) and can't or won't issue name rollover
11766 * Make an explicit check during certificate validation to see that
11767 the CA setting in each certificate on the chain is correct. As a
11768 side effect always do the following basic checks on extensions,
11769 not just when there's an associated purpose to the check:
11771 - if there is an unhandled critical extension (unless the user
11772 has chosen to ignore this fault)
11773 - if the path length has been exceeded (if one is set at all)
11774 - that certain extensions fit the associated purpose (if one has
11779 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11781 * Avoid a race condition when CRLs are checked in a multi threaded
11782 environment. This would happen due to the reordering of the revoked
11783 entries during signature checking and serial number lookup. Now the
11784 encoding is cached and the serial number sort performed under a lock.
11785 Add new STACK function sk_is_sorted().
11789 * Add Delta CRL to the extension code.
11793 * Various fixes to s3_pkt.c so alerts are sent properly.
11795 *David Holmes <d.holmes@f5.com>*
11797 * Reduce the chances of duplicate issuer name and serial numbers (in
11798 violation of RFC3280) using the OpenSSL certificate creation utilities.
11799 This is done by creating a random 64 bit value for the initial serial
11800 number when a serial number file is created or when a self signed
11801 certificate is created using 'openssl req -x509'. The initial serial
11802 number file is created using 'openssl x509 -next_serial' in CA.pl
11803 rather than being initialized to 1.
11807 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11809 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11810 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11812 *Joe Orton, Steve Henson*
11814 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11817 *Joe Orton, Steve Henson*
11819 * Make it possible to have multiple active certificates with the same
11820 subject in the CA index file. This is done only if the keyword
11821 'unique_subject' is set to 'no' in the main CA section (default
11822 if 'CA_default') of the configuration file. The value is saved
11823 with the database itself in a separate index attribute file,
11824 named like the index file with '.attr' appended to the name.
11828 * X509 verify fixes. Disable broken certificate workarounds when
11829 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11830 keyUsage extension present. Don't accept CRLs with unhandled critical
11831 extensions: since verify currently doesn't process CRL extensions this
11832 rejects a CRL with *any* critical extensions. Add new verify error codes
11837 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11838 A clarification of RFC2560 will require the use of OCTET STRINGs and
11839 some implementations cannot handle the current raw format. Since OpenSSL
11840 copies and compares OCSP nonces as opaque blobs without any attempt at
11841 parsing them this should not create any compatibility issues.
11845 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11846 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11847 this HMAC (and other) operations are several times slower than OpenSSL
11852 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11854 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11856 * Use the correct content when signing type "other".
11860 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11862 * Fix various bugs revealed by running the NISCC test suite:
11864 Stop out of bounds reads in the ASN1 code when presented with
11865 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11867 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11869 If verify callback ignores invalid public key errors don't try to check
11870 certificate signature with the NULL public key.
11874 * New -ignore_err option in ocsp application to stop the server
11875 exiting on the first error in a request.
11879 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11880 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11885 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11886 extra data after the compression methods not only for TLS 1.0
11887 but also for SSL 3.0 (as required by the specification).
11889 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11891 * Change X509_certificate_type() to mark the key as exported/exportable
11892 when it's 512 *bits* long, not 512 bytes.
11896 * Change AES_cbc_encrypt() so it outputs exact multiple of
11897 blocks during encryption.
11901 * Various fixes to base64 BIO and non blocking I/O. On write
11902 flushes were not handled properly if the BIO retried. On read
11903 data was not being buffered properly and had various logic bugs.
11904 This also affects blocking I/O when the data being decoded is a
11909 * Various S/MIME bugfixes and compatibility changes:
11910 output correct application/pkcs7 MIME type if
11911 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11912 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11913 of files as .eml work). Correctly handle very long lines in MIME
11918 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11920 * Countermeasure against the Klima-Pokorny-Rosa extension of
11921 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11922 a protocol version number mismatch like a decryption error
11923 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11927 * Turn on RSA blinding by default in the default implementation
11928 to avoid a timing attack. Applications that don't want it can call
11929 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11930 They would be ill-advised to do so in most cases.
11932 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11934 * Change RSA blinding code so that it works when the PRNG is not
11935 seeded (in this case, the secret RSA exponent is abused as
11936 an unpredictable seed -- if it is not unpredictable, there
11937 is no point in blinding anyway). Make RSA blinding thread-safe
11938 by remembering the creator's thread ID in rsa->blinding and
11939 having all other threads use local one-time blinding factors
11940 (this requires more computation than sharing rsa->blinding, but
11941 avoids excessive locking; and if an RSA object is not shared
11942 between threads, blinding will still be very fast).
11946 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11947 ENGINE as defaults for all supported algorithms irrespective of
11948 the 'flags' parameter. 'flags' is now honoured, so applications
11949 should make sure they are passing it correctly.
11953 * Target "mingw" now allows native Windows code to be generated in
11954 the Cygwin environment as well as with the MinGW compiler.
11958 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11960 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11961 via timing by performing a MAC computation even if incorrect
11962 block cipher padding has been found. This is a countermeasure
11963 against active attacks where the attacker has to distinguish
11964 between bad padding and a MAC verification error. ([CVE-2003-0078])
11966 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11967 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11968 Martin Vuagnoux (EPFL, Ilion)*
11970 * Make the no-err option work as intended. The intention with no-err
11971 is not to have the whole error stack handling routines removed from
11972 libcrypto, it's only intended to remove all the function name and
11973 reason texts, thereby removing some of the footprint that may not
11974 be interesting if those errors aren't displayed anyway.
11976 NOTE: it's still possible for any application or module to have its
11977 own set of error texts inserted. The routines are there, just not
11978 used by default when no-err is given.
11982 * Add support for FreeBSD on IA64.
11984 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11986 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11987 Kerberos function mit_des_cbc_cksum(). Before this change,
11988 the value returned by DES_cbc_cksum() was like the one from
11989 mit_des_cbc_cksum(), except the bytes were swapped.
11991 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11993 * Allow an application to disable the automatic SSL chain building.
11994 Before this a rather primitive chain build was always performed in
11995 ssl3_output_cert_chain(): an application had no way to send the
11996 correct chain if the automatic operation produced an incorrect result.
11998 Now the chain builder is disabled if either:
12000 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12002 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12004 The reasoning behind this is that an application would not want the
12005 auto chain building to take place if extra chain certificates are
12006 present and it might also want a means of sending no additional
12007 certificates (for example the chain has two certificates and the
12012 * Add the possibility to build without the ENGINE framework.
12014 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12016 * Under Win32 gmtime() can return NULL: check return value in
12017 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12021 * DSA routines: under certain error conditions uninitialized BN objects
12022 could be freed. Solution: make sure initialization is performed early
12023 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12024 Nils Larsch <nla@trustcenter.de> via PR#459)
12028 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12029 checked on reconnect on the client side, therefore session resumption
12030 could still fail with a "ssl session id is different" error. This
12031 behaviour is masked when SSL_OP_ALL is used due to
12032 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12033 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12034 followup to PR #377.
12038 * IA-32 assembler support enhancements: unified ELF targets, support
12039 for SCO/Caldera platforms, fix for Cygwin shared build.
12043 * Add support for FreeBSD on sparc64. As a consequence, support for
12044 FreeBSD on non-x86 processors is separate from x86 processors on
12045 the config script, much like the NetBSD support.
12047 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12049 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12051 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12054 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12055 code (06) was taken as the first octet of the session ID and the last
12056 octet was ignored consequently. As a result SSLv2 client side session
12057 caching could not have worked due to the session ID mismatch between
12059 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12064 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12065 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12070 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12071 seems that in spite of existing for more than a year, many application
12072 author have done nothing to provide the necessary callbacks, which
12073 means that this particular engine will not work properly anywhere.
12074 This is a very unfortunate situation which forces us, in the name
12075 of usability, to give the hw_ncipher.c a static lock, which is part
12077 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12078 appear in 0.9.8 or later. We EXPECT application authors to have
12079 dealt properly with this when 0.9.8 is released (unless we actually
12080 make such changes in the libcrypto locking code that changes will
12081 have to be made anyway).
12085 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12086 octets have been read, EOF or an error occurs. Without this change
12087 some truncated ASN1 structures will not produce an error.
12091 * Disable Heimdal support, since it hasn't been fully implemented.
12092 Still give the possibility to force the use of Heimdal, but with
12093 warnings and a request that patches get sent to openssl-dev.
12097 * Add the VC-CE target, introduce the WINCE sysname, and add
12098 INSTALL.WCE and appropriate conditionals to make it build.
12100 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12102 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12103 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12104 edit numbers of the version.
12106 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12108 * Introduce safe string copy and catenation functions
12109 (BUF_strlcpy() and BUF_strlcat()).
12111 *Ben Laurie (CHATS) and Richard Levitte*
12113 * Avoid using fixed-size buffers for one-line DNs.
12115 *Ben Laurie (CHATS)*
12117 * Add BUF_MEM_grow_clean() to avoid information leakage when
12118 resizing buffers containing secrets, and use where appropriate.
12120 *Ben Laurie (CHATS)*
12122 * Avoid using fixed size buffers for configuration file location.
12124 *Ben Laurie (CHATS)*
12126 * Avoid filename truncation for various CA files.
12128 *Ben Laurie (CHATS)*
12130 * Use sizeof in preference to magic numbers.
12132 *Ben Laurie (CHATS)*
12134 * Avoid filename truncation in cert requests.
12136 *Ben Laurie (CHATS)*
12138 * Add assertions to check for (supposedly impossible) buffer
12141 *Ben Laurie (CHATS)*
12143 * Don't cache truncated DNS entries in the local cache (this could
12144 potentially lead to a spoofing attack).
12146 *Ben Laurie (CHATS)*
12148 * Fix various buffers to be large enough for hex/decimal
12149 representations in a platform independent manner.
12151 *Ben Laurie (CHATS)*
12153 * Add CRYPTO_realloc_clean() to avoid information leakage when
12154 resizing buffers containing secrets, and use where appropriate.
12156 *Ben Laurie (CHATS)*
12158 * Add BIO_indent() to avoid much slightly worrying code to do
12161 *Ben Laurie (CHATS)*
12163 * Convert sprintf()/BIO_puts() to BIO_printf().
12165 *Ben Laurie (CHATS)*
12167 * buffer_gets() could terminate with the buffer only half
12170 *Ben Laurie (CHATS)*
12172 * Add assertions to prevent user-supplied crypto functions from
12173 overflowing internal buffers by having large block sizes, etc.
12175 *Ben Laurie (CHATS)*
12177 * New OPENSSL_assert() macro (similar to assert(), but enabled
12180 *Ben Laurie (CHATS)*
12182 * Eliminate unused copy of key in RC4.
12184 *Ben Laurie (CHATS)*
12186 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12188 *Ben Laurie (CHATS)*
12190 * Fix off-by-one error in EGD path.
12192 *Ben Laurie (CHATS)*
12194 * If RANDFILE path is too long, ignore instead of truncating.
12196 *Ben Laurie (CHATS)*
12198 * Eliminate unused and incorrectly sized X.509 structure
12201 *Ben Laurie (CHATS)*
12203 * Eliminate unused and dangerous function knumber().
12205 *Ben Laurie (CHATS)*
12207 * Eliminate unused and dangerous structure, KSSL_ERR.
12209 *Ben Laurie (CHATS)*
12211 * Protect against overlong session ID context length in an encoded
12212 session object. Since these are local, this does not appear to be
12215 *Ben Laurie (CHATS)*
12217 * Change from security patch (see 0.9.6e below) that did not affect
12218 the 0.9.6 release series:
12220 Remote buffer overflow in SSL3 protocol - an attacker could
12221 supply an oversized master key in Kerberos-enabled versions.
12224 *Ben Laurie (CHATS)*
12226 * Change the SSL kerb5 codes to match RFC 2712.
12230 * Make -nameopt work fully for req and add -reqopt switch.
12232 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12234 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12236 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12238 * Make sure tests can be performed even if the corresponding algorithms
12239 have been removed entirely. This was also the last step to make
12240 OpenSSL compilable with DJGPP under all reasonable conditions.
12242 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12244 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12245 to allow version independent disabling of normally unselected ciphers,
12246 which may be activated as a side-effect of selecting a single cipher.
12248 (E.g., cipher list string "RSA" enables ciphersuites that are left
12249 out of "ALL" because they do not provide symmetric encryption.
12250 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12252 *Lutz Jaenicke, Bodo Moeller*
12254 * Add appropriate support for separate platform-dependent build
12255 directories. The recommended way to make a platform-dependent
12256 build directory is the following (tested on Linux), maybe with
12259 # Place yourself outside of the OpenSSL source tree. In
12260 # this example, the environment variable OPENSSL_SOURCE
12261 # is assumed to contain the absolute OpenSSL source directory.
12262 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12263 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12264 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12265 mkdir -p `dirname $F`
12266 ln -s $OPENSSL_SOURCE/$F $F
12269 To be absolutely sure not to disturb the source tree, a "make clean"
12270 is a good thing. If it isn't successful, don't worry about it,
12271 it probably means the source directory is very clean.
12275 * Make sure any ENGINE control commands make local copies of string
12276 pointers passed to them whenever necessary. Otherwise it is possible
12277 the caller may have overwritten (or deallocated) the original string
12278 data when a later ENGINE operation tries to use the stored values.
12280 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12282 * Improve diagnostics in file reading and command-line digests.
12284 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12286 * Add AES modes CFB and OFB to the object database. Correct an
12287 error in AES-CFB decryption.
12291 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12292 allows existing EVP_CIPHER_CTX structures to be reused after
12293 calling `EVP_*Final()`. This behaviour is used by encryption
12294 BIOs and some applications. This has the side effect that
12295 applications must explicitly clean up cipher contexts with
12296 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12300 * Check the values of dna and dnb in bn_mul_recursive before calling
12301 bn_mul_comba (a non zero value means the a or b arrays do not contain
12302 n2 elements) and fallback to bn_mul_normal if either is not zero.
12306 * Fix escaping of non-ASCII characters when using the -subj option
12307 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12311 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12312 form for "surname", serialNumber has no short form.
12313 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12314 therefore remove "mail" short name for "internet 7".
12315 The OID for unique identifiers in X509 certificates is
12316 x500UniqueIdentifier, not uniqueIdentifier.
12317 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12321 * Add an "init" command to the ENGINE config module and auto initialize
12322 ENGINEs. Without any "init" command the ENGINE will be initialized
12323 after all ctrl commands have been executed on it. If init=1 the
12324 ENGINE is initialized at that point (ctrls before that point are run
12325 on the uninitialized ENGINE and after on the initialized one). If
12326 init=0 then the ENGINE will not be initialized at all.
12330 * Fix the 'app_verify_callback' interface so that the user-defined
12331 argument is actually passed to the callback: In the
12332 SSL_CTX_set_cert_verify_callback() prototype, the callback
12333 declaration has been changed from
12336 int (*cb)(X509_STORE_CTX *,void *);
12337 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12338 i=s->ctx->app_verify_callback(&ctx)
12339 has been changed into
12340 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12342 To update applications using SSL_CTX_set_cert_verify_callback(),
12343 a dummy argument can be added to their callback functions.
12345 *D. K. Smetters <smetters@parc.xerox.com>*
12347 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12349 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12351 * Add and OPENSSL_LOAD_CONF define which will cause
12352 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12353 This allows older applications to transparently support certain
12354 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12355 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12356 load the config file and OPENSSL_add_all_algorithms_conf() which will
12357 always load it have also been added.
12361 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12362 Adjust NIDs and EVP layer.
12364 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12366 * Config modules support in openssl utility.
12368 Most commands now load modules from the config file,
12369 though in a few (such as version) this isn't done
12370 because it couldn't be used for anything.
12372 In the case of ca and req the config file used is
12373 the same as the utility itself: that is the -config
12374 command line option can be used to specify an
12379 * Move default behaviour from OPENSSL_config(). If appname is NULL
12380 use "openssl_conf" if filename is NULL use default openssl config file.
12384 * Add an argument to OPENSSL_config() to allow the use of an alternative
12385 config section name. Add a new flag to tolerate a missing config file
12386 and move code to CONF_modules_load_file().
12390 * Support for crypto accelerator cards from Accelerated Encryption
12391 Processing, www.aep.ie. (Use engine 'aep')
12392 The support was copied from 0.9.6c [engine] and adapted/corrected
12393 to work with the new engine framework.
12395 *AEP Inc. and Richard Levitte*
12397 * Support for SureWare crypto accelerator cards from Baltimore
12398 Technologies. (Use engine 'sureware')
12399 The support was copied from 0.9.6c [engine] and adapted
12400 to work with the new engine framework.
12404 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12405 make the newer ENGINE framework commands for the CHIL engine work.
12407 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12409 * Make it possible to produce shared libraries on ReliantUNIX.
12411 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12413 * Add the configuration target debug-linux-ppro.
12414 Make 'openssl rsa' use the general key loading routines
12415 implemented in `apps.c`, and make those routines able to
12416 handle the key format FORMAT_NETSCAPE and the variant
12419 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12421 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12423 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12425 * Add -keyform to rsautl, and document -engine.
12427 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12429 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12430 BIO_R_NO_SUCH_FILE error code rather than the generic
12431 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12435 * Add new functions
12436 ERR_peek_last_error
12437 ERR_peek_last_error_line
12438 ERR_peek_last_error_line_data.
12439 These are similar to
12441 ERR_peek_error_line
12442 ERR_peek_error_line_data,
12443 but report on the latest error recorded rather than the first one
12444 still in the error queue.
12446 *Ben Laurie, Bodo Moeller*
12448 * default_algorithms option in ENGINE config module. This allows things
12450 default_algorithms = ALL
12451 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12455 * Preliminary ENGINE config module.
12459 * New experimental application configuration code.
12463 * Change the AES code to follow the same name structure as all other
12464 symmetric ciphers, and behave the same way. Move everything to
12465 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12467 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12469 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12471 *Ben Laurie and Theo de Raadt*
12473 * Add option to output public keys in req command.
12475 *Massimiliano Pala madwolf@openca.org*
12477 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12478 (up to about 10% better than before for P-192 and P-224).
12482 * New functions/macros
12484 SSL_CTX_set_msg_callback(ctx, cb)
12485 SSL_CTX_set_msg_callback_arg(ctx, arg)
12486 SSL_set_msg_callback(ssl, cb)
12487 SSL_set_msg_callback_arg(ssl, arg)
12489 to request calling a callback function
12491 void cb(int write_p, int version, int content_type,
12492 const void *buf, size_t len, SSL *ssl, void *arg)
12494 whenever a protocol message has been completely received
12495 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12496 protocol version according to which the SSL library interprets
12497 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12498 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12499 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12500 specification (change_cipher_spec(20), alert(21), handshake(22)).
12501 'buf' and 'len' point to the actual message, 'ssl' to the
12502 SSL object, and 'arg' is the application-defined value set by
12503 SSL[_CTX]_set_msg_callback_arg().
12505 'openssl s_client' and 'openssl s_server' have new '-msg' options
12506 to enable a callback that displays all protocol messages.
12510 * Change the shared library support so shared libraries are built as
12511 soon as the corresponding static library is finished, and thereby get
12512 openssl and the test programs linked against the shared library.
12513 This still only happens when the keyword "shard" has been given to
12514 the configuration scripts.
12516 NOTE: shared library support is still an experimental thing, and
12517 backward binary compatibility is still not guaranteed.
12519 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12521 * Add support for Subject Information Access extension.
12523 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12525 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12526 additional bytes when new memory had to be allocated, not just
12527 when reusing an existing buffer.
12531 * New command line and configuration option 'utf8' for the req command.
12532 This allows field values to be specified as UTF8 strings.
12536 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12537 runs for the former and machine-readable output for the latter.
12541 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12542 of the e-mail address in the DN (i.e., it will go into a certificate
12543 extension only). The new configuration file option 'email_in_dn = no'
12544 has the same effect.
12546 *Massimiliano Pala madwolf@openca.org*
12548 * Change all functions with names starting with `des_` to be starting
12549 with `DES_` instead. Add wrappers that are compatible with libdes,
12550 but are named `_ossl_old_des_*`. Finally, add macros that map the
12551 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12552 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12553 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12556 Since we provide two compatibility mappings, the user needs to
12557 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12558 compatibility is desired. The default (i.e., when that macro
12559 isn't defined) is OpenSSL 0.9.6c compatibility.
12561 There are also macros that enable and disable the support of old
12562 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12563 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12564 are defined, the default will apply: to support the old des routines.
12566 In either case, one must include openssl/des.h to get the correct
12567 definitions. Do not try to just include openssl/des_old.h, that
12570 NOTE: This is a major break of an old API into a new one. Software
12571 authors are encouraged to switch to the `DES_` style functions. Some
12572 time in the future, des_old.h and the libdes compatibility functions
12573 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12574 default), and then completely removed.
12578 * Test for certificates which contain unsupported critical extensions.
12579 If such a certificate is found during a verify operation it is
12580 rejected by default: this behaviour can be overridden by either
12581 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12582 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12583 X509_supported_extension() has also been added which returns 1 if a
12584 particular extension is supported.
12588 * Modify the behaviour of EVP cipher functions in similar way to digests
12589 to retain compatibility with existing code.
12593 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12594 compatibility with existing code. In particular the 'ctx' parameter does
12595 not have to be to be initialized before the call to EVP_DigestInit() and
12596 it is tidied up after a call to EVP_DigestFinal(). New function
12597 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12598 EVP_MD_CTX_copy() changed to not require the destination to be
12599 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12600 requires the destination to be valid.
12602 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12603 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12607 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12608 so that complete 'Handshake' protocol structures are kept in memory
12609 instead of overwriting 'msg_type' and 'length' with 'body' data.
12613 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12615 *Massimo Santin via Richard Levitte*
12617 * Major restructuring to the underlying ENGINE code. This includes
12618 reduction of linker bloat, separation of pure "ENGINE" manipulation
12619 (initialisation, etc) from functionality dealing with implementations
12620 of specific crypto interfaces. This change also introduces integrated
12621 support for symmetric ciphers and digest implementations - so ENGINEs
12622 can now accelerate these by providing EVP_CIPHER and EVP_MD
12623 implementations of their own. This is detailed in
12624 [crypto/engine/README.md](crypto/engine/README.md)
12625 as it couldn't be adequately described here. However, there are a few
12626 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12627 were changed in the original introduction of ENGINE code have now
12628 reverted back - the hooking from this code to ENGINE is now a good
12629 deal more passive and at run-time, operations deal directly with
12630 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12631 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12632 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12633 they were not being used by the framework as there is no concept of a
12634 BIGNUM_METHOD and they could not be generalised to the new
12635 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12636 ENGINE_cpy() has been removed as it cannot be consistently defined in
12641 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12645 * Change mkdef.pl to sort symbols that get the same entry number,
12646 and make sure the automatically generated functions `ERR_load_*`
12647 become part of libeay.num as well.
12651 * New function SSL_renegotiate_pending(). This returns true once
12652 renegotiation has been requested (either SSL_renegotiate() call
12653 or HelloRequest/ClientHello received from the peer) and becomes
12654 false once a handshake has been completed.
12655 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12656 sends a HelloRequest, but does not ensure that a handshake takes
12657 place. SSL_renegotiate_pending() is useful for checking if the
12658 client has followed the request.)
12662 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12663 By default, clients may request session resumption even during
12664 renegotiation (if session ID contexts permit); with this option,
12665 session resumption is possible only in the first handshake.
12667 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12668 more bits available for options that should not be part of
12669 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12673 * Add some demos for certificate and certificate request creation.
12677 * Make maximum certificate chain size accepted from the peer application
12678 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12679 "Douglas E. Engert" <deengert@anl.gov>.
12683 * Add support for shared libraries for Unixware-7
12684 (Boyd Lynn Gerber <gerberb@zenez.com>).
12688 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12689 be done prior to destruction. Use this to unload error strings from
12690 ENGINEs that load their own error strings. NB: This adds two new API
12691 functions to "get" and "set" this destroy handler in an ENGINE.
12695 * Alter all existing ENGINE implementations (except "openssl" and
12696 "openbsd") to dynamically instantiate their own error strings. This
12697 makes them more flexible to be built both as statically-linked ENGINEs
12698 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12699 Also, add stub code to each that makes building them as self-contained
12700 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12704 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12705 implementations into applications that are completely implemented in
12706 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12707 commands that can be used to configure what shared-library to load and
12708 to control aspects of the way it is handled. Also, made an update to
12709 the [README-Engine.md](README-Engine.md) file
12710 that brings its information up-to-date and
12711 provides some information and instructions on the "dynamic" ENGINE
12712 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12716 * Make it possible to unload ranges of ERR strings with a new
12717 "ERR_unload_strings" function.
12721 * Add a copy() function to EVP_MD.
12725 * Make EVP_MD routines take a context pointer instead of just the
12726 md_data void pointer.
12730 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12731 that the digest can only process a single chunk of data
12732 (typically because it is provided by a piece of
12733 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12734 is only going to provide a single chunk of data, and hence the
12735 framework needn't accumulate the data for oneshot drivers.
12739 * As with "ERR", make it possible to replace the underlying "ex_data"
12740 functions. This change also alters the storage and management of global
12741 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12742 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12743 index counters. The API functions that use this state have been changed
12744 to take a "class_index" rather than pointers to the class's local STACK
12745 and counter, and there is now an API function to dynamically create new
12746 classes. This centralisation allows us to (a) plug a lot of the
12747 thread-safety problems that existed, and (b) makes it possible to clean
12748 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12749 such data would previously have always leaked in application code and
12750 workarounds were in place to make the memory debugging turn a blind eye
12751 to it. Application code that doesn't use this new function will still
12752 leak as before, but their memory debugging output will announce it now
12753 rather than letting it slide.
12755 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12756 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12757 has a return value to indicate success or failure.
12761 * Make it possible to replace the underlying "ERR" functions such that the
12762 global state (2 LHASH tables and 2 locks) is only used by the "default"
12763 implementation. This change also adds two functions to "get" and "set"
12764 the implementation prior to it being automatically set the first time
12765 any other ERR function takes place. Ie. an application can call "get",
12766 pass the return value to a module it has just loaded, and that module
12767 can call its own "set" function using that value. This means the
12768 module's "ERR" operations will use (and modify) the error state in the
12769 application and not in its own statically linked copy of OpenSSL code.
12773 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12774 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12775 the operation, and provides a more encapsulated way for external code
12776 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12777 to use these functions rather than manually incrementing the counts.
12779 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12783 * Add EVP test program.
12787 * Add symmetric cipher support to ENGINE. Expect the API to change!
12791 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12792 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12793 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12794 These allow a CRL to be built without having to access X509_CRL fields
12795 directly. Modify 'ca' application to use new functions.
12799 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12800 bug workarounds. Rollback attack detection is a security feature.
12801 The problem will only arise on OpenSSL servers when TLSv1 is not
12802 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12803 Software authors not wanting to support TLSv1 will have special reasons
12804 for their choice and can explicitly enable this option.
12806 *Bodo Moeller, Lutz Jaenicke*
12808 * Rationalise EVP so it can be extended: don't include a union of
12809 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12810 (similar to those existing for EVP_CIPHER_CTX).
12815 EVP_MD_CTX_init(&md); /* new function call */
12816 EVP_DigestInit(&md, EVP_sha1());
12817 EVP_DigestUpdate(&md, in, len);
12818 EVP_DigestFinal(&md, out, NULL);
12819 EVP_MD_CTX_cleanup(&md); /* new function call */
12823 * Make DES key schedule conform to the usual scheme, as well as
12824 correcting its structure. This means that calls to DES functions
12825 now have to pass a pointer to a des_key_schedule instead of a
12826 plain des_key_schedule (which was actually always a pointer
12829 des_key_schedule ks;
12831 des_set_key_checked(..., &ks);
12832 des_ncbc_encrypt(..., &ks, ...);
12834 (Note that a later change renames 'des_...' into 'DES_...'.)
12838 * Initial reduction of linker bloat: the use of some functions, such as
12839 PEM causes large amounts of unused functions to be linked in due to
12840 poor organisation. For example pem_all.c contains every PEM function
12841 which has a knock on effect of linking in large amounts of (unused)
12842 ASN1 code. Grouping together similar functions and splitting unrelated
12843 functions prevents this.
12847 * Cleanup of EVP macros.
12851 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12852 correct `_ecb suffix`.
12856 * Add initial OCSP responder support to ocsp application. The
12857 revocation information is handled using the text based index
12858 use by the ca application. The responder can either handle
12859 requests generated internally, supplied in files (for example
12860 via a CGI script) or using an internal minimal server.
12864 * Add configuration choices to get zlib compression for TLS.
12868 * Changes to Kerberos SSL for RFC 2712 compliance:
12869 1. Implemented real KerberosWrapper, instead of just using
12870 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12871 2. Implemented optional authenticator field of KerberosWrapper.
12873 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12874 and authenticator structs; see crypto/krb5/.
12876 Generalized Kerberos calls to support multiple Kerberos libraries.
12877 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12878 via Richard Levitte*
12880 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12881 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12882 values for each of the key sizes rather than having just
12883 parameters (and 'speed' generating keys each time).
12887 * Speed up EVP routines.
12890 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12891 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12892 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12893 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12895 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12896 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12897 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12900 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12902 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12906 * Added the OS2-EMX target.
12908 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12910 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12911 New functions to support `NCONF` routines in extension code.
12912 New function `CONF_set_nconf()`
12913 to allow functions which take an `NCONF` to also handle the old `LHASH`
12914 structure: this means that the old `CONF` compatible routines can be
12915 retained (in particular w.rt. extensions) without having to duplicate the
12916 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12920 * Enhance the general user interface with mechanisms for inner control
12921 and with possibilities to have yes/no kind of prompts.
12925 * Change all calls to low-level digest routines in the library and
12926 applications to use EVP. Add missing calls to HMAC_cleanup() and
12927 don't assume HMAC_CTX can be copied using memcpy().
12929 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12931 * Add the possibility to control engines through control names but with
12932 arbitrary arguments instead of just a string.
12933 Change the key loaders to take a UI_METHOD instead of a callback
12934 function pointer. NOTE: this breaks binary compatibility with earlier
12935 versions of OpenSSL [engine].
12936 Adapt the nCipher code for these new conditions and add a card insertion
12941 * Enhance the general user interface with mechanisms to better support
12942 dialog box interfaces, application-defined prompts, the possibility
12943 to use defaults (for example default passwords from somewhere else)
12944 and interrupts/cancellations.
12948 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12949 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12953 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12954 tidy up some unnecessarily weird code in 'sk_new()').
12956 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12958 * Change the key loading routines for ENGINEs to use the same kind
12959 callback (pem_password_cb) as all other routines that need this
12964 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12965 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12966 than this minimum value is recommended.
12970 * New random seeder for OpenVMS, using the system process statistics
12971 that are easily reachable.
12975 * Windows apparently can't transparently handle global
12976 variables defined in DLLs. Initialisations such as:
12978 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12980 won't compile. This is used by the any applications that need to
12981 declare their own ASN1 modules. This was fixed by adding the option
12982 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12983 needed for static libraries under Win32.
12987 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12988 setting of purpose and trust fields. New X509_STORE trust and
12989 purpose functions and tidy up setting in other SSL functions.
12993 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12994 structure. These are inherited by X509_STORE_CTX when it is
12995 initialised. This allows various defaults to be set in the
12996 X509_STORE structure (such as flags for CRL checking and custom
12997 purpose or trust settings) for functions which only use X509_STORE_CTX
12998 internally such as S/MIME.
13000 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13001 trust settings if they are not set in X509_STORE. This allows X509_STORE
13002 purposes and trust (in S/MIME for example) to override any set by default.
13004 Add command line options for CRL checking to smime, s_client and s_server
13009 * Initial CRL based revocation checking. If the CRL checking flag(s)
13010 are set then the CRL is looked up in the X509_STORE structure and
13011 its validity and signature checked, then if the certificate is found
13012 in the CRL the verify fails with a revoked error.
13014 Various new CRL related callbacks added to X509_STORE_CTX structure.
13016 Command line options added to 'verify' application to support this.
13018 This needs some additional work, such as being able to handle multiple
13019 CRLs with different times, extension based lookup (rather than just
13020 by subject name) and ultimately more complete V2 CRL extension
13025 * Add a general user interface API (crypto/ui/). This is designed
13026 to replace things like des_read_password and friends (backward
13027 compatibility functions using this new API are provided).
13028 The purpose is to remove prompting functions from the DES code
13029 section as well as provide for prompting through dialog boxes in
13030 a window system and the like.
13034 * Add "ex_data" support to ENGINE so implementations can add state at a
13035 per-structure level rather than having to store it globally.
13039 * Make it possible for ENGINE structures to be copied when retrieved by
13040 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13041 This causes the "original" ENGINE structure to act like a template,
13042 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13043 operational state can be localised to each ENGINE structure, despite the
13044 fact they all share the same "methods". New ENGINE structures returned in
13045 this case have no functional references and the return value is the single
13046 structural reference. This matches the single structural reference returned
13047 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13052 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13053 needs to match any other type at all we need to manually clear the
13058 * Changes to the "openssl engine" utility to include;
13059 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13060 about an ENGINE's available control commands.
13061 - executing control commands from command line arguments using the
13062 '-pre' and '-post' switches. '-post' is only used if '-t' is
13063 specified and the ENGINE is successfully initialised. The syntax for
13064 the individual commands are colon-separated, for example;
13065 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13069 * New dynamic control command support for ENGINEs. ENGINEs can now
13070 declare their own commands (numbers), names (strings), descriptions,
13071 and input types for run-time discovery by calling applications. A
13072 subset of these commands are implicitly classed as "executable"
13073 depending on their input type, and only these can be invoked through
13074 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13075 can be based on user input, config files, etc). The distinction is
13076 that "executable" commands cannot return anything other than a boolean
13077 result and can only support numeric or string input, whereas some
13078 discoverable commands may only be for direct use through
13079 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13080 pointers, or other custom uses. The "executable" commands are to
13081 support parameterisations of ENGINE behaviour that can be
13082 unambiguously defined by ENGINEs and used consistently across any
13083 OpenSSL-based application. Commands have been added to all the
13084 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13085 control over shared-library paths without source code alterations.
13089 * Changed all ENGINE implementations to dynamically allocate their
13090 ENGINEs rather than declaring them statically. Apart from this being
13091 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13092 this also allows the implementations to compile without using the
13093 internal engine_int.h header.
13097 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13098 'const' value. Any code that should be able to modify a RAND_METHOD
13099 should already have non-const pointers to it (ie. they should only
13100 modify their own ones).
13104 * Made a variety of little tweaks to the ENGINE code.
13105 - "atalla" and "ubsec" string definitions were moved from header files
13106 to C code. "nuron" string definitions were placed in variables
13107 rather than hard-coded - allowing parameterisation of these values
13108 later on via ctrl() commands.
13109 - Removed unused "#if 0"'d code.
13110 - Fixed engine list iteration code so it uses ENGINE_free() to release
13111 structural references.
13112 - Constified the RAND_METHOD element of ENGINE structures.
13113 - Constified various get/set functions as appropriate and added
13114 missing functions (including a catch-all ENGINE_cpy that duplicates
13115 all ENGINE values onto a new ENGINE except reference counts/state).
13116 - Removed NULL parameter checks in get/set functions. Setting a method
13117 or function to NULL is a way of cancelling out a previously set
13118 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13119 and doesn't justify the extra error symbols and code.
13120 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13121 flags from engine_int.h to engine.h.
13122 - Changed prototypes for ENGINE handler functions (init(), finish(),
13123 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13127 * Implement binary inversion algorithm for BN_mod_inverse in addition
13128 to the algorithm using long division. The binary algorithm can be
13129 used only if the modulus is odd. On 32-bit systems, it is faster
13130 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13131 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13132 up to 450 bits. In 64-bit environments, the binary algorithm
13133 appears to be advantageous for much longer moduli; here we use it
13134 for moduli up to 2048 bits.
13138 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13139 could not support the combine flag in choice fields.
13143 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13144 extensions from a certificate request to the certificate.
13148 * Allow multiple 'certopt' and 'nameopt' options to be separated
13149 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13150 file: this allows the display of the certificate about to be
13151 signed to be customised, to allow certain fields to be included
13152 or excluded and extension details. The old system didn't display
13153 multicharacter strings properly, omitted fields not in the policy
13154 and couldn't display additional details such as extensions.
13158 * Function EC_POINTs_mul for multiple scalar multiplication
13159 of an arbitrary number of elliptic curve points
13160 \sum scalars[i]*points[i],
13161 optionally including the generator defined for the EC_GROUP:
13162 scalar*generator + \sum scalars[i]*points[i].
13164 EC_POINT_mul is a simple wrapper function for the typical case
13165 that the point list has just one item (besides the optional
13170 * First EC_METHODs for curves over GF(p):
13172 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13173 operations and provides various method functions that can also
13174 operate with faster implementations of modular arithmetic.
13176 EC_GFp_mont_method() reuses most functions that are part of
13177 EC_GFp_simple_method, but uses Montgomery arithmetic.
13179 *Bodo Moeller; point addition and point doubling
13180 implementation directly derived from source code provided by
13181 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13183 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13184 crypto/ec/ec_lib.c):
13186 Curves are EC_GROUP objects (with an optional group generator)
13187 based on EC_METHODs that are built into the library.
13189 Points are EC_POINT objects based on EC_GROUP objects.
13191 Most of the framework would be able to handle curves over arbitrary
13192 finite fields, but as there are no obvious types for fields other
13193 than GF(p), some functions are limited to that for now.
13197 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13198 that the file contains a complete HTTP response.
13202 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13203 change the def and num file printf format specifier from "%-40sXXX"
13204 to "%-39s XXX". The latter will always guarantee a space after the
13205 field while the former will cause them to run together if the field
13206 is 40 of more characters long.
13210 * Constify the cipher and digest 'method' functions and structures
13211 and modify related functions to take constant EVP_MD and EVP_CIPHER
13216 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13217 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13221 * Modify `EVP_Digest*()` routines so they now return values. Although the
13222 internal software routines can never fail additional hardware versions
13227 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13229 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13230 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13233 ERR_R_NESTED_ASN1_ERROR
13235 ERR_R_MISSING_ASN1_EOS
13236 were 4 .. 9, conflicting with
13237 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13239 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13240 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13242 Add new error code 'ERR_R_INTERNAL_ERROR'.
13246 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13251 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13252 sets the subject name for a new request or supersedes the
13253 subject name in a given request. Formats that can be parsed are
13254 'CN=Some Name, OU=myOU, C=IT'
13256 'CN=Some Name/OU=myOU/C=IT'.
13258 Add options '-batch' and '-verbose' to 'openssl req'.
13260 *Massimiliano Pala <madwolf@hackmasters.net>*
13262 * Introduce the possibility to access global variables through
13263 functions on platform were that's the best way to handle exporting
13264 global variables in shared libraries. To enable this functionality,
13265 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13266 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13267 is normally done by Configure or something similar).
13269 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13270 in the source file (foo.c) like this:
13272 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13273 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13275 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13276 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13278 OPENSSL_DECLARE_GLOBAL(int,foo);
13279 #define foo OPENSSL_GLOBAL_REF(foo)
13280 OPENSSL_DECLARE_GLOBAL(double,bar);
13281 #define bar OPENSSL_GLOBAL_REF(bar)
13283 The #defines are very important, and therefore so is including the
13284 header file everywhere where the defined globals are used.
13286 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13287 of ASN.1 items, but that structure is a bit different.
13289 The largest change is in util/mkdef.pl which has been enhanced with
13290 better and easier to understand logic to choose which symbols should
13291 go into the Windows .def files as well as a number of fixes and code
13292 cleanup (among others, algorithm keywords are now sorted
13293 lexicographically to avoid constant rewrites).
13297 * In BN_div() keep a copy of the sign of 'num' before writing the
13298 result to 'rm' because if rm==num the value will be overwritten
13299 and produce the wrong result if 'num' is negative: this caused
13300 problems with BN_mod() and BN_nnmod().
13304 * Function OCSP_request_verify(). This checks the signature on an
13305 OCSP request and verifies the signer certificate. The signer
13306 certificate is just checked for a generic purpose and OCSP request
13311 * Add OCSP_check_validity() function to check the validity of OCSP
13312 responses. OCSP responses are prepared in real time and may only
13313 be a few seconds old. Simply checking that the current time lies
13314 between thisUpdate and nextUpdate max reject otherwise valid responses
13315 caused by either OCSP responder or client clock inaccuracy. Instead
13316 we allow thisUpdate and nextUpdate to fall within a certain period of
13317 the current time. The age of the response can also optionally be
13318 checked. Two new options -validity_period and -status_age added to
13323 * If signature or public key algorithm is unrecognized print out its
13324 OID rather that just UNKNOWN.
13328 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13329 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13330 ID to be generated from the issuer certificate alone which can then be
13331 passed to OCSP_id_issuer_cmp().
13335 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13336 ASN1 modules to export functions returning ASN1_ITEM pointers
13337 instead of the ASN1_ITEM structures themselves. This adds several
13338 new macros which allow the underlying ASN1 function/structure to
13339 be accessed transparently. As a result code should not use ASN1_ITEM
13340 references directly (such as &X509_it) but instead use the relevant
13341 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13342 use of the new ASN1 code on platforms where exporting structures
13343 is problematical (for example in shared libraries) but exporting
13344 functions returning pointers to structures is not.
13348 * Add support for overriding the generation of SSL/TLS session IDs.
13349 These callbacks can be registered either in an SSL_CTX or per SSL.
13350 The purpose of this is to allow applications to control, if they wish,
13351 the arbitrary values chosen for use as session IDs, particularly as it
13352 can be useful for session caching in multiple-server environments. A
13353 command-line switch for testing this (and any client code that wishes
13354 to use such a feature) has been added to "s_server".
13356 *Geoff Thorpe, Lutz Jaenicke*
13358 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13359 of the form `#if defined(...) || defined(...) || ...` and
13360 `#if !defined(...) && !defined(...) && ...`. This also avoids
13361 the growing number of special cases it was previously handling.
13365 * Make all configuration macros available for application by making
13366 sure they are available in opensslconf.h, by giving them names starting
13367 with `OPENSSL_` to avoid conflicts with other packages and by making
13368 sure e_os2.h will cover all platform-specific cases together with
13370 Additionally, it is now possible to define configuration/platform-
13371 specific names (called "system identities"). In the C code, these
13372 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13373 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13374 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13379 * New option -set_serial to 'req' and 'x509' this allows the serial
13380 number to use to be specified on the command line. Previously self
13381 signed certificates were hard coded with serial number 0 and the
13382 CA options of 'x509' had to use a serial number in a file which was
13387 * New options to 'ca' utility to support V2 CRL entry extensions.
13388 Currently CRL reason, invalidity date and hold instruction are
13389 supported. Add new CRL extensions to V3 code and some new objects.
13393 * New function EVP_CIPHER_CTX_set_padding() this is used to
13394 disable standard block padding (aka PKCS#5 padding) in the EVP
13395 API, which was previously mandatory. This means that the data is
13396 not padded in any way and so the total length much be a multiple
13397 of the block size, otherwise an error occurs.
13401 * Initial (incomplete) OCSP SSL support.
13405 * New function OCSP_parse_url(). This splits up a URL into its host,
13406 port and path components: primarily to parse OCSP URLs. New -url
13407 option to ocsp utility.
13411 * New nonce behavior. The return value of OCSP_check_nonce() now
13412 reflects the various checks performed. Applications can decide
13413 whether to tolerate certain situations such as an absent nonce
13414 in a response when one was present in a request: the ocsp application
13415 just prints out a warning. New function OCSP_add1_basic_nonce()
13416 this is to allow responders to include a nonce in a response even if
13417 the request is nonce-less.
13421 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13422 skipped when using openssl x509 multiple times on a single input file,
13423 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13427 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13428 set string type: to handle setting ASN1_TIME structures. Fix ca
13429 utility to correctly initialize revocation date of CRLs.
13433 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13434 the clients preferred ciphersuites and rather use its own preferences.
13435 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13436 Internet Explorer by ensuring unchanged hash method during stepup.
13437 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13441 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13442 to aes and add a new 'exist' option to print out symbols that don't
13447 * Additional options to ocsp utility to allow flags to be set and
13448 additional certificates supplied.
13452 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13453 OCSP client a number of certificate to only verify the response
13458 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13459 handle the new API. Currently only ECB, CBC modes supported. Add new
13462 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13463 Encryption Standard (AES) Ciphersuites for Transport Layer
13464 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13465 not enabled by default and were not part of the "ALL" ciphersuite
13466 alias because they were not yet official; they could be
13467 explicitly requested by specifying the "AESdraft" ciphersuite
13468 group alias. In the final release of OpenSSL 0.9.7, the group
13469 alias is called "AES" and is part of "ALL".)
13471 *Ben Laurie, Steve Henson, Bodo Moeller*
13473 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13474 request to response.
13478 * Functions for OCSP responders. OCSP_request_onereq_count(),
13479 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13480 extract information from a certificate request. OCSP_response_create()
13481 creates a response and optionally adds a basic response structure.
13482 OCSP_basic_add1_status() adds a complete single response to a basic
13483 response and returns the OCSP_SINGLERESP structure just added (to allow
13484 extensions to be included for example). OCSP_basic_add1_cert() adds a
13485 certificate to a basic response and OCSP_basic_sign() signs a basic
13486 response with various flags. New helper functions ASN1_TIME_check()
13487 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13488 (converts ASN1_TIME to GeneralizedTime).
13492 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13493 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13494 structure from a certificate. X509_pubkey_digest() digests the public_key
13495 contents: this is used in various key identifiers.
13499 * Make sk_sort() tolerate a NULL argument.
13501 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13503 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13504 passed by the function are trusted implicitly. If any of them signed the
13505 response then it is assumed to be valid and is not verified.
13509 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13510 to data. This was previously part of the PKCS7 ASN1 code. This
13511 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13512 *Steve Henson, reported by Kenneth R. Robinette
13513 <support@securenetterm.com>*
13515 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13516 routines: without these tracing memory leaks is very painful.
13517 Fix leaks in PKCS12 and PKCS7 routines.
13521 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13522 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13523 effectively meant GeneralizedTime would never be used. Now it
13524 is initialised to -1 but X509_time_adj() now has to check the value
13525 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13526 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13527 *Steve Henson, reported by Kenneth R. Robinette
13528 <support@securenetterm.com>*
13530 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13531 result in a zero length in the ASN1_INTEGER structure which was
13532 not consistent with the structure when d2i_ASN1_INTEGER() was used
13533 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13534 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13535 where it did not print out a minus for negative ASN1_INTEGER.
13539 * Add summary printout to ocsp utility. The various functions which
13540 convert status values to strings have been renamed to:
13541 OCSP_response_status_str(), OCSP_cert_status_str() and
13542 OCSP_crl_reason_str() and are no longer static. New options
13543 to verify nonce values and to disable verification. OCSP response
13544 printout format cleaned up.
13548 * Add additional OCSP certificate checks. These are those specified
13549 in RFC2560. This consists of two separate checks: the CA of the
13550 certificate being checked must either be the OCSP signer certificate
13551 or the issuer of the OCSP signer certificate. In the latter case the
13552 OCSP signer certificate must contain the OCSP signing extended key
13553 usage. This check is performed by attempting to match the OCSP
13554 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13555 in the OCSP_CERTID structures of the response.
13559 * Initial OCSP certificate verification added to OCSP_basic_verify()
13560 and related routines. This uses the standard OpenSSL certificate
13561 verify routines to perform initial checks (just CA validity) and
13562 to obtain the certificate chain. Then additional checks will be
13563 performed on the chain. Currently the root CA is checked to see
13564 if it is explicitly trusted for OCSP signing. This is used to set
13565 a root CA as a global signing root: that is any certificate that
13566 chains to that CA is an acceptable OCSP signing certificate.
13570 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13571 extensions from a separate configuration file.
13572 As when reading extensions from the main configuration file,
13573 the '-extensions ...' option may be used for specifying the
13576 *Massimiliano Pala <madwolf@comune.modena.it>*
13578 * New OCSP utility. Allows OCSP requests to be generated or
13579 read. The request can be sent to a responder and the output
13580 parsed, outputted or printed in text form. Not complete yet:
13581 still needs to check the OCSP response validity.
13585 * New subcommands for 'openssl ca':
13586 `openssl ca -status <serial>` prints the status of the cert with
13587 the given serial number (according to the index file).
13588 `openssl ca -updatedb` updates the expiry status of certificates
13591 *Massimiliano Pala <madwolf@comune.modena.it>*
13593 * New '-newreq-nodes' command option to CA.pl. This is like
13594 '-newreq', but calls 'openssl req' with the '-nodes' option
13595 so that the resulting key is not encrypted.
13597 *Damien Miller <djm@mindrot.org>*
13599 * New configuration for the GNU Hurd.
13601 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13603 * Initial code to implement OCSP basic response verify. This
13604 is currently incomplete. Currently just finds the signer's
13605 certificate and verifies the signature on the response.
13609 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13610 value of OPENSSLDIR. This is available via the new '-d' option
13611 to 'openssl version', and is also included in 'openssl version -a'.
13615 * Allowing defining memory allocation callbacks that will be given
13616 file name and line number information in additional arguments
13617 (a `const char*` and an int). The basic functionality remains, as
13618 well as the original possibility to just replace malloc(),
13619 realloc() and free() by functions that do not know about these
13620 additional arguments. To register and find out the current
13621 settings for extended allocation functions, the following
13622 functions are provided:
13624 CRYPTO_set_mem_ex_functions
13625 CRYPTO_set_locked_mem_ex_functions
13626 CRYPTO_get_mem_ex_functions
13627 CRYPTO_get_locked_mem_ex_functions
13629 These work the same way as CRYPTO_set_mem_functions and friends.
13630 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13631 extended allocation function is enabled.
13632 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13633 a conventional allocation function is enabled.
13635 *Richard Levitte, Bodo Moeller*
13637 * Finish off removing the remaining LHASH function pointer casts.
13638 There should no longer be any prototype-casting required when using
13639 the LHASH abstraction, and any casts that remain are "bugs". See
13640 the callback types and macros at the head of lhash.h for details
13641 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13645 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13646 If /dev/[u]random devices are not available or do not return enough
13647 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13649 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13650 /etc/entropy will be queried once each in this sequence, querying stops
13651 when enough entropy was collected without querying more sockets.
13655 * Change the Unix RAND_poll() variant to be able to poll several
13656 random devices, as specified by DEVRANDOM, until a sufficient amount
13657 of data has been collected. We spend at most 10 ms on each file
13658 (select timeout) and read in non-blocking mode. DEVRANDOM now
13659 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13660 (previously it was just the string "/dev/urandom"), so on typical
13661 platforms the 10 ms delay will never occur.
13662 Also separate out the Unix variant to its own file, rand_unix.c.
13663 For VMS, there's a currently-empty rand_vms.c.
13667 * Move OCSP client related routines to ocsp_cl.c. These
13668 provide utility functions which an application needing
13669 to issue a request to an OCSP responder and analyse the
13670 response will typically need: as opposed to those which an
13671 OCSP responder itself would need which will be added later.
13673 OCSP_request_sign() signs an OCSP request with an API similar
13674 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13675 response. OCSP_response_get1_basic() extracts basic response
13676 from response. OCSP_resp_find_status(): finds and extracts status
13677 information from an OCSP_CERTID structure (which will be created
13678 when the request structure is built). These are built from lower
13679 level functions which work on OCSP_SINGLERESP structures but
13680 won't normally be used unless the application wishes to examine
13681 extensions in the OCSP response for example.
13683 Replace nonce routines with a pair of functions.
13684 OCSP_request_add1_nonce() adds a nonce value and optionally
13685 generates a random value. OCSP_check_nonce() checks the
13686 validity of the nonce in an OCSP response.
13690 * Change function OCSP_request_add() to OCSP_request_add0_id().
13691 This doesn't copy the supplied OCSP_CERTID and avoids the
13692 need to free up the newly created id. Change return type
13693 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13694 This can then be used to add extensions to the request.
13695 Deleted OCSP_request_new(), since most of its functionality
13696 is now in OCSP_REQUEST_new() (and the case insensitive name
13697 clash) apart from the ability to set the request name which
13698 will be added elsewhere.
13702 * Update OCSP API. Remove obsolete extensions argument from
13703 various functions. Extensions are now handled using the new
13704 OCSP extension code. New simple OCSP HTTP function which
13705 can be used to send requests and parse the response.
13709 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13710 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13711 uses the special reorder version of SET OF to sort the attributes
13712 and reorder them to match the encoded order. This resolves a long
13713 standing problem: a verify on a PKCS7 structure just after signing
13714 it used to fail because the attribute order did not match the
13715 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13716 it uses the received order. This is necessary to tolerate some broken
13717 software that does not order SET OF. This is handled by encoding
13718 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13719 to produce the required SET OF.
13723 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13724 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13725 files to get correct declarations of the ASN.1 item variables.
13729 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13730 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13731 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13732 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13733 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13734 ASN1_ITEM and no wrapper functions.
13738 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13739 replace the old function pointer based I/O routines. Change most of
13740 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13744 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13745 lines, recognize more "algorithms" that can be deselected, and make
13746 it complain about algorithm deselection that isn't recognised.
13750 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13751 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13752 to use new functions. Add NO_ASN1_OLD which can be set to remove
13753 some old style ASN1 functions: this can be used to determine if old
13754 code will still work when these eventually go away.
13758 * New extension functions for OCSP structures, these follow the
13759 same conventions as certificates and CRLs.
13763 * New function X509V3_add1_i2d(). This automatically encodes and
13764 adds an extension. Its behaviour can be customised with various
13765 flags to append, replace or delete. Various wrappers added for
13766 certificates and CRLs.
13770 * Fix to avoid calling the underlying ASN1 print routine when
13771 an extension cannot be parsed. Correct a typo in the
13772 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13776 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13777 entries for variables.
13781 * Add functionality to `apps/openssl.c` for detecting locking
13782 problems: As the program is single-threaded, all we have
13783 to do is register a locking callback using an array for
13784 storing which locks are currently held by the program.
13788 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13789 SSL_get_ex_data_X509_STORE_idx(), which is used in
13790 ssl_verify_cert_chain() and thus can be called at any time
13791 during TLS/SSL handshakes so that thread-safety is essential.
13792 Unfortunately, the ex_data design is not at all suited
13793 for multi-threaded use, so it probably should be abolished.
13797 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13799 *Broadcom, tweaked and integrated by Geoff Thorpe*
13801 * Move common extension printing code to new function
13802 X509V3_print_extensions(). Reorganise OCSP print routines and
13803 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13807 * New function X509_signature_print() to remove duplication in some
13812 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13813 set (this was treated exactly the same as SET OF previously). This
13814 is used to reorder the STACK representing the structure to match the
13815 encoding. This will be used to get round a problem where a PKCS7
13816 structure which was signed could not be verified because the STACK
13817 order did not reflect the encoded order.
13821 * Reimplement the OCSP ASN1 module using the new code.
13825 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13826 for its ASN1 operations. The old style function pointers still exist
13827 for now but they will eventually go away.
13831 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13832 completely replaces the old ASN1 functionality with a table driven
13833 encoder and decoder which interprets an ASN1_ITEM structure describing
13834 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13835 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13836 has also been converted to the new form.
13840 * Change BN_mod_exp_recp so that negative moduli are tolerated
13841 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13842 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13843 for negative moduli.
13847 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13848 of not touching the result's sign bit.
13852 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13857 * Changed the LHASH code to use prototypes for callbacks, and created
13858 macros to declare and implement thin (optionally static) functions
13859 that provide type-safety and avoid function pointer casting for the
13860 type-specific callbacks.
13864 * Added Kerberos Cipher Suites to be used with TLS, as written in
13866 *Veers Staats <staatsvr@asc.hpc.mil>,
13867 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13869 * Reformat the FAQ so the different questions and answers can be divided
13870 in sections depending on the subject.
13874 * Have the zlib compression code load ZLIB.DLL dynamically under
13879 * New function BN_mod_sqrt for computing square roots modulo a prime
13880 (using the probabilistic Tonelli-Shanks algorithm unless
13881 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13882 be handled deterministically).
13884 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13886 * Make BN_mod_inverse faster by explicitly handling small quotients
13887 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13888 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13892 * New function BN_kronecker.
13896 * Fix BN_gcd so that it works on negative inputs; the result is
13897 positive unless both parameters are zero.
13898 Previously something reasonably close to an infinite loop was
13899 possible because numbers could be growing instead of shrinking
13900 in the implementation of Euclid's algorithm.
13904 * Fix BN_is_word() and BN_is_one() macros to take into account the
13905 sign of the number in question.
13907 Fix BN_is_word(a,w) to work correctly for w == 0.
13909 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13910 because its test if the absolute value of 'a' equals 'w'.
13911 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13912 it exists mostly for use in the implementations of BN_is_zero(),
13913 BN_is_one(), and BN_is_word().
13917 * New function BN_swap.
13921 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13922 the exponentiation functions are more likely to produce reasonable
13923 results on negative inputs.
13927 * Change BN_mod_mul so that the result is always non-negative.
13928 Previously, it could be negative if one of the factors was negative;
13929 I don't think anyone really wanted that behaviour.
13933 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13934 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13935 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13936 and add new functions:
13945 BN_mod_lshift1_quick
13947 BN_mod_lshift_quick
13949 These functions always generate non-negative results.
13951 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13952 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13954 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13955 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13956 be reduced modulo `m`.
13958 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13961 The following entry accidentally appeared in the CHANGES file
13962 distributed with OpenSSL 0.9.7. The modifications described in
13963 it do *not* apply to OpenSSL 0.9.7.
13965 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13966 was actually never needed) and in BN_mul(). The removal in BN_mul()
13967 required a small change in bn_mul_part_recursive() and the addition
13968 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13969 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13970 bn_sub_words() and bn_add_words() except they take arrays with
13976 * In 'openssl passwd', verify passwords read from the terminal
13977 unless the '-salt' option is used (which usually means that
13978 verification would just waste user's time since the resulting
13979 hash is going to be compared with some given password hash)
13980 or the new '-noverify' option is used.
13982 This is an incompatible change, but it does not affect
13983 non-interactive use of 'openssl passwd' (passwords on the command
13984 line, '-stdin' option, '-in ...' option) and thus should not
13985 cause any problems.
13989 * Remove all references to RSAref, since there's no more need for it.
13993 * Make DSO load along a path given through an environment variable
13994 (SHLIB_PATH) with shl_load().
13998 * Constify the ENGINE code as a result of BIGNUM constification.
13999 Also constify the RSA code and most things related to it. In a
14000 few places, most notable in the depth of the ASN.1 code, ugly
14001 casts back to non-const were required (to be solved at a later
14006 * Make it so the openssl application has all engines loaded by default.
14010 * Constify the BIGNUM routines a little more.
14014 * Add the following functions:
14016 ENGINE_load_cswift()
14018 ENGINE_load_atalla()
14019 ENGINE_load_nuron()
14020 ENGINE_load_builtin_engines()
14022 That way, an application can itself choose if external engines that
14023 are built-in in OpenSSL shall ever be used or not. The benefit is
14024 that applications won't have to be linked with libdl or other dso
14025 libraries unless it's really needed.
14027 Changed 'openssl engine' to load all engines on demand.
14028 Changed the engine header files to avoid the duplication of some
14029 declarations (they differed!).
14033 * 'openssl engine' can now list capabilities.
14037 * Better error reporting in 'openssl engine'.
14041 * Never call load_dh_param(NULL) in s_server.
14045 * Add engine application. It can currently list engines by name and
14046 identity, and test if they are actually available.
14050 * Improve RPM specification file by forcing symbolic linking and making
14051 sure the installed documentation is also owned by root.root.
14053 *Damien Miller <djm@mindrot.org>*
14055 * Give the OpenSSL applications more possibilities to make use of
14056 keys (public as well as private) handled by engines.
14060 * Add OCSP code that comes from CertCo.
14064 * Add VMS support for the Rijndael code.
14068 * Added untested support for Nuron crypto accelerator.
14072 * Add support for external cryptographic devices. This code was
14073 previously distributed separately as the "engine" branch.
14075 *Geoff Thorpe, Richard Levitte*
14077 * Rework the filename-translation in the DSO code. It is now possible to
14078 have far greater control over how a "name" is turned into a filename
14079 depending on the operating environment and any oddities about the
14080 different shared library filenames on each system.
14084 * Support threads on FreeBSD-elf in Configure.
14088 * Fix for SHA1 assembly problem with MASM: it produces
14089 warnings about corrupt line number information when assembling
14090 with debugging information. This is caused by the overlapping
14093 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14096 NCONF_get_number() has no error checking at all. As a replacement,
14097 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14098 promoted strongly. The old NCONF_get_number is kept around for
14099 binary backward compatibility.
14100 Make it possible for methods to load from something other than a BIO,
14101 by providing a function pointer that is given a name instead of a BIO.
14102 For example, this could be used to load configuration data from an
14107 * Fix for non blocking accept BIOs. Added new I/O special reason
14108 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14109 with non blocking I/O was not possible because no retry code was
14110 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14115 * Added the beginnings of Rijndael support.
14119 * Fix for bug in DirectoryString mask setting. Add support for
14120 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14121 to allow certificate printing to more controllable, additional
14122 'certopt' option to 'x509' to allow new printing options to be
14127 * Clean old EAY MD5 hack from e_os.h.
14131 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14133 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14134 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14136 *Joe Orton, Steve Henson*
14138 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14140 * Fix additional bug revealed by the NISCC test suite:
14142 Stop bug triggering large recursion when presented with
14143 certain ASN.1 tags ([CVE-2003-0851])
14147 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14149 * Fix various bugs revealed by running the NISCC test suite:
14151 Stop out of bounds reads in the ASN1 code when presented with
14152 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14154 If verify callback ignores invalid public key errors don't try to check
14155 certificate signature with the NULL public key.
14159 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14160 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14165 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14166 extra data after the compression methods not only for TLS 1.0
14167 but also for SSL 3.0 (as required by the specification).
14169 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14171 * Change X509_certificate_type() to mark the key as exported/exportable
14172 when it's 512 *bits* long, not 512 bytes.
14176 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14178 * Countermeasure against the Klima-Pokorny-Rosa extension of
14179 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14180 a protocol version number mismatch like a decryption error
14181 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14185 * Turn on RSA blinding by default in the default implementation
14186 to avoid a timing attack. Applications that don't want it can call
14187 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14188 They would be ill-advised to do so in most cases.
14190 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14192 * Change RSA blinding code so that it works when the PRNG is not
14193 seeded (in this case, the secret RSA exponent is abused as
14194 an unpredictable seed -- if it is not unpredictable, there
14195 is no point in blinding anyway). Make RSA blinding thread-safe
14196 by remembering the creator's thread ID in rsa->blinding and
14197 having all other threads use local one-time blinding factors
14198 (this requires more computation than sharing rsa->blinding, but
14199 avoids excessive locking; and if an RSA object is not shared
14200 between threads, blinding will still be very fast).
14204 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14206 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14207 via timing by performing a MAC computation even if incorrect
14208 block cipher padding has been found. This is a countermeasure
14209 against active attacks where the attacker has to distinguish
14210 between bad padding and a MAC verification error. ([CVE-2003-0078])
14212 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14213 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14214 Martin Vuagnoux (EPFL, Ilion)*
14216 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14218 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14219 memory from its contents. This is done with a counter that will
14220 place alternating values in each byte. This can be used to solve
14221 two issues: 1) the removal of calls to memset() by highly optimizing
14222 compilers, and 2) cleansing with other values than 0, since those can
14223 be read through on certain media, for example a swap space on disk.
14227 * Bugfix: client side session caching did not work with external caching,
14228 because the session->cipher setting was not restored when reloading
14229 from the external cache. This problem was masked, when
14230 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14231 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14235 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14236 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14238 *Zeev Lieber <zeev-l@yahoo.com>*
14240 * Undo an undocumented change introduced in 0.9.6e which caused
14241 repeated calls to OpenSSL_add_all_ciphers() and
14242 OpenSSL_add_all_digests() to be ignored, even after calling
14247 * Change the default configuration reader to deal with last line not
14248 being properly terminated.
14252 * Change X509_NAME_cmp() so it applies the special rules on handling
14253 DN values that are of type PrintableString, as well as RDNs of type
14254 emailAddress where the value has the type ia5String.
14256 *stefank@valicert.com via Richard Levitte*
14258 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14259 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14260 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14261 the bitwise-OR of the two for use by the majority of applications
14262 wanting this behaviour, and update the docs. The documented
14263 behaviour and actual behaviour were inconsistent and had been
14264 changing anyway, so this is more a bug-fix than a behavioural
14267 *Geoff Thorpe, diagnosed by Nadav Har'El*
14269 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14270 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14274 * Fix initialization code race conditions in
14275 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14276 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14277 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14278 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14279 ssl2_get_cipher_by_char(),
14280 ssl3_get_cipher_by_char().
14282 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14284 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14285 the cached sessions are flushed, as the remove_cb() might use ex_data
14286 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14287 (see [openssl.org #212]).
14289 *Geoff Thorpe, Lutz Jaenicke*
14291 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14292 length, instead of the encoding length to d2i_ASN1_OBJECT.
14296 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14298 * [In 0.9.6g-engine release:]
14299 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14301 *Lynn Gazis <lgazis@rainbow.com>*
14303 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14305 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14306 and get fix the header length calculation.
14307 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14308 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14310 * Use proper error handling instead of 'assertions' in buffer
14311 overflow checks added in 0.9.6e. This prevents DoS (the
14312 assertions could call abort()).
14314 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14316 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14318 * Add various sanity checks to asn1_get_length() to reject
14319 the ASN1 length bytes if they exceed sizeof(long), will appear
14320 negative or the content length exceeds the length of the
14323 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14325 * Fix cipher selection routines: ciphers without encryption had no flags
14326 for the cipher strength set and where therefore not handled correctly
14327 by the selection routines (PR #130).
14331 * Fix EVP_dsa_sha macro.
14336 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14337 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14338 that was added in OpenSSL 0.9.6d.
14340 As the countermeasure turned out to be incompatible with some
14341 broken SSL implementations, the new option is part of SSL_OP_ALL.
14342 SSL_OP_ALL is usually employed when compatibility with weird SSL
14343 implementations is desired (e.g. '-bugs' option to 's_client' and
14344 's_server'), so the new option is automatically set in many
14349 * Changes in security patch:
14351 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14352 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14353 Air Force Materiel Command, USAF, under agreement number
14356 * Add various sanity checks to asn1_get_length() to reject
14357 the ASN1 length bytes if they exceed sizeof(long), will appear
14358 negative or the content length exceeds the length of the
14359 supplied buffer. ([CVE-2002-0659])
14361 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14363 * Assertions for various potential buffer overflows, not known to
14364 happen in practice.
14366 *Ben Laurie (CHATS)*
14368 * Various temporary buffers to hold ASCII versions of integers were
14369 too small for 64 bit platforms. ([CVE-2002-0655])
14370 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14372 * Remote buffer overflow in SSL3 protocol - an attacker could
14373 supply an oversized session ID to a client. ([CVE-2002-0656])
14375 *Ben Laurie (CHATS)*
14377 * Remote buffer overflow in SSL2 protocol - an attacker could
14378 supply an oversized client master key. ([CVE-2002-0656])
14380 *Ben Laurie (CHATS)*
14382 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14384 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14385 encoded as NULL) with id-dsa-with-sha1.
14387 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14389 * Check various `X509_...()` return values in `apps/req.c`.
14391 *Nils Larsch <nla@trustcenter.de>*
14393 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14394 an end-of-file condition would erroneously be flagged, when the CRLF
14395 was just at the end of a processed block. The bug was discovered when
14396 processing data through a buffering memory BIO handing the data to a
14397 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14398 <ptsekov@syntrex.com> and Nedelcho Stanev.
14402 * Implement a countermeasure against a vulnerability recently found
14403 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14404 before application data chunks to avoid the use of known IVs
14405 with data potentially chosen by the attacker.
14409 * Fix length checks in ssl3_get_client_hello().
14413 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14414 to prevent ssl3_read_internal() from incorrectly assuming that
14415 ssl3_read_bytes() found application data while handshake
14416 processing was enabled when in fact s->s3->in_read_app_data was
14417 merely automatically cleared during the initial handshake.
14419 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14421 * Fix object definitions for Private and Enterprise: they were not
14422 recognized in their shortname (=lowercase) representation. Extend
14423 obj_dat.pl to issue an error when using undefined keywords instead
14424 of silently ignoring the problem (Svenning Sorensen
14425 <sss@sss.dnsalias.net>).
14429 * Fix DH_generate_parameters() so that it works for 'non-standard'
14430 generators, i.e. generators other than 2 and 5. (Previously, the
14431 code did not properly initialise the 'add' and 'rem' values to
14432 BN_generate_prime().)
14434 In the new general case, we do not insist that 'generator' is
14435 actually a primitive root: This requirement is rather pointless;
14436 a generator of the order-q subgroup is just as good, if not
14441 * Map new X509 verification errors to alerts. Discovered and submitted by
14442 Tom Wu <tom@arcot.com>.
14446 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14447 returning non-zero before the data has been completely received
14448 when using non-blocking I/O.
14450 *Bodo Moeller; problem pointed out by John Hughes*
14452 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14454 *Ben Laurie, Lutz Jaenicke*
14456 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14457 Yoram Zahavi <YoramZ@gilian.com>).
14461 * Add information about CygWin 1.3 and on, and preserve proper
14462 configuration for the versions before that.
14464 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14466 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14467 check whether we deal with a copy of a session and do not delete from
14468 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14469 <izhar@checkpoint.com>.
14473 * Do not store session data into the internal session cache, if it
14474 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14475 flag is set). Proposed by Aslam <aslam@funk.com>.
14479 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14484 * [In 0.9.6d-engine release:]
14485 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14487 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14489 * Add the configuration target linux-s390x.
14491 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14493 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14494 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14495 variable as an indication that a ClientHello message has been
14496 received. As the flag value will be lost between multiple
14497 invocations of ssl3_accept when using non-blocking I/O, the
14498 function may not be aware that a handshake has actually taken
14499 place, thus preventing a new session from being added to the
14502 To avoid this problem, we now set s->new_session to 2 instead of
14503 using a local variable.
14505 *Lutz Jaenicke, Bodo Moeller*
14507 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14508 if the SSL_R_LENGTH_MISMATCH error is detected.
14510 *Geoff Thorpe, Bodo Moeller*
14512 * New 'shared_ldflag' column in Configure platform table.
14516 * Fix EVP_CIPHER_mode macro.
14518 *"Dan S. Camper" <dan@bti.net>*
14520 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14521 type, we must throw them away by setting rr->length to 0.
14523 *D P Chang <dpc@qualys.com>*
14525 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14527 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14528 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14529 worked incorrectly for those cases where range = `10..._2` and
14530 `3*range` is two bits longer than range.)
14534 * Only add signing time to PKCS7 structures if it is not already
14539 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14540 OBJ_ld_ce should be OBJ_id_ce.
14541 Also some ip-pda OIDs in crypto/objects/objects.txt were
14542 incorrect (cf. RFC 3039).
14544 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14546 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14547 returns early because it has nothing to do.
14549 *Andy Schneider <andy.schneider@bjss.co.uk>*
14551 * [In 0.9.6c-engine release:]
14552 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14554 *Andy Schneider <andy.schneider@bjss.co.uk>*
14556 * [In 0.9.6c-engine release:]
14557 Add support for Cryptographic Appliance's keyserver technology.
14558 (Use engine 'keyclient')
14560 *Cryptographic Appliances and Geoff Thorpe*
14562 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14563 is called via tools/c89.sh because arguments have to be
14564 rearranged (all '-L' options must appear before the first object
14567 *Richard Shapiro <rshapiro@abinitio.com>*
14569 * [In 0.9.6c-engine release:]
14570 Add support for Broadcom crypto accelerator cards, backported
14573 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14575 * [In 0.9.6c-engine release:]
14576 Add support for SureWare crypto accelerator cards from
14577 Baltimore Technologies. (Use engine 'sureware')
14579 *Baltimore Technologies and Mark Cox*
14581 * [In 0.9.6c-engine release:]
14582 Add support for crypto accelerator cards from Accelerated
14583 Encryption Processing, www.aep.ie. (Use engine 'aep')
14585 *AEP Inc. and Mark Cox*
14587 * Add a configuration entry for gcc on UnixWare.
14589 *Gary Benson <gbenson@redhat.com>*
14591 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14592 messages are stored in a single piece (fixed-length part and
14593 variable-length part combined) and fix various bugs found on the way.
14597 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14598 instead. BIO_gethostbyname() does not know what timeouts are
14599 appropriate, so entries would stay in cache even when they have
14601 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14603 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14604 faced with a pathologically small ClientHello fragment that does
14605 not contain client_version: Instead of aborting with an error,
14606 simply choose the highest available protocol version (i.e.,
14607 TLS 1.0 unless it is disabled). In practice, ClientHello
14608 messages are never sent like this, but this change gives us
14609 strictly correct behaviour at least for TLS.
14613 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14614 never resets s->method to s->ctx->method when called from within
14615 one of the SSL handshake functions.
14617 *Bodo Moeller; problem pointed out by Niko Baric*
14619 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14620 (sent using the client's version number) if client_version is
14621 smaller than the protocol version in use. Also change
14622 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14623 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14624 the client will at least see that alert.
14628 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14633 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14634 client receives HelloRequest while in a handshake.
14636 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14638 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14639 should end in 'break', not 'goto end' which circumvents various
14640 cleanups done in state SSL_ST_OK. But session related stuff
14641 must be disabled for SSL_ST_OK in the case that we just sent a
14644 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14645 before just sending a HelloRequest.
14647 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14649 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14650 reveal whether illegal block cipher padding was found or a MAC
14651 verification error occurred. (Neither SSLerr() codes nor alerts
14652 are directly visible to potential attackers, but the information
14653 may leak via logfiles.)
14655 Similar changes are not required for the SSL 2.0 implementation
14656 because the number of padding bytes is sent in clear for SSL 2.0,
14657 and the extra bytes are just ignored. However ssl/s2_pkt.c
14658 failed to verify that the purported number of padding bytes is in
14663 * Add OpenUNIX-8 support including shared libraries
14664 (Boyd Lynn Gerber <gerberb@zenez.com>).
14668 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14669 'wristwatch attack' using huge encoding parameters (cf.
14670 James H. Manger's CRYPTO 2001 paper). Note that the
14671 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14672 encoding parameters and hence was not vulnerable.
14676 * BN_sqr() bug fix.
14678 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14680 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14681 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14682 followed by modular reduction.
14684 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14686 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14687 equivalent based on BN_pseudo_rand() instead of BN_rand().
14691 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14692 This function was broken, as the check for a new client hello message
14693 to handle SGC did not allow these large messages.
14694 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14698 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14702 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14703 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14707 * Rework the configuration and shared library support for Tru64 Unix.
14708 The configuration part makes use of modern compiler features and
14709 still retains old compiler behavior for those that run older versions
14710 of the OS. The shared library support part includes a variant that
14711 uses the RPATH feature, and is available through the special
14712 configuration target "alpha-cc-rpath", which will never be selected
14715 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14717 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14718 with the same message size as in ssl3_get_certificate_request().
14719 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14720 messages might inadvertently be reject as too long.
14722 *Petr Lampa <lampa@fee.vutbr.cz>*
14724 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14728 * Modified SSL library such that the verify_callback that has been set
14729 specifically for an SSL object with SSL_set_verify() is actually being
14730 used. Before the change, a verify_callback set with this function was
14731 ignored and the verify_callback() set in the SSL_CTX at the time of
14732 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14733 to allow the necessary settings.
14737 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14738 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14739 done automatically (in contradiction to the requirements of the C
14740 standard). This made problems when used from OpenSSH.
14744 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14745 dh->length and always used
14747 BN_rand_range(priv_key, dh->p).
14749 BN_rand_range() is not necessary for Diffie-Hellman, and this
14750 specific range makes Diffie-Hellman unnecessarily inefficient if
14751 dh->length (recommended exponent length) is much smaller than the
14752 length of dh->p. We could use BN_rand_range() if the order of
14753 the subgroup was stored in the DH structure, but we only have
14758 BN_rand(priv_key, l, ...)
14760 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14767 RSA_eay_public_encrypt
14768 RSA_eay_private_decrypt
14769 RSA_eay_private_encrypt (signing)
14770 RSA_eay_public_decrypt (signature verification)
14772 (default implementations for RSA_public_encrypt,
14773 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14774 always reject numbers >= n.
14778 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14779 to synchronize access to 'locking_thread'. This is necessary on
14780 systems where access to 'locking_thread' (an 'unsigned long'
14781 variable) is not atomic.
14785 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14786 *before* setting the 'crypto_lock_rand' flag. The previous code had
14787 a race condition if 0 is a valid thread ID.
14789 *Travis Vitek <vitek@roguewave.com>*
14791 * Add support for shared libraries under Irix.
14793 *Albert Chin-A-Young <china@thewrittenword.com>*
14795 * Add configuration option to build on Linux on both big-endian and
14796 little-endian MIPS.
14798 *Ralf Baechle <ralf@uni-koblenz.de>*
14800 * Add the possibility to create shared libraries on HP-UX.
14804 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14806 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14807 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14808 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14809 PRNG state recovery was possible based on the output of
14810 one PRNG request appropriately sized to gain knowledge on
14811 'md' followed by enough consecutive 1-byte PRNG requests
14812 to traverse all of 'state'.
14814 1. When updating 'md_local' (the current thread's copy of 'md')
14815 during PRNG output generation, hash all of the previous
14816 'md_local' value, not just the half used for PRNG output.
14818 2. Make the number of bytes from 'state' included into the hash
14819 independent from the number of PRNG bytes requested.
14821 The first measure alone would be sufficient to avoid
14822 Markku-Juhani's attack. (Actually it had never occurred
14823 to me that the half of 'md_local' used for chaining was the
14824 half from which PRNG output bytes were taken -- I had always
14825 assumed that the secret half would be used.) The second
14826 measure makes sure that additional data from 'state' is never
14827 mixed into 'md_local' in small portions; this heuristically
14828 further strengthens the PRNG.
14832 * Fix crypto/bn/asm/mips3.s.
14836 * When only the key is given to "enc", the IV is undefined. Print out
14837 an error message in this case.
14841 * Handle special case when X509_NAME is empty in X509 printing routines.
14845 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14846 positive and less than q.
14850 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14851 used: it isn't thread safe and the add_lock_callback should handle
14854 *Paul Rose <Paul.Rose@bridge.com>*
14856 * Verify that incoming data obeys the block size in
14857 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14863 *Ulf Möller, Bodo Möller*
14865 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14866 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14867 when fixing the server behaviour for backwards-compatible 'client
14868 hello' messages. (Note that the attack is impractical against
14869 SSL 3.0 and TLS 1.0 anyway because length and version checking
14870 means that the probability of guessing a valid ciphertext is
14871 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14874 Before 0.9.5, the countermeasure (hide the error by generating a
14875 random 'decryption result') did not work properly because
14876 ERR_clear_error() was missing, meaning that SSL_get_error() would
14877 detect the supposedly ignored error.
14879 Both problems are now fixed.
14883 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14884 (previously it was 1024).
14888 * Fix for compatibility mode trust settings: ignore trust settings
14889 unless some valid trust or reject settings are present.
14893 * Fix for blowfish EVP: its a variable length cipher.
14897 * Fix various bugs related to DSA S/MIME verification. Handle missing
14898 parameters in DSA public key structures and return an error in the
14899 DSA routines if parameters are absent.
14903 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14904 in the current directory if neither $RANDFILE nor $HOME was set.
14905 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14906 caused some confusion to Windows users who haven't defined $HOME.
14907 Thus RAND_file_name() is changed again: e_os.h can define a
14908 DEFAULT_HOME, which will be used if $HOME is not set.
14909 For Windows, we use "C:"; on other platforms, we still require
14910 environment variables.
14912 * Move 'if (!initialized) RAND_poll()' into regions protected by
14913 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14914 having multiple threads call RAND_poll() concurrently.
14918 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14919 combination of a flag and a thread ID variable.
14920 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14921 flag), *other* threads can enter ssleay_add_bytes without obeying
14922 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14923 that they do not hold after the first thread unsets add_do_not_lock).
14927 * Change bctest again: '-x' expressions are not available in all
14928 versions of 'test'.
14932 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14934 * Fix a couple of memory leaks in PKCS7_dataDecode()
14936 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14938 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14939 the default extension for executables, if any. Also, make the perl
14940 scripts that use symlink() to test if it really exists and use "cp"
14941 if it doesn't. All this made OpenSSL compilable and installable in
14946 * Fix for asn1_GetSequence() for indefinite length constructed data.
14947 If SEQUENCE is length is indefinite just set c->slen to the total
14948 amount of data available.
14950 *Steve Henson, reported by shige@FreeBSD.org*
14952 *This change does not apply to 0.9.7.*
14954 * Change bctest to avoid here-documents inside command substitution
14955 (workaround for FreeBSD /bin/sh bug).
14956 For compatibility with Ultrix, avoid shell functions (introduced
14957 in the bctest version that searches along $PATH).
14961 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14962 with des_encrypt() defined on some operating systems, like Solaris
14967 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14968 On the Importance of Eliminating Errors in Cryptographic
14969 Computations, J. Cryptology 14 (2001) 2, 101-119,
14970 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14974 * MIPS assembler BIGNUM division bug fix.
14978 * Disabled incorrect Alpha assembler code.
14982 * Fix PKCS#7 decode routines so they correctly update the length
14983 after reading an EOC for the EXPLICIT tag.
14987 *This change does not apply to 0.9.7.*
14989 * Fix bug in PKCS#12 key generation routines. This was triggered
14990 if a 3DES key was generated with a 0 initial byte. Include
14991 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14992 (but broken) behaviour.
14996 * Enhance bctest to search for a working bc along $PATH and print
14999 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15001 * Fix memory leaks in err.c: free err_data string if necessary;
15002 don't write to the wrong index in ERR_set_error_data.
15006 * Implement ssl23_peek (analogous to ssl23_read), which previously
15011 * Replace rdtsc with `_emit` statements for VC++ version 5.
15013 *Jeremy Cooper <jeremy@baymoo.org>*
15015 * Make it possible to reuse SSLv2 sessions.
15019 * In copy_email() check for >= 0 as a return value for
15020 X509_NAME_get_index_by_NID() since 0 is a valid index.
15022 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15024 * Avoid coredump with unsupported or invalid public keys by checking if
15025 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15026 PKCS7_verify() fails with non detached data.
15030 * Don't use getenv in library functions when run as setuid/setgid.
15031 New function OPENSSL_issetugid().
15035 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15036 due to incorrect handling of multi-threading:
15038 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15040 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15042 3. Count how many times MemCheck_off() has been called so that
15043 nested use can be treated correctly. This also avoids
15044 inband-signalling in the previous code (which relied on the
15045 assumption that thread ID 0 is impossible).
15049 * Add "-rand" option also to s_client and s_server.
15053 * Fix CPU detection on Irix 6.x.
15054 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15055 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15057 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15062 *This change does not apply to 0.9.7.*
15064 * Use the cached encoding of an X509_NAME structure rather than
15065 copying it. This is apparently the reason for the libsafe "errors"
15066 but the code is actually correct.
15070 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15071 Bleichenbacher's DSA attack.
15072 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15073 to be set and top=0 forces the highest bit to be set; top=-1 is new
15074 and leaves the highest bit random.
15076 *Ulf Moeller, Bodo Moeller*
15078 * In the `NCONF_...`-based implementations for `CONF_...` queries
15079 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15080 a temporary CONF structure with the data component set to NULL
15081 (which gives segmentation faults in lh_retrieve).
15082 Instead, use NULL for the CONF pointer in CONF_get_string and
15083 CONF_get_number (which may use environment variables) and directly
15084 return NULL from CONF_get_section.
15088 * Fix potential buffer overrun for EBCDIC.
15092 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15093 keyUsage if basicConstraints absent for a CA.
15097 * Make SMIME_write_PKCS7() write mail header values with a format that
15098 is more generally accepted (no spaces before the semicolon), since
15099 some programs can't parse those values properly otherwise. Also make
15100 sure BIO's that break lines after each write do not create invalid
15105 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15106 macros previously used would not encode an empty SEQUENCE OF
15107 and break the signature.
15111 *This change does not apply to 0.9.7.*
15113 * Zero the premaster secret after deriving the master secret in
15118 * Add some EVP_add_digest_alias registrations (as found in
15119 OpenSSL_add_all_digests()) to SSL_library_init()
15120 aka OpenSSL_add_ssl_algorithms(). This provides improved
15121 compatibility with peers using X.509 certificates
15122 with unconventional AlgorithmIdentifier OIDs.
15126 * Fix for Irix with NO_ASM.
15128 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15130 * ./config script fixes.
15132 *Ulf Moeller, Richard Levitte*
15134 * Fix 'openssl passwd -1'.
15138 * Change PKCS12_key_gen_asc() so it can cope with non null
15139 terminated strings whose length is passed in the passlen
15140 parameter, for example from PEM callbacks. This was done
15141 by adding an extra length parameter to asc2uni().
15143 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15145 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15146 call failed, free the DSA structure.
15150 * Fix to uni2asc() to cope with zero length Unicode strings.
15151 These are present in some PKCS#12 files.
15155 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15156 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15157 when writing a 32767 byte record.
15159 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15161 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15162 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15164 (RSA objects have a reference count access to which is protected
15165 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15166 so they are meant to be shared between threads.)
15167 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15168 "Reddie, Steven" <Steven.Reddie@ca.com>*
15170 * Fix a deadlock in CRYPTO_mem_leaks().
15174 * Use better test patterns in bntest.
15178 * rand_win.c fix for Borland C.
15182 * BN_rshift bugfix for n == 0.
15186 * Add a 'bctest' script that checks for some known 'bc' bugs
15187 so that 'make test' does not abort just because 'bc' is broken.
15191 * Store verify_result within SSL_SESSION also for client side to
15192 avoid potential security hole. (Re-used sessions on the client side
15193 always resulted in verify_result==X509_V_OK, not using the original
15194 result of the server certificate verification.)
15198 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15199 SSL3_RT_APPLICATION_DATA, return 0.
15200 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15205 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15206 releases, have been re-implemented by renaming the previous
15207 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15208 and ssl3_read_internal, respectively, and adding 'peek' parameters
15209 to them. The new ssl[23]_{read,peek} functions are calls to
15210 ssl[23]_read_internal with the 'peek' flag set appropriately.
15211 A 'peek' parameter has also been added to ssl3_read_bytes, which
15212 does the actual work for ssl3_read_internal.
15216 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15217 the method-specific "init()" handler. Also clean up ex_data after
15218 calling the method-specific "finish()" handler. Previously, this was
15219 happening the other way round.
15223 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15224 The previous value, 12, was not always sufficient for BN_mod_exp().
15228 * Make sure that shared libraries get the internal name engine with
15229 the full version number and not just 0. This should mark the
15230 shared libraries as not backward compatible. Of course, this should
15231 be changed again when we can guarantee backward binary compatibility.
15235 * Fix typo in get_cert_by_subject() in by_dir.c
15237 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15239 * Rework the system to generate shared libraries:
15241 - Make note of the expected extension for the shared libraries and
15242 if there is a need for symbolic links from for example libcrypto.so.0
15243 to libcrypto.so.0.9.7. There is extended info in Configure for
15246 - Make as few rebuilds of the shared libraries as possible.
15248 - Still avoid linking the OpenSSL programs with the shared libraries.
15250 - When installing, install the shared libraries separately from the
15255 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15257 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15258 and not in SSL_clear because the latter is also used by the
15259 accept/connect functions; previously, the settings made by
15260 SSL_set_read_ahead would be lost during the handshake.
15262 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15264 * Correct util/mkdef.pl to be selective about disabled algorithms.
15265 Previously, it would create entries for disabled algorithms no
15270 * Added several new manual pages for SSL_* function.
15274 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15276 * In ssl23_get_client_hello, generate an error message when faced
15277 with an initial SSL 3.0/TLS record that is too small to contain the
15278 first two bytes of the ClientHello message, i.e. client_version.
15279 (Note that this is a pathologic case that probably has never happened
15280 in real life.) The previous approach was to use the version number
15281 from the record header as a substitute; but our protocol choice
15282 should not depend on that one because it is not authenticated
15283 by the Finished messages.
15287 * More robust randomness gathering functions for Windows.
15289 *Jeffrey Altman <jaltman@columbia.edu>*
15291 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15292 not set then we don't setup the error code for issuer check errors
15293 to avoid possibly overwriting other errors which the callback does
15294 handle. If an application does set the flag then we assume it knows
15295 what it is doing and can handle the new informational codes
15300 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15301 a general "ANY" type, as such it should be able to decode anything
15302 including tagged types. However it didn't check the class so it would
15303 wrongly interpret tagged types in the same way as their universal
15304 counterpart and unknown types were just rejected. Changed so that the
15305 tagged and unknown types are handled in the same way as a SEQUENCE:
15306 that is the encoding is stored intact. There is also a new type
15307 "V_ASN1_OTHER" which is used when the class is not universal, in this
15308 case we have no idea what the actual type is so we just lump them all
15313 * On VMS, stdout may very well lead to a file that is written to
15314 in a record-oriented fashion. That means that every write() will
15315 write a separate record, which will be read separately by the
15316 programs trying to read from it. This can be very confusing.
15318 The solution is to put a BIO filter in the way that will buffer
15319 text until a linefeed is reached, and then write everything a
15320 line at a time, so every record written will be an actual line,
15321 not chunks of lines and not (usually doesn't happen, but I've
15322 seen it once) several lines in one record. BIO_f_linebuffer() is
15325 Currently, it's a VMS-only method, because that's where it has
15326 been tested well enough.
15330 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15331 it can return incorrect results.
15332 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15333 but it was in 0.9.6-beta[12].)
15337 * Disable the check for content being present when verifying detached
15338 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15339 include zero length content when signing messages.
15343 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15344 BIO_ctrl (for BIO pairs).
15348 * Add DSO method for VMS.
15352 * Bug fix: Montgomery multiplication could produce results with the
15357 * Add RPM specification openssl.spec and modify it to build three
15358 packages. The default package contains applications, application
15359 documentation and run-time libraries. The devel package contains
15360 include files, static libraries and function documentation. The
15361 doc package contains the contents of the doc directory. The original
15362 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15366 * Add a large number of documentation files for many SSL routines.
15368 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15370 * Add a configuration entry for Sony News 4.
15372 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15374 * Don't set the two most significant bits to one when generating a
15375 random number < q in the DSA library.
15379 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15380 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15381 the underlying transport is blocking) if a handshake took place.
15382 (The default behaviour is needed by applications such as s_client
15383 and s_server that use select() to determine when to use SSL_read;
15384 but for applications that know in advance when to expect data, it
15385 just makes things more complicated.)
15389 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15394 * Add a few more EBCDIC conditionals that make `req` and `x509`
15395 work better on such systems.
15397 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15399 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15400 Update PKCS12_parse() so it copies the friendlyName and the
15401 keyid to the certificates aux info.
15405 * Fix bug in PKCS7_verify() which caused an infinite loop
15406 if there was more than one signature.
15408 *Sven Uszpelkat <su@celocom.de>*
15410 * Major change in util/mkdef.pl to include extra information
15411 about each symbol, as well as presenting variables as well
15412 as functions. This change means that there's n more need
15413 to rebuild the .num files when some algorithms are excluded.
15417 * Allow the verify time to be set by an application,
15418 rather than always using the current time.
15422 * Phase 2 verify code reorganisation. The certificate
15423 verify code now looks up an issuer certificate by a
15424 number of criteria: subject name, authority key id
15425 and key usage. It also verifies self signed certificates
15426 by the same criteria. The main comparison function is
15427 X509_check_issued() which performs these checks.
15429 Lot of changes were necessary in order to support this
15430 without completely rewriting the lookup code.
15432 Authority and subject key identifier are now cached.
15434 The LHASH 'certs' is X509_STORE has now been replaced
15435 by a STACK_OF(X509_OBJECT). This is mainly because an
15436 LHASH can't store or retrieve multiple objects with
15437 the same hash value.
15439 As a result various functions (which were all internal
15440 use only) have changed to handle the new X509_STORE
15441 structure. This will break anything that messed round
15442 with X509_STORE internally.
15444 The functions X509_STORE_add_cert() now checks for an
15445 exact match, rather than just subject name.
15447 The X509_STORE API doesn't directly support the retrieval
15448 of multiple certificates matching a given criteria, however
15449 this can be worked round by performing a lookup first
15450 (which will fill the cache with candidate certificates)
15451 and then examining the cache for matches. This is probably
15452 the best we can do without throwing out X509_LOOKUP
15453 entirely (maybe later...).
15455 The X509_VERIFY_CTX structure has been enhanced considerably.
15457 All certificate lookup operations now go via a get_issuer()
15458 callback. Although this currently uses an X509_STORE it
15459 can be replaced by custom lookups. This is a simple way
15460 to bypass the X509_STORE hackery necessary to make this
15461 work and makes it possible to use more efficient techniques
15462 in future. A very simple version which uses a simple
15463 STACK for its trusted certificate store is also provided
15464 using X509_STORE_CTX_trusted_stack().
15466 The verify_cb() and verify() callbacks now have equivalents
15467 in the X509_STORE_CTX structure.
15469 X509_STORE_CTX also has a 'flags' field which can be used
15470 to customise the verify behaviour.
15474 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15475 excludes S/MIME capabilities.
15479 * When a certificate request is read in keep a copy of the
15480 original encoding of the signed data and use it when outputting
15481 again. Signatures then use the original encoding rather than
15482 a decoded, encoded version which may cause problems if the
15483 request is improperly encoded.
15487 * For consistency with other BIO_puts implementations, call
15488 buffer_write(b, ...) directly in buffer_puts instead of calling
15491 In BIO_puts, increment b->num_write as in BIO_write.
15493 *Peter.Sylvester@EdelWeb.fr*
15495 * Fix BN_mul_word for the case where the word is 0. (We have to use
15496 BN_zero, we may not return a BIGNUM with an array consisting of
15497 words set to zero.)
15501 * Avoid calling abort() from within the library when problems are
15502 detected, except if preprocessor symbols have been defined
15503 (such as REF_CHECK, BN_DEBUG etc.).
15507 * New openssl application 'rsautl'. This utility can be
15508 used for low-level RSA operations. DER public key
15509 BIO/fp routines also added.
15513 * New Configure entry and patches for compiling on QNX 4.
15515 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15517 * A demo state-machine implementation was sponsored by
15518 Nuron (<http://www.nuron.com/>) and is now available in
15519 demos/state_machine.
15523 * New options added to the 'dgst' utility for signature
15524 generation and verification.
15528 * Unrecognized PKCS#7 content types are now handled via a
15529 catch all ASN1_TYPE structure. This allows unsupported
15530 types to be stored as a "blob" and an application can
15531 encode and decode it manually.
15535 * Fix various signed/unsigned issues to make a_strex.c
15536 compile under VC++.
15538 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15540 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15541 length if passed a buffer. ASN1_INTEGER_to_BN failed
15542 if passed a NULL BN and its argument was negative.
15544 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15546 * Modification to PKCS#7 encoding routines to output definite
15547 length encoding. Since currently the whole structures are in
15548 memory there's not real point in using indefinite length
15549 constructed encoding. However if OpenSSL is compiled with
15550 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15554 * Added BIO_vprintf() and BIO_vsnprintf().
15558 * Added more prefixes to parse for in the strings written
15559 through a logging bio, to cover all the levels that are available
15560 through syslog. The prefixes are now:
15562 PANIC, EMERG, EMR => LOG_EMERG
15563 ALERT, ALR => LOG_ALERT
15564 CRIT, CRI => LOG_CRIT
15565 ERROR, ERR => LOG_ERR
15566 WARNING, WARN, WAR => LOG_WARNING
15567 NOTICE, NOTE, NOT => LOG_NOTICE
15568 INFO, INF => LOG_INFO
15569 DEBUG, DBG => LOG_DEBUG
15571 and as before, if none of those prefixes are present at the
15572 beginning of the string, LOG_ERR is chosen.
15574 On Win32, the `LOG_*` levels are mapped according to this:
15576 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15577 LOG_WARNING => EVENTLOG_WARNING_TYPE
15578 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15582 * Made it possible to reconfigure with just the configuration
15583 argument "reconf" or "reconfigure". The command line arguments
15584 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15585 and are retrieved from there when reconfiguring.
15591 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15593 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15597 * The obj_dat.pl script was messing up the sorting of object
15598 names. The reason was that it compared the quoted version
15599 of strings as a result "OCSP" > "OCSP Signing" because
15600 " > SPACE. Changed script to store unquoted versions of
15601 names and add quotes on output. It was also omitting some
15602 names from the lookup table if they were given a default
15603 value (that is if SN is missing it is given the same
15604 value as LN and vice versa), these are now added on the
15605 grounds that if an object has a name we should be able to
15606 look it up. Finally added warning output when duplicate
15607 short or long names are found.
15611 * Changes needed for Tandem NSK.
15613 *Scott Uroff <scott@xypro.com>*
15615 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15616 RSA_padding_check_SSLv23(), special padding was never detected
15617 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15618 version rollback attacks was not effective.
15620 In s23_clnt.c, don't use special rollback-attack detection padding
15621 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15622 client; similarly, in s23_srvr.c, don't do the rollback check if
15623 SSL 2.0 is the only protocol enabled in the server.
15627 * Make it possible to get hexdumps of unprintable data with 'openssl
15628 asn1parse'. By implication, the functions ASN1_parse_dump() and
15629 BIO_dump_indent() are added.
15633 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15634 these print out strings and name structures based on various
15635 flags including RFC2253 support and proper handling of
15636 multibyte characters. Added options to the 'x509' utility
15637 to allow the various flags to be set.
15641 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15642 Also change the functions X509_cmp_current_time() and
15643 X509_gmtime_adj() work with an ASN1_TIME structure,
15644 this will enable certificates using GeneralizedTime in validity
15645 dates to be checked.
15649 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15650 negative public key encodings) on by default,
15651 NO_NEG_PUBKEY_BUG can be set to disable it.
15655 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15656 content octets. An i2c_ASN1_OBJECT is unnecessary because
15657 the encoding can be trivially obtained from the structure.
15661 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15662 not read locks (`CRYPTO_r_[un]lock`).
15666 * A first attempt at creating official support for shared
15667 libraries through configuration. I've kept it so the
15668 default is static libraries only, and the OpenSSL programs
15669 are always statically linked for now, but there are
15670 preparations for dynamic linking in place.
15671 This has been tested on Linux and Tru64.
15675 * Randomness polling function for Win9x, as described in:
15676 Peter Gutmann, Software Generation of Practically Strong
15681 * Fix so PRNG is seeded in req if using an already existing
15686 * New options to smime application. -inform and -outform
15687 allow alternative formats for the S/MIME message including
15688 PEM and DER. The -content option allows the content to be
15689 specified separately. This should allow things like Netscape
15690 form signing output easier to verify.
15694 * Fix the ASN1 encoding of tags using the 'long form'.
15698 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15699 STRING types. These convert content octets to and from the
15700 underlying type. The actual tag and length octets are
15701 already assumed to have been read in and checked. These
15702 are needed because all other string types have virtually
15703 identical handling apart from the tag. By having versions
15704 of the ASN1 functions that just operate on content octets
15705 IMPLICIT tagging can be handled properly. It also allows
15706 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15707 and ASN1_INTEGER are identical apart from the tag.
15711 * Change the handling of OID objects as follows:
15713 - New object identifiers are inserted in objects.txt, following
15714 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15715 - objects.pl is used to process obj_mac.num and create a new
15717 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15720 This is currently kind of a hack, and the perl code in objects.pl
15721 isn't very elegant, but it works as I intended. The simplest way
15722 to check that it worked correctly is to look in obj_dat.h and
15723 check the array nid_objs and make sure the objects haven't moved
15724 around (this is important!). Additions are OK, as well as
15725 consistent name changes.
15729 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15733 * Addition of the command line parameter '-rand file' to 'openssl req'.
15734 The given file adds to whatever has already been seeded into the
15735 random pool through the RANDFILE configuration file option or
15736 environment variable, or the default random state file.
15740 * mkstack.pl now sorts each macro group into lexical order.
15741 Previously the output order depended on the order the files
15742 appeared in the directory, resulting in needless rewriting
15747 * Patches to make OpenSSL compile under Win32 again. Mostly
15748 work arounds for the VC++ problem that it treats func() as
15749 func(void). Also stripped out the parts of mkdef.pl that
15750 added extra typesafe functions: these no longer exist.
15754 * Reorganisation of the stack code. The macros are now all
15755 collected in safestack.h . Each macro is defined in terms of
15756 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15757 DEBUG_SAFESTACK is now handled in terms of function casts,
15758 this has the advantage of retaining type safety without the
15759 use of additional functions. If DEBUG_SAFESTACK is not defined
15760 then the non typesafe macros are used instead. Also modified the
15761 mkstack.pl script to handle the new form. Needs testing to see
15762 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15763 the default if no major problems. Similar behaviour for ASN1_SET_OF
15764 and PKCS12_STACK_OF.
15768 * When some versions of IIS use the 'NET' form of private key the
15769 key derivation algorithm is different. Normally MD5(password) is
15770 used as a 128 bit RC4 key. In the modified case
15771 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15772 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15773 as the old Netscape_RSA functions except they have an additional
15774 'sgckey' parameter which uses the modified algorithm. Also added
15775 an -sgckey command line option to the rsa utility. Thanks to
15776 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15777 algorithm to openssl-dev.
15781 * The evp_local.h macros were using 'c.##kname' which resulted in
15782 invalid expansion on some systems (SCO 5.0.5 for example).
15783 Corrected to 'c.kname'.
15785 *Phillip Porch <root@theporch.com>*
15787 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15788 a STACK of email addresses from a certificate or request, these look
15789 in the subject name and the subject alternative name extensions and
15790 omit any duplicate addresses.
15794 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15795 This makes DSA verification about 2 % faster.
15799 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15800 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15801 plus overhead for 1024 bit moduli).
15802 This makes exponentiations about 0.5 % faster for 1024 bit
15803 exponents (as measured by "openssl speed rsa2048").
15807 * Rename memory handling macros to avoid conflicts with other
15809 Malloc => OPENSSL_malloc
15810 Malloc_locked => OPENSSL_malloc_locked
15811 Realloc => OPENSSL_realloc
15812 Free => OPENSSL_free
15816 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15817 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15821 * CygWin32 support.
15823 *John Jarvie <jjarvie@newsguy.com>*
15825 * The type-safe stack code has been rejigged. It is now only compiled
15826 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15827 by default all type-specific stack functions are "#define"d back to
15828 standard stack functions. This results in more streamlined output
15829 but retains the type-safety checking possibilities of the original
15834 * The STACK code has been cleaned up, and certain type declarations
15835 that didn't make a lot of sense have been brought in line. This has
15836 also involved a cleanup of sorts in safestack.h to more correctly
15837 map type-safe stack functions onto their plain stack counterparts.
15838 This work has also resulted in a variety of "const"ifications of
15839 lots of the code, especially `_cmp` operations which should normally
15840 be prototyped with "const" parameters anyway.
15844 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15845 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15846 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15847 where all of 'md' is used each time the PRNG is used, but 'state'
15848 is used only indexed by a cyclic counter. As entropy may not be
15849 well distributed from the beginning, 'md' is important as a
15850 chaining variable. However, the output function chains only half
15851 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15852 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15853 in all of 'state' being rewritten, with the new values depending
15854 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15858 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15859 the handshake is continued after ssl_verify_cert_chain();
15860 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15861 can lead to 'unexplainable' connection aborts later.
15863 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15865 * Major EVP API cipher revision.
15866 Add hooks for extra EVP features. This allows various cipher
15867 parameters to be set in the EVP interface. Support added for variable
15868 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15869 setting of RC2 and RC5 parameters.
15871 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15874 Remove lots of duplicated code from the EVP library. For example *every*
15875 cipher init() function handles the 'iv' in the same way according to the
15876 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15877 for CFB and OFB modes they zero ctx->num.
15879 New functionality allows removal of S/MIME code RC2 hack.
15881 Most of the routines have the same form and so can be declared in terms
15884 By shifting this to the top level EVP_CipherInit() it can be removed from
15885 all individual ciphers. If the cipher wants to handle IVs or keys
15886 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15889 Change lots of functions like EVP_EncryptUpdate() to now return a
15890 value: although software versions of the algorithms cannot fail
15891 any installed hardware versions can.
15895 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15896 this option is set, tolerate broken clients that send the negotiated
15897 protocol version number instead of the requested protocol version
15902 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15903 i.e. non-zero for export ciphersuites, zero otherwise.
15904 Previous versions had this flag inverted, inconsistent with
15905 rsa_tmp_cb (..._TMP_RSA_CB).
15907 *Bodo Moeller; problem reported by Amit Chopra*
15909 * Add missing DSA library text string. Work around for some IIS
15910 key files with invalid SEQUENCE encoding.
15914 * Add a document (doc/standards.txt) that list all kinds of standards
15915 and so on that are implemented in OpenSSL.
15919 * Enhance c_rehash script. Old version would mishandle certificates
15920 with the same subject name hash and wouldn't handle CRLs at all.
15921 Added -fingerprint option to crl utility, to support new c_rehash
15926 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15930 * Fix for SSL server purpose checking. Server checking was
15931 rejecting certificates which had extended key usage present
15932 but no ssl client purpose.
15934 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15936 * Make PKCS#12 code work with no password. The PKCS#12 spec
15937 is a little unclear about how a blank password is handled.
15938 Since the password in encoded as a BMPString with terminating
15939 double NULL a zero length password would end up as just the
15940 double NULL. However no password at all is different and is
15941 handled differently in the PKCS#12 key generation code. NS
15942 treats a blank password as zero length. MSIE treats it as no
15943 password on export: but it will try both on import. We now do
15944 the same: PKCS12_parse() tries zero length and no password if
15945 the password is set to "" or NULL (NULL is now a valid password:
15946 it wasn't before) as does the pkcs12 application.
15950 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15951 perror when PEM_read_bio_X509_REQ fails, the error message must
15952 be obtained from the error queue.
15956 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15957 it in ERR_remove_state if appropriate, and change ERR_get_state
15958 accordingly to avoid race conditions (this is necessary because
15959 thread_hash is no longer constant once set).
15963 * Bugfix for linux-elf makefile.one.
15967 * RSA_get_default_method() will now cause a default
15968 RSA_METHOD to be chosen if one doesn't exist already.
15969 Previously this was only set during a call to RSA_new()
15970 or RSA_new_method(NULL) meaning it was possible for
15971 RSA_get_default_method() to return NULL.
15975 * Added native name translation to the existing DSO code
15976 that will convert (if the flag to do so is set) filenames
15977 that are sufficiently small and have no path information
15978 into a canonical native form. Eg. "blah" converted to
15979 "libblah.so" or "blah.dll" etc.
15983 * New function ERR_error_string_n(e, buf, len) which is like
15984 ERR_error_string(e, buf), but writes at most 'len' bytes
15985 including the 0 terminator. For ERR_error_string_n, 'buf'
15988 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15990 * CONF library reworked to become more general. A new CONF
15991 configuration file reader "class" is implemented as well as a
15992 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15993 old `CONF_*` functions are still there, but are reimplemented to
15994 work in terms of the new functions. Also, a set of functions
15995 to handle the internal storage of the configuration data is
15996 provided to make it easier to write new configuration file
15997 reader "classes" (I can definitely see something reading a
15998 configuration file in XML format, for example), called `_CONF_*`,
15999 or "the configuration storage API"...
16001 The new configuration file reading functions are:
16003 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16004 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16006 NCONF_default, NCONF_WIN32
16008 NCONF_dump_fp, NCONF_dump_bio
16010 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16011 NCONF_new creates a new CONF object. This works in the same way
16012 as other interfaces in OpenSSL, like the BIO interface.
16013 `NCONF_dump_*` dump the internal storage of the configuration file,
16014 which is useful for debugging. All other functions take the same
16015 arguments as the old `CONF_*` functions with the exception of the
16016 first that must be a `CONF *` instead of a `LHASH *`.
16018 To make it easier to use the new classes with the old `CONF_*` functions,
16019 the function CONF_set_default_method is provided.
16023 * Add '-tls1' option to 'openssl ciphers', which was already
16024 mentioned in the documentation but had not been implemented.
16025 (This option is not yet really useful because even the additional
16026 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16030 * Initial DSO code added into libcrypto for letting OpenSSL (and
16031 OpenSSL-based applications) load shared libraries and bind to
16032 them in a portable way.
16034 *Geoff Thorpe, with contributions from Richard Levitte*
16036 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16038 * Make sure _lrotl and _lrotr are only used with MSVC.
16040 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16041 (the default implementation of RAND_status).
16043 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16044 to '-clrext' (= clear extensions), as intended and documented.
16045 *Bodo Moeller; inconsistency pointed out by Michael Attili
16046 <attili@amaxo.com>*
16048 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16049 was larger than the MD block size.
16051 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16053 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16054 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16055 using the passed key: if the passed key was a private key the result
16056 of X509_print(), for example, would be to print out all the private key
16061 * des_quad_cksum() byte order bug fix.
16062 *Ulf Möller, using the problem description in krb4-0.9.7, where
16063 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16065 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16068 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16070 * For easily testing in shell scripts whether some command
16071 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16072 returns with exit code 0 iff no command of the given name is available.
16073 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16074 the output goes to stdout and nothing is printed to stderr.
16075 Additional arguments are always ignored.
16077 Since for each cipher there is a command of the same name,
16078 the 'no-cipher' compilation switches can be tested this way.
16080 ('openssl no-XXX' is not able to detect pseudo-commands such
16081 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16085 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16089 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16090 is set; it will be thrown away anyway because each handshake creates
16092 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16093 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16094 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16095 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16099 * New s_client option -ign_eof: EOF at stdin is ignored, and
16100 'Q' and 'R' lose their special meanings (quit/renegotiate).
16101 This is part of what -quiet does; unlike -quiet, -ign_eof
16102 does not suppress any output.
16106 * Add compatibility options to the purpose and trust code. The
16107 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16108 accepts a certificate or CA, this was the previous behaviour,
16109 with all the associated security issues.
16111 X509_TRUST_COMPAT is the old trust behaviour: only and
16112 automatically trust self signed roots in certificate store. A
16113 new trust setting X509_TRUST_DEFAULT is used to specify that
16114 a purpose has no associated trust setting and it should instead
16115 use the value in the default purpose.
16119 * Fix the PKCS#8 DSA private key code so it decodes keys again
16120 and fix a memory leak.
16124 * In util/mkerr.pl (which implements 'make errors'), preserve
16125 reason strings from the previous version of the .c file, as
16126 the default to have only downcase letters (and digits) in
16127 automatically generated reasons codes is not always appropriate.
16131 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16132 using strerror. Previously, ERR_reason_error_string() returned
16133 library names as reason strings for SYSerr; but SYSerr is a special
16134 case where small numbers are errno values, not library numbers.
16138 * Add '-dsaparam' option to 'openssl dhparam' application. This
16139 converts DSA parameters into DH parameters. (When creating parameters,
16140 DSA_generate_parameters is used.)
16144 * Include 'length' (recommended exponent length) in C code generated
16145 by 'openssl dhparam -C'.
16149 * The second argument to set_label in perlasm was already being used
16150 so couldn't be used as a "file scope" flag. Moved to third argument
16155 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16156 instead of RAND_bytes for encryption IVs and salts.
16160 * Include RAND_status() into RAND_METHOD instead of implementing
16161 it only for md_rand.c Otherwise replacing the PRNG by calling
16162 RAND_set_rand_method would be impossible.
16166 * Don't let DSA_generate_key() enter an infinite loop if the random
16167 number generation fails.
16171 * New 'rand' application for creating pseudo-random output.
16175 * Added configuration support for Linux/IA64
16177 *Rolf Haberrecker <rolf@suse.de>*
16179 * Assembler module support for Mingw32.
16183 * Shared library support for HPUX (in shlib/).
16185 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16187 * Shared library support for Solaris gcc.
16189 *Lutz Behnke <behnke@trustcenter.de>*
16191 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16193 * PKCS7_encrypt() was adding text MIME headers twice because they
16194 were added manually and by SMIME_crlf_copy().
16198 * In bntest.c don't call BN_rand with zero bits argument.
16200 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16202 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16203 case was implemented. This caused BN_div_recp() to fail occasionally.
16207 * Add an optional second argument to the set_label() in the perl
16208 assembly language builder. If this argument exists and is set
16209 to 1 it signals that the assembler should use a symbol whose
16210 scope is the entire file, not just the current function. This
16211 is needed with MASM which uses the format label:: for this scope.
16213 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16215 * Change the ASN1 types so they are typedefs by default. Before
16216 almost all types were #define'd to ASN1_STRING which was causing
16217 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16222 * Change names of new functions to the new get1/get0 naming
16223 convention: After 'get1', the caller owns a reference count
16224 and has to call `..._free`; 'get0' returns a pointer to some
16225 data structure without incrementing reference counters.
16226 (Some of the existing 'get' functions increment a reference
16227 counter, some don't.)
16228 Similarly, 'set1' and 'add1' functions increase reference
16229 counters or duplicate objects.
16233 * Allow for the possibility of temp RSA key generation failure:
16234 the code used to assume it always worked and crashed on failure.
16238 * Fix potential buffer overrun problem in BIO_printf().
16239 *Ulf Möller, using public domain code by Patrick Powell; problem
16240 pointed out by David Sacerdote <das33@cornell.edu>*
16242 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16243 RAND_egd() and RAND_status(). In the command line application,
16244 the EGD socket can be specified like a seed file using RANDFILE
16249 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16250 Some CAs (e.g. Verisign) distribute certificates in this form.
16254 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16255 list to exclude them. This means that no special compilation option
16256 is needed to use anonymous DH: it just needs to be included in the
16261 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16262 EVP_MD_type. The old functionality is available in a new macro called
16263 EVP_MD_md(). Change code that uses it and update docs.
16267 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16268 where the `void *` argument is replaced by a function pointer argument.
16269 Previously `void *` was abused to point to functions, which works on
16270 many platforms, but is not correct. As these functions are usually
16271 called by macros defined in OpenSSL header files, most source code
16272 should work without changes.
16276 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16277 sections with information on -D... compiler switches used for
16278 compiling the library so that applications can see them. To enable
16279 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16280 must be defined. E.g.,
16281 #define OPENSSL_ALGORITHM_DEFINES
16282 #include <openssl/opensslconf.h>
16283 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16285 *Richard Levitte, Ulf and Bodo Möller*
16287 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16292 * Change the 'other' type in certificate aux info to a STACK_OF
16293 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16294 the required ASN1 format: arbitrary types determined by an OID.
16298 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16299 argument to 'req'. This is not because the function is newer or
16300 better than others it just uses the work 'NEW' in the certificate
16301 request header lines. Some software needs this.
16305 * Reorganise password command line arguments: now passwords can be
16306 obtained from various sources. Delete the PEM_cb function and make
16307 it the default behaviour: i.e. if the callback is NULL and the
16308 usrdata argument is not NULL interpret it as a null terminated pass
16309 phrase. If usrdata and the callback are NULL then the pass phrase
16310 is prompted for as usual.
16314 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16315 the support is automatically enabled. The resulting binaries will
16316 autodetect the card and use it if present.
16318 *Ben Laurie and Compaq Inc.*
16320 * Work around for Netscape hang bug. This sends certificate request
16321 and server done in one record. Since this is perfectly legal in the
16322 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16323 the bugs/SSLv3 entry for more info.
16327 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16331 * Add -rand argument to smime and pkcs12 applications and read/write
16336 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16340 * Add command line password options to the remaining applications.
16344 * Bug fix for BN_div_recp() for numerators with an even number of
16349 * More tests in bntest.c, and changed test_bn output.
16353 * ./config recognizes MacOS X now.
16357 * Bug fix for BN_div() when the first words of num and divisor are
16358 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16362 * Add support for various broken PKCS#8 formats, and command line
16363 options to produce them.
16367 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16368 get temporary BIGNUMs from a BN_CTX.
16372 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16377 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16378 include a #define from the old name to the new. The original intent
16379 was that statically linked binaries could for example just call
16380 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16381 link with digests. This never worked because SSLeay_add_all_digests()
16382 and SSLeay_add_all_ciphers() were in the same source file so calling
16383 one would link with the other. They are now in separate source files.
16387 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16391 * Use a less unusual form of the Miller-Rabin primality test (it used
16392 a binary algorithm for exponentiation integrated into the Miller-Rabin
16393 loop, our standard modexp algorithms are faster).
16397 * Support for the EBCDIC character set completed.
16399 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16401 * Source code cleanups: use const where appropriate, eliminate casts,
16402 use `void *` instead of `char *` in lhash.
16406 * Bugfix: ssl3_send_server_key_exchange was not restartable
16407 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16408 this the server could overwrite ephemeral keys that the client
16413 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16414 using 50 iterations of the Rabin-Miller test.
16416 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16417 iterations of the Rabin-Miller test as required by the appendix
16418 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16419 As BN_is_prime_fasttest includes trial division, DSA parameter
16420 generation becomes much faster.
16422 This implies a change for the callback functions in DSA_is_prime
16423 and DSA_generate_parameters: The callback function is called once
16424 for each positive witness in the Rabin-Miller test, not just
16425 occasionally in the inner loop; and the parameters to the
16426 callback function now provide an iteration count for the outer
16427 loop rather than for the current invocation of the inner loop.
16428 DSA_generate_parameters additionally can call the callback
16429 function with an 'iteration count' of -1, meaning that a
16430 candidate has passed the trial division test (when q is generated
16431 from an application-provided seed, trial division is skipped).
16435 * New function BN_is_prime_fasttest that optionally does trial
16436 division before starting the Rabin-Miller test and has
16437 an additional BN_CTX * argument (whereas BN_is_prime always
16438 has to allocate at least one BN_CTX).
16439 'callback(1, -1, cb_arg)' is called when a number has passed the
16440 trial division stage.
16444 * Fix for bug in CRL encoding. The validity dates weren't being handled
16449 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16453 * New function BN_pseudo_rand().
16457 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16458 bignum version of BN_from_montgomery() with the working code from
16459 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16464 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16465 made it impossible to use the same SSL_SESSION data structure in
16466 SSL2 clients in multiple threads.
16470 * The return value of RAND_load_file() no longer counts bytes obtained
16471 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16472 to seed the PRNG (previously an explicit byte count was required).
16474 *Ulf Möller, Bodo Möller*
16476 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16477 used `char *` instead of `void *` and had casts all over the place.
16481 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16485 * Retain source code compatibility for BN_prime_checks macro:
16486 BN_is_prime(..., BN_prime_checks, ...) now uses
16487 BN_prime_checks_for_size to determine the appropriate number of
16488 Rabin-Miller iterations.
16492 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16493 DH_CHECK_P_NOT_SAFE_PRIME.
16494 (Check if this is true? OpenPGP calls them "strong".)
16498 * Merge the functionality of "dh" and "gendh" programs into a new program
16499 "dhparam". The old programs are retained for now but will handle DH keys
16500 (instead of parameters) in future.
16504 * Make the ciphers, s_server and s_client programs check the return values
16505 when a new cipher list is set.
16509 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16510 ciphers. Before when the 56bit ciphers were enabled the sorting was
16513 The syntax for the cipher sorting has been extended to support sorting by
16514 cipher-strength (using the strength_bits hard coded in the tables).
16515 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16517 Fix a bug in the cipher-command parser: when supplying a cipher command
16518 string with an "undefined" symbol (neither command nor alphanumeric
16519 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16520 an error is flagged.
16522 Due to the strength-sorting extension, the code of the
16523 ssl_create_cipher_list() function was completely rearranged. I hope that
16524 the readability was also increased :-)
16526 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16528 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16529 for the first serial number and places 2 in the serial number file. This
16530 avoids problems when the root CA is created with serial number zero and
16531 the first user certificate has the same issuer name and serial number
16536 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16537 the new code. Add documentation for this stuff.
16541 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16542 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16543 structures and behave in an analogous way to the X509v3 functions:
16544 they shouldn't be called directly but wrapper functions should be used
16547 So we also now have some wrapper functions that call the X509at functions
16548 when passed certificate requests. (TO DO: similar things can be done with
16549 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16550 things. Some of these need some d2i or i2d and print functionality
16551 because they handle more complex structures.)
16555 * Add missing #ifndefs that caused missing symbols when building libssl
16556 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16557 NO_RSA in `ssl/s2*.c`.
16559 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16561 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16562 has a return value which indicates the quality of the random data
16563 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16564 error queue. New function RAND_pseudo_bytes() generates output that is
16565 guaranteed to be unique but not unpredictable. RAND_add is like
16566 RAND_seed, but takes an extra argument for an entropy estimate
16567 (RAND_seed always assumes full entropy).
16571 * Do more iterations of Rabin-Miller probable prime test (specifically,
16572 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16573 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16574 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16575 false-positive rate of at most 2^-80 for random input.
16579 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16583 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16584 in the 0.9.5 release), this returns the chain
16585 from an X509_CTX structure with a dup of the stack and all
16586 the X509 reference counts upped: so the stack will exist
16587 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16590 Also make SSL_SESSION_print() print out the verify return
16595 * Add manpage for the pkcs12 command. Also change the default
16596 behaviour so MAC iteration counts are used unless the new
16597 -nomaciter option is used. This improves file security and
16598 only older versions of MSIE (4.0 for example) need it.
16602 * Honor the no-xxx Configure options when creating .DEF files.
16606 * Add PKCS#10 attributes to field table: challengePassword,
16607 unstructuredName and unstructuredAddress. These are taken from
16608 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16609 international characters are used.
16611 More changes to X509_ATTRIBUTE code: allow the setting of types
16612 based on strings. Remove the 'loc' parameter when adding
16613 attributes because these will be a SET OF encoding which is sorted
16618 * Initial changes to the 'req' utility to allow request generation
16619 automation. This will allow an application to just generate a template
16620 file containing all the field values and have req construct the
16623 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16624 used all over the place including certificate requests and PKCS#7
16625 structures. They are currently handled manually where necessary with
16626 some primitive wrappers for PKCS#7. The new functions behave in a
16627 manner analogous to the X509 extension functions: they allow
16628 attributes to be looked up by NID and added.
16630 Later something similar to the X509V3 code would be desirable to
16631 automatically handle the encoding, decoding and printing of the
16632 more complex types. The string types like challengePassword can
16633 be handled by the string table functions.
16635 Also modified the multi byte string table handling. Now there is
16636 a 'global mask' which masks out certain types. The table itself
16637 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16638 is useful when for example there is only one permissible type
16639 (as in countryName) and using the mask might result in no valid
16644 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16645 SSL_get_peer_finished to allow applications to obtain the latest
16646 Finished messages sent to the peer or expected from the peer,
16647 respectively. (SSL_get_peer_finished is usually the Finished message
16648 actually received from the peer, otherwise the protocol will be aborted.)
16650 As the Finished message are message digests of the complete handshake
16651 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16652 be used for external authentication procedures when the authentication
16653 provided by SSL/TLS is not desired or is not enough.
16657 * Enhanced support for Alpha Linux is added. Now ./config checks if
16658 the host supports BWX extension and if Compaq C is present on the
16659 $PATH. Just exploiting of the BWX extension results in 20-30%
16660 performance kick for some algorithms, e.g. DES and RC4 to mention
16661 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16666 * Add support for MS "fast SGC". This is arguably a violation of the
16667 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16668 weak crypto and after checking the certificate is SGC a second one
16669 with strong crypto. MS SGC stops the first handshake after receiving
16670 the server certificate message and sends a second client hello. Since
16671 a server will typically do all the time consuming operations before
16672 expecting any further messages from the client (server key exchange
16673 is the most expensive) there is little difference between the two.
16675 To get OpenSSL to support MS SGC we have to permit a second client
16676 hello message after we have sent server done. In addition we have to
16677 reset the MAC if we do get this second client hello.
16681 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16682 if a DER encoded private key is RSA or DSA traditional format. Changed
16683 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16684 format DER encoded private key. Newer code should use PKCS#8 format which
16685 has the key type encoded in the ASN1 structure. Added DER private key
16686 support to pkcs8 application.
16690 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16691 ciphersuites has been selected (as required by the SSL 3/TLS 1
16692 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16693 is set, we interpret this as a request to violate the specification
16694 (the worst that can happen is a handshake failure, and 'correct'
16695 behaviour would result in a handshake failure anyway).
16699 * In SSL_CTX_add_session, take into account that there might be multiple
16700 SSL_SESSION structures with the same session ID (e.g. when two threads
16701 concurrently obtain them from an external cache).
16702 The internal cache can handle only one SSL_SESSION with a given ID,
16703 so if there's a conflict, we now throw out the old one to achieve
16708 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16709 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16710 some routines that use cipher OIDs: some ciphers do not have OIDs
16711 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16716 * Simplify the trust setting structure and code. Now we just have
16717 two sequences of OIDs for trusted and rejected settings. These will
16718 typically have values the same as the extended key usage extension
16719 and any application specific purposes.
16721 The trust checking code now has a default behaviour: it will just
16722 check for an object with the same NID as the passed id. Functions can
16723 be provided to override either the default behaviour or the behaviour
16724 for a given id. SSL client, server and email already have functions
16725 in place for compatibility: they check the NID and also return "trusted"
16726 if the certificate is self signed.
16730 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16731 traditional format into an EVP_PKEY structure.
16735 * Add a password callback function PEM_cb() which either prompts for
16736 a password if usr_data is NULL or otherwise assumes it is a null
16737 terminated password. Allow passwords to be passed on command line
16738 environment or config files in a few more utilities.
16742 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16743 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16744 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16745 Update documentation.
16749 * Support for ASN1 "NULL" type. This could be handled before by using
16750 ASN1_TYPE but there wasn't any function that would try to read a NULL
16751 and produce an error if it couldn't. For compatibility we also have
16752 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16753 don't allocate anything because they don't need to.
16757 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16760 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16762 * Rebuild of the memory allocation routines used by OpenSSL code and
16763 possibly others as well. The purpose is to make an interface that
16764 provide hooks so anyone can build a separate set of allocation and
16765 deallocation routines to be used by OpenSSL, for example memory
16766 pool implementations, or something else, which was previously hard
16767 since Malloc(), Realloc() and Free() were defined as macros having
16768 the values malloc, realloc and free, respectively (except for Win32
16769 compilations). The same is provided for memory debugging code.
16770 OpenSSL already comes with functionality to find memory leaks, but
16771 this gives people a chance to debug other memory problems.
16773 With these changes, a new set of functions and macros have appeared:
16775 CRYPTO_set_mem_debug_functions() [F]
16776 CRYPTO_get_mem_debug_functions() [F]
16777 CRYPTO_dbg_set_options() [F]
16778 CRYPTO_dbg_get_options() [F]
16779 CRYPTO_malloc_debug_init() [M]
16781 The memory debug functions are NULL by default, unless the library
16782 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16783 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16784 gives the standard debugging functions that come with OpenSSL) or
16785 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16786 provided by the library user) must be used. When the standard
16787 debugging functions are used, CRYPTO_dbg_set_options can be used to
16788 request additional information:
16789 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16790 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16792 Also, things like CRYPTO_set_mem_functions will always give the
16793 expected result (the new set of functions is used for allocation
16794 and deallocation) at all times, regardless of platform and compiler
16797 To finish it up, some functions that were never use in any other
16798 way than through macros have a new API and new semantic:
16800 CRYPTO_dbg_malloc()
16801 CRYPTO_dbg_realloc()
16804 All macros of value have retained their old syntax.
16806 *Richard Levitte and Bodo Moeller*
16808 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16809 ordering of SMIMECapabilities wasn't in "strength order" and there
16810 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16815 * Some ASN1 types with illegal zero length encoding (INTEGER,
16816 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16818 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16820 * Merge in my S/MIME library for OpenSSL. This provides a simple
16821 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16822 functionality to handle multipart/signed properly) and a utility
16823 called 'smime' to call all this stuff. This is based on code I
16824 originally wrote for Celo who have kindly allowed it to be
16825 included in OpenSSL.
16829 * Add variants des_set_key_checked and des_set_key_unchecked of
16830 des_set_key (aka des_key_sched). Global variable des_check_key
16831 decides which of these is called by des_set_key; this way
16832 des_check_key behaves as it always did, but applications and
16833 the library itself, which was buggy for des_check_key == 1,
16834 have a cleaner way to pick the version they need.
16838 * New function PKCS12_newpass() which changes the password of a
16843 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16844 dynamic mix. In both cases the ids can be used as an index into the
16845 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16846 functions so they accept a list of the field values and the
16847 application doesn't need to directly manipulate the X509_TRUST
16852 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16857 * Modify the way the V3 extension code looks up extensions. This now
16858 works in a similar way to the object code: we have some "standard"
16859 extensions in a static table which is searched with OBJ_bsearch()
16860 and the application can add dynamic ones if needed. The file
16861 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16862 updated whenever a new extension is added to the core code and kept
16863 in ext_nid order. There is a simple program 'tabtest.c' which checks
16864 this. New extensions are not added too often so this file can readily
16865 be maintained manually.
16867 There are two big advantages in doing things this way. The extensions
16868 can be looked up immediately and no longer need to be "added" using
16869 X509V3_add_standard_extensions(): this function now does nothing.
16870 Side note: I get *lots* of email saying the extension code doesn't
16871 work because people forget to call this function.
16872 Also no dynamic allocation is done unless new extensions are added:
16873 so if we don't add custom extensions there is no need to call
16874 X509V3_EXT_cleanup().
16878 * Modify enc utility's salting as follows: make salting the default. Add a
16879 magic header, so unsalted files fail gracefully instead of just decrypting
16880 to garbage. This is because not salting is a big security hole, so people
16881 should be discouraged from doing it.
16885 * Fixes and enhancements to the 'x509' utility. It allowed a message
16886 digest to be passed on the command line but it only used this
16887 parameter when signing a certificate. Modified so all relevant
16888 operations are affected by the digest parameter including the
16889 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16890 DSA key was used because it didn't fix the digest.
16894 * Initial certificate chain verify code. Currently tests the untrusted
16895 certificates for consistency with the verify purpose (which is set
16896 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16898 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16899 this is because it will reject chains with invalid extensions whereas
16900 every previous version of OpenSSL and SSLeay made no checks at all.
16902 Trust code: checks the root CA for the relevant trust settings. Trust
16903 settings have an initial value consistent with the verify purpose: e.g.
16904 if the verify purpose is for SSL client use it expects the CA to be
16905 trusted for SSL client use. However the default value can be changed to
16906 permit custom trust settings: one example of this would be to only trust
16907 certificates from a specific "secure" set of CAs.
16909 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16910 which should be used for version portability: especially since the
16911 verify structure is likely to change more often now.
16913 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16914 to set them. If not set then assume SSL clients will verify SSL servers
16917 Two new options to the verify program: -untrusted allows a set of
16918 untrusted certificates to be passed in and -purpose which sets the
16919 intended purpose of the certificate. If a purpose is set then the
16920 new chain verify code is used to check extension consistency.
16924 * Support for the authority information access extension.
16928 * Modify RSA and DSA PEM read routines to transparently handle
16929 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16930 public keys in a format compatible with certificate
16931 SubjectPublicKeyInfo structures. Unfortunately there were already
16932 functions called *_PublicKey_* which used various odd formats so
16933 these are retained for compatibility: however the DSA variants were
16934 never in a public release so they have been deleted. Changed dsa/rsa
16935 utilities to handle the new format: note no releases ever handled public
16936 keys so we should be OK.
16938 The primary motivation for this change is to avoid the same fiasco
16939 that dogs private keys: there are several incompatible private key
16940 formats some of which are standard and some OpenSSL specific and
16941 require various evil hacks to allow partial transparent handling and
16942 even then it doesn't work with DER formats. Given the option anything
16943 other than PKCS#8 should be dumped: but the other formats have to
16944 stay in the name of compatibility.
16946 With public keys and the benefit of hindsight one standard format
16947 is used which works with EVP_PKEY, RSA or DSA structures: though
16948 it clearly returns an error if you try to read the wrong kind of key.
16950 Added a -pubkey option to the 'x509' utility to output the public key.
16951 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16952 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16953 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16954 that do the same as the `EVP_PKEY_assign_*()` except they up the
16955 reference count of the added key (they don't "swallow" the
16960 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16961 CRLs would fail if the file contained no certificates or no CRLs:
16962 added a new function to read in both types and return the number
16963 read: this means that if none are read it will be an error. The
16964 DER versions of the certificate and CRL reader would always fail
16965 because it isn't possible to mix certificates and CRLs in DER format
16966 without choking one or the other routine. Changed this to just read
16967 a certificate: this is the best we can do. Also modified the code
16968 in `apps/verify.c` to take notice of return codes: it was previously
16969 attempting to read in certificates from NULL pointers and ignoring
16970 any errors: this is one reason why the cert and CRL reader seemed
16971 to work. It doesn't check return codes from the default certificate
16972 routines: these may well fail if the certificates aren't installed.
16976 * Code to support otherName option in GeneralName.
16980 * First update to verify code. Change the verify utility
16981 so it warns if it is passed a self signed certificate:
16982 for consistency with the normal behaviour. X509_verify
16983 has been modified to it will now verify a self signed
16984 certificate if *exactly* the same certificate appears
16985 in the store: it was previously impossible to trust a
16986 single self signed certificate. This means that:
16987 openssl verify ss.pem
16988 now gives a warning about a self signed certificate but
16989 openssl verify -CAfile ss.pem ss.pem
16994 * For servers, store verify_result in SSL_SESSION data structure
16995 (and add it to external session representation).
16996 This is needed when client certificate verifications fails,
16997 but an application-provided verification callback (set by
16998 SSL_CTX_set_cert_verify_callback) allows accepting the session
16999 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17000 but returns 1): When the session is reused, we have to set
17001 ssl->verify_result to the appropriate error code to avoid
17004 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17006 * Fix a bug in the new PKCS#7 code: it didn't consider the
17007 case in PKCS7_dataInit() where the signed PKCS7 structure
17008 didn't contain any existing data because it was being created.
17010 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17012 * Add a salt to the key derivation routines in enc.c. This
17013 forms the first 8 bytes of the encrypted file. Also add a
17014 -S option to allow a salt to be input on the command line.
17018 * New function X509_cmp(). Oddly enough there wasn't a function
17019 to compare two certificates. We do this by working out the SHA1
17020 hash and comparing that. X509_cmp() will be needed by the trust
17025 * SSL_get1_session() is like SSL_get_session(), but increments
17026 the reference count in the SSL_SESSION returned.
17028 *Geoff Thorpe <geoff@eu.c2.net>*
17030 * Fix for 'req': it was adding a null to request attributes.
17031 Also change the X509_LOOKUP and X509_INFO code to handle
17032 certificate auxiliary information.
17036 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17041 * Add the possibility to add extra information to the memory leak
17042 detecting output, to form tracebacks, showing from where each
17043 allocation was originated: CRYPTO_push_info("constant string") adds
17044 the string plus current file name and line number to a per-thread
17045 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17046 is like calling CYRPTO_pop_info() until the stack is empty.
17047 Also updated memory leak detection code to be multi-thread-safe.
17051 * Add options -text and -noout to pkcs7 utility and delete the
17052 encryption options which never did anything. Update docs.
17056 * Add options to some of the utilities to allow the pass phrase
17057 to be included on either the command line (not recommended on
17058 OSes like Unix) or read from the environment. Update the
17059 manpages and fix a few bugs.
17063 * Add a few manpages for some of the openssl commands.
17067 * Fix the -revoke option in ca. It was freeing up memory twice,
17068 leaking and not finding already revoked certificates.
17072 * Extensive changes to support certificate auxiliary information.
17073 This involves the use of X509_CERT_AUX structure and X509_AUX
17074 functions. An X509_AUX function such as PEM_read_X509_AUX()
17075 can still read in a certificate file in the usual way but it
17076 will also read in any additional "auxiliary information". By
17077 doing things this way a fair degree of compatibility can be
17078 retained: existing certificates can have this information added
17079 using the new 'x509' options.
17081 Current auxiliary information includes an "alias" and some trust
17082 settings. The trust settings will ultimately be used in enhanced
17083 certificate chain verification routines: currently a certificate
17084 can only be trusted if it is self signed and then it is trusted
17089 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17090 The problem was that one of the replacement routines had not been working
17091 since SSLeay releases. For now the offending routine has been replaced
17092 with non-optimised assembler. Even so, this now gives around 95%
17093 performance improvement for 1024 bit RSA signs.
17097 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17098 handling. Most clients have the effective key size in bits equal to
17099 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17100 A few however don't do this and instead use the size of the decrypted key
17101 to determine the RC2 key length and the AlgorithmIdentifier to determine
17102 the effective key length. In this case the effective key length can still
17103 be 40 bits but the key length can be 168 bits for example. This is fixed
17104 by manually forcing an RC2 key into the EVP_PKEY structure because the
17105 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17106 the key length and effective key length are equal.
17110 * Add a bunch of functions that should simplify the creation of
17111 X509_NAME structures. Now you should be able to do:
17112 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17113 and have it automatically work out the correct field type and fill in
17114 the structures. The more adventurous can try:
17115 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17116 and it will (hopefully) work out the correct multibyte encoding.
17120 * Change the 'req' utility to use the new field handling and multibyte
17121 copy routines. Before the DN field creation was handled in an ad hoc
17122 way in req, ca, and x509 which was rather broken and didn't support
17123 BMPStrings or UTF8Strings. Since some software doesn't implement
17124 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17125 using the dirstring_type option. See the new comment in the default
17126 openssl.cnf for more info.
17130 * Make crypto/rand/md_rand.c more robust:
17131 - Assure unique random numbers after fork().
17132 - Make sure that concurrent threads access the global counter and
17133 md serializably so that we never lose entropy in them
17134 or use exactly the same state in multiple threads.
17135 Access to the large state is not always serializable because
17136 the additional locking could be a performance killer, and
17137 md should be large enough anyway.
17141 * New file `apps/app_rand.c` with commonly needed functionality
17142 for handling the random seed file.
17144 Use the random seed file in some applications that previously did not:
17146 dsaparam -genkey (which also ignored its '-rand' option),
17149 x509 (when signing).
17150 Except on systems with /dev/urandom, it is crucial to have a random
17151 seed file at least for key creation, DSA signing, and for DH exchanges;
17152 for RSA signatures we could do without one.
17154 gendh and gendsa (unlike genrsa) used to read only the first byte
17155 of each file listed in the '-rand' option. The function as previously
17156 found in genrsa is now in app_rand.c and is used by all programs
17157 that support '-rand'.
17161 * In RAND_write_file, use mode 0600 for creating files;
17162 don't just chmod when it may be too late.
17166 * Report an error from X509_STORE_load_locations
17167 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17171 * New function ASN1_mbstring_copy() this copies a string in either
17172 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17173 into an ASN1_STRING type. A mask of permissible types is passed
17174 and it chooses the "minimal" type to use or an error if not type
17179 * Add function equivalents to the various macros in asn1.h. The old
17180 macros are retained with an `M_` prefix. Code inside the library can
17181 use the `M_` macros. External code (including the openssl utility)
17182 should *NOT* in order to be "shared library friendly".
17186 * Add various functions that can check a certificate's extensions
17187 to see if it usable for various purposes such as SSL client,
17188 server or S/MIME and CAs of these types. This is currently
17189 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17190 verification. Also added a -purpose flag to x509 utility to
17191 print out all the purposes.
17195 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17200 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17201 for, obtain and decode and extension and obtain its critical flag.
17202 This allows all the necessary extension code to be handled in a
17203 single function call.
17207 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17208 platforms. See crypto/rc4/rc4_enc.c for further details.
17212 * New -noout option to asn1parse. This causes no output to be produced
17213 its main use is when combined with -strparse and -out to extract data
17214 from a file (which may not be in ASN.1 format).
17218 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17219 when producing the local key id.
17221 *Richard Levitte <levitte@stacken.kth.se>*
17223 * New option -dhparam in s_server. This allows a DH parameter file to be
17224 stated explicitly. If it is not stated then it tries the first server
17225 certificate file. The previous behaviour hard coded the filename
17230 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17231 a public key to be input or output. For example:
17232 openssl rsa -in key.pem -pubout -out pubkey.pem
17233 Also added necessary DSA public key functions to handle this.
17237 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17238 in the message. This was handled by allowing
17239 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17241 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17243 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17244 to the end of the strings whereas this didn't. This would cause problems
17245 if strings read with d2i_ASN1_bytes() were later modified.
17247 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17249 * Fix for base64 decode bug. When a base64 bio reads only one line of
17250 data and it contains EOF it will end up returning an error. This is
17251 caused by input 46 bytes long. The cause is due to the way base64
17252 BIOs find the start of base64 encoded data. They do this by trying a
17253 trial decode on each line until they find one that works. When they
17254 do a flag is set and it starts again knowing it can pass all the
17255 data directly through the decoder. Unfortunately it doesn't reset
17256 the context it uses. This means that if EOF is reached an attempt
17257 is made to pass two EOFs through the context and this causes the
17258 resulting error. This can also cause other problems as well. As is
17259 usual with these problems it takes *ages* to find and the fix is
17260 trivial: move one line.
17262 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17264 * Ugly workaround to get s_client and s_server working under Windows. The
17265 old code wouldn't work because it needed to select() on sockets and the
17266 tty (for keypresses and to see if data could be written). Win32 only
17267 supports select() on sockets so we select() with a 1s timeout on the
17268 sockets and then see if any characters are waiting to be read, if none
17269 are present then we retry, we also assume we can always write data to
17270 the tty. This isn't nice because the code then blocks until we've
17271 received a complete line of data and it is effectively polling the
17272 keyboard at 1s intervals: however it's quite a bit better than not
17273 working at all :-) A dedicated Windows application might handle this
17274 with an event loop for example.
17278 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17279 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17280 will be called when RSA_sign() and RSA_verify() are used. This is useful
17281 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17282 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17283 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17284 This necessitated the support of an extra signature type NID_md5_sha1
17285 for SSL signatures and modifications to the SSL library to use it instead
17286 of calling RSA_public_decrypt() and RSA_private_encrypt().
17290 * Add new -verify -CAfile and -CApath options to the crl program, these
17291 will lookup a CRL issuers certificate and verify the signature in a
17292 similar way to the verify program. Tidy up the crl program so it
17293 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17294 less strict. It will now permit CRL extensions even if it is not
17295 a V2 CRL: this will allow it to tolerate some broken CRLs.
17299 * Initialize all non-automatic variables each time one of the openssl
17300 sub-programs is started (this is necessary as they may be started
17301 multiple times from the "OpenSSL>" prompt).
17303 *Lennart Bang, Bodo Moeller*
17305 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17306 removing all other RSA functionality (this is what NO_RSA does). This
17307 is so (for example) those in the US can disable those operations covered
17308 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17313 * Non-copying interface to BIO pairs.
17314 (still largely untested)
17318 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17319 ASCII string. This was handled independently in various places before.
17323 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17324 UTF8 strings a character at a time.
17328 * Use client_version from client hello to select the protocol
17329 (s23_srvr.c) and for RSA client key exchange verification
17330 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17334 * Add various utility functions to handle SPKACs, these were previously
17335 handled by poking round in the structure internals. Added new function
17336 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17337 print, verify and generate SPKACs. Based on an original idea from
17338 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17342 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17346 * Allow the config file extension section to be overwritten on the
17347 command line. Based on an original idea from Massimiliano Pala
17348 <madwolf@comune.modena.it>. The new option is called -extensions
17349 and can be applied to ca, req and x509. Also -reqexts to override
17350 the request extensions in req and -crlexts to override the crl extensions
17355 * Add new feature to the SPKAC handling in ca. Now you can include
17356 the same field multiple times by preceding it by "XXXX." for example:
17359 this is the same syntax as used in the req config file.
17363 * Allow certificate extensions to be added to certificate requests. These
17364 are specified in a 'req_extensions' option of the req section of the
17365 config file. They can be printed out with the -text option to req but
17366 are otherwise ignored at present.
17370 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17371 data read consists of only the final block it would not decrypted because
17372 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17373 A misplaced 'break' also meant the decrypted final block might not be
17374 copied until the next read.
17378 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17379 a few extra parameters to the DH structure: these will be useful if
17380 for example we want the value of 'q' or implement X9.42 DH.
17384 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17385 provides hooks that allow the default DSA functions or functions on a
17386 "per key" basis to be replaced. This allows hardware acceleration and
17387 hardware key storage to be handled without major modification to the
17388 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17389 associated functions.
17393 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17394 as "read only": it can't be written to and the buffer it points to will
17395 not be freed. Reading from a read only BIO is much more efficient than
17396 a normal memory BIO. This was added because there are several times when
17397 an area of memory needs to be read from a BIO. The previous method was
17398 to create a memory BIO and write the data to it, this results in two
17399 copies of the data and an O(n^2) reading algorithm. There is a new
17400 function BIO_new_mem_buf() which creates a read only memory BIO from
17401 an area of memory. Also modified the PKCS#7 routines to use read only
17406 * Bugfix: ssl23_get_client_hello did not work properly when called in
17407 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17408 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17409 but a retry condition occurred while trying to read the rest.
17413 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17414 NID_pkcs7_encrypted by default: this was wrong since this should almost
17415 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17416 the encrypted data type: this is a more sensible place to put it and it
17417 allows the PKCS#12 code to be tidied up that duplicated this
17422 * Changed obj_dat.pl script so it takes its input and output files on
17423 the command line. This should avoid shell escape redirection problems
17428 * Initial support for certificate extension requests, these are included
17429 in things like Xenroll certificate requests. Included functions to allow
17430 extensions to be obtained and added.
17434 * -crlf option to s_client and s_server for sending newlines as
17435 CRLF (as required by many protocols).
17439 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17441 * Install libRSAglue.a when OpenSSL is built with RSAref.
17443 *Ralf S. Engelschall*
17445 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17447 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17449 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17454 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17455 DH parameters/keys (q is lost during that conversion, but the resulting
17456 DH parameters contain its length).
17458 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17459 much faster than DH_generate_parameters (which creates parameters
17460 where `p = 2*q + 1`), and also the smaller q makes DH computations
17461 much more efficient (160-bit exponentiation instead of 1024-bit
17462 exponentiation); so this provides a convenient way to support DHE
17463 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17464 utter importance to use
17465 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17467 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17468 when such DH parameters are used, because otherwise small subgroup
17469 attacks may become possible!
17473 * Avoid memory leak in i2d_DHparams.
17477 * Allow the -k option to be used more than once in the enc program:
17478 this allows the same encrypted message to be read by multiple recipients.
17482 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17483 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17484 it will always use the numerical form of the OID, even if it has a short
17489 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17490 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17491 otherwise bn_mod_exp was called. In the case of hardware keys for example
17492 no private key components need be present and it might store extra data
17493 in the RSA structure, which cannot be accessed from bn_mod_exp.
17494 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17495 private key operations.
17499 * Added support for SPARC Linux.
17503 * pem_password_cb function type incompatibly changed from
17504 typedef int pem_password_cb(char *buf, int size, int rwflag);
17506 ....(char *buf, int size, int rwflag, void *userdata);
17507 so that applications can pass data to their callbacks:
17508 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17509 additional void * argument, which is just handed through whenever
17510 the password callback is called.
17512 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17514 New function SSL_CTX_set_default_passwd_cb_userdata.
17516 Compatibility note: As many C implementations push function arguments
17517 onto the stack in reverse order, the new library version is likely to
17518 interoperate with programs that have been compiled with the old
17519 pem_password_cb definition (PEM_whatever takes some data that
17520 happens to be on the stack as its last argument, and the callback
17521 just ignores this garbage); but there is no guarantee whatsoever that
17524 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17525 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17526 problems not only on Windows, but also on some Unix platforms.
17527 To avoid problematic command lines, these definitions are now in an
17528 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17529 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17533 * MIPS III/IV assembler module is reimplemented.
17537 * More DES library cleanups: remove references to srand/rand and
17538 delete an unused file.
17542 * Add support for the free Netwide assembler (NASM) under Win32,
17543 since not many people have MASM (ml) and it can be hard to obtain.
17544 This is currently experimental but it seems to work OK and pass all
17545 the tests. Check out INSTALL.W32 for info.
17549 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17550 without temporary keys kept an extra copy of the server key,
17551 and connections with temporary keys did not free everything in case
17556 * New function RSA_check_key and new openssl rsa option -check
17557 for verifying the consistency of RSA keys.
17559 *Ulf Moeller, Bodo Moeller*
17561 * Various changes to make Win32 compile work:
17562 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17563 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17564 comparison" warnings.
17565 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17569 * Add a debugging option to PKCS#5 v2 key generation function: when
17570 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17571 derived keys are printed to stderr.
17575 * Copy the flags in ASN1_STRING_dup().
17577 *Roman E. Pavlov <pre@mo.msk.ru>*
17579 * The x509 application mishandled signing requests containing DSA
17580 keys when the signing key was also DSA and the parameters didn't match.
17582 It was supposed to omit the parameters when they matched the signing key:
17583 the verifying software was then supposed to automatically use the CA's
17584 parameters if they were absent from the end user certificate.
17586 Omitting parameters is no longer recommended. The test was also
17587 the wrong way round! This was probably due to unusual behaviour in
17588 EVP_cmp_parameters() which returns 1 if the parameters match.
17589 This meant that parameters were omitted when they *didn't* match and
17590 the certificate was useless. Certificates signed with 'ca' didn't have
17593 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17595 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17596 The interface is as follows:
17597 Applications can use
17598 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17599 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17600 "off" is now the default.
17601 The library internally uses
17602 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17603 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17604 to disable memory-checking temporarily.
17606 Some inconsistent states that previously were possible (and were
17607 even the default) are now avoided.
17609 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17610 with each memory chunk allocated; this is occasionally more helpful
17611 than just having a counter.
17613 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17615 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17620 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17621 which largely parallels "options", but is for changing API behaviour,
17622 whereas "options" are about protocol behaviour.
17623 Initial "mode" flags are:
17625 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17626 a single record has been written.
17627 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17628 retries use the same buffer location.
17629 (But all of the contents must be
17634 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17637 * Fix problems with no-hmac etc.
17639 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17641 * New functions RSA_get_default_method(), RSA_set_method() and
17642 RSA_get_method(). These allows replacement of RSA_METHODs without having
17643 to mess around with the internals of an RSA structure.
17647 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17648 Also really enable memory leak checks in openssl.c and in some
17651 *Chad C. Mulligan, Bodo Moeller*
17653 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17654 up the length of negative integers. This has now been simplified to just
17655 store the length when it is first determined and use it later, rather
17656 than trying to keep track of where data is copied and updating it to
17658 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17660 * Add a new function PKCS7_signatureVerify. This allows the verification
17661 of a PKCS#7 signature but with the signing certificate passed to the
17662 function itself. This contrasts with PKCS7_dataVerify which assumes the
17663 certificate is present in the PKCS#7 structure. This isn't always the
17664 case: certificates can be omitted from a PKCS#7 structure and be
17665 distributed by "out of band" means (such as a certificate database).
17669 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17670 function prototypes in pem.h, also change util/mkdef.pl to add the
17671 necessary function names.
17675 * mk1mf.pl (used by Windows builds) did not properly read the
17676 options set by Configure in the top level Makefile, and Configure
17677 was not even able to write more than one option correctly.
17678 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17682 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17683 file to be loaded from a BIO or FILE pointer. The BIO version will
17684 for example allow memory BIOs to contain config info.
17688 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17689 Whoever hopes to achieve shared-library compatibility across versions
17690 must use this, not the compile-time macro.
17691 (Exercise 0.9.4: Which is the minimum library version required by
17693 Note: All this applies only to multi-threaded programs, others don't
17698 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17699 through a BIO pair triggered the default case, i.e.
17700 SSLerr(...,SSL_R_UNKNOWN_STATE).
17704 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17705 can use the SSL library even if none of the specific BIOs is
17710 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17711 for the encoded length.
17713 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17715 * Add initial documentation of the X509V3 functions.
17719 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17720 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17721 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17722 secure PKCS#8 private key format with a high iteration count.
17726 * Fix determination of Perl interpreter: A perl or perl5
17727 *directory* in $PATH was also accepted as the interpreter.
17729 *Ralf S. Engelschall*
17731 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17732 wrong with it but it was very old and did things like calling
17733 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17734 unusual formatting.
17738 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17739 to use the new extension code.
17743 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17744 with macros. This should make it easier to change their form, add extra
17745 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17750 * Add to configuration table a new entry that can specify an alternative
17751 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17752 according to Mark Crispin <MRC@Panda.COM>.
17756 * DES CBC did not update the IV. Weird.
17760 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17761 Changing the behaviour of the former might break existing programs --
17762 where IV updating is needed, des_ncbc_encrypt can be used.
17765 * When bntest is run from "make test" it drives bc to check its
17766 calculations, as well as internally checking them. If an internal check
17767 fails, it needs to cause bc to give a non-zero result or make test carries
17768 on without noticing the failure. Fixed.
17772 * DES library cleanups.
17776 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17777 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17778 ciphers. NOTE: although the key derivation function has been verified
17779 against some published test vectors it has not been extensively tested
17780 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17785 * Instead of "mkdir -p", which is not fully portable, use new
17786 Perl script "util/mkdir-p.pl".
17790 * Rewrite the way password based encryption (PBE) is handled. It used to
17791 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17792 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17793 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17794 the 'parameter' field of the AlgorithmIdentifier is passed to the
17795 underlying key generation function so it must do its own ASN1 parsing.
17796 This has also changed the EVP_PBE_CipherInit() function which now has a
17797 'parameter' argument instead of literal salt and iteration count values
17798 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17802 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17803 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17804 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17805 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17806 value was just used as a "magic string" and not used directly its
17807 value doesn't matter.
17811 * Introduce some semblance of const correctness to BN. Shame C doesn't
17816 * "linux-sparc64" configuration (ultrapenguin).
17818 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17819 "linux-sparc" configuration.
17821 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17823 * config now generates no-xxx options for missing ciphers.
17827 * Support the EBCDIC character set (work in progress).
17828 File ebcdic.c not yet included because it has a different license.
17830 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17832 * Support BS2000/OSD-POSIX.
17834 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17836 * Make callbacks for key generation use `void *` instead of `char *`.
17840 * Make S/MIME samples compile (not yet tested).
17844 * Additional typesafe stacks.
17848 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17852 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17854 * New configuration variant "sco5-gcc".
17856 * Updated some demos.
17858 *Sean O Riordain, Wade Scholine*
17860 * Add missing BIO_free at exit of pkcs12 application.
17864 * Fix memory leak in conf.c.
17868 * Updates for Win32 to assembler version of MD5.
17872 * Set #! path to perl in `apps/der_chop` to where we found it
17873 instead of using a fixed path.
17877 * SHA library changes for irix64-mips4-cc.
17881 * Improvements for VMS support.
17885 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17887 * Bignum library bug fix. IRIX 6 passes "make test" now!
17888 This also avoids the problems with SC4.2 and unpatched SC5.
17890 *Andy Polyakov <appro@fy.chalmers.se>*
17892 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17893 These are required because of the typesafe stack would otherwise break
17894 existing code. If old code used a structure member which used to be STACK
17895 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17896 sk_num or sk_value it would produce an error because the num, data members
17897 are not present in STACK_OF. Now it just produces a warning. sk_set
17898 replaces the old method of assigning a value to sk_value
17899 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17900 that does this will no longer work (and should use sk_set instead) but
17901 this could be regarded as a "questionable" behaviour anyway.
17905 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17906 correctly handle encrypted S/MIME data.
17910 * Change type of various DES function arguments from des_cblock
17911 (which means, in function argument declarations, pointer to char)
17912 to des_cblock * (meaning pointer to array with 8 char elements),
17913 which allows the compiler to do more typechecking; it was like
17914 that back in SSLeay, but with lots of ugly casts.
17916 Introduce new type const_des_cblock.
17920 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17921 problems: find RecipientInfo structure that matches recipient certificate
17922 and initialise the ASN1 structures properly based on passed cipher.
17926 * Belatedly make the BN tests actually check the results.
17930 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17931 to and from BNs: it was completely broken. New compilation option
17932 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17933 key elements as negative integers.
17937 * Reorganize and speed up MD5.
17939 *Andy Polyakov <appro@fy.chalmers.se>*
17943 *Richard Levitte <richard@levitte.org>*
17945 * New option -out to asn1parse to allow the parsed structure to be
17946 output to a file. This is most useful when combined with the -strparse
17947 option to examine the output of things like OCTET STRINGS.
17951 * Make SSL library a little more fool-proof by not requiring any longer
17952 that `SSL_set_{accept,connect}_state` be called before
17953 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17954 in many applications because usually everything *appeared* to work as
17955 intended anyway -- now it really works as intended).
17959 * Move openssl.cnf out of lib/.
17963 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17964 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17965 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17967 *Ralf S. Engelschall*
17969 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17970 handle PKCS#7 enveloped data properly.
17972 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17974 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17975 copying pointers. The cert_st handling is changed by this in
17976 various ways (and thus what used to be known as ctx->default_cert
17977 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17978 any longer when s->cert does not give us what we need).
17979 ssl_cert_instantiate becomes obsolete by this change.
17980 As soon as we've got the new code right (possibly it already is?),
17981 we have solved a couple of bugs of the earlier code where s->cert
17982 was used as if it could not have been shared with other SSL structures.
17984 Note that using the SSL API in certain dirty ways now will result
17985 in different behaviour than observed with earlier library versions:
17986 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17987 does not influence s as it used to.
17989 In order to clean up things more thoroughly, inside SSL_SESSION
17990 we don't use CERT any longer, but a new structure SESS_CERT
17991 that holds per-session data (if available); currently, this is
17992 the peer's certificate chain and, for clients, the server's certificate
17993 and temporary key. CERT holds only those values that can have
17994 meaningful defaults in an SSL_CTX.
17998 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17999 from the internal representation. Various PKCS#7 fixes: remove some
18000 evil casts and set the enc_dig_alg field properly based on the signing
18005 * Allow PKCS#12 password to be set from the command line or the
18006 environment. Let 'ca' get its config file name from the environment
18007 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18012 * Allow certificate policies extension to use an IA5STRING for the
18013 organization field. This is contrary to the PKIX definition but
18014 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18019 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18020 without disallowing inline assembler and the like for non-pedantic builds.
18024 * Support Borland C++ builder.
18026 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18032 * SHA-1 cleanups and performance enhancements.
18034 *Andy Polyakov <appro@fy.chalmers.se>*
18036 * Sparc v8plus assembler for the bignum library.
18038 *Andy Polyakov <appro@fy.chalmers.se>*
18040 * Accept any -xxx and +xxx compiler options in Configure.
18044 * Update HPUX configuration.
18048 * Add missing `sk_<type>_unshift()` function to safestack.h
18050 *Ralf S. Engelschall*
18052 * New function SSL_CTX_use_certificate_chain_file that sets the
18053 "extra_cert"s in addition to the certificate. (This makes sense
18054 only for "PEM" format files, as chains as a whole are not
18059 * Support verify_depth from the SSL API.
18060 x509_vfy.c had what can be considered an off-by-one-error:
18061 Its depth (which was not part of the external interface)
18062 was actually counting the number of certificates in a chain;
18063 now it really counts the depth.
18067 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18068 instead of X509err, which often resulted in confusing error
18069 messages since the error codes are not globally unique
18070 (e.g. an alleged error in ssl3_accept when a certificate
18071 didn't match the private key).
18073 * New function SSL_CTX_set_session_id_context that allows to set a default
18074 value (so that you don't need SSL_set_session_id_context for each
18075 connection using the SSL_CTX).
18079 * OAEP decoding bug fix.
18083 * Support INSTALL_PREFIX for package builders, as proposed by
18088 * New Configure options "threads" and "no-threads". For systems
18089 where the proper compiler options are known (currently Solaris
18090 and Linux), "threads" is the default.
18094 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18098 * Install various scripts to $(OPENSSLDIR)/misc, not to
18099 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18100 such as /usr/local/bin.
18104 * "make linux-shared" to build shared libraries.
18106 *Niels Poppe <niels@netbox.org>*
18108 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18112 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18113 extension adding in x509 utility.
18117 * Remove NOPROTO sections and error code comments.
18121 * Partial rewrite of the DEF file generator to now parse the ANSI
18126 * New Configure options --prefix=DIR and --openssldir=DIR.
18130 * Complete rewrite of the error code script(s). It is all now handled
18131 by one script at the top level which handles error code gathering,
18132 header rewriting and C source file generation. It should be much better
18133 than the old method: it now uses a modified version of Ulf's parser to
18134 read the ANSI prototypes in all header files (thus the old K&R definitions
18135 aren't needed for error creation any more) and do a better job of
18136 translating function codes into names. The old 'ASN1 error code embedded
18137 in a comment' is no longer necessary and it doesn't use .err files which
18138 have now been deleted. Also the error code call doesn't have to appear all
18139 on one line (which resulted in some large lines...).
18143 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18147 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18148 0 (which usually indicates a closed connection), but continue reading.
18152 * Fix some race conditions.
18156 * Add support for CRL distribution points extension. Add Certificate
18157 Policies and CRL distribution points documentation.
18161 * Move the autogenerated header file parts to crypto/opensslconf.h.
18165 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18166 8 of keying material. Merlin has also confirmed interop with this fix
18167 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18169 *Merlin Hughes <merlin@baltimore.ie>*
18171 * Fix lots of warnings.
18173 *Richard Levitte <levitte@stacken.kth.se>*
18175 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18176 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18178 *Richard Levitte <levitte@stacken.kth.se>*
18180 * Fix problems with sizeof(long) == 8.
18182 *Andy Polyakov <appro@fy.chalmers.se>*
18184 * Change functions to ANSI C.
18188 * Fix typos in error codes.
18190 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18192 * Remove defunct assembler files from Configure.
18196 * SPARC v8 assembler BIGNUM implementation.
18198 *Andy Polyakov <appro@fy.chalmers.se>*
18200 * Support for Certificate Policies extension: both print and set.
18201 Various additions to support the r2i method this uses.
18205 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18206 return a const string when you are expecting an allocated buffer.
18210 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18211 types DirectoryString and DisplayText.
18215 * Add code to allow r2i extensions to access the configuration database,
18216 add an LHASH database driver and add several ctx helper functions.
18220 * Fix an evil bug in bn_expand2() which caused various BN functions to
18221 fail when they extended the size of a BIGNUM.
18225 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18226 support typesafe stack.
18230 * Fix typo in SSL_[gs]et_options().
18232 *Nils Frostberg <nils@medcom.se>*
18234 * Delete various functions and files that belonged to the (now obsolete)
18235 old X509V3 handling code.
18239 * New Configure option "rsaref".
18243 * Don't auto-generate pem.h.
18247 * Introduce type-safe ASN.1 SETs.
18251 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18253 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18255 * Introduce type-safe STACKs. This will almost certainly break lots of code
18256 that links with OpenSSL (well at least cause lots of warnings), but fear
18257 not: the conversion is trivial, and it eliminates loads of evil casts. A
18258 few STACKed things have been converted already. Feel free to convert more.
18259 In the fullness of time, I'll do away with the STACK type altogether.
18263 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18264 specified in `<certfile>` by updating the entry in the index.txt file.
18265 This way one no longer has to edit the index.txt file manually for
18266 revoking a certificate. The -revoke option does the gory details now.
18268 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18270 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18271 `-text` option at all and this way the `-noout -text` combination was
18272 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18274 *Ralf S. Engelschall*
18276 * Make sure a corresponding plain text error message exists for the
18277 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18278 verify callback function determined that a certificate was revoked.
18280 *Ralf S. Engelschall*
18282 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18283 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18284 all available ciphers including rc5, which was forgotten until now.
18285 In order to let the testing shell script know which algorithms
18286 are available, a new (up to now undocumented) command
18287 `openssl list-cipher-commands` is used.
18291 * Bugfix: s_client occasionally would sleep in select() when
18292 it should have checked SSL_pending() first.
18296 * New functions DSA_do_sign and DSA_do_verify to provide access to
18297 the raw DSA values prior to ASN.1 encoding.
18301 * Tweaks to Configure
18303 *Niels Poppe <niels@netbox.org>*
18305 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18310 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18314 * New config option to avoid instructions that are illegal on the 80386.
18315 The default code is faster, but requires at least a 486.
18319 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18320 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18321 same as SSL2_VERSION anyway.
18325 * New "-showcerts" option for s_client.
18329 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18330 application. Various cleanups and fixes.
18334 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18335 modify error routines to work internally. Add error codes and PBE init
18336 to library startup routines.
18340 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18341 packing functions to asn1 and evp. Changed function names and error
18342 codes along the way.
18346 * PKCS12 integration: and so it begins... First of several patches to
18347 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18348 objects to objects.h
18352 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18353 and display support for Thawte strong extranet extension.
18357 * Add LinuxPPC support.
18359 *Jeff Dubrule <igor@pobox.org>*
18361 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18362 bn_div_words in alpha.s.
18364 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18366 * Make sure the RSA OAEP test is skipped under -DRSAref because
18367 OAEP isn't supported when OpenSSL is built with RSAref.
18369 *Ulf Moeller <ulf@fitug.de>*
18371 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18372 so they no longer are missing under -DNOPROTO.
18374 *Soren S. Jorvang <soren@t.dk>*
18376 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18378 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18379 doesn't work when the session is reused. Coming soon!
18383 * Fix a security hole, that allows sessions to be reused in the wrong
18384 context thus bypassing client cert protection! All software that uses
18385 client certs and session caches in multiple contexts NEEDS PATCHING to
18386 allow session reuse! A fuller solution is in the works.
18388 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18390 * Some more source tree cleanups (removed obsolete files
18391 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18392 permission on "config" script to be executable) and a fix for the INSTALL
18395 *Ulf Moeller <ulf@fitug.de>*
18397 * Remove some legacy and erroneous uses of malloc, free instead of
18400 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18402 * Make rsa_oaep_test return non-zero on error.
18404 *Ulf Moeller <ulf@fitug.de>*
18406 * Add support for native Solaris shared libraries. Configure
18407 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18408 if someone would make that last step automatic.
18410 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18412 * ctx_size was not built with the right compiler during "make links". Fixed.
18416 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18417 except NULL ciphers". This means the default cipher list will no longer
18418 enable NULL ciphers. They need to be specifically enabled e.g. with
18419 the string "DEFAULT:eNULL".
18423 * Fix to RSA private encryption routines: if p < q then it would
18424 occasionally produce an invalid result. This will only happen with
18425 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18429 * Be less restrictive and allow also `perl util/perlpath.pl
18430 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18431 because this way one can also use an interpreter named `perl5` (which is
18432 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18433 installed as `perl`).
18435 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18437 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18439 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18441 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18442 advapi32.lib to Win32 build and change the pem test comparison
18443 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18444 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18445 and crypto/des/ede_cbcm_enc.c.
18449 * DES quad checksum was broken on big-endian architectures. Fixed.
18453 * Comment out two functions in bio.h that aren't implemented. Fix up the
18454 Win32 test batch file so it (might) work again. The Win32 test batch file
18455 is horrible: I feel ill....
18459 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18460 in e_os.h. Audit of header files to check ANSI and non ANSI
18461 sections: 10 functions were absent from non ANSI section and not exported
18462 from Windows DLLs. Fixed up libeay.num for new functions.
18466 * Make `openssl version` output lines consistent.
18468 *Ralf S. Engelschall*
18470 * Fix Win32 symbol export lists for BIO functions: Added
18471 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18472 to ms/libeay{16,32}.def.
18474 *Ralf S. Engelschall*
18476 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18477 fine under Unix and passes some trivial tests I've now added. But the
18478 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18479 added to make sure no one expects that this stuff really works in the
18480 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18481 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18484 *Ralf S. Engelschall*
18486 * Fix the generation of two part addresses in perl.
18488 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18490 * Add config entry for Linux on MIPS.
18492 *John Tobey <jtobey@channel1.com>*
18494 * Make links whenever Configure is run, unless we are on Windoze.
18498 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18499 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18504 * Add a useful kludge to allow package maintainers to specify compiler and
18505 other platforms details on the command line without having to patch the
18506 Configure script every time: One now can use
18507 `perl Configure <id>:<details>`,
18508 i.e. platform ids are allowed to have details appended
18509 to them (separated by colons). This is treated as there would be a static
18510 pre-configured entry in Configure's %table under key `<id>` with value
18511 `<details>` and `perl Configure <id>` is called. So, when you want to
18512 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18513 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18514 now, which overrides the FreeBSD-elf entry on-the-fly.
18516 *Ralf S. Engelschall*
18518 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18522 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18523 on the `perl Configure ...` command line. This way one can compile
18524 OpenSSL libraries with Position Independent Code (PIC) which is needed
18525 for linking it into DSOs.
18527 *Ralf S. Engelschall*
18529 * Remarkably, export ciphers were totally broken and no-one had noticed!
18534 * Cleaned up the LICENSE document: The official contact for any license
18535 questions now is the OpenSSL core team under openssl-core@openssl.org.
18536 And add a paragraph about the dual-license situation to make sure people
18537 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18538 to the OpenSSL toolkit.
18540 *Ralf S. Engelschall*
18542 * General source tree makefile cleanups: Made `making xxx in yyy...`
18543 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18544 Additionally cleaned up the `make links` target: Remove unnecessary
18545 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18546 to speed processing and no longer clutter the display with confusing
18547 stuff. Instead only the actually done links are displayed.
18549 *Ralf S. Engelschall*
18551 * Permit null encryption ciphersuites, used for authentication only. It used
18552 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18553 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18558 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18559 signed attributes when verifying signatures (this would break them),
18560 the detached data encoding was wrong and public keys obtained using
18561 X509_get_pubkey() weren't freed.
18565 * Add text documentation for the BUFFER functions. Also added a work around
18566 to a Win95 console bug. This was triggered by the password read stuff: the
18567 last character typed gets carried over to the next fread(). If you were
18568 generating a new cert request using 'req' for example then the last
18569 character of the passphrase would be CR which would then enter the first
18574 * Added the new 'Includes OpenSSL Cryptography Software' button as
18575 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18576 button and can be used by applications based on OpenSSL to show the
18577 relationship to the OpenSSL project.
18579 *Ralf S. Engelschall*
18581 * Remove confusing variables in function signatures in files
18582 ssl/ssl_lib.c and ssl/ssl.h.
18584 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18586 * Don't install bss_file.c under PREFIX/include/
18588 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18590 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18591 functions that return function pointers and has support for NT specific
18592 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18593 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18594 unsigned to signed types: this was killing the Win32 compile.
18598 * Add new certificate file to stack functions,
18599 SSL_add_dir_cert_subjects_to_stack() and
18600 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18601 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18602 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18603 This means that Apache-SSL and similar packages don't have to mess around
18604 to add as many CAs as they want to the preferred list.
18608 * Experiment with doxygen documentation. Currently only partially applied to
18610 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18611 openssl.doxy as the configuration file.
18615 * Get rid of remaining C++-style comments which strict C compilers hate.
18617 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18619 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18620 compiled in by default: it has problems with large keys.
18624 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18625 DH private keys and/or callback functions which directly correspond to
18626 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18627 is needed for applications which have to configure certificates on a
18628 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18630 For the RSA certificate situation is makes no difference, but
18631 for the DSA certificate situation this fixes the "no shared cipher"
18632 problem where the OpenSSL cipher selection procedure failed because the
18633 temporary keys were not overtaken from the context and the API provided
18634 no way to reconfigure them.
18635 The new functions now let applications reconfigure the stuff and they
18636 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18637 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18638 non-public-API function ssl_cert_instantiate() is used as a helper
18639 function and also to reduce code redundancy inside ssl_rsa.c.
18641 *Ralf S. Engelschall*
18643 * Move s_server -dcert and -dkey options out of the undocumented feature
18644 area because they are useful for the DSA situation and should be
18645 recognized by the users.
18647 *Ralf S. Engelschall*
18649 * Fix the cipher decision scheme for export ciphers: the export bits are
18650 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18651 SSL_EXP_MASK. So, the original variable has to be used instead of the
18652 already masked variable.
18654 *Richard Levitte <levitte@stacken.kth.se>*
18656 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18658 *Richard Levitte <levitte@stacken.kth.se>*
18660 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18661 from `int` to `unsigned int` because it is a length and initialized by
18662 EVP_DigestFinal() which expects an `unsigned int *`.
18664 *Richard Levitte <levitte@stacken.kth.se>*
18666 * Don't hard-code path to Perl interpreter on shebang line of Configure
18667 script. Instead use the usual Shell->Perl transition trick.
18669 *Ralf S. Engelschall*
18671 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18672 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18673 -noout -modulus` as it's already the case for `openssl rsa -noout
18674 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18675 currently the public key is printed (a decision which was already done by
18676 `openssl dsa -modulus` in the past) which serves a similar purpose.
18677 Additionally the NO_RSA no longer completely removes the whole -modulus
18678 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18681 *Ralf S. Engelschall*
18683 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18684 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18686 *Arne Ansper <arne@ats.cyber.ee>*
18688 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18689 to be added. Now both 'req' and 'ca' can use new objects defined in the
18694 * Add cool BIO that does syslog (or event log on NT).
18696 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18698 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18699 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18700 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18701 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18705 * Add preliminary config info for new extension code.
18709 * Make RSA_NO_PADDING really use no padding.
18711 *Ulf Moeller <ulf@fitug.de>*
18713 * Generate errors when private/public key check is done.
18717 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18718 for some CRL extensions and new objects added.
18722 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18723 key usage extension and fuller support for authority key id.
18727 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18728 padding method for RSA, which is recommended for new applications in PKCS
18729 #1 v2.0 (RFC 2437, October 1998).
18730 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18731 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18732 against Bleichbacher's attack on RSA.
18733 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18736 * Updates to the new SSL compression code
18738 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18740 * Fix so that the version number in the master secret, when passed
18741 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18742 (because the server will not accept higher), that the version number
18743 is 0x03,0x01, not 0x03,0x00
18745 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18747 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18748 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18749 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18753 * Support for RAW extensions where an arbitrary extension can be
18754 created by including its DER encoding. See `apps/openssl.cnf` for
18759 * Make sure latest Perl versions don't interpret some generated C array
18760 code as Perl array code in the crypto/err/err_genc.pl script.
18762 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18764 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18765 not many people have the assembler. Various Win32 compilation fixes and
18766 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18767 build instructions.
18771 * Modify configure script 'Configure' to automatically create crypto/date.h
18772 file under Win32 and also build pem.h from pem.org. New script
18773 util/mkfiles.pl to create the MINFO file on environments that can't do a
18774 'make files': perl util/mkfiles.pl >MINFO should work.
18778 * Major rework of DES function declarations, in the pursuit of correctness
18779 and purity. As a result, many evil casts evaporated, and some weirdness,
18780 too. You may find this causes warnings in your code. Zapping your evil
18781 casts will probably fix them. Mostly.
18785 * Fix for a typo in asn1.h. Bug fix to object creation script
18786 obj_dat.pl. It considered a zero in an object definition to mean
18787 "end of object": none of the objects in objects.h have any zeros
18788 so it wasn't spotted.
18790 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18792 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18793 Masking (CBCM). In the absence of test vectors, the best I have been able
18794 to do is check that the decrypt undoes the encrypt, so far. Send me test
18795 vectors if you have them.
18799 * Correct calculation of key length for export ciphers (too much space was
18800 allocated for null ciphers). This has not been tested!
18804 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18805 message is now correct (it understands "crypto" and "ssl" on its
18806 command line). There is also now an "update" option. This will update
18807 the util/ssleay.num and util/libeay.num files with any new functions.
18809 perl util/mkdef.pl crypto ssl update
18810 it will update them.
18814 * Overhauled the Perl interface:
18815 - ported BN stuff to OpenSSL's different BN library
18816 - made the perl/ source tree CVS-aware
18817 - renamed the package from SSLeay to OpenSSL (the files still contain
18818 their history because I've copied them in the repository)
18819 - removed obsolete files (the test scripts will be replaced
18820 by better Test::Harness variants in the future)
18822 *Ralf S. Engelschall*
18824 * First cut for a very conservative source tree cleanup:
18825 1. merge various obsolete readme texts into doc/ssleay.txt
18826 where we collect the old documents and readme texts.
18827 2. remove the first part of files where I'm already sure that we no
18828 longer need them because of three reasons: either they are just temporary
18829 files which were left by Eric or they are preserved original files where
18830 I've verified that the diff is also available in the CVS via "cvs diff
18831 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18832 the crypto/md/ stuff).
18834 *Ralf S. Engelschall*
18836 * More extension code. Incomplete support for subject and issuer alt
18837 name, issuer and authority key id. Change the i2v function parameters
18838 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18839 what that's for :-) Fix to ASN1 macro which messed up
18840 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18844 * Preliminary support for ENUMERATED type. This is largely copied from the
18849 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18851 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18853 * Make sure `make rehash` target really finds the `openssl` program.
18855 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18857 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18858 like to hear about it if this slows down other processors.
18862 * Add CygWin32 platform information to Configure script.
18864 *Alan Batie <batie@aahz.jf.intel.com>*
18866 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18868 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18870 * New program nseq to manipulate netscape certificate sequences
18874 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18879 * Fixes to BN code. Previously the default was to define BN_RECURSION
18880 but the BN code had some problems that would cause failures when
18881 doing certificate verification and some other functions.
18883 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18885 * Add ASN1 and PEM code to support netscape certificate sequences.
18889 * Add ASN1 and PEM code to support netscape certificate sequences.
18893 * Add several PKIX and private extended key usage OIDs.
18897 * Modify the 'ca' program to handle the new extension code. Modify
18898 openssl.cnf for new extension format, add comments.
18902 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18903 and add a sample to openssl.cnf so req -x509 now adds appropriate
18908 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18909 error code, add initial support to X509_print() and x509 application.
18913 * Takes a deep breath and start adding X509 V3 extension support code. Add
18914 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18915 stuff is currently isolated and isn't even compiled yet.
18919 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18920 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18921 Removed the versions check from X509 routines when loading extensions:
18922 this allows certain broken certificates that don't set the version
18923 properly to be processed.
18927 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18928 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18929 can still be regenerated with "make depend".
18933 * Spelling mistake in C version of CAST-128.
18935 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18937 * Changes to the error generation code. The perl script err-code.pl
18938 now reads in the old error codes and retains the old numbers, only
18939 adding new ones if necessary. It also only changes the .err files if new
18940 codes are added. The makefiles have been modified to only insert errors
18941 when needed (to avoid needlessly modifying header files). This is done
18942 by only inserting errors if the .err file is newer than the auto generated
18943 C file. To rebuild all the error codes from scratch (the old behaviour)
18944 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18945 or delete all the .err files.
18949 * CAST-128 was incorrectly implemented for short keys. The C version has
18950 been fixed, but is untested. The assembler versions are also fixed, but
18951 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18952 to regenerate it if needed.
18953 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18954 Hagino <itojun@kame.net>*
18956 * File was opened incorrectly in randfile.c.
18958 *Ulf Möller <ulf@fitug.de>*
18960 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18961 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18962 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18963 al: it's just almost always a UTCTime. Note this patch adds new error
18964 codes so do a "make errors" if there are problems.
18968 * Correct Linux 1 recognition in config.
18970 *Ulf Möller <ulf@fitug.de>*
18972 * Remove pointless MD5 hash when using DSA keys in ca.
18974 *Anonymous <nobody@replay.com>*
18976 * Generate an error if given an empty string as a cert directory. Also
18977 generate an error if handed NULL (previously returned 0 to indicate an
18978 error, but didn't set one).
18980 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18982 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18986 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18987 parameters. This was causing a warning which killed off the Win32 compile.
18991 * Remove C++ style comments from crypto/bn/bn_local.h.
18993 *Neil Costigan <neil.costigan@celocom.com>*
18995 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18996 based on a text string, looking up short and long names and finally
18997 "dot" format. The "dot" format stuff didn't work. Added new function
18998 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18999 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19000 OID is not part of the table.
19004 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19005 X509_LOOKUP_by_alias().
19009 * Sort openssl functions by name.
19013 * Get the `gendsa` command working and add it to the `list` command. Remove
19014 encryption from sample DSA keys (in case anyone is interested the password
19019 * Make *all* `*_free` functions accept a NULL pointer.
19021 *Frans Heymans <fheymans@isaserver.be>*
19023 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19026 *Anonymous <nobody@replay.com>*
19028 * s_server should send the CAfile as acceptable CAs, not its own cert.
19030 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19032 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19034 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19036 * Temp key "for export" tests were wrong in s3_srvr.c.
19038 *Anonymous <nobody@replay.com>*
19040 * Add prototype for temp key callback functions
19041 SSL_CTX_set_tmp_{rsa,dh}_callback().
19045 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19046 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19050 * X509_name_add_entry() freed the wrong thing after an error.
19052 *Arne Ansper <arne@ats.cyber.ee>*
19054 * rsa_eay.c would attempt to free a NULL context.
19056 *Arne Ansper <arne@ats.cyber.ee>*
19058 * BIO_s_socket() had a broken should_retry() on Windoze.
19060 *Arne Ansper <arne@ats.cyber.ee>*
19062 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19064 *Arne Ansper <arne@ats.cyber.ee>*
19066 * Make sure the already existing X509_STORE->depth variable is initialized
19067 in X509_STORE_new(), but document the fact that this variable is still
19068 unused in the certificate verification process.
19070 *Ralf S. Engelschall*
19072 * Fix the various library and `apps/` files to free up pkeys obtained from
19073 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19077 * Fix reference counting in X509_PUBKEY_get(). This makes
19078 demos/maurice/example2.c work, amongst others, probably.
19080 *Steve Henson and Ben Laurie*
19082 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19083 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19084 are no longer created. This way we have a single and consistent command
19085 line interface `openssl <command>`, similar to `cvs <command>`.
19087 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19089 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19090 BIT STRING wrapper always have zero unused bits.
19094 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19098 * Make the top-level INSTALL documentation easier to understand.
19102 * Makefiles updated to exit if an error occurs in a sub-directory
19103 make (including if user presses ^C) [Paul Sutton]
19105 * Make Montgomery context stuff explicit in RSA data structure.
19109 * Fix build order of pem and err to allow for generated pem.h.
19113 * Fix renumbering bug in X509_NAME_delete_entry().
19117 * Enhanced the err-ins.pl script so it makes the error library number
19118 global and can add a library name. This is needed for external ASN1 and
19119 other error libraries.
19123 * Fixed sk_insert which never worked properly.
19127 * Fix ASN1 macros so they can handle indefinite length constructed
19128 EXPLICIT tags. Some non standard certificates use these: they can now
19133 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19134 into a single doc/ssleay.txt bundle. This way the information is still
19135 preserved but no longer messes up this directory. Now it's new room for
19136 the new set of documentation files.
19138 *Ralf S. Engelschall*
19140 * SETs were incorrectly DER encoded. This was a major pain, because they
19141 shared code with SEQUENCEs, which aren't coded the same. This means that
19142 almost everything to do with SETs or SEQUENCEs has either changed name or
19143 number of arguments.
19145 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19147 * Fix test data to work with the above.
19151 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19152 was already fixed by Eric for 0.9.1 it seems.
19154 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19156 * Autodetect FreeBSD3.
19160 * Fix various bugs in Configure. This affects the following platforms:
19164 unixware-2.0-pentium
19169 * Eliminate generated files from CVS. Reorder tests to regenerate files
19170 before they are needed.
19174 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19178 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19180 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19181 changed SSLeay to OpenSSL in version strings.
19183 *Ralf S. Engelschall*
19185 * Some fixups to the top-level documents.
19189 * Fixed the nasty bug where rsaref.h was not found under compile-time
19190 because the symlink to include/ was missing.
19192 *Ralf S. Engelschall*
19194 * Incorporated the popular no-RSA/DSA-only patches
19195 which allow to compile a RSA-free SSLeay.
19197 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19199 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19200 when "ssleay" is still not found.
19202 *Ralf S. Engelschall*
19204 * Added more platforms to Configure: Cray T3E, HPUX 11,
19206 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19208 * Updated the README file.
19210 *Ralf S. Engelschall*
19212 * Added various .cvsignore files in the CVS repository subdirs
19213 to make a "cvs update" really silent.
19215 *Ralf S. Engelschall*
19217 * Recompiled the error-definition header files and added
19218 missing symbols to the Win32 linker tables.
19220 *Ralf S. Engelschall*
19222 * Cleaned up the top-level documents;
19223 o new files: CHANGES and LICENSE
19224 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19225 o merged COPYRIGHT into LICENSE
19226 o removed obsolete TODO file
19227 o renamed MICROSOFT to INSTALL.W32
19229 *Ralf S. Engelschall*
19231 * Removed dummy files from the 0.9.1b source tree:
19232 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19233 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19234 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19235 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19236 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19238 *Ralf S. Engelschall*
19240 * Added various platform portability fixes.
19244 * The Genesis of the OpenSSL rpject:
19245 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19246 Young and Tim J. Hudson created while they were working for C2Net until
19249 *The OpenSSL Project*
19251 ### Changes between 0.9.0b and 0.9.1b [not released]
19253 * Updated a few CA certificates under certs/
19257 * Changed some BIGNUM api stuff.
19261 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19262 DGUX x86, Linux Alpha, etc.
19266 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19267 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19272 * Add -strparse option to asn1pars program which parses nested
19275 *Dr Stephen Henson <shenson@bigfoot.com>*
19277 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19281 * DSA fix for "ca" program.
19285 * Added "-genkey" option to "dsaparam" program.
19289 * Added RIPE MD160 (rmd160) message digest.
19293 * Added -a (all) option to "ssleay version" command.
19297 * Added PLATFORM define which is the id given to Configure.
19301 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19305 * Extended the ASN.1 parser routines.
19309 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19313 * Added a BN_CTX to the BN library.
19317 * Fixed the weak key values in DES library
19321 * Changed API in EVP library for cipher aliases.
19325 * Added support for RC2/64bit cipher.
19329 * Converted the lhash library to the crypto/mem.c functions.
19333 * Added more recognized ASN.1 object ids.
19337 * Added more RSA padding checks for SSL/TLS.
19341 * Added BIO proxy/filter functionality.
19345 * Added extra_certs to SSL_CTX which can be used
19346 send extra CA certificates to the client in the CA cert chain sending
19347 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19351 * Now Fortezza is denied in the authentication phase because
19352 this is key exchange mechanism is not supported by SSLeay at all.
19356 * Additional PKCS1 checks.
19360 * Support the string "TLSv1" for all TLS v1 ciphers.
19364 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19365 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19369 * Fixed a few memory leaks.
19373 * Fixed various code and comment typos.
19377 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19378 bytes sent in the client random.
19380 *Edward Bishop <ebishop@spyglass.com>*
19384 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19385 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19386 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19387 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19388 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19389 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19390 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19391 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19392 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19393 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19394 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19395 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19396 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19397 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19398 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19399 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19400 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19401 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19402 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19403 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19404 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19405 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19406 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19407 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19408 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19409 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19410 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19411 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19412 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19413 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19414 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19415 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19416 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19417 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19418 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19419 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19420 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19421 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19422 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19423 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19424 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19425 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19426 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19427 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19428 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19429 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19430 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19431 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19432 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19433 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19434 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19435 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19436 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19437 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19438 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19439 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19440 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19441 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19442 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19443 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19444 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19445 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19446 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19447 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19448 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19449 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19450 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19451 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19452 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19453 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19454 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19455 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19456 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19457 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19458 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19459 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19460 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19461 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19462 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19463 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19464 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19465 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19466 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19467 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19468 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19469 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19470 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19471 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19472 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19473 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19474 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19475 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19476 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19477 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19478 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19479 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19480 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19481 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19482 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19483 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19484 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19485 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19486 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19487 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19488 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19489 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19490 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19491 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19492 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19493 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19494 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19495 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19496 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19497 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19498 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19499 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19500 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19501 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19502 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19503 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19504 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19505 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19506 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19507 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19508 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19509 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19510 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19511 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19512 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19513 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19514 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19515 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19516 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19517 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19518 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19519 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19520 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19521 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19522 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19523 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19524 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19525 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19526 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19527 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19528 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19529 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19530 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19531 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19532 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19533 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19534 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19535 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19536 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19537 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19538 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19539 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19540 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19541 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19542 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19543 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19544 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19545 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19546 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19547 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655