4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
27 * The various OBJ_* functions have been made thread safe.
34 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
35 listed here are only a brief description.
36 The migration guide contains more detailed information related to new features,
37 breaking changes, and mappings for the large list of deprecated functions.
39 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
41 ### Changes between 3.0.6 and 3.0.7 [xx XXX xxxx]
43 * Fixed two buffer overflows in punycode decoding functions.
45 A buffer overrun can be triggered in X.509 certificate verification,
46 specifically in name constraint checking. Note that this occurs after
47 certificate chain signature verification and requires either a CA to
48 have signed the malicious certificate or for the application to continue
49 certificate verification despite failure to construct a path to a trusted
52 In a TLS client, this can be triggered by connecting to a malicious
53 server. In a TLS server, this can be triggered if the server requests
54 client authentication and a malicious client connects.
56 An attacker can craft a malicious email address to overflow
57 an arbitrary number of bytes containing the `.` character (decimal 46)
58 on the stack. This buffer overflow could result in a crash (causing a
62 An attacker can craft a malicious email address to overflow four
63 attacker-controlled bytes on the stack. This buffer overflow could
64 result in a crash (causing a denial of service) or potentially remote code
65 execution depending on stack layout for any given platform/compiler.
70 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
71 parameters in OpenSSL code.
72 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
73 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
74 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
75 Using these invalid names may cause algorithms to use slower methods
76 that ignore the CRT parameters.
80 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
85 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
86 data to be signed before signing the certificate.
90 * Added RIPEMD160 to the default provider.
94 * Ensured that the key share group sent or accepted for the key exchange
95 is allowed for the protocol version.
99 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
101 * OpenSSL supports creating a custom cipher via the legacy
102 EVP_CIPHER_meth_new() function and associated function calls. This function
103 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
104 to use the new provider mechanism in order to implement custom ciphers.
106 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
107 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
108 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
109 and decryption initialisation functions). Instead of using the custom cipher
110 directly it incorrectly tries to fetch an equivalent cipher from the
111 available providers. An equivalent cipher is found based on the NID passed to
112 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
113 given cipher. However it is possible for an application to incorrectly pass
114 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
115 is used in this way the OpenSSL encryption/decryption initialisation function
116 will match the NULL cipher as being equivalent and will fetch this from the
117 available providers. This will succeed if the default provider has been
118 loaded (or if a third party provider has been loaded that offers this
119 cipher). Using the NULL cipher means that the plaintext is emitted as the
122 Applications are only affected by this issue if they call
123 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
124 encryption/decryption initialisation function. Applications that only use
125 SSL/TLS are not impacted by this issue.
130 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
135 * Fixed the linux-mips64 Configure target which was missing the
136 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
141 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
146 * Correctly handle a retransmitted ClientHello in DTLS
150 * Fixed detection of ktls support in cross-compile environment on Linux
154 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
159 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
160 report correct results in some cases
164 * Fix UWP builds by defining VirtualLock
168 * For known safe primes use the minimum key length according to RFC 7919.
169 Longer private key sizes unnecessarily raise the cycles needed to compute the
170 shared secret without any increase of the real security. This fixes a
171 regression from 1.1.1 where these shorter keys were generated for the known
176 * Added the loongarch64 target
180 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
181 only passed to the FIPS provider and not to the default or legacy provider.
185 * Fixed reported performance degradation on aarch64. Restored the
186 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
187 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
188 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
189 The new algorithm is still used for 32 bit targets.
193 * Added a missing header for memcmp that caused compilation failure on some
198 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
200 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
201 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
202 This issue makes the RSA implementation with 2048 bit private keys
203 incorrect on such machines and memory corruption will happen during
204 the computation. As a consequence of the memory corruption an attacker
205 may be able to trigger a remote code execution on the machine performing
208 SSL/TLS servers or other servers using 2048 bit RSA private keys running
209 on machines supporting AVX512IFMA instructions of the X86_64 architecture
210 are affected by this issue.
215 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
216 implementation would not encrypt the entirety of the data under some
217 circumstances. This could reveal sixteen bytes of data that was
218 preexisting in the memory that wasn't written. In the special case of
219 "in place" encryption, sixteen bytes of the plaintext would be revealed.
221 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
222 they are both unaffected.
225 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
227 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
229 * In addition to the c_rehash shell command injection identified in
230 CVE-2022-1292, further bugs where the c_rehash script does not
231 properly sanitise shell metacharacters to prevent command injection have been
234 When the CVE-2022-1292 was fixed it was not discovered that there
235 are other places in the script where the file names of certificates
236 being hashed were possibly passed to a command executed through the shell.
238 This script is distributed by some operating systems in a manner where
239 it is automatically executed. On such operating systems, an attacker
240 could execute arbitrary commands with the privileges of the script.
242 Use of the c_rehash script is considered obsolete and should be replaced
243 by the OpenSSL rehash command line tool.
246 *Daniel Fiala, Tomáš Mráz*
248 * Case insensitive string comparison no longer uses locales. It has instead
249 been directly implemented.
253 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
255 * Case insensitive string comparison is reimplemented via new locale-agnostic
256 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
257 comparison. The previous implementation had problems when the Turkish locale
262 * Fixed a bug in the c_rehash script which was not properly sanitising shell
263 metacharacters to prevent command injection. This script is distributed by
264 some operating systems in a manner where it is automatically executed. On
265 such operating systems, an attacker could execute arbitrary commands with the
266 privileges of the script.
268 Use of the c_rehash script is considered obsolete and should be replaced
269 by the OpenSSL rehash command line tool.
274 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
275 certificate on an OCSP response. The bug caused the function in the case
276 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
277 response (meaning a successful verification) even in the case where the
278 response signing certificate fails to verify.
280 It is anticipated that most users of `OCSP_basic_verify` will not use the
281 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
282 a negative value (indicating a fatal error) in the case of a certificate
283 verification failure. The normal expected return value in this case would be
286 This issue also impacts the command line OpenSSL "ocsp" application. When
287 verifying an ocsp response with the "-no_cert_checks" option the command line
288 application will report that the verification is successful even though it
289 has in fact failed. In this case the incorrect successful response will also
290 be accompanied by error messages showing the failure and contradicting the
291 apparently successful result.
296 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
297 AAD data as the MAC key. This made the MAC key trivially predictable.
299 An attacker could exploit this issue by performing a man-in-the-middle attack
300 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
301 that the modified data would still pass the MAC integrity check.
303 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
304 endpoint will always be rejected by the recipient and the connection will
305 fail at that point. Many application protocols require data to be sent from
306 the client to the server first. Therefore, in such a case, only an OpenSSL
307 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
309 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
310 sent in both directions. In this case both clients and servers could be
311 affected, regardless of the application protocol.
313 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
314 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
315 the handshake when using this ciphersuite.
317 The confidentiality of data is not impacted by this issue, i.e. an attacker
318 cannot decrypt data that has been encrypted using this ciphersuite - they can
321 In order for this attack to work both endpoints must legitimately negotiate
322 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
323 OpenSSL 3.0, and is not available within the default provider or the default
324 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
325 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
326 following must have occurred:
328 1) OpenSSL must have been compiled with the (non-default) compile time option
329 enable-weak-ssl-ciphers
331 2) OpenSSL must have had the legacy provider explicitly loaded (either
332 through application code or via configuration)
334 3) The ciphersuite must have been explicitly added to the ciphersuite list
336 4) The libssl security level must have been set to 0 (default is 1)
338 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
340 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
341 others that both endpoints have in common
346 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
347 occuppied by the removed hash table entries.
349 This function is used when decoding certificates or keys. If a long lived
350 process periodically decodes certificates or keys its memory usage will
351 expand without bounds and the process might be terminated by the operating
352 system causing a denial of service. Also traversing the empty hash table
353 entries will take increasingly more time.
355 Typically such long lived processes might be TLS clients or TLS servers
356 configured to accept client certificate authentication.
359 *Hugo Landau, Aliaksei Levin*
361 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
362 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
363 statistics are no longer supported. For compatibility, these statistics are
364 still listed in the output but are now always reported as zero.
368 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
370 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
371 for non-prime moduli.
373 Internally this function is used when parsing certificates that contain
374 elliptic curve public keys in compressed form or explicit elliptic curve
375 parameters with a base point encoded in compressed form.
377 It is possible to trigger the infinite loop by crafting a certificate that
378 has invalid explicit curve parameters.
380 Since certificate parsing happens prior to verification of the certificate
381 signature, any process that parses an externally supplied certificate may thus
382 be subject to a denial of service attack. The infinite loop can also be
383 reached when parsing crafted private keys as they can contain explicit
384 elliptic curve parameters.
386 Thus vulnerable situations include:
388 - TLS clients consuming server certificates
389 - TLS servers consuming client certificates
390 - Hosting providers taking certificates or private keys from customers
391 - Certificate authorities parsing certification requests from subscribers
392 - Anything else which parses ASN.1 elliptic curve parameters
394 Also any other applications that use the BN_mod_sqrt() where the attacker
395 can control the parameter values are vulnerable to this DoS issue.
400 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
401 to the list of ciphersuites providing Perfect Forward Secrecy as
402 required by SECLEVEL >= 3.
404 *Dmitry Belyavskiy, Nicola Tuveri*
406 * Made the AES constant time code for no-asm configurations
407 optional due to the resulting 95% performance degradation.
408 The AES constant time code can be enabled, for no assembly
409 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
413 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
418 * The negative return value handling of the certificate verification callback
419 was reverted. The replacement is to set the verification retry state with
420 the SSL_set_retry_verify() function.
424 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
426 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
427 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
428 verify a certificate supplied by a server. That function may return a
429 negative return value to indicate an internal error (for example out of
430 memory). Such a negative return value is mishandled by OpenSSL and will cause
431 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
432 success and a subsequent call to SSL_get_error() to return the value
433 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
434 returned by OpenSSL if the application has previously called
435 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
436 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
437 totally unexpected and applications may not behave correctly as a result. The
438 exact behaviour will depend on the application but it could result in
439 crashes, infinite loops or other similar incorrect responses.
441 This issue is made more serious in combination with a separate bug in OpenSSL
442 3.0 that will cause X509_verify_cert() to indicate an internal error when
443 processing a certificate chain. This will occur where a certificate does not
444 include the Subject Alternative Name extension but where a Certificate
445 Authority has enforced name constraints. This issue can occur even with valid
451 * Corrected a few file name and file reference bugs in the build,
452 installation and setup scripts, which lead to installation verification
453 failures. Slightly enhanced the installation verification script.
457 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
462 * Fixed PVK encoder to properly query for the passphrase.
466 * Multiple fixes in the OSSL_HTTP API functions.
470 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
471 OSSL_PARAM_INTEGER data type and return error on negative numbers
472 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
473 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
477 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
481 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
485 * Multiple threading fixes.
489 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
493 * Allow fetching an operation from the provider that owns an unexportable key
494 as a fallback if that is still allowed by the property query.
498 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
500 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
505 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
506 S390X capability vector to zero. This simplifies testing of different code
507 paths on S390X architecture.
511 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
512 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
513 SP 800-38D". The communication will fail at this point.
517 * The EC_GROUP_clear_free() function is deprecated as there is nothing
518 confidential in EC_GROUP data.
522 * The byte order mark (BOM) character is ignored if encountered at the
523 beginning of a PEM-formatted file.
527 * Added CMS support for the Russian GOST algorithms.
531 * Due to move of the implementation of cryptographic operations
532 to the providers, validation of various operation parameters can
533 be postponed until the actual operation is executed where previously
534 it happened immediately when an operation parameter was set.
536 For example when setting an unsupported curve with
537 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
538 fail but later keygen operations with the EVP_PKEY_CTX will fail.
540 *OpenSSL team members and many third party contributors*
542 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
543 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
544 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
545 instead to retrieve these algorithms from a provider.
549 * On build targets where the multilib postfix is set in the build
550 configuration the libdir directory was changing based on whether
551 the lib directory with the multilib postfix exists on the system
552 or not. This unpredictable behavior was removed and eventual
553 multilib postfix is now always added to the default libdir. Use
554 `--libdir=lib` to override the libdir if adding the postfix is
559 * The triple DES key wrap functionality now conforms to RFC 3217 but is
560 no longer interoperable with OpenSSL 1.1.1.
564 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
565 function codes, this function can only cause problems for calling
570 * Add a configurable flag to output date formats as ISO 8601. Does not
571 change the default date format.
575 * Version of MSVC earlier than 1300 could get link warnings, which could
576 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
577 Support for this flag has been removed.
581 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
582 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
583 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
584 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
585 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
589 * The signatures of the functions to get and set options on SSL and
590 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
591 Some source code changes may be required.
595 * The public definitions of conf_method_st and conf_st have been
596 deprecated. They will be made opaque in a future release.
598 *Rich Salz and Tomáš Mráz*
600 * Client-initiated renegotiation is disabled by default. To allow it, use
601 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
602 flag, or the "ClientRenegotiation" config parameter as appropriate.
606 * Add "abspath" and "includedir" pragma's to config files, to prevent,
607 or modify relative pathname inclusion.
611 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
612 validated. Please consult the README-FIPS and
613 README-PROVIDERS files, as well as the migration guide.
615 *OpenSSL team members and many third party contributors*
617 * For the key types DH and DHX the allowed settable parameters are now different.
621 * The openssl commands that read keys, certificates, and CRLs now
622 automatically detect the PEM or DER format of the input files.
624 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
626 * Added enhanced PKCS#12 APIs which accept a library context.
630 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
634 * Added support for Kernel TLS (KTLS).
636 *Boris Pismenny, John Baldwin and Andrew Gallatin*
638 * Support for RFC 5746 secure renegotiation is now required by default for
639 SSL or TLS connections to succeed.
643 * The signature of the `copy` functional parameter of the
644 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
645 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
646 the signature of the `pub_decode` functional parameter of the
647 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
648 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
652 * The error return values from some control calls (ctrl) have changed.
656 * A public key check is now performed during EVP_PKEY_derive_set_peer().
660 * Many functions in the EVP_ namespace that are getters of values from
661 implementations or contexts were renamed to include get or get0 in their
662 names. Old names are provided as macro aliases for compatibility and
667 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
668 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
669 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
674 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
677 * The output from the command line applications may have minor
682 * The output from numerous "printing" may have minor changes.
686 * Windows thread synchronization uses read/write primitives (SRWLock) when
687 supported by the OS, otherwise CriticalSection continues to be used.
691 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
692 work on read only BIO source/sinks that do not support these functions.
693 This allows piping or redirection of a file BIO using stdin to be buffered
694 into memory. This is used internally in OSSL_DECODER_from_bio().
698 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
699 this function would return one of the values OSSL_STORE_INFO_NAME,
700 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
701 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
702 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
703 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
704 using this function should be amended to handle the changed return value.
708 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
709 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
710 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
711 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
712 contains more than one certificate identifier: This means that all
713 certificates referenced there MUST be part of the validation chain.
717 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
718 RC5, DESX and DES have been moved to the legacy provider.
722 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
723 RIPEMD-160 have been moved to the legacy provider.
727 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
732 * The deprecated functions EVP_PKEY_get0_RSA(),
733 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
734 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
735 well as the similarly named "get1" functions behave differently in
740 * A number of functions handling low-level keys or engines were deprecated
741 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
742 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
743 EVP_PKEY_get0_siphash().
747 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
748 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
749 will need to load the legacy crypto provider. This includes these PBE
750 algorithms which use this KDF:
751 - NID_pbeWithMD2AndDES_CBC
752 - NID_pbeWithMD5AndDES_CBC
753 - NID_pbeWithSHA1AndRC2_CBC
754 - NID_pbeWithMD2AndRC2_CBC
755 - NID_pbeWithMD5AndRC2_CBC
756 - NID_pbeWithSHA1AndDES_CBC
760 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
761 BIO_debug_callback() functions.
765 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
766 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
770 * The RAND_METHOD APIs have been deprecated.
774 * The SRP APIs have been deprecated.
778 * Add a compile time option to prevent the caching of provider fetched
779 algorithms. This is enabled by including the no-cached-fetch option
780 at configuration time.
784 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
785 count of PKCS12_DEFAULT_ITER.
787 *Tomáš Mráz and Sahana Prasad*
789 * The openssl speed command does not use low-level API calls anymore.
793 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
796 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
798 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
802 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
803 providers may supply their own group implementations (using either the "key
804 exchange" or the "key encapsulation" methods) which will automatically be
805 detected and used by libssl.
807 *Matt Caswell, Nicola Tuveri*
809 * The undocumented function X509_certificate_type() has been deprecated;
813 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
817 * Removed RSA padding mode for SSLv23 (which was only used for
818 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
819 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
824 * Deprecated the obsolete X9.31 RSA key generation related functions.
826 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
827 is not allowed to return a value > 1, this is no more taken as failure.
829 *Viktor Dukhovni and David von Oheimb*
831 * Deprecated the obsolete X9.31 RSA key generation related functions
832 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
833 BN_X931_generate_prime_ex().
837 * The default key generation method for the regular 2-prime RSA keys was
838 changed to the FIPS 186-4 B.3.6 method.
842 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
846 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
850 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
851 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
853 *Rich Salz, Richard Levitte, and David von Oheimb*
855 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
859 * Deprecated `OCSP_parse_url()`.
863 * Validation of SM2 keys has been separated from the validation of regular EC
868 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
869 switches: a validation failure triggers an early exit, returning a failure
870 exit status to the parent process.
874 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
875 to ignore unknown ciphers.
879 * The `-cipher-commands` and `-digest-commands` options
880 of the command line utility `list` have been deprecated.
881 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
885 * Added convenience functions for generating asymmetric key pairs:
886 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
887 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
891 * All of the low level EC_KEY functions have been deprecated.
893 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
895 * Deprecated all the libcrypto and libssl error string loading
900 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
901 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
906 * The `-crypt` option to the `passwd` command line tool has been removed.
910 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
915 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
919 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
920 EVP_PKEY_get1_tls_encodedpoint().
924 * The security callback, which can be customised by application code, supports
925 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
926 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
930 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
931 interface. Their functionality remains unchanged.
935 * Added new option for 'openssl list', '-providers', which will display the
936 list of loaded providers, their names, version and status. It optionally
937 displays their gettable parameters.
941 * Removed EVP_PKEY_set_alias_type().
945 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
946 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
950 * Changed all "STACK" functions to be macros instead of inline functions. Macro
951 parameters are still checked for type safety at compile time via helper
956 * Remove the RAND_DRBG API
958 *Paul Dale and Matthias St. Pierre*
960 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
961 as well as actual hostnames.
965 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
966 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
967 conversely, silently ignore DTLS protocol version bounds when configuring
968 TLS-based contexts. The commands can be repeated to set bounds of both
969 types. The same applies with the corresponding "min_protocol" and
970 "max_protocol" command-line switches, in case some application uses both TLS
973 SSL_CTX instances that are created for a fixed protocol version (e.g.
974 `TLSv1_server_method()`) also silently ignore version bounds. Previously
975 attempts to apply bounds to these protocol versions would result in an
976 error. Now only the "version-flexible" SSL_CTX instances are subject to
977 limits in configuration files in command-line options.
981 * Deprecated the `ENGINE` API. Engines should be replaced with providers
986 * Reworked the recorded ERR codes to make better space for system errors.
987 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
988 given code is a system error (true) or an OpenSSL error (false).
992 * Reworked the test perl framework to better allow parallel testing.
994 *Nicola Tuveri and David von Oheimb*
996 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
997 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1001 * 'Configure' has been changed to figure out the configuration target if
1002 none is given on the command line. Consequently, the 'config' script is
1003 now only a mere wrapper. All documentation is changed to only mention
1006 *Rich Salz and Richard Levitte*
1008 * Added a library context `OSSL_LIB_CTX` that applications as well as
1009 other libraries can use to form a separate context within which
1010 libcrypto operations are performed.
1014 * Added various `_ex` functions to the OpenSSL API that support using
1015 a non-default `OSSL_LIB_CTX`.
1019 * Handshake now fails if Extended Master Secret extension is dropped
1024 * Dropped interactive mode from the `openssl` program.
1028 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1030 *David von Oheimb and Shane Lontis*
1032 * Deprecated `EC_METHOD_get_field_type()`.
1036 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1037 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1038 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1042 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1046 * Add CAdES-BES signature verification support, mostly derived
1047 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1049 *Filipe Raimundo da Silva*
1051 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1055 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1056 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1060 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1064 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1065 EC_KEY_precompute_mult().
1069 * Deprecated EC_POINTs_mul().
1073 * Removed FIPS_mode() and FIPS_mode_set().
1077 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1081 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1082 EC_POINT_get_Jprojective_coordinates_GFp().
1086 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1087 arrays to be more easily constructed via a series of utility functions.
1088 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1089 the various push functions and finally convert to a passable OSSL_PARAM
1090 array using OSSL_PARAM_BLD_to_param().
1094 * The security strength of SHA1 and MD5 based signatures in TLS has been
1099 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1100 contain a provider side internal key.
1104 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1108 * Project text documents not yet having a proper file name extension
1109 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1110 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1114 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1115 have been converted to Markdown with the goal to produce documents
1116 which not only look pretty when viewed online in the browser, but
1117 remain well readable inside a plain text editor.
1119 To achieve this goal, a 'minimalistic' Markdown style has been applied
1120 which avoids formatting elements that interfere too much with the
1121 reading flow in the text file. For example, it
1123 * avoids [ATX headings][] and uses [setext headings][] instead
1124 (which works for `<h1>` and `<h2>` headings only).
1125 * avoids [inline links][] and uses [reference links][] instead.
1126 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1128 [ATX headings]: https://github.github.com/gfm/#atx-headings
1129 [setext headings]: https://github.github.com/gfm/#setext-headings
1130 [inline links]: https://github.github.com/gfm/#inline-link
1131 [reference links]: https://github.github.com/gfm/#reference-link
1132 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1133 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1135 *Matthias St. Pierre*
1137 * The test suite is changed to preserve results of each test recipe.
1138 A new directory test-runs/ with subdirectories named like the
1139 test recipes are created in the build tree for this purpose.
1143 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1144 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1145 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1147 *David von Oheimb, Martin Peylo*
1149 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1150 It supports arbitrary request and response content types, GET redirection,
1151 TLS, connections via HTTP(S) proxies, connections and exchange via
1152 user-defined BIOs (allowing implicit connections), persistent connections,
1153 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1154 The legacy OCSP-focused (and only partly documented) API
1155 is retained for backward compatibility, while most of it is deprecated.
1159 * Added `util/check-format.pl`, a tool for checking adherence to the
1160 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1161 The checks performed are incomplete and yield some false positives.
1162 Still the tool should be useful for detecting most typical glitches.
1166 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1167 If domain name resolution yields multiple IP addresses all of them are tried
1168 after `connect()` failures.
1172 * All of the low level RSA functions have been deprecated.
1176 * X509 certificates signed using SHA1 are no longer allowed at security
1181 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1182 modified to use PKEY APIs. These commands are now in maintenance mode
1183 and no new features will be added to them.
1187 * The command line utility rsautl has been deprecated.
1191 * The command line utilities genrsa and rsa have been modified to use PKEY
1192 APIs. They now write PKCS#8 keys by default. These commands are now in
1193 maintenance mode and no new features will be added to them.
1197 * All of the low level DH functions have been deprecated.
1199 *Paul Dale and Matt Caswell*
1201 * All of the low level DSA functions have been deprecated.
1205 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1206 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1210 * Deprecated low level ECDH and ECDSA functions.
1214 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1218 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1219 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1220 a new formulation to include all the things it can be used for,
1221 as well as words of caution.
1225 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1229 * All of the low level HMAC functions have been deprecated.
1231 *Paul Dale and David von Oheimb*
1233 * Over two thousand fixes were made to the documentation, including:
1234 - Common options (such as -rand/-writerand, TLS version control, etc)
1235 were refactored and point to newly-enhanced descriptions in openssl.pod.
1236 - Added style conformance for all options (with help from Richard Levitte),
1237 documented all reported missing options, added a CI build to check
1238 that all options are documented and that no unimplemented options
1240 - Documented some internals, such as all use of environment variables.
1241 - Addressed all internal broken L<> references.
1245 * All of the low level CMAC functions have been deprecated.
1249 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1250 functions have been deprecated.
1252 *Paul Dale and David von Oheimb*
1254 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1255 set of functions. The documentation mentioned negative values for some
1256 errors, but this was never the case, so the mention of negative values
1259 Code that followed the documentation and thereby check with something
1260 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1264 * All of the low level cipher functions have been deprecated.
1266 *Matt Caswell and Paul Dale*
1268 * Removed include/openssl/opensslconf.h.in and replaced it with
1269 include/openssl/configuration.h.in, which differs in not including
1270 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1271 was added to include both.
1273 This allows internal hacks where one might need to modify the set
1274 of configured macros, for example this if deprecated symbols are
1275 still supposed to be available internally:
1277 #include <openssl/configuration.h>
1279 #undef OPENSSL_NO_DEPRECATED
1280 #define OPENSSL_SUPPRESS_DEPRECATED
1282 #include <openssl/macros.h>
1284 This should not be used by applications that use the exported
1285 symbols, as that will lead to linking errors.
1289 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1290 used in exponentiation with 512-bit moduli. No EC algorithms are
1291 affected. Analysis suggests that attacks against 2-prime RSA1024,
1292 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1293 difficult to perform and are not believed likely. Attacks against DH512
1294 are considered just feasible. However, for an attack the target would
1295 have to re-use the DH512 private key, which is not recommended anyway.
1296 Also applications directly using the low-level API BN_mod_exp may be
1297 affected if they use BN_FLG_CONSTTIME.
1302 * Most memory-debug features have been deprecated, and the functionality
1303 replaced with no-ops.
1307 * Added documentation for the STACK API.
1311 * Introduced a new method type and API, OSSL_ENCODER, to represent
1312 generic encoders. These do the same sort of job that PEM writers
1313 and d2i functions do, but with support for methods supplied by
1314 providers, and the possibility for providers to support other
1319 * Introduced a new method type and API, OSSL_DECODER, to represent
1320 generic decoders. These do the same sort of job that PEM readers
1321 and i2d functions do, but with support for methods supplied by
1322 providers, and the possibility for providers to support other
1327 * Added a .pragma directive to the syntax of configuration files, to
1328 allow varying behavior in a supported and predictable manner.
1329 Currently added pragma:
1333 This allows dollar signs to be a keyword character unless it's
1334 followed by a opening brace or parenthesis. This is useful for
1335 platforms where dollar signs are commonly used in names, such as
1336 volume names and system directory names on VMS.
1340 * Added functionality to create an EVP_PKEY from user data.
1344 * Change the interpretation of the '--api' configuration option to
1345 mean that this is a desired API compatibility level with no
1346 further meaning. The previous interpretation, that this would
1347 also mean to remove all deprecated symbols up to and including
1348 the given version, no requires that 'no-deprecated' is also used
1349 in the configuration.
1351 When building applications, the desired API compatibility level
1352 can be set with the OPENSSL_API_COMPAT macro like before. For
1353 API compatibility version below 3.0, the old style numerical
1354 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1355 For version 3.0 and on, the value is expected to be the decimal
1356 value calculated from the major and minor version like this:
1358 MAJOR * 10000 + MINOR * 100
1362 -DOPENSSL_API_COMPAT=30000 For 3.0
1363 -DOPENSSL_API_COMPAT=30200 For 3.2
1365 To hide declarations that are deprecated up to and including the
1366 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1367 given when building the application as well.
1371 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1372 access to certificate and CRL stores via URIs and OSSL_STORE
1375 This adds the following functions:
1377 - X509_LOOKUP_store()
1378 - X509_STORE_load_file()
1379 - X509_STORE_load_path()
1380 - X509_STORE_load_store()
1381 - SSL_add_store_cert_subjects_to_stack()
1382 - SSL_CTX_set_default_verify_store()
1383 - SSL_CTX_load_verify_file()
1384 - SSL_CTX_load_verify_dir()
1385 - SSL_CTX_load_verify_store()
1389 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1390 The presence of this system service is determined at run-time.
1394 * Added functionality to create an EVP_PKEY context based on data
1395 for methods from providers. This takes an algorithm name and a
1396 property query string and simply stores them, with the intent
1397 that any operation that uses this context will use those strings
1398 to fetch the needed methods implicitly, thereby making the port
1399 of application written for pre-3.0 OpenSSL easier.
1403 * The undocumented function NCONF_WIN32() has been deprecated; for
1404 conversion details see the HISTORY section of doc/man5/config.pod
1408 * Introduced the new functions EVP_DigestSignInit_ex() and
1409 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1410 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1411 pages for further details.
1415 * Over two thousand fixes were made to the documentation, including:
1416 adding missing command flags, better style conformance, documentation
1419 *Rich Salz, Richard Levitte*
1421 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1422 X25519, X448, Ed25519 and Ed448.
1426 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1431 * Deprecated the public definition of `ERR_STATE` as well as the function
1432 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1437 * Added ERR functionality to give callers access to the stored function
1438 names that have replaced the older function code based functions.
1440 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1441 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1442 ERR_peek_error_all() and ERR_peek_last_error_all().
1444 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1445 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1446 ERR_func_error_string().
1450 * Extended testing to be verbose for failing tests only. The make variables
1451 VERBOSE_FAILURE or VF can be used to enable this:
1453 $ make VF=1 test # Unix
1454 $ mms /macro=(VF=1) test ! OpenVMS
1455 $ nmake VF=1 test # Windows
1459 * Added the `-copy_extensions` option to the `x509` command for use with
1460 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1461 all extensions in the request are copied to the certificate or vice versa.
1463 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1465 * Added the `-copy_extensions` option to the `req` command for use with
1466 `-x509`. When given with the `copy` or `copyall` argument,
1467 all extensions in the certification request are copied to the certificate.
1471 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1472 they generate are by default RFC 5280 compliant in the following sense:
1473 There is a subjectKeyIdentifier extension with a hash value of the public key
1474 and for not self-signed certs there is an authorityKeyIdentifier extension
1475 with a keyIdentifier field or issuer information identifying the signing key.
1476 This is done unless some configuration overrides the new default behavior,
1477 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1481 * Added several checks to `X509_verify_cert()` according to requirements in
1482 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1483 (which may be done by using the CLI option `-x509_strict`):
1484 * The basicConstraints of CA certificates must be marked critical.
1485 * CA certificates must explicitly include the keyUsage extension.
1486 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1487 * The issuer name of any certificate must not be empty.
1488 * The subject name of CA certs, certs with keyUsage crlSign,
1489 and certs without subjectAlternativeName must not be empty.
1490 * If a subjectAlternativeName extension is given it must not be empty.
1491 * The signatureAlgorithm field and the cert signature must be consistent.
1492 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1493 must not be marked critical.
1494 * The authorityKeyIdentifier must be given for X.509v3 certs
1495 unless they are self-signed.
1496 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1500 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1501 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1505 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1506 used even when parsing explicit parameters, when loading a encoded key
1507 or calling `EC_GROUP_new_from_ecpkparameters()`/
1508 `EC_GROUP_new_from_ecparameters()`.
1509 This prevents bypass of security hardening and performance gains,
1510 especially for curves with specialized EC_METHODs.
1511 By default, if a key encoded with explicit parameters is loaded and later
1512 encoded, the output is still encoded with explicit parameters, even if
1513 internally a "named" EC_GROUP is used for computation.
1517 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1518 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1519 NULL. After this change, only the cofactor parameter can be NULL. It also
1520 does some minimal sanity checks on the passed order.
1525 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1526 An attack is simple, if the first CMS_recipientInfo is valid but the
1527 second CMS_recipientInfo is chosen ciphertext. If the second
1528 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1529 encryption key will be replaced by garbage, and the message cannot be
1530 decoded, but if the RSA decryption fails, the correct encryption key is
1531 used and the recipient will not notice the attack.
1532 As a work around for this potential attack the length of the decrypted
1533 key must be equal to the cipher default key length, in case the
1534 certifiate is not given and all recipientInfo are tried out.
1535 The old behaviour can be re-enabled in the CMS code by setting the
1536 CMS_DEBUG_DECRYPT flag.
1540 * Early start up entropy quality from the DEVRANDOM seed source has been
1541 improved for older Linux systems. The RAND subsystem will wait for
1542 /dev/random to be producing output before seeding from /dev/urandom.
1543 The seeded state is stored for future library initialisations using
1544 a system global shared memory segment. The shared memory identifier
1545 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1546 the desired value. The default identifier is 114.
1550 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1551 when primes for RSA keys are computed.
1552 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1553 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1554 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1555 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1556 This avoids possible fingerprinting of newly generated RSA modules.
1560 * Correct the extended master secret constant on EBCDIC systems. Without this
1561 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1562 negotiate EMS will fail. Unfortunately this also means that TLS connections
1563 between EBCDIC systems with this fix, and EBCDIC systems without this
1564 fix will fail if they negotiate EMS.
1568 * Changed the library initialisation so that the config file is now loaded
1569 by default. This was already the case for libssl. It now occurs for both
1570 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1571 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1575 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1576 where the former acts as a replacement for `ERR_put_error()`, and the
1577 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1578 `ERR_raise_data()` adds more flexibility by taking a format string and
1579 an arbitrary number of arguments following it, to be processed with
1584 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1585 to check if a named provider is loaded and available. When called, it
1586 will also activate all fallback providers if such are still present.
1590 * Enforce a minimum DH modulus size of 512 bits.
1594 * Changed DH parameters to generate the order q subgroup instead of 2q.
1595 Previously generated DH parameters are still accepted by DH_check
1596 but DH_generate_key works around that by clearing bit 0 of the
1597 private key for those. This avoids leaking bit 0 of the private key.
1601 * Significantly reduce secure memory usage by the randomness pools.
1605 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1610 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1611 algorithms. An implementation of a key exchange algorithm can be obtained
1612 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1613 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1614 the older EVP_PKEY_derive_init() function. See the man pages for the new
1615 functions for further details.
1619 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1623 * Removed the function names from error messages and deprecated the
1628 * Removed NextStep support and the macro OPENSSL_UNISTD
1632 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1633 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1634 Also removed "export var as function" capability; we do not export
1635 variables, only functions.
1639 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1640 an error and 1 indicating success. In previous versions of OpenSSL this
1641 was a void type. If a key was set longer than the maximum possible this
1646 * Support SM2 signing and verification schemes with X509 certificate.
1650 * Use SHA256 as the default digest for TS query in the `ts` app.
1654 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1658 * Default cipher lists/suites are now available via a function, the
1659 #defines are deprecated.
1663 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1664 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1665 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1669 * Join the directories crypto/x509 and crypto/x509v3
1673 * Added command 'openssl kdf' that uses the EVP_KDF API.
1677 * Added command 'openssl mac' that uses the EVP_MAC API.
1681 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1682 as default directories. Also added the command 'openssl info'
1683 for scripting purposes.
1687 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1692 * Add prediction resistance to the DRBG reseeding process.
1696 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1697 mandated by IEEE Std 1619-2018.
1701 * Added newline escaping functionality to a filename when using openssl dgst.
1702 This output format is to replicate the output format found in the `*sum`
1703 checksum programs. This aims to preserve backward compatibility.
1705 *Matt Eaton, Richard Levitte, and Paul Dale*
1707 * Removed the heartbeat message in DTLS feature, as it has very
1708 little usage and doesn't seem to fulfill a valuable purpose.
1709 The configuration option is now deprecated.
1713 * Changed the output of 'openssl {digestname} < file' to display the
1714 digest name in its output.
1718 * Added a new generic trace API which provides support for enabling
1719 instrumentation through trace output.
1721 *Richard Levitte & Matthias St. Pierre*
1723 * Added build tests for C++. These are generated files that only do one
1724 thing, to include one public OpenSSL head file each. This tests that
1725 the public header files can be usefully included in a C++ application.
1727 This test isn't enabled by default. It can be enabled with the option
1728 'enable-buildtest-c++'.
1732 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1736 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1740 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1744 * Added KMAC to EVP_MAC.
1748 * Added property based algorithm implementation selection framework to
1753 * Added SCA hardening for modular field inversion in EC_GROUP through
1754 a new dedicated field_inv() pointer in EC_METHOD.
1755 This also addresses a leakage affecting conversions from projective
1756 to affine coordinates.
1758 *Billy Bob Brumley, Nicola Tuveri*
1760 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1761 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1762 those algorithms that were already supported through the EVP_PKEY API
1763 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1764 and scrypt are now wrappers that call EVP_KDF.
1768 * Build devcrypto engine as a dynamic engine.
1770 *Eneas U de Queiroz*
1772 * Add keyed BLAKE2 to EVP_MAC.
1776 * Fix a bug in the computation of the endpoint-pair shared secret used
1777 by DTLS over SCTP. This breaks interoperability with older versions
1778 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1779 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1780 interoperability with such broken implementations. However, enabling
1781 this switch breaks interoperability with correct implementations.
1783 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1784 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1788 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1792 * Changed the license to the Apache License v2.0.
1796 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1798 - Major releases (indicated by incrementing the MAJOR release number)
1799 may introduce incompatible API/ABI changes.
1800 - Minor releases (indicated by incrementing the MINOR release number)
1801 may introduce new features but retain API/ABI compatibility.
1802 - Patch releases (indicated by incrementing the PATCH number)
1803 are intended for bug fixes and other improvements of existing
1804 features only (like improving performance or adding documentation)
1805 and retain API/ABI compatibility.
1809 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1813 * Remove the 'dist' target and add a tarball building script. The
1814 'dist' target has fallen out of use, and it shouldn't be
1815 necessary to configure just to create a source distribution.
1819 * Recreate the OS390-Unix config target. It no longer relies on a
1820 special script like it did for OpenSSL pre-1.1.0.
1824 * Instead of having the source directories listed in Configure, add
1825 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1830 * Add GMAC to EVP_MAC.
1834 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1838 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1839 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1840 to facilitate the continued use of MACs through raw private keys in
1841 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1845 * Deprecate ECDH_KDF_X9_62().
1849 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1850 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1851 are retained for backwards compatibility.
1855 * AES-XTS mode now enforces that its two keys are different to mitigate
1856 the attacked described in "Efficient Instantiations of Tweakable
1857 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1858 Details of this attack can be obtained from:
1859 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1863 * Rename the object files, i.e. give them other names than in previous
1864 versions. Their names now include the name of the final product, as
1865 well as its type mnemonic (bin, lib, shlib).
1869 * Added new option for 'openssl list', '-objects', which will display the
1870 list of built in objects, i.e. OIDs with names.
1874 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1875 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1880 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1881 improves application performance by removing data copies and providing
1882 applications with zero-copy system calls such as sendfile and splice.
1886 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1890 * `PKCS12_parse` now maintains the order of the parsed certificates
1891 when outputting them via `*ca` (rather than reversing it).
1895 * Deprecated pthread fork support methods.
1899 * Added support for FFDHE key exchange in TLS 1.3.
1903 * Added a new concept for OpenSSL plugability: providers. This
1904 functionality is designed to replace the ENGINE API and ENGINE
1905 implementations, and to be much more dynamic, allowing provider
1906 authors to introduce new algorithms among other things, as long as
1907 there's an API that supports the algorithm type.
1909 With this concept comes a new core API for interaction between
1910 libcrypto and provider implementations. Public libcrypto functions
1911 that want to use providers do so through this core API.
1913 The main documentation for this core API is found in
1914 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1915 refer to other manuals describing the API specific for supported
1916 algorithm types (also called operations).
1923 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1925 * Avoid loading of a dynamic engine twice.
1929 * Prioritise DANE TLSA issuer certs over peer certs
1933 * Fixed random API for MacOS prior to 10.12
1935 These MacOS versions don't support the CommonCrypto APIs
1939 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1941 * Fixed an SM2 Decryption Buffer Overflow.
1943 In order to decrypt SM2 encrypted data an application is expected to
1944 call the API function EVP_PKEY_decrypt(). Typically an application will
1945 call this function twice. The first time, on entry, the "out" parameter
1946 can be NULL and, on exit, the "outlen" parameter is populated with the
1947 buffer size required to hold the decrypted plaintext. The application
1948 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1949 again, but this time passing a non-NULL value for the "out" parameter.
1951 A bug in the implementation of the SM2 decryption code means that the
1952 calculation of the buffer size required to hold the plaintext returned
1953 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1954 size required by the second call. This can lead to a buffer overflow
1955 when EVP_PKEY_decrypt() is called by the application a second time with
1956 a buffer that is too small.
1958 A malicious attacker who is able present SM2 content for decryption to
1959 an application could cause attacker chosen data to overflow the buffer
1960 by up to a maximum of 62 bytes altering the contents of other data held
1961 after the buffer, possibly changing application behaviour or causing
1962 the application to crash. The location of the buffer is application
1963 dependent but is typically heap allocated.
1968 * Fixed various read buffer overruns processing ASN.1 strings
1970 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1971 structure which contains a buffer holding the string data and a field
1972 holding the buffer length. This contrasts with normal C strings which
1973 are repesented as a buffer for the string data which is terminated
1974 with a NUL (0) byte.
1976 Although not a strict requirement, ASN.1 strings that are parsed using
1977 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1978 well as any string whose value has been set with the ASN1_STRING_set()
1979 function will additionally NUL terminate the byte array in the
1980 ASN1_STRING structure.
1982 However, it is possible for applications to directly construct valid
1983 ASN1_STRING structures which do not NUL terminate the byte array by
1984 directly setting the "data" and "length" fields in the ASN1_STRING
1985 array. This can also happen by using the ASN1_STRING_set0() function.
1987 Numerous OpenSSL functions that print ASN.1 data have been found to
1988 assume that the ASN1_STRING byte array will be NUL terminated, even
1989 though this is not guaranteed for strings that have been directly
1990 constructed. Where an application requests an ASN.1 structure to be
1991 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1992 been directly constructed by the application without NUL terminating
1993 the "data" field, then a read buffer overrun can occur.
1995 The same thing can also occur during name constraints processing
1996 of certificates (for example if a certificate has been directly
1997 constructed by the application instead of loading it via the OpenSSL
1998 parsing functions, and the certificate contains non NUL terminated
1999 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2000 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2002 If a malicious actor can cause an application to directly construct an
2003 ASN1_STRING and then process it through one of the affected OpenSSL
2004 functions then this issue could be hit. This might result in a crash
2005 (causing a Denial of Service attack). It could also result in the
2006 disclosure of private memory contents (such as private keys, or
2007 sensitive plaintext).
2012 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2014 * Fixed a problem with verifying a certificate chain when using the
2015 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2016 the certificates present in a certificate chain. It is not set by default.
2018 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2019 the chain that have explicitly encoded elliptic curve parameters was added
2020 as an additional strict check.
2022 An error in the implementation of this check meant that the result of a
2023 previous check to confirm that certificates in the chain are valid CA
2024 certificates was overwritten. This effectively bypasses the check
2025 that non-CA certificates must not be able to issue other certificates.
2027 If a "purpose" has been configured then there is a subsequent opportunity
2028 for checks that the certificate is a valid CA. All of the named "purpose"
2029 values implemented in libcrypto perform this check. Therefore, where
2030 a purpose is set the certificate chain will still be rejected even when the
2031 strict flag has been used. A purpose is set by default in libssl client and
2032 server certificate verification routines, but it can be overridden or
2033 removed by an application.
2035 In order to be affected, an application must explicitly set the
2036 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2037 for the certificate verification or, in the case of TLS client or server
2038 applications, override the default purpose.
2043 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2044 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2045 renegotiation ClientHello omits the signature_algorithms extension (where it
2046 was present in the initial ClientHello), but includes a
2047 signature_algorithms_cert extension then a NULL pointer dereference will
2048 result, leading to a crash and a denial of service attack.
2050 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2051 (which is the default configuration). OpenSSL TLS clients are not impacted by
2055 *Peter Kästle and Samuel Sapalski*
2057 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2059 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2060 create a unique hash value based on the issuer and serial number data
2061 contained within an X509 certificate. However it was failing to correctly
2062 handle any errors that may occur while parsing the issuer field (which might
2063 occur if the issuer field is maliciously constructed). This may subsequently
2064 result in a NULL pointer deref and a crash leading to a potential denial of
2070 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2071 padding mode to correctly check for rollback attacks. This is considered a
2072 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2077 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2078 functions. Previously they could overflow the output length argument in some
2079 cases where the input length is close to the maximum permissable length for
2080 an integer on the platform. In such cases the return value from the function
2081 call would be 1 (indicating success), but the output length value would be
2082 negative. This could cause applications to behave incorrectly or crash.
2087 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2088 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2089 could be exploited in a side channel attack to recover the password. Since
2090 the attack is local host only this is outside of the current OpenSSL
2091 threat model and therefore no CVE is assigned.
2093 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2098 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2100 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2101 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2102 If an attacker can control both items being compared then this could lead
2103 to a possible denial of service attack. OpenSSL itself uses the
2104 GENERAL_NAME_cmp function for two purposes:
2105 1) Comparing CRL distribution point names between an available CRL and a
2106 CRL distribution point embedded in an X509 certificate
2107 2) When verifying that a timestamp response token signer matches the
2108 timestamp authority name (exposed via the API functions
2109 TS_RESP_verify_response and TS_RESP_verify_token)
2114 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2116 * Certificates with explicit curve parameters are now disallowed in
2117 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2121 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2122 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2123 conversely, silently ignore DTLS protocol version bounds when configuring
2124 TLS-based contexts. The commands can be repeated to set bounds of both
2125 types. The same applies with the corresponding "min_protocol" and
2126 "max_protocol" command-line switches, in case some application uses both TLS
2129 SSL_CTX instances that are created for a fixed protocol version (e.g.
2130 TLSv1_server_method()) also silently ignore version bounds. Previously
2131 attempts to apply bounds to these protocol versions would result in an
2132 error. Now only the "version-flexible" SSL_CTX instances are subject to
2133 limits in configuration files in command-line options.
2137 * Handshake now fails if Extended Master Secret extension is dropped
2142 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2144 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2146 * Fixed segmentation fault in SSL_check_chain()
2147 Server or client applications that call the SSL_check_chain() function
2148 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2149 dereference as a result of incorrect handling of the
2150 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2151 or unrecognised signature algorithm is received from the peer. This could
2152 be exploited by a malicious peer in a Denial of Service attack.
2157 * Added AES consttime code for no-asm configurations
2158 an optional constant time support for AES was added
2159 when building openssl for no-asm.
2160 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2161 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2162 At this time this feature is by default disabled.
2163 It will be enabled by default in 3.0.
2167 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2169 * Revert the change of EOF detection while reading in libssl to avoid
2170 regressions in applications depending on the current way of reporting
2171 the EOF. As the existing method is not fully accurate the change to
2172 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2173 branch and will be present in the 3.0 release.
2177 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2178 when primes for RSA keys are computed.
2179 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2180 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2181 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2182 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2183 This avoids possible fingerprinting of newly generated RSA modules.
2187 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2189 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2190 while reading in libssl then we would report an error back to the
2191 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2192 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2193 therefore give a hint as to what went wrong.
2197 * Check that ed25519 and ed448 are allowed by the security level. Previously
2198 signature algorithms not using an MD were not being checked that they were
2199 allowed by the security level.
2203 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2204 was not quite right. The behaviour was not consistent between resumption
2205 and normal handshakes, and also not quite consistent with historical
2206 behaviour. The behaviour in various scenarios has been clarified and
2207 it has been updated to make it match historical behaviour as closely as
2212 * *[VMS only]* The header files that the VMS compilers include automatically,
2213 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2214 that the C++ compiler doesn't understand. This is a shortcoming in the
2215 compiler, but can be worked around with `__cplusplus` guards.
2217 C++ applications that use OpenSSL libraries must be compiled using the
2218 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2219 functions. Otherwise, only functions with symbols of less than 31
2220 characters can be used, as the linker will not be able to successfully
2221 resolve symbols with longer names.
2225 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2226 The presence of this system service is determined at run-time.
2230 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2235 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2237 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2238 number generator (RNG). This was intended to include protection in the
2239 event of a fork() system call in order to ensure that the parent and child
2240 processes did not share the same RNG state. However this protection was not
2241 being used in the default case.
2243 A partial mitigation for this issue is that the output from a high
2244 precision timer is mixed into the RNG state so the likelihood of a parent
2245 and child process sharing state is significantly reduced.
2247 If an application already calls OPENSSL_init_crypto() explicitly using
2248 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2251 *Matthias St. Pierre*
2253 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2254 used even when parsing explicit parameters, when loading a encoded key
2255 or calling `EC_GROUP_new_from_ecpkparameters()`/
2256 `EC_GROUP_new_from_ecparameters()`.
2257 This prevents bypass of security hardening and performance gains,
2258 especially for curves with specialized EC_METHODs.
2259 By default, if a key encoded with explicit parameters is loaded and later
2260 encoded, the output is still encoded with explicit parameters, even if
2261 internally a "named" EC_GROUP is used for computation.
2265 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2266 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2267 NULL. After this change, only the cofactor parameter can be NULL. It also
2268 does some minimal sanity checks on the passed order.
2273 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2274 An attack is simple, if the first CMS_recipientInfo is valid but the
2275 second CMS_recipientInfo is chosen ciphertext. If the second
2276 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2277 encryption key will be replaced by garbage, and the message cannot be
2278 decoded, but if the RSA decryption fails, the correct encryption key is
2279 used and the recipient will not notice the attack.
2280 As a work around for this potential attack the length of the decrypted
2281 key must be equal to the cipher default key length, in case the
2282 certifiate is not given and all recipientInfo are tried out.
2283 The old behaviour can be re-enabled in the CMS code by setting the
2284 CMS_DEBUG_DECRYPT flag.
2289 * Early start up entropy quality from the DEVRANDOM seed source has been
2290 improved for older Linux systems. The RAND subsystem will wait for
2291 /dev/random to be producing output before seeding from /dev/urandom.
2292 The seeded state is stored for future library initialisations using
2293 a system global shared memory segment. The shared memory identifier
2294 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2295 the desired value. The default identifier is 114.
2299 * Correct the extended master secret constant on EBCDIC systems. Without this
2300 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2301 negotiate EMS will fail. Unfortunately this also means that TLS connections
2302 between EBCDIC systems with this fix, and EBCDIC systems without this
2303 fix will fail if they negotiate EMS.
2307 * Use Windows installation paths in the mingw builds
2309 Mingw isn't a POSIX environment per se, which means that Windows
2310 paths should be used for installation.
2315 * Changed DH_check to accept parameters with order q and 2q subgroups.
2316 With order 2q subgroups the bit 0 of the private key is not secret
2317 but DH_generate_key works around that by clearing bit 0 of the
2318 private key for those. This avoids leaking bit 0 of the private key.
2322 * Significantly reduce secure memory usage by the randomness pools.
2326 * Revert the DEVRANDOM_WAIT feature for Linux systems
2328 The DEVRANDOM_WAIT feature added a select() call to wait for the
2329 /dev/random device to become readable before reading from the
2330 /dev/urandom device.
2332 It turned out that this change had negative side effects on
2333 performance which were not acceptable. After some discussion it
2334 was decided to revert this feature and leave it up to the OS
2335 resp. the platform maintainer to ensure a proper initialization
2336 during early boot time.
2338 *Matthias St. Pierre*
2340 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2342 * Add build tests for C++. These are generated files that only do one
2343 thing, to include one public OpenSSL head file each. This tests that
2344 the public header files can be usefully included in a C++ application.
2346 This test isn't enabled by default. It can be enabled with the option
2347 'enable-buildtest-c++'.
2351 * Enable SHA3 pre-hashing for ECDSA and DSA.
2355 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2356 This changes the size when using the `genpkey` command when no size is given.
2357 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2358 generation commands to use 2048 bits by default.
2362 * Reorganize the manual pages to consistently have RETURN VALUES,
2363 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2364 util/fix-doc-nits accordingly.
2366 *Paul Yang, Joshua Lock*
2368 * Add the missing accessor EVP_PKEY_get0_engine()
2372 * Have commands like `s_client` and `s_server` output the signature scheme
2373 along with other cipher suite parameters when debugging.
2377 * Make OPENSSL_config() error agnostic again.
2381 * Do the error handling in RSA decryption constant time.
2385 * Prevent over long nonces in ChaCha20-Poly1305.
2387 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2388 for every encryption operation. RFC 7539 specifies that the nonce value
2389 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2390 and front pads the nonce with 0 bytes if it is less than 12
2391 bytes. However it also incorrectly allows a nonce to be set of up to 16
2392 bytes. In this case only the last 12 bytes are significant and any
2393 additional leading bytes are ignored.
2395 It is a requirement of using this cipher that nonce values are
2396 unique. Messages encrypted using a reused nonce value are susceptible to
2397 serious confidentiality and integrity attacks. If an application changes
2398 the default nonce length to be longer than 12 bytes and then makes a
2399 change to the leading bytes of the nonce expecting the new value to be a
2400 new unique nonce then such an application could inadvertently encrypt
2401 messages with a reused nonce.
2403 Additionally the ignored bytes in a long nonce are not covered by the
2404 integrity guarantee of this cipher. Any application that relies on the
2405 integrity of these ignored leading bytes of a long nonce may be further
2406 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2407 is safe because no such use sets such a long nonce value. However user
2408 applications that use this cipher directly and set a non-default nonce
2409 length to be longer than 12 bytes may be vulnerable.
2411 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2417 * Add DEVRANDOM_WAIT feature for Linux systems
2419 On older Linux systems where the getrandom() system call is not available,
2420 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2421 Contrary to getrandom(), the /dev/urandom device will not block during
2422 early boot when the kernel CSPRNG has not been seeded yet.
2424 To mitigate this known weakness, use select() to wait for /dev/random to
2425 become readable before reading from /dev/urandom.
2427 * Ensure that SM2 only uses SM3 as digest algorithm
2431 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2433 * Change the info callback signals for the start and end of a post-handshake
2434 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2435 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2436 confused by this and assume that a TLSv1.2 renegotiation has started. This
2437 can break KeyUpdate handling. Instead we no longer signal the start and end
2438 of a post handshake message exchange (although the messages themselves are
2439 still signalled). This could break some applications that were expecting
2440 the old signals. However without this KeyUpdate is not usable for many
2445 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2447 * Timing vulnerability in DSA signature generation
2449 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2450 timing side channel attack. An attacker could use variations in the signing
2451 algorithm to recover the private key.
2453 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2458 * Timing vulnerability in ECDSA signature generation
2460 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2461 timing side channel attack. An attacker could use variations in the signing
2462 algorithm to recover the private key.
2464 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2469 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2470 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2471 of two gigabytes and the error handling improved.
2473 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2474 categorized as a normal bug, not a security issue, because the DRBG reseeds
2475 automatically and is fully functional even without additional randomness
2476 provided by the application.
2478 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2480 * Add a new ClientHello callback. Provides a callback interface that gives
2481 the application the ability to adjust the nascent SSL object at the
2482 earliest stage of ClientHello processing, immediately after extensions have
2483 been collected but before they have been processed. In particular, this
2484 callback can adjust the supported TLS versions in response to the contents
2489 * Add SM2 base algorithm support.
2493 * s390x assembly pack: add (improved) hardware-support for the following
2494 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2495 aes-cfb/cfb8, aes-ecb.
2499 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2500 parameter is no longer accepted, as it leads to a corrupt table. NULL
2501 pem_str is reserved for alias entries only.
2505 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2506 step for prime curves. The new implementation is based on formulae from
2507 differential addition-and-doubling in homogeneous projective coordinates
2508 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2509 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2510 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2511 to work in projective coordinates.
2513 *Billy Bob Brumley, Nicola Tuveri*
2515 * Change generating and checking of primes so that the error rate of not
2516 being prime depends on the intended use based on the size of the input.
2517 For larger primes this will result in more rounds of Miller-Rabin.
2518 The maximal error rate for primes with more than 1080 bits is lowered
2521 *Kurt Roeckx, Annie Yousar*
2523 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2527 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2528 moving between systems, and to avoid confusion when a Windows build is
2529 done with mingw vs with MSVC. For POSIX installs, there's still a
2530 symlink or copy named 'tsget' to avoid that confusion as well.
2534 * Revert blinding in ECDSA sign and instead make problematic addition
2535 length-invariant. Switch even to fixed-length Montgomery multiplication.
2539 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2540 step for binary curves. The new implementation is based on formulae from
2541 differential addition-and-doubling in mixed Lopez-Dahab projective
2542 coordinates, modified to independently blind the operands.
2544 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2546 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2547 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2548 EC_METHODs to implement their own specialized "ladder step", to take
2549 advantage of more favorable coordinate systems or more efficient
2550 differential addition-and-doubling algorithms.
2552 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2554 * Modified the random device based seed sources to keep the relevant
2555 file descriptors open rather than reopening them on each access.
2556 This allows such sources to operate in a chroot() jail without
2557 the associated device nodes being available. This behaviour can be
2558 controlled using RAND_keep_random_devices_open().
2562 * Numerous side-channel attack mitigations have been applied. This may have
2563 performance impacts for some algorithms for the benefit of improved
2564 security. Specific changes are noted in this change log by their respective
2569 * AIX shared library support overhaul. Switch to AIX "natural" way of
2570 handling shared libraries, which means collecting shared objects of
2571 different versions and bitnesses in one common archive. This allows to
2572 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2573 doesn't affect the way 3rd party applications are linked, only how
2574 multi-version installation is managed.
2578 * Make ec_group_do_inverse_ord() more robust and available to other
2579 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2580 mitigations are applied to the fallback BN_mod_inverse().
2581 When using this function rather than BN_mod_inverse() directly, new
2582 EC cryptosystem implementations are then safer-by-default.
2586 * Add coordinate blinding for EC_POINT and implement projective
2587 coordinate blinding for generic prime curves as a countermeasure to
2588 chosen point SCA attacks.
2590 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2592 * Add blinding to ECDSA and DSA signatures to protect against side channel
2593 attacks discovered by Keegan Ryan (NCC Group).
2597 * Enforce checking in the `pkeyutl` command to ensure that the input
2598 length does not exceed the maximum supported digest length when performing
2599 a sign, verify or verifyrecover operation.
2603 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2604 I/O in combination with something like select() or poll() will hang. This
2605 can be turned off again using SSL_CTX_clear_mode().
2606 Many applications do not properly handle non-application data records, and
2607 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2608 around the problems in those applications, but can also break some.
2609 It's recommended to read the manpages about SSL_read(), SSL_write(),
2610 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2611 SSL_CTX_set_read_ahead() again.
2615 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2616 now allow empty (zero character) pass phrases.
2620 * Apply blinding to binary field modular inversion and remove patent
2621 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2625 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2626 binary and prime elliptic curves.
2630 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2631 constant time fixed point multiplication.
2635 * Revise elliptic curve scalar multiplication with timing attack
2636 defenses: ec_wNAF_mul redirects to a constant time implementation
2637 when computing fixed point and variable point multiplication (which
2638 in OpenSSL are mostly used with secret scalars in keygen, sign,
2639 ECDH derive operations).
2640 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2643 * Updated CONTRIBUTING
2647 * Updated DRBG / RAND to request nonce and additional low entropy
2648 randomness from the system.
2650 *Matthias St. Pierre*
2652 * Updated 'openssl rehash' to use OpenSSL consistent default.
2656 * Moved the load of the ssl_conf module to libcrypto, which helps
2657 loading engines that libssl uses before libssl is initialised.
2661 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2665 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2667 *Ingo Schwarze, Rich Salz*
2669 * Added output of accepting IP address and port for 'openssl s_server'
2673 * Added a new API for TLSv1.3 ciphersuites:
2674 SSL_CTX_set_ciphersuites()
2675 SSL_set_ciphersuites()
2679 * Memory allocation failures consistently add an error to the error
2684 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2685 in libcrypto when run as setuid/setgid.
2689 * Load any config file by default when libssl is used.
2693 * Added new public header file <openssl/rand_drbg.h> and documentation
2694 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2696 *Matthias St. Pierre*
2698 * QNX support removed (cannot find contributors to get their approval
2699 for the license change).
2703 * TLSv1.3 replay protection for early data has been implemented. See the
2704 SSL_read_early_data() man page for further details.
2708 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2709 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2710 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2711 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2712 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2713 configuration has been separated out. See the ciphers man page or the
2714 SSL_CTX_set_ciphersuites() man page for more information.
2718 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2719 in responder mode now supports the new "-multi" option, which
2720 spawns the specified number of child processes to handle OCSP
2721 requests. The "-timeout" option now also limits the OCSP
2722 responder's patience to wait to receive the full client request
2723 on a newly accepted connection. Child processes are respawned
2724 as needed, and the CA index file is automatically reloaded
2725 when changed. This makes it possible to run the "ocsp" responder
2726 as a long-running service, making the OpenSSL CA somewhat more
2727 feature-complete. In this mode, most diagnostic messages logged
2728 after entering the event loop are logged via syslog(3) rather than
2733 * Added support for X448 and Ed448. Heavily based on original work by
2738 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2739 objects loaded. This adds the functions OSSL_STORE_expect() and
2740 OSSL_STORE_find() as well as needed tools to construct searches and
2741 get the search data out of them.
2745 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2746 version of OpenSSL should review their configuration settings to ensure
2747 that they are still appropriate for TLSv1.3. For further information see:
2748 <https://wiki.openssl.org/index.php/TLS1.3>
2752 * Grand redesign of the OpenSSL random generator
2754 The default RAND method now utilizes an AES-CTR DRBG according to
2755 NIST standard SP 800-90Ar1. The new random generator is essentially
2756 a port of the default random generator from the OpenSSL FIPS 2.0
2757 object module. It is a hybrid deterministic random bit generator
2758 using an AES-CTR bit stream and which seeds and reseeds itself
2759 automatically using trusted system entropy sources.
2761 Some of its new features are:
2762 - Support for multiple DRBG instances with seed chaining.
2763 - The default RAND method makes use of a DRBG.
2764 - There is a public and private DRBG instance.
2765 - The DRBG instances are fork-safe.
2766 - Keep all global DRBG instances on the secure heap if it is enabled.
2767 - The public and private DRBG instance are per thread for lock free
2770 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2772 * Changed Configure so it only says what it does and doesn't dump
2773 so much data. Instead, ./configdata.pm should be used as a script
2774 to display all sorts of configuration data.
2778 * Added processing of "make variables" to Configure.
2782 * Added SHA512/224 and SHA512/256 algorithm support.
2786 * The last traces of Netware support, first removed in 1.1.0, have
2791 * Get rid of Makefile.shared, and in the process, make the processing
2792 of certain files (rc.obj, or the .def/.map/.opt files produced from
2793 the ordinal files) more visible and hopefully easier to trace and
2794 debug (or make silent).
2798 * Make it possible to have environment variable assignments as
2799 arguments to config / Configure.
2803 * Add multi-prime RSA (RFC 8017) support.
2807 * Add SM3 implemented according to GB/T 32905-2016
2808 *Jack Lloyd <jack.lloyd@ribose.com>,*
2809 *Ronald Tse <ronald.tse@ribose.com>,*
2810 *Erick Borsboom <erick.borsboom@ribose.com>*
2812 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2813 as documented in RFC6066.
2814 Based on a patch from Tomasz Moń
2816 *Filipe Raimundo da Silva*
2818 * Add SM4 implemented according to GB/T 32907-2016.
2819 *Jack Lloyd <jack.lloyd@ribose.com>,*
2820 *Ronald Tse <ronald.tse@ribose.com>,*
2821 *Erick Borsboom <erick.borsboom@ribose.com>*
2823 * Reimplement -newreq-nodes and ERR_error_string_n; the
2824 original author does not agree with the license change.
2828 * Add ARIA AEAD TLS support.
2832 * Some macro definitions to support VS6 have been removed. Visual
2833 Studio 6 has not worked since 1.1.0
2837 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2838 without clearing the errors.
2842 * Add "atfork" functions. If building on a system that without
2843 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2844 requirements. The RAND facility now uses/requires this.
2852 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2853 not possible to disable entirely. However, it's still possible to
2854 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2857 To disable, configure with 'no-ui-console'. 'no-ui' is still
2858 possible to use as an alias. Check at compile time with the
2859 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2860 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2864 * Add a STORE module, which implements a uniform and URI based reader of
2865 stores that can contain keys, certificates, CRLs and numerous other
2866 objects. The main API is loosely based on a few stdio functions,
2867 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2868 OSSL_STORE_error and OSSL_STORE_close.
2869 The implementation uses backends called "loaders" to implement arbitrary
2870 URI schemes. There is one built in "loader" for the 'file' scheme.
2874 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2875 then adjusted to work on FreeBSD 8.4 as well.
2876 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2877 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2881 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2882 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2883 error code calls like this:
2885 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2887 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2888 that can be encoded in C. For the foreseeable future, this will only
2891 *Richard Levitte and Tim Hudson*
2893 * Removed BSD cryptodev engine.
2897 * Add a build target 'build_all_generated', to build all generated files
2898 and only that. This can be used to prepare everything that requires
2899 things like perl for a system that lacks perl and then move everything
2900 to that system and do the rest of the build there.
2904 * In the UI interface, make it possible to duplicate the user data. This
2905 can be used by engines that need to retain the data for a longer time
2906 than just the call where this user data is passed.
2910 * Ignore the '-named_curve auto' value for compatibility of applications
2913 *Tomáš Mráz <tmraz@fedoraproject.org>*
2915 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2916 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2917 alerts across multiple records (some of which could be empty). In practice
2918 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2919 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2920 support this at all. Supporting it adds significant complexity to the
2921 record layer, and its removal is unlikely to cause interoperability
2926 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2927 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2928 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2933 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2934 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2936 *Richard Levitte, Andy Polyakov*
2938 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2943 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2944 platform rather than 'mingw'.
2948 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2949 success if they are asked to add an object which already exists
2950 in the store. This change cascades to other functions which load
2951 certificates and CRLs.
2955 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2956 facilitate stack unwinding even from assembly subroutines.
2960 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2961 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2965 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2966 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2967 which is the minimum version we support.
2971 * Certificate time validation (X509_cmp_time) enforces stricter
2972 compliance with RFC 5280. Fractional seconds and timezone offsets
2973 are no longer allowed.
2977 * Add support for ARIA
2981 * s_client will now send the Server Name Indication (SNI) extension by
2982 default unless the new "-noservername" option is used. The server name is
2983 based on the host provided to the "-connect" option unless overridden by
2984 using "-servername".
2988 * Add support for SipHash
2992 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2993 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2994 prevent issues where no progress is being made and the peer continually
2995 sends unrecognised record types, using up resources processing them.
2999 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3000 using the algorithm defined in
3001 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3005 * Heartbeat support has been removed; the ABI is changed for now.
3007 *Richard Levitte, Rich Salz*
3009 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3013 * The RSA "null" method, which was partially supported to avoid patent
3014 issues, has been replaced to always returns NULL.
3021 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3023 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3024 used even when parsing explicit parameters, when loading a encoded key
3025 or calling `EC_GROUP_new_from_ecpkparameters()`/
3026 `EC_GROUP_new_from_ecparameters()`.
3027 This prevents bypass of security hardening and performance gains,
3028 especially for curves with specialized EC_METHODs.
3029 By default, if a key encoded with explicit parameters is loaded and later
3030 encoded, the output is still encoded with explicit parameters, even if
3031 internally a "named" EC_GROUP is used for computation.
3035 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3036 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3037 NULL. After this change, only the cofactor parameter can be NULL. It also
3038 does some minimal sanity checks on the passed order.
3043 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3044 An attack is simple, if the first CMS_recipientInfo is valid but the
3045 second CMS_recipientInfo is chosen ciphertext. If the second
3046 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3047 encryption key will be replaced by garbage, and the message cannot be
3048 decoded, but if the RSA decryption fails, the correct encryption key is
3049 used and the recipient will not notice the attack.
3050 As a work around for this potential attack the length of the decrypted
3051 key must be equal to the cipher default key length, in case the
3052 certifiate is not given and all recipientInfo are tried out.
3053 The old behaviour can be re-enabled in the CMS code by setting the
3054 CMS_DEBUG_DECRYPT flag.
3059 * Use Windows installation paths in the mingw builds
3061 Mingw isn't a POSIX environment per se, which means that Windows
3062 paths should be used for installation.
3067 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3069 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3070 This changes the size when using the `genpkey` command when no size is given.
3071 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3072 generation commands to use 2048 bits by default.
3076 * Prevent over long nonces in ChaCha20-Poly1305.
3078 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3079 for every encryption operation. RFC 7539 specifies that the nonce value
3080 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3081 and front pads the nonce with 0 bytes if it is less than 12
3082 bytes. However it also incorrectly allows a nonce to be set of up to 16
3083 bytes. In this case only the last 12 bytes are significant and any
3084 additional leading bytes are ignored.
3086 It is a requirement of using this cipher that nonce values are
3087 unique. Messages encrypted using a reused nonce value are susceptible to
3088 serious confidentiality and integrity attacks. If an application changes
3089 the default nonce length to be longer than 12 bytes and then makes a
3090 change to the leading bytes of the nonce expecting the new value to be a
3091 new unique nonce then such an application could inadvertently encrypt
3092 messages with a reused nonce.
3094 Additionally the ignored bytes in a long nonce are not covered by the
3095 integrity guarantee of this cipher. Any application that relies on the
3096 integrity of these ignored leading bytes of a long nonce may be further
3097 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3098 is safe because no such use sets such a long nonce value. However user
3099 applications that use this cipher directly and set a non-default nonce
3100 length to be longer than 12 bytes may be vulnerable.
3102 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3108 * Added SCA hardening for modular field inversion in EC_GROUP through
3109 a new dedicated field_inv() pointer in EC_METHOD.
3110 This also addresses a leakage affecting conversions from projective
3111 to affine coordinates.
3113 *Billy Bob Brumley, Nicola Tuveri*
3115 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3116 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3120 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3124 * Remove the 'dist' target and add a tarball building script. The
3125 'dist' target has fallen out of use, and it shouldn't be
3126 necessary to configure just to create a source distribution.
3130 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3132 * Timing vulnerability in DSA signature generation
3134 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3135 timing side channel attack. An attacker could use variations in the signing
3136 algorithm to recover the private key.
3138 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3143 * Timing vulnerability in ECDSA signature generation
3145 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3146 timing side channel attack. An attacker could use variations in the signing
3147 algorithm to recover the private key.
3149 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3154 * Add coordinate blinding for EC_POINT and implement projective
3155 coordinate blinding for generic prime curves as a countermeasure to
3156 chosen point SCA attacks.
3158 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3160 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3162 * Client DoS due to large DH parameter
3164 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3165 malicious server can send a very large prime value to the client. This will
3166 cause the client to spend an unreasonably long period of time generating a
3167 key for this prime resulting in a hang until the client has finished. This
3168 could be exploited in a Denial Of Service attack.
3170 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3175 * Cache timing vulnerability in RSA Key Generation
3177 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3178 a cache timing side channel attack. An attacker with sufficient access to
3179 mount cache timing attacks during the RSA key generation process could
3180 recover the private key.
3182 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3183 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3188 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3189 parameter is no longer accepted, as it leads to a corrupt table. NULL
3190 pem_str is reserved for alias entries only.
3194 * Revert blinding in ECDSA sign and instead make problematic addition
3195 length-invariant. Switch even to fixed-length Montgomery multiplication.
3199 * Change generating and checking of primes so that the error rate of not
3200 being prime depends on the intended use based on the size of the input.
3201 For larger primes this will result in more rounds of Miller-Rabin.
3202 The maximal error rate for primes with more than 1080 bits is lowered
3205 *Kurt Roeckx, Annie Yousar*
3207 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3211 * Add blinding to ECDSA and DSA signatures to protect against side channel
3212 attacks discovered by Keegan Ryan (NCC Group).
3216 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3217 now allow empty (zero character) pass phrases.
3221 * Certificate time validation (X509_cmp_time) enforces stricter
3222 compliance with RFC 5280. Fractional seconds and timezone offsets
3223 are no longer allowed.
3227 * Fixed a text canonicalisation bug in CMS
3229 Where a CMS detached signature is used with text content the text goes
3230 through a canonicalisation process first prior to signing or verifying a
3231 signature. This process strips trailing space at the end of lines, converts
3232 line terminators to CRLF and removes additional trailing line terminators
3233 at the end of a file. A bug in the canonicalisation process meant that
3234 some characters, such as form-feed, were incorrectly treated as whitespace
3235 and removed. This is contrary to the specification (RFC5485). This fix
3236 could mean that detached text data signed with an earlier version of
3237 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3238 signed with a fixed OpenSSL may fail to verify with an earlier version of
3239 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3240 and use the "-binary" flag (for the "cms" command line application) or set
3241 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3245 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3247 * Constructed ASN.1 types with a recursive definition could exceed the stack
3249 Constructed ASN.1 types with a recursive definition (such as can be found
3250 in PKCS7) could eventually exceed the stack given malicious input with
3251 excessive recursion. This could result in a Denial Of Service attack. There
3252 are no such structures used within SSL/TLS that come from untrusted sources
3253 so this is considered safe.
3255 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3261 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3263 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3264 effectively reduced to only comparing the least significant bit of each
3265 byte. This allows an attacker to forge messages that would be considered as
3266 authenticated in an amount of tries lower than that guaranteed by the
3267 security claims of the scheme. The module can only be compiled by the
3268 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3270 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3276 * Add a build target 'build_all_generated', to build all generated files
3277 and only that. This can be used to prepare everything that requires
3278 things like perl for a system that lacks perl and then move everything
3279 to that system and do the rest of the build there.
3283 * Backport SSL_OP_NO_RENGOTIATION
3285 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3286 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3287 changes this is no longer possible in 1.1.0. Therefore the new
3288 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3289 1.1.0 to provide equivalent functionality.
3291 Note that if an application built against 1.1.0h headers (or above) is run
3292 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3293 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3297 * Removed the OS390-Unix config target. It relied on a script that doesn't
3302 * rsaz_1024_mul_avx2 overflow bug on x86_64
3304 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3305 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3306 Analysis suggests that attacks against RSA and DSA as a result of this
3307 defect would be very difficult to perform and are not believed likely.
3308 Attacks against DH1024 are considered just feasible, because most of the
3309 work necessary to deduce information about a private key may be performed
3310 offline. The amount of resources required for such an attack would be
3311 significant. However, for an attack on TLS to be meaningful, the server
3312 would have to share the DH1024 private key among multiple clients, which is
3313 no longer an option since CVE-2016-0701.
3315 This only affects processors that support the AVX2 but not ADX extensions
3316 like Intel Haswell (4th generation).
3318 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3319 was originally found via the OSS-Fuzz project.
3324 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3326 * bn_sqrx8x_internal carry bug on x86_64
3328 There is a carry propagating bug in the x86_64 Montgomery squaring
3329 procedure. No EC algorithms are affected. Analysis suggests that attacks
3330 against RSA and DSA as a result of this defect would be very difficult to
3331 perform and are not believed likely. Attacks against DH are considered just
3332 feasible (although very difficult) because most of the work necessary to
3333 deduce information about a private key may be performed offline. The amount
3334 of resources required for such an attack would be very significant and
3335 likely only accessible to a limited number of attackers. An attacker would
3336 additionally need online access to an unpatched system using the target
3337 private key in a scenario with persistent DH parameters and a private
3338 key that is shared between multiple clients.
3340 This only affects processors that support the BMI1, BMI2 and ADX extensions
3341 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3343 This issue was reported to OpenSSL by the OSS-Fuzz project.
3348 * Malformed X.509 IPAddressFamily could cause OOB read
3350 If an X.509 certificate has a malformed IPAddressFamily extension,
3351 OpenSSL could do a one-byte buffer overread. The most likely result
3352 would be an erroneous display of the certificate in text format.
3354 This issue was reported to OpenSSL by the OSS-Fuzz project.
3359 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3361 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3362 platform rather than 'mingw'.
3366 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3367 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3368 which is the minimum version we support.
3372 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3374 * Encrypt-Then-Mac renegotiation crash
3376 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3377 negotiated where it was not in the original handshake (or vice-versa) then
3378 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3379 and servers are affected.
3381 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3386 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3388 * Truncated packet could crash via OOB read
3390 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3391 cipher is being used, then a truncated packet can cause that host to
3392 perform an out-of-bounds read, usually resulting in a crash.
3394 This issue was reported to OpenSSL by Robert Święcki of Google.
3399 * Bad (EC)DHE parameters cause a client crash
3401 If a malicious server supplies bad parameters for a DHE or ECDHE key
3402 exchange then this can result in the client attempting to dereference a
3403 NULL pointer leading to a client crash. This could be exploited in a Denial
3406 This issue was reported to OpenSSL by Guido Vranken.
3411 * BN_mod_exp may produce incorrect results on x86_64
3413 There is a carry propagating bug in the x86_64 Montgomery squaring
3414 procedure. No EC algorithms are affected. Analysis suggests that attacks
3415 against RSA and DSA as a result of this defect would be very difficult to
3416 perform and are not believed likely. Attacks against DH are considered just
3417 feasible (although very difficult) because most of the work necessary to
3418 deduce information about a private key may be performed offline. The amount
3419 of resources required for such an attack would be very significant and
3420 likely only accessible to a limited number of attackers. An attacker would
3421 additionally need online access to an unpatched system using the target
3422 private key in a scenario with persistent DH parameters and a private
3423 key that is shared between multiple clients. For example this can occur by
3424 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3425 similar to CVE-2015-3193 but must be treated as a separate problem.
3427 This issue was reported to OpenSSL by the OSS-Fuzz project.
3432 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3434 * ChaCha20/Poly1305 heap-buffer-overflow
3436 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3437 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3438 crash. This issue is not considered to be exploitable beyond a DoS.
3440 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3445 * CMS Null dereference
3447 Applications parsing invalid CMS structures can crash with a NULL pointer
3448 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3449 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3450 structure callback if an attempt is made to free certain invalid encodings.
3451 Only CHOICE structures using a callback which do not handle NULL value are
3454 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3459 * Montgomery multiplication may produce incorrect results
3461 There is a carry propagating bug in the Broadwell-specific Montgomery
3462 multiplication procedure that handles input lengths divisible by, but
3463 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3464 and DH private keys are impossible. This is because the subroutine in
3465 question is not used in operations with the private key itself and an input
3466 of the attacker's direct choice. Otherwise the bug can manifest itself as
3467 transient authentication and key negotiation failures or reproducible
3468 erroneous outcome of public-key operations with specially crafted input.
3469 Among EC algorithms only Brainpool P-512 curves are affected and one
3470 presumably can attack ECDH key negotiation. Impact was not analyzed in
3471 detail, because pre-requisites for attack are considered unlikely. Namely
3472 multiple clients have to choose the curve in question and the server has to
3473 share the private key among them, neither of which is default behaviour.
3474 Even then only clients that chose the curve will be affected.
3476 This issue was publicly reported as transient failures and was not
3477 initially recognized as a security issue. Thanks to Richard Morgan for
3478 providing reproducible case.
3483 * Removed automatic addition of RPATH in shared libraries and executables,
3484 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3488 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3490 * Fix Use After Free for large message sizes
3492 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3493 message larger than approx 16k is received then the underlying buffer to
3494 store the incoming message is reallocated and moved. Unfortunately a
3495 dangling pointer to the old location is left which results in an attempt to
3496 write to the previously freed location. This is likely to result in a
3497 crash, however it could potentially lead to execution of arbitrary code.
3499 This issue only affects OpenSSL 1.1.0a.
3501 This issue was reported to OpenSSL by Robert Święcki.
3506 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3508 * OCSP Status Request extension unbounded memory growth
3510 A malicious client can send an excessively large OCSP Status Request
3511 extension. If that client continually requests renegotiation, sending a
3512 large OCSP Status Request extension each time, then there will be unbounded
3513 memory growth on the server. This will eventually lead to a Denial Of
3514 Service attack through memory exhaustion. Servers with a default
3515 configuration are vulnerable even if they do not support OCSP. Builds using
3516 the "no-ocsp" build time option are not affected.
3518 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3523 * SSL_peek() hang on empty record
3525 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3526 sends an empty record. This could be exploited by a malicious peer in a
3527 Denial Of Service attack.
3529 This issue was reported to OpenSSL by Alex Gaynor.
3534 * Excessive allocation of memory in tls_get_message_header() and
3535 dtls1_preprocess_fragment()
3537 A (D)TLS message includes 3 bytes for its length in the header for the
3538 message. This would allow for messages up to 16Mb in length. Messages of
3539 this length are excessive and OpenSSL includes a check to ensure that a
3540 peer is sending reasonably sized messages in order to avoid too much memory
3541 being consumed to service a connection. A flaw in the logic of version
3542 1.1.0 means that memory for the message is allocated too early, prior to
3543 the excessive message length check. Due to way memory is allocated in
3544 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3545 to service a connection. This could lead to a Denial of Service through
3546 memory exhaustion. However, the excessive message length check still takes
3547 place, and this would cause the connection to immediately fail. Assuming
3548 that the application calls SSL_free() on the failed connection in a timely
3549 manner then the 21Mb of allocated memory will then be immediately freed
3550 again. Therefore the excessive memory allocation will be transitory in
3551 nature. This then means that there is only a security impact if:
3553 1) The application does not call SSL_free() in a timely manner in the event
3554 that the connection fails
3556 2) The application is working in a constrained environment where there is
3557 very little free memory
3559 3) The attacker initiates multiple connection attempts such that there are
3560 multiple connections in a state where memory has been allocated for the
3561 connection; SSL_free() has not yet been called; and there is insufficient
3562 memory to service the multiple requests.
3564 Except in the instance of (1) above any Denial Of Service is likely to be
3565 transitory because as soon as the connection fails the memory is
3566 subsequently freed again in the SSL_free() call. However there is an
3567 increased risk during this period of application crashes due to the lack of
3568 memory - which would then mean a more serious Denial of Service.
3570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3571 (CVE-2016-6307 and CVE-2016-6308)
3575 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3576 had to be removed. Primary reason is that vendor assembler can't
3577 assemble our modules with -KPIC flag. As result it, assembly
3578 support, was not even available as option. But its lack means
3579 lack of side-channel resistant code, which is incompatible with
3580 security by todays standards. Fortunately gcc is readily available
3581 prepackaged option, which we firmly point at...
3585 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3587 * Windows command-line tool supports UTF-8 opt-in option for arguments
3588 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3589 (to any value) allows Windows user to access PKCS#12 file generated
3590 with Windows CryptoAPI and protected with non-ASCII password, as well
3591 as files generated under UTF-8 locale on Linux also protected with
3596 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3597 have been disabled by default and removed from DEFAULT, just like RC4.
3598 See the RC4 item below to re-enable both.
3602 * The method for finding the storage location for the Windows RAND seed file
3603 has changed. First we check %RANDFILE%. If that is not set then we check
3604 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3605 all else fails we fall back to C:\.
3609 * The EVP_EncryptUpdate() function has had its return type changed from void
3610 to int. A return of 0 indicates and error while a return of 1 indicates
3615 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3616 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3617 off the constant time implementation for RSA, DSA and DH have been made
3618 no-ops and deprecated.
3622 * Windows RAND implementation was simplified to only get entropy by
3623 calling CryptGenRandom(). Various other RAND-related tickets
3626 *Joseph Wylie Yandle, Rich Salz*
3628 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3629 and `OPENSSL_LH_`, respectively. The old names are available
3630 with API compatibility. They new names are now completely documented.
3634 * Unify TYPE_up_ref(obj) methods signature.
3635 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3636 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3637 int (instead of void) like all others TYPE_up_ref() methods.
3638 So now these methods also check the return value of CRYPTO_atomic_add(),
3639 and the validity of object reference counter.
3641 *fdasilvayy@gmail.com*
3643 * With Windows Visual Studio builds, the .pdb files are installed
3644 alongside the installed libraries and executables. For a static
3645 library installation, ossl_static.pdb is the associate compiler
3646 generated .pdb file to be used when linking programs.
3650 * Remove openssl.spec. Packaging files belong with the packagers.
3654 * Automatic Darwin/OSX configuration has had a refresh, it will now
3655 recognise x86_64 architectures automatically. You can still decide
3656 to build for a different bitness with the environment variable
3657 KERNEL_BITS (can be 32 or 64), for example:
3659 KERNEL_BITS=32 ./config
3663 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3664 256 bit AES and HMAC with SHA256.
3668 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3672 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3676 * To enable users to have their own config files and build file templates,
3677 Configure looks in the directory indicated by the environment variable
3678 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3679 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3680 name and is used as is.
3684 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3685 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3686 X509_CERT_FILE_CTX was removed.
3690 * "shared" builds are now the default. To create only static libraries use
3691 the "no-shared" Configure option.
3695 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3696 All of these option have not worked for some while and are fundamental
3701 * Make various cleanup routines no-ops and mark them as deprecated. Most
3702 global cleanup functions are no longer required because they are handled
3703 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3704 Explicitly de-initing can cause problems (e.g. where a library that uses
3705 OpenSSL de-inits, but an application is still using it). The affected
3706 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3707 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3708 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3709 COMP_zlib_cleanup().
3713 * --strict-warnings no longer enables runtime debugging options
3714 such as REF_DEBUG. Instead, debug options are automatically
3715 enabled with '--debug' builds.
3717 *Andy Polyakov, Emilia Käsper*
3719 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3720 have been moved out of the public header files. New functions for managing
3721 these have been added.
3725 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3726 objects have been moved out of the public header files. New
3727 functions for managing these have been added.
3731 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3732 have been moved out of the public header files. New functions for managing
3733 these have been added.
3737 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3738 moved out of the public header files. New functions for managing these
3743 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3747 * Removed the mk1mf build scripts.
3751 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3752 it is always safe to #include a header now.
3756 * Removed the aged BC-32 config and all its supporting scripts
3760 * Removed support for Ultrix, Netware, and OS/2.
3764 * Add support for HKDF.
3766 *Alessandro Ghedini*
3768 * Add support for blake2b and blake2s
3772 * Added support for "pipelining". Ciphers that have the
3773 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3774 encryptions/decryptions simultaneously. There are currently no built-in
3775 ciphers with this property but the expectation is that engines will be able
3776 to offer it to significantly improve throughput. Support has been extended
3777 into libssl so that multiple records for a single connection can be
3778 processed in one go (for >=TLS 1.1).
3782 * Added the AFALG engine. This is an async capable engine which is able to
3783 offload work to the Linux kernel. In this initial version it only supports
3784 AES128-CBC. The kernel must be version 4.1.0 or greater.
3788 * OpenSSL now uses a new threading API. It is no longer necessary to
3789 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3790 are two supported threading models: pthreads and windows threads. It is
3791 also possible to configure OpenSSL at compile time for "no-threads". The
3792 old threading API should no longer be used. The functions have been
3793 replaced with "no-op" compatibility macros.
3795 *Alessandro Ghedini, Matt Caswell*
3797 * Modify behavior of ALPN to invoke callback after SNI/servername
3798 callback, such that updates to the SSL_CTX affect ALPN.
3802 * Add SSL_CIPHER queries for authentication and key-exchange.
3806 * Changes to the DEFAULT cipherlist:
3807 - Prefer (EC)DHE handshakes over plain RSA.
3808 - Prefer AEAD ciphers over legacy ciphers.
3809 - Prefer ECDSA over RSA when both certificates are available.
3810 - Prefer TLSv1.2 ciphers/PRF.
3811 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3816 * Change the ECC default curve list to be this, in order: x25519,
3817 secp256r1, secp521r1, secp384r1.
3821 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3822 disabled by default. They can be re-enabled using the
3823 enable-weak-ssl-ciphers option to Configure.
3827 * If the server has ALPN configured, but supports no protocols that the
3828 client advertises, send a fatal "no_application_protocol" alert.
3829 This behaviour is SHALL in RFC 7301, though it isn't universally
3830 implemented by other servers.
3834 * Add X25519 support.
3835 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3836 for public and private key encoding using the format documented in
3837 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3838 key generation and key derivation.
3840 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3845 * Deprecate SRP_VBASE_get_by_user.
3846 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3847 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3848 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3849 seed, even if the seed is configured.
3851 Users should use SRP_VBASE_get1_by_user instead. Note that in
3852 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3853 also that even though configuring the SRP seed attempts to hide
3854 invalid usernames by continuing the handshake with fake
3855 credentials, this behaviour is not constant time and no strong
3856 guarantees are made that the handshake is indistinguishable from
3857 that of a valid user.
3861 * Configuration change; it's now possible to build dynamic engines
3862 without having to build shared libraries and vice versa. This
3863 only applies to the engines in `engines/`, those in `crypto/engine/`
3864 will always be built into libcrypto (i.e. "static").
3866 Building dynamic engines is enabled by default; to disable, use
3867 the configuration option "disable-dynamic-engine".
3869 The only requirements for building dynamic engines are the
3870 presence of the DSO module and building with position independent
3871 code, so they will also automatically be disabled if configuring
3872 with "disable-dso" or "disable-pic".
3874 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3875 are also taken away from openssl/opensslconf.h, as they are
3880 * Configuration change; if there is a known flag to compile
3881 position independent code, it will always be applied on the
3882 libcrypto and libssl object files, and never on the application
3883 object files. This means other libraries that use routines from
3884 libcrypto / libssl can be made into shared libraries regardless
3885 of how OpenSSL was configured.
3887 If this isn't desirable, the configuration options "disable-pic"
3888 or "no-pic" can be used to disable the use of PIC. This will
3889 also disable building shared libraries and dynamic engines.
3893 * Removed JPAKE code. It was experimental and has no wide use.
3897 * The INSTALL_PREFIX Makefile variable has been renamed to
3898 DESTDIR. That makes for less confusion on what this variable
3899 is for. Also, the configuration option --install_prefix is
3904 * Heartbeat for TLS has been removed and is disabled by default
3905 for DTLS; configure with enable-heartbeats. Code that uses the
3906 old #define's might need to be updated.
3908 *Emilia Käsper, Rich Salz*
3910 * Rename REF_CHECK to REF_DEBUG.
3914 * New "unified" build system
3916 The "unified" build system is aimed to be a common system for all
3917 platforms we support. With it comes new support for VMS.
3919 This system builds supports building in a different directory tree
3920 than the source tree. It produces one Makefile (for unix family
3921 or lookalikes), or one descrip.mms (for VMS).
3923 The source of information to make the Makefile / descrip.mms is
3924 small files called 'build.info', holding the necessary
3925 information for each directory with source to compile, and a
3926 template in Configurations, like unix-Makefile.tmpl or
3929 With this change, the library names were also renamed on Windows
3930 and on VMS. They now have names that are closer to the standard
3931 on Unix, and include the major version number, and in certain
3932 cases, the architecture they are built for. See "Notes on shared
3933 libraries" in INSTALL.
3935 We rely heavily on the perl module Text::Template.
3939 * Added support for auto-initialisation and de-initialisation of the library.
3940 OpenSSL no longer requires explicit init or deinit routines to be called,
3941 except in certain circumstances. See the OPENSSL_init_crypto() and
3942 OPENSSL_init_ssl() man pages for further information.
3946 * The arguments to the DTLSv1_listen function have changed. Specifically the
3947 "peer" argument is now expected to be a BIO_ADDR object.
3949 * Rewrite of BIO networking library. The BIO library lacked consistent
3950 support of IPv6, and adding it required some more extensive
3951 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3952 which hold all types of addresses and chains of address information.
3953 It also introduces a new API, with functions like BIO_socket,
3954 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3955 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3956 have been adapted accordingly.
3960 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3965 * CRIME protection: disable compression by default, even if OpenSSL is
3966 compiled with zlib enabled. Applications can still enable compression
3967 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3968 using the SSL_CONF library to configure compression.
3972 * The signature of the session callback configured with
3973 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3974 was explicitly marked as `const unsigned char*` instead of
3979 * Always DPURIFY. Remove the use of uninitialized memory in the
3980 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3984 * Removed many obsolete configuration items, including
3985 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3986 MD2_CHAR, MD2_INT, MD2_LONG
3988 IDEA_SHORT, IDEA_LONG
3989 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3991 *Rich Salz, with advice from Andy Polyakov*
3993 * Many BN internals have been moved to an internal header file.
3995 *Rich Salz with help from Andy Polyakov*
3997 * Configuration and writing out the results from it has changed.
3998 Files such as Makefile include/openssl/opensslconf.h and are now
3999 produced through general templates, such as Makefile.in and
4000 crypto/opensslconf.h.in and some help from the perl module
4003 Also, the center of configuration information is no longer
4004 Makefile. Instead, Configure produces a perl module in
4005 configdata.pm which holds most of the config data (in the hash
4006 table %config), the target data that comes from the target
4007 configuration in one of the `Configurations/*.conf` files (in
4012 * To clarify their intended purposes, the Configure options
4013 --prefix and --openssldir change their semantics, and become more
4014 straightforward and less interdependent.
4016 --prefix shall be used exclusively to give the location INSTALLTOP
4017 where programs, scripts, libraries, include files and manuals are
4018 going to be installed. The default is now /usr/local.
4020 --openssldir shall be used exclusively to give the default
4021 location OPENSSLDIR where certificates, private keys, CRLs are
4022 managed. This is also where the default openssl.cnf gets
4024 If the directory given with this option is a relative path, the
4025 values of both the --prefix value and the --openssldir value will
4026 be combined to become OPENSSLDIR.
4027 The default for --openssldir is INSTALLTOP/ssl.
4029 Anyone who uses --openssldir to specify where OpenSSL is to be
4030 installed MUST change to use --prefix instead.
4034 * The GOST engine was out of date and therefore it has been removed. An up
4035 to date GOST engine is now being maintained in an external repository.
4036 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4037 support for GOST ciphersuites (these are only activated if a GOST engine
4042 * EGD is no longer supported by default; use enable-egd when
4045 *Ben Kaduk and Rich Salz*
4047 * The distribution now has Makefile.in files, which are used to
4048 create Makefile's when Configure is run. *Configure must be run
4049 before trying to build now.*
4053 * The return value for SSL_CIPHER_description() for error conditions
4058 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4060 Obtaining and performing DNSSEC validation of TLSA records is
4061 the application's responsibility. The application provides
4062 the TLSA records of its choice to OpenSSL, and these are then
4063 used to authenticate the peer.
4065 The TLSA records need not even come from DNS. They can, for
4066 example, be used to implement local end-entity certificate or
4067 trust-anchor "pinning", where the "pin" data takes the form
4068 of TLSA records, which can augment or replace verification
4069 based on the usual WebPKI public certification authorities.
4073 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4074 continues to support deprecated interfaces in default builds.
4075 However, applications are strongly advised to compile their
4076 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4077 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4078 or the 1.1.0 releases.
4080 In environments in which all applications have been ported to
4081 not use any deprecated interfaces OpenSSL's Configure script
4082 should be used with the --api=1.1.0 option to entirely remove
4083 support for the deprecated features from the library and
4084 unconditionally disable them in the installed headers.
4085 Essentially the same effect can be achieved with the "no-deprecated"
4086 argument to Configure, except that this will always restrict
4087 the build to just the latest API, rather than a fixed API
4090 As applications are ported to future revisions of the API,
4091 they should update their compile-time OPENSSL_API_COMPAT define
4092 accordingly, but in most cases should be able to continue to
4093 compile with later releases.
4095 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4096 0x10000000L and 0x00908000L, respectively. However those
4097 versions did not support the OPENSSL_API_COMPAT feature, and
4098 so applications are not typically tested for explicit support
4099 of just the undeprecated features of either release.
4103 * Add support for setting the minimum and maximum supported protocol.
4104 It can bet set via the SSL_set_min_proto_version() and
4105 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4106 MaxProtocol. It's recommended to use the new APIs to disable
4107 protocols instead of disabling individual protocols using
4108 SSL_set_options() or SSL_CONF's Protocol. This change also
4109 removes support for disabling TLS 1.2 in the OpenSSL TLS
4110 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4114 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4118 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4119 and integrates ECDSA and ECDH functionality into EC. Implementations can
4120 now redirect key generation and no longer need to convert to or from
4123 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4124 include the ec.h header file instead.
4128 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4129 ciphers who are no longer supported and drops support the ephemeral RSA key
4130 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4134 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4135 opaque. For HMAC_CTX, the following constructors and destructors
4138 HMAC_CTX *HMAC_CTX_new(void);
4139 void HMAC_CTX_free(HMAC_CTX *ctx);
4141 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4142 destroy such methods has been added. See EVP_MD_meth_new(3) and
4143 EVP_CIPHER_meth_new(3) for documentation.
4146 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4147 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4148 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4149 an already created structure.
4150 2) For consistency with the majority of our object creators and
4151 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4152 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4153 for deprecated builds.
4157 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4158 cryptographic operations to be performed asynchronously as long as an
4159 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4160 further details. Libssl has also had this capability integrated with the
4161 introduction of the new mode SSL_MODE_ASYNC and associated error
4162 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4163 pages. This work was developed in partnership with Intel Corp.
4167 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4168 always enabled now. If you want to disable the support you should
4169 exclude it using the list of supported ciphers. This also means that the
4170 "-no_ecdhe" option has been removed from s_server.
4174 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4175 SSL_{CTX_}set1_curves() which can set a list.
4179 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4180 curve you want to support using SSL_{CTX_}set1_curves().
4184 * State machine rewrite. The state machine code has been significantly
4185 refactored in order to remove much duplication of code and solve issues
4186 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4187 further details). This change does have some associated API changes.
4188 Notably the SSL_state() function has been removed and replaced by
4189 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4190 SSL_set_state() has been removed altogether. The previous handshake states
4191 defined in ssl.h and ssl3.h have also been removed.
4195 * All instances of the string "ssleay" in the public API were replaced
4196 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4197 Some error codes related to internal RSA_eay API's were renamed.
4201 * The demo files in crypto/threads were moved to demo/threads.
4205 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4208 *Matt Caswell, Rich Salz*
4210 * New ASN.1 embed macro.
4212 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4213 structure is not allocated: it is part of the parent. That is instead of
4221 This reduces memory fragmentation and make it impossible to accidentally
4222 set a mandatory field to NULL.
4224 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4225 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4226 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4231 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4235 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4236 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4237 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4238 DES and RC4 ciphersuites.
4242 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4243 This changes the decoding behaviour for some invalid messages,
4244 though the change is mostly in the more lenient direction, and
4245 legacy behaviour is preserved as much as possible.
4249 * Fix no-stdio build.
4250 *David Woodhouse <David.Woodhouse@intel.com> and also*
4251 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4253 * New testing framework
4254 The testing framework has been largely rewritten and is now using
4255 perl and the perl modules Test::Harness and an extended variant of
4256 Test::More called OpenSSL::Test to do its work. All test scripts in
4257 test/ have been rewritten into test recipes, and all direct calls to
4258 executables in test/Makefile have become individual recipes using the
4259 simplified testing OpenSSL::Test::Simple.
4261 For documentation on our testing modules, do:
4263 perldoc test/testlib/OpenSSL/Test/Simple.pm
4264 perldoc test/testlib/OpenSSL/Test.pm
4268 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4269 are used; the latter aborts on memory leaks (usually checked on exit).
4270 Some undocumented "set malloc, etc., hooks" functions were removed
4271 and others were changed. All are now documented.
4275 * In DSA_generate_parameters_ex, if the provided seed is too short,
4278 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4280 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4281 from RFC4279, RFC4785, RFC5487, RFC5489.
4283 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4284 original RSA_PSK patch.
4288 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4289 era flag was never set throughout the codebase (only read). Also removed
4290 SSL3_FLAGS_POP_BUFFER which was only used if
4291 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4295 * Changed the default name options in the "ca", "crl", "req" and "x509"
4296 to be "oneline" instead of "compat".
4300 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4301 not aware of clients that still exhibit this bug, and the workaround
4302 hasn't been working properly for a while.
4306 * The return type of BIO_number_read() and BIO_number_written() as well as
4307 the corresponding num_read and num_write members in the BIO structure has
4308 changed from unsigned long to uint64_t. On platforms where an unsigned
4309 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4314 * Given the pervasive nature of TLS extensions it is inadvisable to run
4315 OpenSSL without support for them. It also means that maintaining
4316 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4317 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4321 * Removed support for the two export grade static DH ciphersuites
4322 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4323 were newly added (along with a number of other static DH ciphersuites) to
4324 1.0.2. However the two export ones have *never* worked since they were
4325 introduced. It seems strange in any case to be adding new export
4326 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4330 * Version negotiation has been rewritten. In particular SSLv23_method(),
4331 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4332 and turned into macros which simply call the new preferred function names
4333 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4334 should use the new names instead. Also as part of this change the ssl23.h
4335 header file has been removed.
4339 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4340 code and the associated standard is no longer considered fit-for-purpose.
4344 * RT2547 was closed. When generating a private key, try to make the
4345 output file readable only by the owner. This behavior change might
4346 be noticeable when interacting with other software.
4348 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4353 * Added HTTP GET support to the ocsp command.
4357 * Changed default digest for the dgst and enc commands from MD5 to
4362 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4366 * Added support for TLS extended master secret from
4367 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4368 initial patch which was a great help during development.
4372 * All libssl internal structures have been removed from the public header
4373 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4374 now redundant). Users should not attempt to access internal structures
4375 directly. Instead they should use the provided API functions.
4379 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4380 Access to deprecated functions can be re-enabled by running config with
4381 "enable-deprecated". In addition applications wishing to use deprecated
4382 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4383 will, by default, disable some transitive includes that previously existed
4384 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4388 * Added support for OCB mode. OpenSSL has been granted a patent license
4389 compatible with the OpenSSL license for use of OCB. Details are available
4390 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4391 for OCB can be removed by calling config with no-ocb.
4395 * SSLv2 support has been removed. It still supports receiving a SSLv2
4396 compatible client hello.
4400 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4401 done while fixing the error code for the key-too-small case.
4403 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4405 * CA.sh has been removed; use CA.pl instead.
4409 * Removed old DES API.
4413 * Remove various unsupported platforms:
4419 Sinix/ReliantUNIX RM400
4424 16-bit platforms such as WIN16
4428 * Clean up OPENSSL_NO_xxx #define's
4429 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4430 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4431 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4432 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4433 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4434 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4435 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4436 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4437 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4438 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4442 * Cleaned up dead code
4443 Remove all but one '#ifdef undef' which is to be looked at.
4447 * Clean up calling of xxx_free routines.
4448 Just like free(), fix most of the xxx_free routines to accept
4449 NULL. Remove the non-null checks from callers. Save much code.
4453 * Add secure heap for storage of private keys (when possible).
4454 Add BIO_s_secmem(), CBIGNUM, etc.
4455 Contributed by Akamai Technologies under our Corporate CLA.
4459 * Experimental support for a new, fast, unbiased prime candidate generator,
4460 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4462 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4464 * New output format NSS in the sess_id command line tool. This allows
4465 exporting the session id and the master key in NSS keylog format.
4467 *Martin Kaiser <martin@kaiser.cx>*
4469 * Harmonize version and its documentation. -f flag is used to display
4472 *mancha <mancha1@zoho.com>*
4474 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4475 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4477 *mancha <mancha1@zoho.com>*
4479 * Fix some double frees. These are not thought to be exploitable.
4481 *mancha <mancha1@zoho.com>*
4483 * A missing bounds check in the handling of the TLS heartbeat extension
4484 can be used to reveal up to 64k of memory to a connected client or
4487 Thanks for Neel Mehta of Google Security for discovering this bug and to
4488 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4489 preparing the fix ([CVE-2014-0160])
4491 *Adam Langley, Bodo Moeller*
4493 * Fix for the attack described in the paper "Recovering OpenSSL
4494 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4495 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4496 <http://eprint.iacr.org/2014/140>
4498 Thanks to Yuval Yarom and Naomi Benger for discovering this
4499 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4501 *Yuval Yarom and Naomi Benger*
4503 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4504 this fixes a limitation in previous versions of OpenSSL.
4508 * Experimental encrypt-then-mac support.
4510 Experimental support for encrypt then mac from
4511 draft-gutmann-tls-encrypt-then-mac-02.txt
4513 To enable it set the appropriate extension number (0x42 for the test
4514 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4516 For non-compliant peers (i.e. just about everything) this should have no
4519 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4523 * Add EVP support for key wrapping algorithms, to avoid problems with
4524 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4525 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4526 algorithms and include tests cases.
4530 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4535 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4536 MGF1 digest and OAEP label.
4540 * Make openssl verify return errors.
4542 *Chris Palmer <palmer@google.com> and Ben Laurie*
4544 * New function ASN1_TIME_diff to calculate the difference between two
4545 ASN1_TIME structures or one structure and the current time.
4549 * Update fips_test_suite to support multiple command line options. New
4550 test to induce all self test errors in sequence and check expected
4555 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4556 sign or verify all in one operation.
4560 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4561 test programs and fips_test_suite. Includes functionality to parse
4562 the minimal script output of fipsalgest.pl directly.
4566 * Add authorisation parameter to FIPS_module_mode_set().
4570 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4574 * Use separate DRBG fields for internal and external flags. New function
4575 FIPS_drbg_health_check() to perform on demand health checking. Add
4576 generation tests to fips_test_suite with reduced health check interval to
4577 demonstrate periodic health checking. Add "nodh" option to
4578 fips_test_suite to skip very slow DH test.
4582 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4587 * More extensive health check for DRBG checking many more failure modes.
4588 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4589 combination: call this in fips_test_suite.
4593 * Add support for canonical generation of DSA parameter 'g'. See
4596 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4597 POST to handle HMAC cases.
4601 * Add functions FIPS_module_version() and FIPS_module_version_text()
4602 to return numerical and string versions of the FIPS module number.
4606 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4607 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4608 outside the validated module in the FIPS capable OpenSSL.
4612 * Minor change to DRBG entropy callback semantics. In some cases
4613 there is no multiple of the block length between min_len and
4614 max_len. Allow the callback to return more than max_len bytes
4615 of entropy but discard any extra: it is the callback's responsibility
4616 to ensure that the extra data discarded does not impact the
4617 requested amount of entropy.
4621 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4622 information in FIPS186-3, SP800-57 and SP800-131A.
4626 * CCM support via EVP. Interface is very similar to GCM case except we
4627 must supply all data in one chunk (i.e. no update, final) and the
4628 message length must be supplied if AAD is used. Add algorithm test
4633 * Initial version of POST overhaul. Add POST callback to allow the status
4634 of POST to be monitored and/or failures induced. Modify fips_test_suite
4635 to use callback. Always run all selftests even if one fails.
4639 * XTS support including algorithm test driver in the fips_gcmtest program.
4640 Note: this does increase the maximum key length from 32 to 64 bytes but
4641 there should be no binary compatibility issues as existing applications
4642 will never use XTS mode.
4646 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4647 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4648 performs algorithm blocking for unapproved PRNG types. Also do not
4649 set PRNG type in FIPS_mode_set(): leave this to the application.
4650 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4651 the standard OpenSSL PRNG: set additional data to a date time vector.
4655 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4656 This shouldn't present any incompatibility problems because applications
4657 shouldn't be using these directly and any that are will need to rethink
4658 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4662 * Extensive self tests and health checking required by SP800-90 DRBG.
4663 Remove strength parameter from FIPS_drbg_instantiate and always
4664 instantiate at maximum supported strength.
4668 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4672 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4676 * New function DH_compute_key_padded() to compute a DH key and pad with
4677 leading zeroes if needed: this complies with SP800-56A et al.
4681 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4682 anything, incomplete, subject to change and largely untested at present.
4686 * Modify fipscanisteronly build option to only build the necessary object
4687 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4691 * Add experimental option FIPSSYMS to give all symbols in
4692 fipscanister.o and FIPS or fips prefix. This will avoid
4693 conflicts with future versions of OpenSSL. Add perl script
4694 util/fipsas.pl to preprocess assembly language source files
4695 and rename any affected symbols.
4699 * Add selftest checks and algorithm block of non-fips algorithms in
4700 FIPS mode. Remove DES2 from selftests.
4704 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4705 return internal method without any ENGINE dependencies. Add new
4706 tiny fips sign and verify functions.
4710 * New build option no-ec2m to disable characteristic 2 code.
4714 * New build option "fipscanisteronly". This only builds fipscanister.o
4715 and (currently) associated fips utilities. Uses the file Makefile.fips
4716 instead of Makefile.org as the prototype.
4720 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4721 Update fips_gcmtest to use IV generator.
4725 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4726 setting output buffer to NULL. The `*Final` function must be
4727 called although it will not retrieve any additional data. The tag
4728 can be set or retrieved with a ctrl. The IV length is by default 12
4729 bytes (96 bits) but can be set to an alternative value. If the IV
4730 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4735 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4736 underlying do_cipher function handles all cipher semantics itself
4737 including padding and finalisation. This is useful if (for example)
4738 an ENGINE cipher handles block padding itself. The behaviour of
4739 do_cipher is subtly changed if this flag is set: the return value
4740 is the number of characters written to the output buffer (zero is
4741 no longer an error code) or a negative error code. Also if the
4742 input buffer is NULL and length 0 finalisation should be performed.
4746 * If a candidate issuer certificate is already part of the constructed
4747 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4751 * Improve forward-security support: add functions
4753 void SSL_CTX_set_not_resumable_session_callback(
4754 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4755 void SSL_set_not_resumable_session_callback(
4756 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4758 for use by SSL/TLS servers; the callback function will be called whenever a
4759 new session is created, and gets to decide whether the session may be
4760 cached to make it resumable (return 0) or not (return 1). (As by the
4761 SSL/TLS protocol specifications, the session_id sent by the server will be
4762 empty to indicate that the session is not resumable; also, the server will
4763 not generate RFC 4507 (RFC 5077) session tickets.)
4765 A simple reasonable callback implementation is to return is_forward_secure.
4766 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4767 by the SSL/TLS server library, indicating whether it can provide forward
4770 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4772 * New -verify_name option in command line utilities to set verification
4777 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4778 Add CMAC pkey methods.
4782 * Experimental renegotiation in s_server -www mode. If the client
4783 browses /reneg connection is renegotiated. If /renegcert it is
4784 renegotiated requesting a certificate.
4788 * Add an "external" session cache for debugging purposes to s_server. This
4789 should help trace issues which normally are only apparent in deployed
4790 multi-process servers.
4794 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4795 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4796 BIO_set_cipher() and some obscure PEM functions were changed so they
4797 can now return an error. The RAND changes required a change to the
4798 RAND_METHOD structure.
4802 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4803 a gcc attribute to warn if the result of a function is ignored. This
4804 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4805 whose return value is often ignored.
4809 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4810 These allow SCTs (signed certificate timestamps) to be requested and
4811 validated when establishing a connection.
4813 *Rob Percival <robpercival@google.com>*
4818 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4820 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4821 used even when parsing explicit parameters, when loading a encoded key
4822 or calling `EC_GROUP_new_from_ecpkparameters()`/
4823 `EC_GROUP_new_from_ecparameters()`.
4824 This prevents bypass of security hardening and performance gains,
4825 especially for curves with specialized EC_METHODs.
4826 By default, if a key encoded with explicit parameters is loaded and later
4827 encoded, the output is still encoded with explicit parameters, even if
4828 internally a "named" EC_GROUP is used for computation.
4832 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4833 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4834 NULL. After this change, only the cofactor parameter can be NULL. It also
4835 does some minimal sanity checks on the passed order.
4840 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4841 An attack is simple, if the first CMS_recipientInfo is valid but the
4842 second CMS_recipientInfo is chosen ciphertext. If the second
4843 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4844 encryption key will be replaced by garbage, and the message cannot be
4845 decoded, but if the RSA decryption fails, the correct encryption key is
4846 used and the recipient will not notice the attack.
4847 As a work around for this potential attack the length of the decrypted
4848 key must be equal to the cipher default key length, in case the
4849 certifiate is not given and all recipientInfo are tried out.
4850 The old behaviour can be re-enabled in the CMS code by setting the
4851 CMS_DEBUG_DECRYPT flag.
4856 * Document issue with installation paths in diverse Windows builds
4858 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4859 binaries and run-time config file.
4864 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4866 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4867 This changes the size when using the `genpkey` command when no size is given.
4868 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4869 generation commands to use 2048 bits by default.
4873 * Add FIPS support for Android Arm 64-bit
4875 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4876 Module in Version 2.0.10. For some reason, the corresponding target
4877 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4878 built with FIPS support on Android Arm 64-bit. This omission has been
4881 *Matthias St. Pierre*
4883 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4885 * 0-byte record padding oracle
4887 If an application encounters a fatal protocol error and then calls
4888 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4889 then OpenSSL can respond differently to the calling application if a 0 byte
4890 record is received with invalid padding compared to if a 0 byte record is
4891 received with an invalid MAC. If the application then behaves differently
4892 based on that in a way that is detectable to the remote peer, then this
4893 amounts to a padding oracle that could be used to decrypt data.
4895 In order for this to be exploitable "non-stitched" ciphersuites must be in
4896 use. Stitched ciphersuites are optimised implementations of certain
4897 commonly used ciphersuites. Also the application must call SSL_shutdown()
4898 twice even if a protocol error has occurred (applications should not do
4899 this but some do anyway).
4901 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4902 Aviram, with additional investigation by Steven Collison and Andrew
4903 Hourselt. It was reported to OpenSSL on 10th December 2018.
4908 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4912 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4914 * Microarchitecture timing vulnerability in ECC scalar multiplication
4916 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4917 shown to be vulnerable to a microarchitecture timing side channel attack.
4918 An attacker with sufficient access to mount local timing attacks during
4919 ECDSA signature generation could recover the private key.
4921 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4922 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4928 * Timing vulnerability in DSA signature generation
4930 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4931 timing side channel attack. An attacker could use variations in the signing
4932 algorithm to recover the private key.
4934 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4939 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4940 Module, accidentally introduced while backporting security fixes from the
4941 development branch and hindering the use of ECC in FIPS mode.
4945 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4947 * Client DoS due to large DH parameter
4949 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4950 malicious server can send a very large prime value to the client. This will
4951 cause the client to spend an unreasonably long period of time generating a
4952 key for this prime resulting in a hang until the client has finished. This
4953 could be exploited in a Denial Of Service attack.
4955 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4960 * Cache timing vulnerability in RSA Key Generation
4962 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4963 a cache timing side channel attack. An attacker with sufficient access to
4964 mount cache timing attacks during the RSA key generation process could
4965 recover the private key.
4967 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4968 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4973 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4974 parameter is no longer accepted, as it leads to a corrupt table. NULL
4975 pem_str is reserved for alias entries only.
4979 * Revert blinding in ECDSA sign and instead make problematic addition
4980 length-invariant. Switch even to fixed-length Montgomery multiplication.
4984 * Change generating and checking of primes so that the error rate of not
4985 being prime depends on the intended use based on the size of the input.
4986 For larger primes this will result in more rounds of Miller-Rabin.
4987 The maximal error rate for primes with more than 1080 bits is lowered
4990 *Kurt Roeckx, Annie Yousar*
4992 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4996 * Add blinding to ECDSA and DSA signatures to protect against side channel
4997 attacks discovered by Keegan Ryan (NCC Group).
5001 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5002 now allow empty (zero character) pass phrases.
5006 * Certificate time validation (X509_cmp_time) enforces stricter
5007 compliance with RFC 5280. Fractional seconds and timezone offsets
5008 are no longer allowed.
5012 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5014 * Constructed ASN.1 types with a recursive definition could exceed the stack
5016 Constructed ASN.1 types with a recursive definition (such as can be found
5017 in PKCS7) could eventually exceed the stack given malicious input with
5018 excessive recursion. This could result in a Denial Of Service attack. There
5019 are no such structures used within SSL/TLS that come from untrusted sources
5020 so this is considered safe.
5022 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5028 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5030 * Read/write after SSL object in error state
5032 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5033 mechanism. The intent was that if a fatal error occurred during a handshake
5034 then OpenSSL would move into the error state and would immediately fail if
5035 you attempted to continue the handshake. This works as designed for the
5036 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5037 SSL_connect()), however due to a bug it does not work correctly if
5038 SSL_read() or SSL_write() is called directly. In that scenario, if the
5039 handshake fails then a fatal error will be returned in the initial function
5040 call. If SSL_read()/SSL_write() is subsequently called by the application
5041 for the same SSL object then it will succeed and the data is passed without
5042 being decrypted/encrypted directly from the SSL/TLS record layer.
5044 In order to exploit this issue an application bug would have to be present
5045 that resulted in a call to SSL_read()/SSL_write() being issued after having
5046 already received a fatal error.
5048 This issue was reported to OpenSSL by David Benjamin (Google).
5053 * rsaz_1024_mul_avx2 overflow bug on x86_64
5055 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5056 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5057 Analysis suggests that attacks against RSA and DSA as a result of this
5058 defect would be very difficult to perform and are not believed likely.
5059 Attacks against DH1024 are considered just feasible, because most of the
5060 work necessary to deduce information about a private key may be performed
5061 offline. The amount of resources required for such an attack would be
5062 significant. However, for an attack on TLS to be meaningful, the server
5063 would have to share the DH1024 private key among multiple clients, which is
5064 no longer an option since CVE-2016-0701.
5066 This only affects processors that support the AVX2 but not ADX extensions
5067 like Intel Haswell (4th generation).
5069 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5070 was originally found via the OSS-Fuzz project.
5075 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5077 * bn_sqrx8x_internal carry bug on x86_64
5079 There is a carry propagating bug in the x86_64 Montgomery squaring
5080 procedure. No EC algorithms are affected. Analysis suggests that attacks
5081 against RSA and DSA as a result of this defect would be very difficult to
5082 perform and are not believed likely. Attacks against DH are considered just
5083 feasible (although very difficult) because most of the work necessary to
5084 deduce information about a private key may be performed offline. The amount
5085 of resources required for such an attack would be very significant and
5086 likely only accessible to a limited number of attackers. An attacker would
5087 additionally need online access to an unpatched system using the target
5088 private key in a scenario with persistent DH parameters and a private
5089 key that is shared between multiple clients.
5091 This only affects processors that support the BMI1, BMI2 and ADX extensions
5092 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5094 This issue was reported to OpenSSL by the OSS-Fuzz project.
5099 * Malformed X.509 IPAddressFamily could cause OOB read
5101 If an X.509 certificate has a malformed IPAddressFamily extension,
5102 OpenSSL could do a one-byte buffer overread. The most likely result
5103 would be an erroneous display of the certificate in text format.
5105 This issue was reported to OpenSSL by the OSS-Fuzz project.
5109 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5111 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5112 platform rather than 'mingw'.
5116 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5118 * Truncated packet could crash via OOB read
5120 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5121 cipher is being used, then a truncated packet can cause that host to
5122 perform an out-of-bounds read, usually resulting in a crash.
5124 This issue was reported to OpenSSL by Robert Święcki of Google.
5129 * BN_mod_exp may produce incorrect results on x86_64
5131 There is a carry propagating bug in the x86_64 Montgomery squaring
5132 procedure. No EC algorithms are affected. Analysis suggests that attacks
5133 against RSA and DSA as a result of this defect would be very difficult to
5134 perform and are not believed likely. Attacks against DH are considered just
5135 feasible (although very difficult) because most of the work necessary to
5136 deduce information about a private key may be performed offline. The amount
5137 of resources required for such an attack would be very significant and
5138 likely only accessible to a limited number of attackers. An attacker would
5139 additionally need online access to an unpatched system using the target
5140 private key in a scenario with persistent DH parameters and a private
5141 key that is shared between multiple clients. For example this can occur by
5142 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5143 similar to CVE-2015-3193 but must be treated as a separate problem.
5145 This issue was reported to OpenSSL by the OSS-Fuzz project.
5150 * Montgomery multiplication may produce incorrect results
5152 There is a carry propagating bug in the Broadwell-specific Montgomery
5153 multiplication procedure that handles input lengths divisible by, but
5154 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5155 and DH private keys are impossible. This is because the subroutine in
5156 question is not used in operations with the private key itself and an input
5157 of the attacker's direct choice. Otherwise the bug can manifest itself as
5158 transient authentication and key negotiation failures or reproducible
5159 erroneous outcome of public-key operations with specially crafted input.
5160 Among EC algorithms only Brainpool P-512 curves are affected and one
5161 presumably can attack ECDH key negotiation. Impact was not analyzed in
5162 detail, because pre-requisites for attack are considered unlikely. Namely
5163 multiple clients have to choose the curve in question and the server has to
5164 share the private key among them, neither of which is default behaviour.
5165 Even then only clients that chose the curve will be affected.
5167 This issue was publicly reported as transient failures and was not
5168 initially recognized as a security issue. Thanks to Richard Morgan for
5169 providing reproducible case.
5174 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5175 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5176 prevent issues where no progress is being made and the peer continually
5177 sends unrecognised record types, using up resources processing them.
5181 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5183 * Missing CRL sanity check
5185 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5186 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5187 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5189 This issue only affects the OpenSSL 1.0.2i
5194 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5196 * OCSP Status Request extension unbounded memory growth
5198 A malicious client can send an excessively large OCSP Status Request
5199 extension. If that client continually requests renegotiation, sending a
5200 large OCSP Status Request extension each time, then there will be unbounded
5201 memory growth on the server. This will eventually lead to a Denial Of
5202 Service attack through memory exhaustion. Servers with a default
5203 configuration are vulnerable even if they do not support OCSP. Builds using
5204 the "no-ocsp" build time option are not affected.
5206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5211 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5214 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5220 * OOB write in MDC2_Update()
5222 An overflow can occur in MDC2_Update() either if called directly or
5223 through the EVP_DigestUpdate() function using MDC2. If an attacker
5224 is able to supply very large amounts of input data after a previous
5225 call to EVP_EncryptUpdate() with a partial block then a length check
5226 can overflow resulting in a heap corruption.
5228 The amount of data needed is comparable to SIZE_MAX which is impractical
5231 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5236 * Malformed SHA512 ticket DoS
5238 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5239 DoS attack where a malformed ticket will result in an OOB read which will
5242 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5243 a custom server callback and ticket lookup mechanism.
5245 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5250 * OOB write in BN_bn2dec()
5252 The function BN_bn2dec() does not check the return value of BN_div_word().
5253 This can cause an OOB write if an application uses this function with an
5254 overly large BIGNUM. This could be a problem if an overly large certificate
5255 or CRL is printed out from an untrusted source. TLS is not affected because
5256 record limits will reject an oversized certificate before it is parsed.
5258 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5263 * OOB read in TS_OBJ_print_bio()
5265 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5266 the total length the OID text representation would use and not the amount
5267 of data written. This will result in OOB reads when large OIDs are
5270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5275 * Pointer arithmetic undefined behaviour
5277 Avoid some undefined pointer arithmetic
5279 A common idiom in the codebase is to check limits in the following manner:
5282 Where "p" points to some malloc'd data of SIZE bytes and
5285 "len" here could be from some externally supplied data (e.g. from a TLS
5288 The rules of C pointer arithmetic are such that "p + len" is only well
5289 defined where len <= SIZE. Therefore the above idiom is actually
5290 undefined behaviour.
5292 For example this could cause problems if some malloc implementation
5293 provides an address for "p" such that "p + len" actually overflows for
5294 values of len that are too big and therefore p + len < limit.
5296 This issue was reported to OpenSSL by Guido Vranken
5301 * Constant time flag not preserved in DSA signing
5303 Operations in the DSA signing algorithm should run in constant time in
5304 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5305 implementation means that a non-constant time codepath is followed for
5306 certain operations. This has been demonstrated through a cache-timing
5307 attack to be sufficient for an attacker to recover the private DSA key.
5309 This issue was reported by César Pereida (Aalto University), Billy Brumley
5310 (Tampere University of Technology), and Yuval Yarom (The University of
5311 Adelaide and NICTA).
5316 * DTLS buffered message DoS
5318 In a DTLS connection where handshake messages are delivered out-of-order
5319 those messages that OpenSSL is not yet ready to process will be buffered
5320 for later use. Under certain circumstances, a flaw in the logic means that
5321 those messages do not get removed from the buffer even though the handshake
5322 has been completed. An attacker could force up to approx. 15 messages to
5323 remain in the buffer when they are no longer required. These messages will
5324 be cleared when the DTLS connection is closed. The default maximum size for
5325 a message is 100k. Therefore the attacker could force an additional 1500k
5326 to be consumed per connection. By opening many simulataneous connections an
5327 attacker could cause a DoS attack through memory exhaustion.
5329 This issue was reported to OpenSSL by Quan Luo.
5334 * DTLS replay protection DoS
5336 A flaw in the DTLS replay attack protection mechanism means that records
5337 that arrive for future epochs update the replay protection "window" before
5338 the MAC for the record has been validated. This could be exploited by an
5339 attacker by sending a record for the next epoch (which does not have to
5340 decrypt or have a valid MAC), with a very large sequence number. This means
5341 that all subsequent legitimate packets are dropped causing a denial of
5342 service for a specific DTLS connection.
5344 This issue was reported to OpenSSL by the OCAP audit team.
5349 * Certificate message OOB reads
5351 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5352 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5353 theoretical DoS risk but this has not been observed in practice on common
5356 The messages affected are client certificate, client certificate request
5357 and server certificate. As a result the attack can only be performed
5358 against a client or a server which enables client authentication.
5360 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5365 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5367 * Prevent padding oracle in AES-NI CBC MAC check
5369 A MITM attacker can use a padding oracle attack to decrypt traffic
5370 when the connection uses an AES CBC cipher and the server support
5373 This issue was introduced as part of the fix for Lucky 13 padding
5374 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5375 constant time by making sure that always the same bytes are read and
5376 compared against either the MAC or padding bytes. But it no longer
5377 checked that there was enough data to have both the MAC and padding
5380 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5384 * Fix EVP_EncodeUpdate overflow
5386 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5387 Base64 encoding of binary data. If an attacker is able to supply very large
5388 amounts of input data then a length check can overflow resulting in a heap
5391 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5392 the `PEM_write_bio*` family of functions. These are mainly used within the
5393 OpenSSL command line applications, so any application which processes data
5394 from an untrusted source and outputs it as a PEM file should be considered
5395 vulnerable to this issue. User applications that call these APIs directly
5396 with large amounts of untrusted data may also be vulnerable.
5398 This issue was reported by Guido Vranken.
5403 * Fix EVP_EncryptUpdate overflow
5405 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5406 is able to supply very large amounts of input data after a previous call to
5407 EVP_EncryptUpdate() with a partial block then a length check can overflow
5408 resulting in a heap corruption. Following an analysis of all OpenSSL
5409 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5410 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5411 the first called function after an EVP_EncryptInit(), and therefore that
5412 specific call must be safe. The second form is where the length passed to
5413 EVP_EncryptUpdate() can be seen from the code to be some small value and
5414 therefore there is no possibility of an overflow. Since all instances are
5415 one of these two forms, it is believed that there can be no overflows in
5416 internal code due to this problem. It should be noted that
5417 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5418 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5419 of these calls have also been analysed too and it is believed there are no
5420 instances in internal usage where an overflow could occur.
5422 This issue was reported by Guido Vranken.
5427 * Prevent ASN.1 BIO excessive memory allocation
5429 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5430 a short invalid encoding can cause allocation of large amounts of memory
5431 potentially consuming excessive resources or exhausting memory.
5433 Any application parsing untrusted data through d2i BIO functions is
5434 affected. The memory based functions such as d2i_X509() are *not* affected.
5435 Since the memory based functions are used by the TLS library, TLS
5436 applications are not affected.
5438 This issue was reported by Brian Carpenter.
5445 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5446 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5447 in arbitrary stack data being returned in the buffer.
5449 This issue was reported by Guido Vranken.
5454 * Modify behavior of ALPN to invoke callback after SNI/servername
5455 callback, such that updates to the SSL_CTX affect ALPN.
5459 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5464 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5465 methods are enabled and ssl2 is disabled the methods return NULL.
5469 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5471 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5472 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5473 provide any "EXPORT" or "LOW" strength ciphers.
5477 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5478 is by default disabled at build-time. Builds that are not configured with
5479 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5480 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5481 will need to explicitly call either of:
5483 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5485 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5487 as appropriate. Even if either of those is used, or the application
5488 explicitly uses the version-specific SSLv2_method() or its client and
5489 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5490 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5491 ciphers, and SSLv2 56-bit DES are no longer available.
5496 * Fix a double-free in DSA code
5498 A double free bug was discovered when OpenSSL parses malformed DSA private
5499 keys and could lead to a DoS attack or memory corruption for applications
5500 that receive DSA private keys from untrusted sources. This scenario is
5503 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5509 * Disable SRP fake user seed to address a server memory leak.
5511 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5513 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5514 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5515 was changed to ignore the "fake user" SRP seed, even if the seed
5518 Users should use SRP_VBASE_get1_by_user instead. Note that in
5519 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5520 also that even though configuring the SRP seed attempts to hide
5521 invalid usernames by continuing the handshake with fake
5522 credentials, this behaviour is not constant time and no strong
5523 guarantees are made that the handshake is indistinguishable from
5524 that of a valid user.
5529 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5531 In the BN_hex2bn function the number of hex digits is calculated using an
5532 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5533 large values of `i` this can result in `bn_expand` not allocating any
5534 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5535 field as NULL leading to a subsequent NULL ptr deref. For very large values
5536 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5537 In this case memory is allocated to the internal BIGNUM data field, but it
5538 is insufficiently sized leading to heap corruption. A similar issue exists
5539 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5540 is ever called by user applications with very large untrusted hex/dec data.
5541 This is anticipated to be a rare occurrence.
5543 All OpenSSL internal usage of these functions use data that is not expected
5544 to be untrusted, e.g. config file data or application command line
5545 arguments. If user developed applications generate config file data based
5546 on untrusted data then it is possible that this could also lead to security
5547 consequences. This is also anticipated to be rare.
5549 This issue was reported to OpenSSL by Guido Vranken.
5554 * Fix memory issues in `BIO_*printf` functions
5556 The internal `fmtstr` function used in processing a "%s" format string in
5557 the `BIO_*printf` functions could overflow while calculating the length of a
5558 string and cause an OOB read when printing very long strings.
5560 Additionally the internal `doapr_outch` function can attempt to write to an
5561 OOB memory location (at an offset from the NULL pointer) in the event of a
5562 memory allocation failure. In 1.0.2 and below this could be caused where
5563 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5564 could be in processing a very long "%s" format string. Memory leaks can
5567 The first issue may mask the second issue dependent on compiler behaviour.
5568 These problems could enable attacks where large amounts of untrusted data
5569 is passed to the `BIO_*printf` functions. If applications use these functions
5570 in this way then they could be vulnerable. OpenSSL itself uses these
5571 functions when printing out human-readable dumps of ASN.1 data. Therefore
5572 applications that print this data could be vulnerable if the data is from
5573 untrusted sources. OpenSSL command line applications could also be
5574 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5575 as command line arguments.
5577 Libssl is not considered directly vulnerable. Additionally certificates etc
5578 received via remote connections via libssl are also unlikely to be able to
5579 trigger these issues because of message size limits enforced within libssl.
5581 This issue was reported to OpenSSL Guido Vranken.
5586 * Side channel attack on modular exponentiation
5588 A side-channel attack was found which makes use of cache-bank conflicts on
5589 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5590 of RSA keys. The ability to exploit this issue is limited as it relies on
5591 an attacker who has control of code in a thread running on the same
5592 hyper-threaded core as the victim thread which is performing decryptions.
5594 This issue was reported to OpenSSL by Yuval Yarom, The University of
5595 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5596 Nadia Heninger, University of Pennsylvania with more information at
5597 <http://cachebleed.info>.
5602 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5603 if no keysize is specified with default_bits. This fixes an
5604 omission in an earlier change that changed all RSA/DSA key generation
5605 commands to use 2048 bits by default.
5609 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5611 * DH small subgroups
5613 Historically OpenSSL only ever generated DH parameters based on "safe"
5614 primes. More recently (in version 1.0.2) support was provided for
5615 generating X9.42 style parameter files such as those required for RFC 5114
5616 support. The primes used in such files may not be "safe". Where an
5617 application is using DH configured with parameters based on primes that are
5618 not "safe" then an attacker could use this fact to find a peer's private
5619 DH exponent. This attack requires that the attacker complete multiple
5620 handshakes in which the peer uses the same private DH exponent. For example
5621 this could be used to discover a TLS server's private DH exponent if it's
5622 reusing the private DH exponent or it's using a static DH ciphersuite.
5624 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5625 TLS. It is not on by default. If the option is not set then the server
5626 reuses the same private DH exponent for the life of the server process and
5627 would be vulnerable to this attack. It is believed that many popular
5628 applications do set this option and would therefore not be at risk.
5630 The fix for this issue adds an additional check where a "q" parameter is
5631 available (as is the case in X9.42 based parameters). This detects the
5632 only known attack, and is the only possible defense for static DH
5633 ciphersuites. This could have some performance impact.
5635 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5636 default and cannot be disabled. This could have some performance impact.
5638 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5643 * SSLv2 doesn't block disabled ciphers
5645 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5646 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5647 been disabled, provided that the SSLv2 protocol was not also disabled via
5650 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5651 and Sebastian Schinzel.
5656 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5658 * BN_mod_exp may produce incorrect results on x86_64
5660 There is a carry propagating bug in the x86_64 Montgomery squaring
5661 procedure. No EC algorithms are affected. Analysis suggests that attacks
5662 against RSA and DSA as a result of this defect would be very difficult to
5663 perform and are not believed likely. Attacks against DH are considered just
5664 feasible (although very difficult) because most of the work necessary to
5665 deduce information about a private key may be performed offline. The amount
5666 of resources required for such an attack would be very significant and
5667 likely only accessible to a limited number of attackers. An attacker would
5668 additionally need online access to an unpatched system using the target
5669 private key in a scenario with persistent DH parameters and a private
5670 key that is shared between multiple clients. For example this can occur by
5671 default in OpenSSL DHE based SSL/TLS ciphersuites.
5673 This issue was reported to OpenSSL by Hanno Böck.
5678 * Certificate verify crash with missing PSS parameter
5680 The signature verification routines will crash with a NULL pointer
5681 dereference if presented with an ASN.1 signature using the RSA PSS
5682 algorithm and absent mask generation function parameter. Since these
5683 routines are used to verify certificate signature algorithms this can be
5684 used to crash any certificate verification operation and exploited in a
5685 DoS attack. Any application which performs certificate verification is
5686 vulnerable including OpenSSL clients and servers which enable client
5689 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5694 * X509_ATTRIBUTE memory leak
5696 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5697 memory. This structure is used by the PKCS#7 and CMS routines so any
5698 application which reads PKCS#7 or CMS data from untrusted sources is
5699 affected. SSL/TLS is not affected.
5701 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5707 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5708 This changes the decoding behaviour for some invalid messages,
5709 though the change is mostly in the more lenient direction, and
5710 legacy behaviour is preserved as much as possible.
5714 * In DSA_generate_parameters_ex, if the provided seed is too short,
5717 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5719 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5721 * Alternate chains certificate forgery
5723 During certificate verification, OpenSSL will attempt to find an
5724 alternative certificate chain if the first attempt to build such a chain
5725 fails. An error in the implementation of this logic can mean that an
5726 attacker could cause certain checks on untrusted certificates to be
5727 bypassed, such as the CA flag, enabling them to use a valid leaf
5728 certificate to act as a CA and "issue" an invalid certificate.
5730 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5735 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5737 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5738 incompatibility in the handling of HMAC. The previous ABI has now been
5743 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5745 * Malformed ECParameters causes infinite loop
5747 When processing an ECParameters structure OpenSSL enters an infinite loop
5748 if the curve specified is over a specially malformed binary polynomial
5751 This can be used to perform denial of service against any
5752 system which processes public keys, certificate requests or
5753 certificates. This includes TLS clients and TLS servers with
5754 client authentication enabled.
5756 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5761 * Exploitable out-of-bounds read in X509_cmp_time
5763 X509_cmp_time does not properly check the length of the ASN1_TIME
5764 string and can read a few bytes out of bounds. In addition,
5765 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5768 An attacker can use this to craft malformed certificates and CRLs of
5769 various sizes and potentially cause a segmentation fault, resulting in
5770 a DoS on applications that verify certificates or CRLs. TLS clients
5771 that verify CRLs are affected. TLS clients and servers with client
5772 authentication enabled may be affected if they use custom verification
5775 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5776 independently by Hanno Böck.
5781 * PKCS7 crash with missing EnvelopedContent
5783 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5784 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5785 with missing content and trigger a NULL pointer dereference on parsing.
5787 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5788 structures from untrusted sources are affected. OpenSSL clients and
5789 servers are not affected.
5791 This issue was reported to OpenSSL by Michal Zalewski (Google).
5796 * CMS verify infinite loop with unknown hash function
5798 When verifying a signedData message the CMS code can enter an infinite loop
5799 if presented with an unknown hash function OID. This can be used to perform
5800 denial of service against any system which verifies signedData messages using
5802 This issue was reported to OpenSSL by Johannes Bauer.
5807 * Race condition handling NewSessionTicket
5809 If a NewSessionTicket is received by a multi-threaded client when attempting to
5810 reuse a previous ticket then a race condition can occur potentially leading to
5811 a double free of the ticket data.
5816 * Only support 256-bit or stronger elliptic curves with the
5817 'ecdh_auto' setting (server) or by default (client). Of supported
5818 curves, prefer P-256 (both).
5822 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5824 * ClientHello sigalgs DoS fix
5826 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5827 invalid signature algorithms extension a NULL pointer dereference will
5828 occur. This can be exploited in a DoS attack against the server.
5830 This issue was was reported to OpenSSL by David Ramos of Stanford
5834 *Stephen Henson and Matt Caswell*
5836 * Multiblock corrupted pointer fix
5838 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5839 feature only applies on 64 bit x86 architecture platforms that support AES
5840 NI instructions. A defect in the implementation of "multiblock" can cause
5841 OpenSSL's internal write buffer to become incorrectly set to NULL when
5842 using non-blocking IO. Typically, when the user application is using a
5843 socket BIO for writing, this will only result in a failed connection.
5844 However if some other BIO is used then it is likely that a segmentation
5845 fault will be triggered, thus enabling a potential DoS attack.
5847 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5852 * Segmentation fault in DTLSv1_listen fix
5854 The DTLSv1_listen function is intended to be stateless and processes the
5855 initial ClientHello from many peers. It is common for user code to loop
5856 over the call to DTLSv1_listen until a valid ClientHello is received with
5857 an associated cookie. A defect in the implementation of DTLSv1_listen means
5858 that state is preserved in the SSL object from one invocation to the next
5859 that can lead to a segmentation fault. Errors processing the initial
5860 ClientHello can trigger this scenario. An example of such an error could be
5861 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5864 This issue was reported to OpenSSL by Per Allansson.
5869 * Segmentation fault in ASN1_TYPE_cmp fix
5871 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5872 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5873 certificate signature algorithm consistency this can be used to crash any
5874 certificate verification operation and exploited in a DoS attack. Any
5875 application which performs certificate verification is vulnerable including
5876 OpenSSL clients and servers which enable client authentication.
5881 * Segmentation fault for invalid PSS parameters fix
5883 The signature verification routines will crash with a NULL pointer
5884 dereference if presented with an ASN.1 signature using the RSA PSS
5885 algorithm and invalid parameters. Since these routines are used to verify
5886 certificate signature algorithms this can be used to crash any
5887 certificate verification operation and exploited in a DoS attack. Any
5888 application which performs certificate verification is vulnerable including
5889 OpenSSL clients and servers which enable client authentication.
5891 This issue was was reported to OpenSSL by Brian Carpenter.
5896 * ASN.1 structure reuse memory corruption fix
5898 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5899 memory corruption via an invalid write. Such reuse is and has been
5900 strongly discouraged and is believed to be rare.
5902 Applications that parse structures containing CHOICE or ANY DEFINED BY
5903 components may be affected. Certificate parsing (d2i_X509 and related
5904 functions) are however not affected. OpenSSL clients and servers are
5910 * PKCS7 NULL pointer dereferences fix
5912 The PKCS#7 parsing code does not handle missing outer ContentInfo
5913 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5914 missing content and trigger a NULL pointer dereference on parsing.
5916 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5917 otherwise parse PKCS#7 structures from untrusted sources are
5918 affected. OpenSSL clients and servers are not affected.
5920 This issue was reported to OpenSSL by Michal Zalewski (Google).
5925 * DoS via reachable assert in SSLv2 servers fix
5927 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5928 servers that both support SSLv2 and enable export cipher suites by sending
5929 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5931 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5932 (OpenSSL development team).
5937 * Empty CKE with client auth and DHE fix
5939 If client auth is used then a server can seg fault in the event of a DHE
5940 ciphersuite being selected and a zero length ClientKeyExchange message
5941 being sent by the client. This could be exploited in a DoS attack.
5946 * Handshake with unseeded PRNG fix
5948 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5949 with an unseeded PRNG. The conditions are:
5950 - The client is on a platform where the PRNG has not been seeded
5951 automatically, and the user has not seeded manually
5952 - A protocol specific client method version has been used (i.e. not
5953 SSL_client_methodv23)
5954 - A ciphersuite is used that does not require additional random data from
5955 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5957 If the handshake succeeds then the client random that has been used will
5958 have been generated from a PRNG with insufficient entropy and therefore the
5959 output may be predictable.
5961 For example using the following command with an unseeded openssl will
5962 succeed on an unpatched platform:
5964 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5969 * Use After Free following d2i_ECPrivatekey error fix
5971 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5972 could cause a use after free condition. This, in turn, could cause a double
5973 free in several private key parsing functions (such as d2i_PrivateKey
5974 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5975 for applications that receive EC private keys from untrusted
5976 sources. This scenario is considered rare.
5978 This issue was discovered by the BoringSSL project and fixed in their
5984 * X509_to_X509_REQ NULL pointer deref fix
5986 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5987 the certificate key is invalid. This function is rarely used in practice.
5989 This issue was discovered by Brian Carpenter.
5994 * Removed the export ciphers from the DEFAULT ciphers
5998 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6000 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6001 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6002 So far those who have to target multiple platforms would compromise
6003 and argue that binary targeting say ARMv5 would still execute on
6004 ARMv8. "Universal" build resolves this compromise by providing
6005 near-optimal performance even on newer platforms.
6009 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6010 (other platforms pending).
6012 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6014 * Add support for the SignedCertificateTimestampList certificate and
6015 OCSP response extensions from RFC6962.
6019 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6020 for corner cases. (Certain input points at infinity could lead to
6021 bogus results, with non-infinity inputs mapped to infinity too.)
6025 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6026 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6027 common cases are optimized and there still is room for further
6028 improvements. Vector Permutation AES for Altivec is also added.
6032 * Add support for little-endian ppc64 Linux target.
6034 *Marcelo Cerri (IBM)*
6036 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6037 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6038 are optimized and there still is room for further improvements.
6039 Both 32- and 64-bit modes are supported.
6041 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6043 * Improved ARMv7 NEON support.
6047 * Support for SPARC Architecture 2011 crypto extensions, first
6048 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6049 SHA256/512, MD5, GHASH and modular exponentiation.
6051 *Andy Polyakov, David Miller*
6053 * Accelerated modular exponentiation for Intel processors, a.k.a.
6056 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6058 * Support for new and upcoming Intel processors, including AVX2,
6059 BMI and SHA ISA extensions. This includes additional "stitched"
6060 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6063 This work was sponsored by Intel Corp.
6067 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6068 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6069 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6073 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6074 this fixes a limitation in previous versions of OpenSSL.
6078 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6079 MGF1 digest and OAEP label.
6083 * Add EVP support for key wrapping algorithms, to avoid problems with
6084 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6085 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6086 algorithms and include tests cases.
6090 * Add functions to allocate and set the fields of an ECDSA_METHOD
6093 *Douglas E. Engert, Steve Henson*
6095 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6096 difference in days and seconds between two tm or ASN1_TIME structures.
6100 * Add -rev test option to s_server to just reverse order of characters
6101 received by client and send back to server. Also prints an abbreviated
6102 summary of the connection parameters.
6106 * New option -brief for s_client and s_server to print out a brief summary
6107 of connection parameters.
6111 * Add callbacks for arbitrary TLS extensions.
6113 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6115 * New option -crl_download in several openssl utilities to download CRLs
6116 from CRLDP extension in certificates.
6120 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6124 * New function X509_CRL_diff to generate a delta CRL from the difference
6125 of two full CRLs. Add support to "crl" utility.
6129 * New functions to set lookup_crls function and to retrieve
6130 X509_STORE from X509_STORE_CTX.
6134 * Print out deprecated issuer and subject unique ID fields in
6139 * Extend OCSP I/O functions so they can be used for simple general purpose
6140 HTTP as well as OCSP. New wrapper function which can be used to download
6141 CRLs using the OCSP API.
6145 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6149 * `SSL_CONF*` functions. These provide a common framework for application
6150 configuration using configuration files or command lines.
6154 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6155 message callback and prints the results. Needs compile time option
6156 "enable-ssl-trace". New options to s_client and s_server to enable
6161 * New ctrl and macro to retrieve supported points extensions.
6162 Print out extension in s_server and s_client.
6166 * New functions to retrieve certificate signature and signature
6171 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6176 * New Suite B modes for TLS code. These use and enforce the requirements
6177 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6178 only use Suite B curves. The Suite B modes can be set by using the
6179 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6183 * New chain verification flags for Suite B levels of security. Check
6184 algorithms are acceptable when flags are set in X509_verify_cert.
6188 * Make tls1_check_chain return a set of flags indicating checks passed
6189 by a certificate chain. Add additional tests to handle client
6190 certificates: checks for matching certificate type and issuer name
6195 * If an attempt is made to use a signature algorithm not in the peer
6196 preference list abort the handshake. If client has no suitable
6197 signature algorithms in response to a certificate request do not
6198 use the certificate.
6202 * If server EC tmp key is not in client preference list abort handshake.
6206 * Add support for certificate stores in CERT structure. This makes it
6207 possible to have different stores per SSL structure or one store in
6208 the parent SSL_CTX. Include distinct stores for certificate chain
6209 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6210 to build and store a certificate chain in CERT structure: returning
6211 an error if the chain cannot be built: this will allow applications
6212 to test if a chain is correctly configured.
6214 Note: if the CERT based stores are not set then the parent SSL_CTX
6215 store is used to retain compatibility with existing behaviour.
6219 * New function ssl_set_client_disabled to set a ciphersuite disabled
6220 mask based on the current session, check mask when sending client
6221 hello and checking the requested ciphersuite.
6225 * New ctrls to retrieve and set certificate types in a certificate
6226 request message. Print out received values in s_client. If certificate
6227 types is not set with custom values set sensible values based on
6228 supported signature algorithms.
6232 * Support for distinct client and server supported signature algorithms.
6236 * Add certificate callback. If set this is called whenever a certificate
6237 is required by client or server. An application can decide which
6238 certificate chain to present based on arbitrary criteria: for example
6239 supported signature algorithms. Add very simple example to s_server.
6240 This fixes many of the problems and restrictions of the existing client
6241 certificate callback: for example you can now clear an existing
6242 certificate and specify the whole chain.
6246 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6247 the certificate can be used for (if anything). Set valid_flags field
6248 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6249 to have similar checks in it.
6251 Add new "cert_flags" field to CERT structure and include a "strict mode".
6252 This enforces some TLS certificate requirements (such as only permitting
6253 certificate signature algorithms contained in the supported algorithms
6254 extension) which some implementations ignore: this option should be used
6255 with caution as it could cause interoperability issues.
6259 * Update and tidy signature algorithm extension processing. Work out
6260 shared signature algorithms based on preferences and peer algorithms
6261 and print them out in s_client and s_server. Abort handshake if no
6262 shared signature algorithms.
6266 * Add new functions to allow customised supported signature algorithms
6267 for SSL and SSL_CTX structures. Add options to s_client and s_server
6272 * New function SSL_certs_clear() to delete all references to certificates
6273 from an SSL structure. Before this once a certificate had been added
6274 it couldn't be removed.
6278 * Integrate hostname, email address and IP address checking with certificate
6279 verification. New verify options supporting checking in openssl utility.
6283 * Fixes and wildcard matching support to hostname and email checking
6284 functions. Add manual page.
6286 *Florian Weimer (Red Hat Product Security Team)*
6288 * New functions to check a hostname email or IP address against a
6289 certificate. Add options x509 utility to print results of checks against
6294 * Fix OCSP checking.
6296 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6298 * Initial experimental support for explicitly trusted non-root CAs.
6299 OpenSSL still tries to build a complete chain to a root but if an
6300 intermediate CA has a trust setting included that is used. The first
6301 setting is used: whether to trust (e.g., -addtrust option to the x509
6306 * Add -trusted_first option which attempts to find certificates in the
6307 trusted store even if an untrusted chain is also supplied.
6311 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6312 platform support for Linux and Android.
6316 * Support for linux-x32, ILP32 environment in x86_64 framework.
6320 * Experimental multi-implementation support for FIPS capable OpenSSL.
6321 When in FIPS mode the approved implementations are used as normal,
6322 when not in FIPS mode the internal unapproved versions are used instead.
6323 This means that the FIPS capable OpenSSL isn't forced to use the
6324 (often lower performance) FIPS implementations outside FIPS mode.
6328 * Transparently support X9.42 DH parameters when calling
6329 PEM_read_bio_DHparameters. This means existing applications can handle
6330 the new parameter format automatically.
6334 * Initial experimental support for X9.42 DH parameter format: mainly
6335 to support use of 'q' parameter for RFC5114 parameters.
6339 * Add DH parameters from RFC5114 including test data to dhtest.
6343 * Support for automatic EC temporary key parameter selection. If enabled
6344 the most preferred EC parameters are automatically used instead of
6345 hardcoded fixed parameters. Now a server just has to call:
6346 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6347 support ECDH and use the most appropriate parameters.
6351 * Enhance and tidy EC curve and point format TLS extension code. Use
6352 static structures instead of allocation if default values are used.
6353 New ctrls to set curves we wish to support and to retrieve shared curves.
6354 Print out shared curves in s_server. New options to s_server and s_client
6355 to set list of supported curves.
6359 * New ctrls to retrieve supported signature algorithms and
6360 supported curve values as an array of NIDs. Extend openssl utility
6361 to print out received values.
6365 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6366 between NIDs and the more common NIST names such as "P-256". Enhance
6367 ecparam utility and ECC method to recognise the NIST names for curves.
6371 * Enhance SSL/TLS certificate chain handling to support different
6372 chains for each certificate instead of one chain in the parent SSL_CTX.
6376 * Support for fixed DH ciphersuite client authentication: where both
6377 server and client use DH certificates with common parameters.
6381 * Support for fixed DH ciphersuites: those requiring DH server
6386 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6388 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6389 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6390 X509_CINF_get_signature were reverted post internal team review.
6395 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6397 * OCSP Status Request extension unbounded memory growth
6399 A malicious client can send an excessively large OCSP Status Request
6400 extension. If that client continually requests renegotiation, sending a
6401 large OCSP Status Request extension each time, then there will be unbounded
6402 memory growth on the server. This will eventually lead to a Denial Of
6403 Service attack through memory exhaustion. Servers with a default
6404 configuration are vulnerable even if they do not support OCSP. Builds using
6405 the "no-ocsp" build time option are not affected.
6407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6412 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6415 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6421 * OOB write in MDC2_Update()
6423 An overflow can occur in MDC2_Update() either if called directly or
6424 through the EVP_DigestUpdate() function using MDC2. If an attacker
6425 is able to supply very large amounts of input data after a previous
6426 call to EVP_EncryptUpdate() with a partial block then a length check
6427 can overflow resulting in a heap corruption.
6429 The amount of data needed is comparable to SIZE_MAX which is impractical
6432 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6437 * Malformed SHA512 ticket DoS
6439 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6440 DoS attack where a malformed ticket will result in an OOB read which will
6443 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6444 a custom server callback and ticket lookup mechanism.
6446 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6451 * OOB write in BN_bn2dec()
6453 The function BN_bn2dec() does not check the return value of BN_div_word().
6454 This can cause an OOB write if an application uses this function with an
6455 overly large BIGNUM. This could be a problem if an overly large certificate
6456 or CRL is printed out from an untrusted source. TLS is not affected because
6457 record limits will reject an oversized certificate before it is parsed.
6459 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6464 * OOB read in TS_OBJ_print_bio()
6466 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6467 the total length the OID text representation would use and not the amount
6468 of data written. This will result in OOB reads when large OIDs are
6471 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6476 * Pointer arithmetic undefined behaviour
6478 Avoid some undefined pointer arithmetic
6480 A common idiom in the codebase is to check limits in the following manner:
6483 Where "p" points to some malloc'd data of SIZE bytes and
6486 "len" here could be from some externally supplied data (e.g. from a TLS
6489 The rules of C pointer arithmetic are such that "p + len" is only well
6490 defined where len <= SIZE. Therefore the above idiom is actually
6491 undefined behaviour.
6493 For example this could cause problems if some malloc implementation
6494 provides an address for "p" such that "p + len" actually overflows for
6495 values of len that are too big and therefore p + len < limit.
6497 This issue was reported to OpenSSL by Guido Vranken
6502 * Constant time flag not preserved in DSA signing
6504 Operations in the DSA signing algorithm should run in constant time in
6505 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6506 implementation means that a non-constant time codepath is followed for
6507 certain operations. This has been demonstrated through a cache-timing
6508 attack to be sufficient for an attacker to recover the private DSA key.
6510 This issue was reported by César Pereida (Aalto University), Billy Brumley
6511 (Tampere University of Technology), and Yuval Yarom (The University of
6512 Adelaide and NICTA).
6517 * DTLS buffered message DoS
6519 In a DTLS connection where handshake messages are delivered out-of-order
6520 those messages that OpenSSL is not yet ready to process will be buffered
6521 for later use. Under certain circumstances, a flaw in the logic means that
6522 those messages do not get removed from the buffer even though the handshake
6523 has been completed. An attacker could force up to approx. 15 messages to
6524 remain in the buffer when they are no longer required. These messages will
6525 be cleared when the DTLS connection is closed. The default maximum size for
6526 a message is 100k. Therefore the attacker could force an additional 1500k
6527 to be consumed per connection. By opening many simulataneous connections an
6528 attacker could cause a DoS attack through memory exhaustion.
6530 This issue was reported to OpenSSL by Quan Luo.
6535 * DTLS replay protection DoS
6537 A flaw in the DTLS replay attack protection mechanism means that records
6538 that arrive for future epochs update the replay protection "window" before
6539 the MAC for the record has been validated. This could be exploited by an
6540 attacker by sending a record for the next epoch (which does not have to
6541 decrypt or have a valid MAC), with a very large sequence number. This means
6542 that all subsequent legitimate packets are dropped causing a denial of
6543 service for a specific DTLS connection.
6545 This issue was reported to OpenSSL by the OCAP audit team.
6550 * Certificate message OOB reads
6552 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6553 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6554 theoretical DoS risk but this has not been observed in practice on common
6557 The messages affected are client certificate, client certificate request
6558 and server certificate. As a result the attack can only be performed
6559 against a client or a server which enables client authentication.
6561 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6566 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6568 * Prevent padding oracle in AES-NI CBC MAC check
6570 A MITM attacker can use a padding oracle attack to decrypt traffic
6571 when the connection uses an AES CBC cipher and the server support
6574 This issue was introduced as part of the fix for Lucky 13 padding
6575 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6576 constant time by making sure that always the same bytes are read and
6577 compared against either the MAC or padding bytes. But it no longer
6578 checked that there was enough data to have both the MAC and padding
6581 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6586 * Fix EVP_EncodeUpdate overflow
6588 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6589 Base64 encoding of binary data. If an attacker is able to supply very large
6590 amounts of input data then a length check can overflow resulting in a heap
6593 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6594 the `PEM_write_bio*` family of functions. These are mainly used within the
6595 OpenSSL command line applications, so any application which processes data
6596 from an untrusted source and outputs it as a PEM file should be considered
6597 vulnerable to this issue. User applications that call these APIs directly
6598 with large amounts of untrusted data may also be vulnerable.
6600 This issue was reported by Guido Vranken.
6605 * Fix EVP_EncryptUpdate overflow
6607 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6608 is able to supply very large amounts of input data after a previous call to
6609 EVP_EncryptUpdate() with a partial block then a length check can overflow
6610 resulting in a heap corruption. Following an analysis of all OpenSSL
6611 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6612 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6613 the first called function after an EVP_EncryptInit(), and therefore that
6614 specific call must be safe. The second form is where the length passed to
6615 EVP_EncryptUpdate() can be seen from the code to be some small value and
6616 therefore there is no possibility of an overflow. Since all instances are
6617 one of these two forms, it is believed that there can be no overflows in
6618 internal code due to this problem. It should be noted that
6619 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6620 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6621 of these calls have also been analysed too and it is believed there are no
6622 instances in internal usage where an overflow could occur.
6624 This issue was reported by Guido Vranken.
6629 * Prevent ASN.1 BIO excessive memory allocation
6631 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6632 a short invalid encoding can casuse allocation of large amounts of memory
6633 potentially consuming excessive resources or exhausting memory.
6635 Any application parsing untrusted data through d2i BIO functions is
6636 affected. The memory based functions such as d2i_X509() are *not* affected.
6637 Since the memory based functions are used by the TLS library, TLS
6638 applications are not affected.
6640 This issue was reported by Brian Carpenter.
6647 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6648 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6649 in arbitrary stack data being returned in the buffer.
6651 This issue was reported by Guido Vranken.
6656 * Modify behavior of ALPN to invoke callback after SNI/servername
6657 callback, such that updates to the SSL_CTX affect ALPN.
6661 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6666 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6667 methods are enabled and ssl2 is disabled the methods return NULL.
6671 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6673 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6674 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6675 provide any "EXPORT" or "LOW" strength ciphers.
6679 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6680 is by default disabled at build-time. Builds that are not configured with
6681 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6682 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6683 will need to explicitly call either of:
6685 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6687 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6689 as appropriate. Even if either of those is used, or the application
6690 explicitly uses the version-specific SSLv2_method() or its client and
6691 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6692 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6693 ciphers, and SSLv2 56-bit DES are no longer available.
6698 * Fix a double-free in DSA code
6700 A double free bug was discovered when OpenSSL parses malformed DSA private
6701 keys and could lead to a DoS attack or memory corruption for applications
6702 that receive DSA private keys from untrusted sources. This scenario is
6705 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6711 * Disable SRP fake user seed to address a server memory leak.
6713 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6715 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6716 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6717 was changed to ignore the "fake user" SRP seed, even if the seed
6720 Users should use SRP_VBASE_get1_by_user instead. Note that in
6721 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6722 also that even though configuring the SRP seed attempts to hide
6723 invalid usernames by continuing the handshake with fake
6724 credentials, this behaviour is not constant time and no strong
6725 guarantees are made that the handshake is indistinguishable from
6726 that of a valid user.
6731 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6733 In the BN_hex2bn function the number of hex digits is calculated using an
6734 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6735 large values of `i` this can result in `bn_expand` not allocating any
6736 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6737 field as NULL leading to a subsequent NULL ptr deref. For very large values
6738 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6739 In this case memory is allocated to the internal BIGNUM data field, but it
6740 is insufficiently sized leading to heap corruption. A similar issue exists
6741 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6742 is ever called by user applications with very large untrusted hex/dec data.
6743 This is anticipated to be a rare occurrence.
6745 All OpenSSL internal usage of these functions use data that is not expected
6746 to be untrusted, e.g. config file data or application command line
6747 arguments. If user developed applications generate config file data based
6748 on untrusted data then it is possible that this could also lead to security
6749 consequences. This is also anticipated to be rare.
6751 This issue was reported to OpenSSL by Guido Vranken.
6756 * Fix memory issues in `BIO_*printf` functions
6758 The internal `fmtstr` function used in processing a "%s" format string in
6759 the `BIO_*printf` functions could overflow while calculating the length of a
6760 string and cause an OOB read when printing very long strings.
6762 Additionally the internal `doapr_outch` function can attempt to write to an
6763 OOB memory location (at an offset from the NULL pointer) in the event of a
6764 memory allocation failure. In 1.0.2 and below this could be caused where
6765 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6766 could be in processing a very long "%s" format string. Memory leaks can
6769 The first issue may mask the second issue dependent on compiler behaviour.
6770 These problems could enable attacks where large amounts of untrusted data
6771 is passed to the `BIO_*printf` functions. If applications use these functions
6772 in this way then they could be vulnerable. OpenSSL itself uses these
6773 functions when printing out human-readable dumps of ASN.1 data. Therefore
6774 applications that print this data could be vulnerable if the data is from
6775 untrusted sources. OpenSSL command line applications could also be
6776 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6777 as command line arguments.
6779 Libssl is not considered directly vulnerable. Additionally certificates etc
6780 received via remote connections via libssl are also unlikely to be able to
6781 trigger these issues because of message size limits enforced within libssl.
6783 This issue was reported to OpenSSL Guido Vranken.
6788 * Side channel attack on modular exponentiation
6790 A side-channel attack was found which makes use of cache-bank conflicts on
6791 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6792 of RSA keys. The ability to exploit this issue is limited as it relies on
6793 an attacker who has control of code in a thread running on the same
6794 hyper-threaded core as the victim thread which is performing decryptions.
6796 This issue was reported to OpenSSL by Yuval Yarom, The University of
6797 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6798 Nadia Heninger, University of Pennsylvania with more information at
6799 <http://cachebleed.info>.
6804 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6805 if no keysize is specified with default_bits. This fixes an
6806 omission in an earlier change that changed all RSA/DSA key generation
6807 commands to use 2048 bits by default.
6811 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6813 * Protection for DH small subgroup attacks
6815 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6816 switched on by default and cannot be disabled. This could have some
6821 * SSLv2 doesn't block disabled ciphers
6823 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6824 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6825 been disabled, provided that the SSLv2 protocol was not also disabled via
6828 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6829 and Sebastian Schinzel.
6834 * Reject DH handshakes with parameters shorter than 1024 bits.
6838 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6840 * Certificate verify crash with missing PSS parameter
6842 The signature verification routines will crash with a NULL pointer
6843 dereference if presented with an ASN.1 signature using the RSA PSS
6844 algorithm and absent mask generation function parameter. Since these
6845 routines are used to verify certificate signature algorithms this can be
6846 used to crash any certificate verification operation and exploited in a
6847 DoS attack. Any application which performs certificate verification is
6848 vulnerable including OpenSSL clients and servers which enable client
6851 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6856 * X509_ATTRIBUTE memory leak
6858 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6859 memory. This structure is used by the PKCS#7 and CMS routines so any
6860 application which reads PKCS#7 or CMS data from untrusted sources is
6861 affected. SSL/TLS is not affected.
6863 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6869 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6870 This changes the decoding behaviour for some invalid messages,
6871 though the change is mostly in the more lenient direction, and
6872 legacy behaviour is preserved as much as possible.
6876 * In DSA_generate_parameters_ex, if the provided seed is too short,
6877 use a random seed, as already documented.
6879 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6881 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6883 * Alternate chains certificate forgery
6885 During certificate verfification, OpenSSL will attempt to find an
6886 alternative certificate chain if the first attempt to build such a chain
6887 fails. An error in the implementation of this logic can mean that an
6888 attacker could cause certain checks on untrusted certificates to be
6889 bypassed, such as the CA flag, enabling them to use a valid leaf
6890 certificate to act as a CA and "issue" an invalid certificate.
6892 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6898 * Race condition handling PSK identify hint
6900 If PSK identity hints are received by a multi-threaded client then
6901 the values are wrongly updated in the parent SSL_CTX structure. This can
6902 result in a race condition potentially leading to a double free of the
6908 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6910 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6911 incompatibility in the handling of HMAC. The previous ABI has now been
6914 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6916 * Malformed ECParameters causes infinite loop
6918 When processing an ECParameters structure OpenSSL enters an infinite loop
6919 if the curve specified is over a specially malformed binary polynomial
6922 This can be used to perform denial of service against any
6923 system which processes public keys, certificate requests or
6924 certificates. This includes TLS clients and TLS servers with
6925 client authentication enabled.
6927 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6932 * Exploitable out-of-bounds read in X509_cmp_time
6934 X509_cmp_time does not properly check the length of the ASN1_TIME
6935 string and can read a few bytes out of bounds. In addition,
6936 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6939 An attacker can use this to craft malformed certificates and CRLs of
6940 various sizes and potentially cause a segmentation fault, resulting in
6941 a DoS on applications that verify certificates or CRLs. TLS clients
6942 that verify CRLs are affected. TLS clients and servers with client
6943 authentication enabled may be affected if they use custom verification
6946 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6947 independently by Hanno Böck.
6952 * PKCS7 crash with missing EnvelopedContent
6954 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6955 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6956 with missing content and trigger a NULL pointer dereference on parsing.
6958 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6959 structures from untrusted sources are affected. OpenSSL clients and
6960 servers are not affected.
6962 This issue was reported to OpenSSL by Michal Zalewski (Google).
6967 * CMS verify infinite loop with unknown hash function
6969 When verifying a signedData message the CMS code can enter an infinite loop
6970 if presented with an unknown hash function OID. This can be used to perform
6971 denial of service against any system which verifies signedData messages using
6973 This issue was reported to OpenSSL by Johannes Bauer.
6978 * Race condition handling NewSessionTicket
6980 If a NewSessionTicket is received by a multi-threaded client when attempting to
6981 reuse a previous ticket then a race condition can occur potentially leading to
6982 a double free of the ticket data.
6987 * Reject DH handshakes with parameters shorter than 768 bits.
6989 *Kurt Roeckx and Emilia Kasper*
6991 * dhparam: generate 2048-bit parameters by default.
6993 *Kurt Roeckx and Emilia Kasper*
6995 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6997 * Segmentation fault in ASN1_TYPE_cmp fix
6999 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7000 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7001 certificate signature algorithm consistency this can be used to crash any
7002 certificate verification operation and exploited in a DoS attack. Any
7003 application which performs certificate verification is vulnerable including
7004 OpenSSL clients and servers which enable client authentication.
7009 * ASN.1 structure reuse memory corruption fix
7011 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7012 memory corruption via an invalid write. Such reuse is and has been
7013 strongly discouraged and is believed to be rare.
7015 Applications that parse structures containing CHOICE or ANY DEFINED BY
7016 components may be affected. Certificate parsing (d2i_X509 and related
7017 functions) are however not affected. OpenSSL clients and servers are
7023 * PKCS7 NULL pointer dereferences fix
7025 The PKCS#7 parsing code does not handle missing outer ContentInfo
7026 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7027 missing content and trigger a NULL pointer dereference on parsing.
7029 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7030 otherwise parse PKCS#7 structures from untrusted sources are
7031 affected. OpenSSL clients and servers are not affected.
7033 This issue was reported to OpenSSL by Michal Zalewski (Google).
7038 * DoS via reachable assert in SSLv2 servers fix
7040 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7041 servers that both support SSLv2 and enable export cipher suites by sending
7042 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7044 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7045 (OpenSSL development team).
7050 * Use After Free following d2i_ECPrivatekey error fix
7052 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7053 could cause a use after free condition. This, in turn, could cause a double
7054 free in several private key parsing functions (such as d2i_PrivateKey
7055 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7056 for applications that receive EC private keys from untrusted
7057 sources. This scenario is considered rare.
7059 This issue was discovered by the BoringSSL project and fixed in their
7065 * X509_to_X509_REQ NULL pointer deref fix
7067 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7068 the certificate key is invalid. This function is rarely used in practice.
7070 This issue was discovered by Brian Carpenter.
7075 * Removed the export ciphers from the DEFAULT ciphers
7079 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7081 * Build fixes for the Windows and OpenVMS platforms
7083 *Matt Caswell and Richard Levitte*
7085 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7087 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7088 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7089 dereference. This could lead to a Denial Of Service attack. Thanks to
7090 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7095 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7096 dtls1_buffer_record function under certain conditions. In particular this
7097 could occur if an attacker sent repeated DTLS records with the same
7098 sequence number but for the next epoch. The memory leak could be exploited
7099 by an attacker in a Denial of Service attack through memory exhaustion.
7100 Thanks to Chris Mueller for reporting this issue.
7105 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7106 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7107 method would be set to NULL which could later result in a NULL pointer
7108 dereference. Thanks to Frank Schmirler for reporting this issue.
7113 * Abort handshake if server key exchange message is omitted for ephemeral
7116 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7117 reporting this issue.
7122 * Remove non-export ephemeral RSA code on client and server. This code
7123 violated the TLS standard by allowing the use of temporary RSA keys in
7124 non-export ciphersuites and could be used by a server to effectively
7125 downgrade the RSA key length used to a value smaller than the server
7126 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7127 INRIA or reporting this issue.
7132 * Fixed issue where DH client certificates are accepted without verification.
7133 An OpenSSL server will accept a DH certificate for client authentication
7134 without the certificate verify message. This effectively allows a client to
7135 authenticate without the use of a private key. This only affects servers
7136 which trust a client certificate authority which issues certificates
7137 containing DH keys: these are extremely rare and hardly ever encountered.
7138 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7144 * Ensure that the session ID context of an SSL is updated when its
7145 SSL_CTX is updated via SSL_set_SSL_CTX.
7147 The session ID context is typically set from the parent SSL_CTX,
7148 and can vary with the CTX.
7152 * Fix various certificate fingerprint issues.
7154 By using non-DER or invalid encodings outside the signed portion of a
7155 certificate the fingerprint can be changed without breaking the signature.
7156 Although no details of the signed portion of the certificate can be changed
7157 this can cause problems with some applications: e.g. those using the
7158 certificate fingerprint for blacklists.
7160 1. Reject signatures with non zero unused bits.
7162 If the BIT STRING containing the signature has non zero unused bits reject
7163 the signature. All current signature algorithms require zero unused bits.
7165 2. Check certificate algorithm consistency.
7167 Check the AlgorithmIdentifier inside TBS matches the one in the
7168 certificate signature. NB: this will result in signature failure
7169 errors for some broken certificates.
7171 Thanks to Konrad Kraszewski from Google for reporting this issue.
7173 3. Check DSA/ECDSA signatures use DER.
7175 Re-encode DSA/ECDSA signatures and compare with the original received
7176 signature. Return an error if there is a mismatch.
7178 This will reject various cases including garbage after signature
7179 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7180 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7181 (negative or with leading zeroes).
7183 Further analysis was conducted and fixes were developed by Stephen Henson
7184 of the OpenSSL core team.
7190 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7191 results on some platforms, including x86_64. This bug occurs at random
7192 with a very low probability, and is not known to be exploitable in any
7193 way, though its exact impact is difficult to determine. Thanks to Pieter
7194 Wuille (Blockstream) who reported this issue and also suggested an initial
7195 fix. Further analysis was conducted by the OpenSSL development team and
7196 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7197 the OpenSSL core team.
7202 * Do not resume sessions on the server if the negotiated protocol
7203 version does not match the session's version. Resuming with a different
7204 version, while not strictly forbidden by the RFC, is of questionable
7205 sanity and breaks all known clients.
7207 *David Benjamin, Emilia Käsper*
7209 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7210 early CCS messages during renegotiation. (Note that because
7211 renegotiation is encrypted, this early CCS was not exploitable.)
7215 * Tighten client-side session ticket handling during renegotiation:
7216 ensure that the client only accepts a session ticket if the server sends
7217 the extension anew in the ServerHello. Previously, a TLS client would
7218 reuse the old extension state and thus accept a session ticket if one was
7219 announced in the initial ServerHello.
7221 Similarly, ensure that the client requires a session ticket if one
7222 was advertised in the ServerHello. Previously, a TLS client would
7223 ignore a missing NewSessionTicket message.
7227 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7231 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7232 sends a carefully crafted handshake message, to cause OpenSSL to fail
7233 to free up to 64k of memory causing a memory leak. This could be
7234 exploited in a Denial Of Service attack. This issue affects OpenSSL
7235 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7236 whether SRTP is used or configured. Implementations of OpenSSL that
7237 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7239 The fix was developed by the OpenSSL team.
7244 * Session Ticket Memory Leak.
7246 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7247 integrity of that ticket is first verified. In the event of a session
7248 ticket integrity check failing, OpenSSL will fail to free memory
7249 causing a memory leak. By sending a large number of invalid session
7250 tickets an attacker could exploit this issue in a Denial Of Service
7256 * Build option no-ssl3 is incomplete.
7258 When OpenSSL is configured with "no-ssl3" as a build option, servers
7259 could accept and complete a SSL 3.0 handshake, and clients could be
7260 configured to send them.
7263 *Akamai and the OpenSSL team*
7265 * Add support for TLS_FALLBACK_SCSV.
7266 Client applications doing fallback retries should call
7267 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7270 *Adam Langley, Bodo Moeller*
7272 * Add additional DigestInfo checks.
7274 Re-encode DigestInto in DER and check against the original when
7275 verifying RSA signature: this will reject any improperly encoded
7276 DigestInfo structures.
7278 Note: this is a precautionary measure and no attacks are currently known.
7282 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7284 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7285 SRP code can be overrun an internal buffer. Add sanity check that
7286 g, A, B < N to SRP code.
7288 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7289 Group for discovering this issue.
7294 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7295 TLS 1.0 instead of higher protocol versions when the ClientHello message
7296 is badly fragmented. This allows a man-in-the-middle attacker to force a
7297 downgrade to TLS 1.0 even if both the server and the client support a
7298 higher protocol version, by modifying the client's TLS records.
7300 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7301 researching this issue.
7306 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7307 to a denial of service attack. A malicious server can crash the client
7308 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7309 ciphersuite and sending carefully crafted handshake messages.
7311 Thanks to Felix Gröbert (Google) for discovering and researching this
7317 * By sending carefully crafted DTLS packets an attacker could cause openssl
7318 to leak memory. This can be exploited through a Denial of Service attack.
7319 Thanks to Adam Langley for discovering and researching this issue.
7324 * An attacker can force openssl to consume large amounts of memory whilst
7325 processing DTLS handshake messages. This can be exploited through a
7326 Denial of Service attack.
7327 Thanks to Adam Langley for discovering and researching this issue.
7332 * An attacker can force an error condition which causes openssl to crash
7333 whilst processing DTLS packets due to memory being freed twice. This
7334 can be exploited through a Denial of Service attack.
7335 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7341 * If a multithreaded client connects to a malicious server using a resumed
7342 session and the server sends an ec point format extension it could write
7343 up to 255 bytes to freed memory.
7345 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7351 * A malicious server can crash an OpenSSL client with a null pointer
7352 dereference (read) by specifying an SRP ciphersuite even though it was not
7353 properly negotiated with the client. This can be exploited through a
7354 Denial of Service attack.
7356 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7357 discovering and researching this issue.
7362 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7363 X509_name_oneline, X509_name_print_ex et al. to leak some information
7364 from the stack. Applications may be affected if they echo pretty printing
7365 output to the attacker.
7367 Thanks to Ivan Fratric (Google) for discovering this issue.
7370 *Emilia Käsper, and Steve Henson*
7372 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7373 for corner cases. (Certain input points at infinity could lead to
7374 bogus results, with non-infinity inputs mapped to infinity too.)
7378 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7380 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7381 handshake can force the use of weak keying material in OpenSSL
7382 SSL/TLS clients and servers.
7384 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7385 researching this issue. ([CVE-2014-0224])
7387 *KIKUCHI Masashi, Steve Henson*
7389 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7390 OpenSSL DTLS client the code can be made to recurse eventually crashing
7393 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7396 *Imre Rad, Steve Henson*
7398 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7399 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7400 client or server. This is potentially exploitable to run arbitrary
7401 code on a vulnerable client or server.
7403 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7405 *Jüri Aedla, Steve Henson*
7407 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7408 are subject to a denial of service attack.
7410 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7411 this issue. ([CVE-2014-3470])
7413 *Felix Gröbert, Ivan Fratric, Steve Henson*
7415 * Harmonize version and its documentation. -f flag is used to display
7418 *mancha <mancha1@zoho.com>*
7420 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7421 in i2d_ECPrivateKey.
7423 *mancha <mancha1@zoho.com>*
7425 * Fix some double frees. These are not thought to be exploitable.
7427 *mancha <mancha1@zoho.com>*
7429 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7431 * A missing bounds check in the handling of the TLS heartbeat extension
7432 can be used to reveal up to 64k of memory to a connected client or
7435 Thanks for Neel Mehta of Google Security for discovering this bug and to
7436 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7437 preparing the fix ([CVE-2014-0160])
7439 *Adam Langley, Bodo Moeller*
7441 * Fix for the attack described in the paper "Recovering OpenSSL
7442 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7443 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7444 <http://eprint.iacr.org/2014/140>
7446 Thanks to Yuval Yarom and Naomi Benger for discovering this
7447 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7449 *Yuval Yarom and Naomi Benger*
7451 * TLS pad extension: draft-agl-tls-padding-03
7453 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7454 TLS client Hello record length value would otherwise be > 255 and
7455 less that 512 pad with a dummy extension containing zeroes so it
7456 is at least 512 bytes long.
7458 *Adam Langley, Steve Henson*
7460 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7462 * Fix for TLS record tampering bug. A carefully crafted invalid
7463 handshake could crash OpenSSL with a NULL pointer exception.
7464 Thanks to Anton Johansson for reporting this issues.
7467 * Keep original DTLS digest and encryption contexts in retransmission
7468 structures so we can use the previous session parameters if they need
7469 to be resent. ([CVE-2013-6450])
7473 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7474 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7475 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7476 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7477 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7478 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7480 *Rob Stradling, Adam Langley*
7482 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7484 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7485 supporting platforms or when small records were transferred.
7487 *Andy Polyakov, Steve Henson*
7489 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7491 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7493 This addresses the flaw in CBC record processing discovered by
7494 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7495 at: <http://www.isg.rhul.ac.uk/tls/>
7497 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7498 Security Group at Royal Holloway, University of London
7499 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7500 Emilia Käsper for the initial patch.
7503 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7505 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7506 ciphersuites which can be exploited in a denial of service attack.
7507 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7508 and detecting this bug and to Wolfgang Ettlinger
7509 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7514 * Return an error when checking OCSP signatures when key is NULL.
7515 This fixes a DoS attack. ([CVE-2013-0166])
7519 * Make openssl verify return errors.
7521 *Chris Palmer <palmer@google.com> and Ben Laurie*
7523 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7524 the right response is stapled. Also change SSL_get_certificate()
7525 so it returns the certificate actually sent.
7526 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7528 *Rob Stradling <rob.stradling@comodo.com>*
7530 * Fix possible deadlock when decoding public keys.
7534 * Don't use TLS 1.0 record version number in initial client hello
7539 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7541 * Sanity check record length before skipping explicit IV in TLS
7542 1.2, 1.1 and DTLS to fix DoS attack.
7544 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7545 fuzzing as a service testing platform.
7550 * Initialise tkeylen properly when encrypting CMS messages.
7551 Thanks to Solar Designer of Openwall for reporting this issue.
7555 * In FIPS mode don't try to use composite ciphers as they are not
7560 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7562 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7563 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7564 mean any application compiled against OpenSSL 1.0.0 headers setting
7565 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7566 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7567 0x10000000L Any application which was previously compiled against
7568 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7569 will need to be recompiled as a result. Letting be results in
7570 inability to disable specifically TLS 1.1 and in client context,
7571 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7575 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7576 disable just protocol X, but all protocols above X *if* there are
7577 protocols *below* X still enabled. In more practical terms it means
7578 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7579 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7580 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7585 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7587 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7588 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7589 in CRYPTO_realloc_clean.
7591 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7592 issue and to Adam Langley <agl@chromium.org> for fixing it.
7595 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7597 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7601 * Workarounds for some broken servers that "hang" if a client hello
7602 record length exceeds 255 bytes.
7604 1. Do not use record version number > TLS 1.0 in initial client
7605 hello: some (but not all) hanging servers will now work.
7606 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7607 the number of ciphers sent in the client hello. This should be
7608 set to an even number, such as 50, for example by passing:
7609 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7610 Most broken servers should now work.
7611 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7612 TLS 1.2 client support entirely.
7616 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7620 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7622 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7623 STRING form instead of a DigestInfo.
7627 * The format used for MDC2 RSA signatures is inconsistent between EVP
7628 and the RSA_sign/RSA_verify functions. This was made more apparent when
7629 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7630 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7631 the correct format in RSA_verify so both forms transparently work.
7635 * Some servers which support TLS 1.0 can choke if we initially indicate
7636 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7637 encrypted premaster secret. As a workaround use the maximum permitted
7638 client version in client hello, this should keep such servers happy
7639 and still work with previous versions of OpenSSL.
7643 * Add support for TLS/DTLS heartbeats.
7645 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7647 * Add support for SCTP.
7649 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7651 * Improved PRNG seeding for VOS.
7653 *Paul Green <Paul.Green@stratus.com>*
7655 * Extensive assembler packs updates, most notably:
7657 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7658 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7659 - x86_64: bit-sliced AES implementation;
7660 - ARM: NEON support, contemporary platforms optimizations;
7661 - s390x: z196 support;
7662 - `*`: GHASH and GF(2^m) multiplication implementations;
7666 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7667 (removal of unnecessary code)
7669 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7671 * Add TLS key material exporter from RFC 5705.
7675 * Add DTLS-SRTP negotiation from RFC 5764.
7679 * Add Next Protocol Negotiation,
7680 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7681 disabled with a no-npn flag to config or Configure. Code donated
7684 *Adam Langley <agl@google.com> and Ben Laurie*
7686 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7687 NIST-P256, NIST-P521, with constant-time single point multiplication on
7688 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7689 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7690 Code made available under Apache License version 2.0.
7692 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7693 line to include this in your build of OpenSSL, and run "make depend" (or
7694 "make update"). This enables the following EC_METHODs:
7696 EC_GFp_nistp224_method()
7697 EC_GFp_nistp256_method()
7698 EC_GFp_nistp521_method()
7700 EC_GROUP_new_by_curve_name() will automatically use these (while
7701 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7704 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7706 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7707 all platforms. Move ssize_t definition from e_os.h to the public
7708 header file e_os2.h as it now appears in public header file cms.h
7712 * New -sigopt option to the ca, req and x509 utilities. Additional
7713 signature parameters can be passed using this option and in
7718 * Add RSA PSS signing function. This will generate and set the
7719 appropriate AlgorithmIdentifiers for PSS based on those in the
7720 corresponding EVP_MD_CTX structure. No application support yet.
7724 * Support for companion algorithm specific ASN1 signing routines.
7725 New function ASN1_item_sign_ctx() signs a pre-initialised
7726 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7727 the appropriate parameters.
7731 * Add new algorithm specific ASN1 verification initialisation function
7732 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7733 handling will be the same no matter what EVP_PKEY_METHOD is used.
7734 Add a PSS handler to support verification of PSS signatures: checked
7735 against a number of sample certificates.
7739 * Add signature printing for PSS. Add PSS OIDs.
7741 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7743 * Add algorithm specific signature printing. An individual ASN1 method
7744 can now print out signatures instead of the standard hex dump.
7746 More complex signatures (e.g. PSS) can print out more meaningful
7747 information. Include DSA version that prints out the signature
7752 * Password based recipient info support for CMS library: implementing
7757 * Split password based encryption into PBES2 and PBKDF2 functions. This
7758 neatly separates the code into cipher and PBE sections and is required
7759 for some algorithms that split PBES2 into separate pieces (such as
7760 password based CMS).
7764 * Session-handling fixes:
7765 - Fix handling of connections that are resuming with a session ID,
7766 but also support Session Tickets.
7767 - Fix a bug that suppressed issuing of a new ticket if the client
7768 presented a ticket with an expired session.
7769 - Try to set the ticket lifetime hint to something reasonable.
7770 - Make tickets shorter by excluding irrelevant information.
7771 - On the client side, don't ignore renewed tickets.
7773 *Adam Langley, Bodo Moeller (Google)*
7775 * Fix PSK session representation.
7779 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7781 This work was sponsored by Intel.
7785 * Add GCM support to TLS library. Some custom code is needed to split
7786 the IV between the fixed (from PRF) and explicit (from TLS record)
7787 portions. This adds all GCM ciphersuites supported by RFC5288 and
7788 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7789 add a special AESGCM string for GCM only.
7793 * Expand range of ctrls for AES GCM. Permit setting invocation
7794 field on decrypt and retrieval of invocation field only on encrypt.
7798 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7799 As required by RFC5289 these ciphersuites cannot be used if for
7800 versions of TLS earlier than 1.2.
7804 * For FIPS capable OpenSSL interpret a NULL default public key method
7805 as unset and return the appropriate default but do *not* set the default.
7806 This means we can return the appropriate method in applications that
7807 switch between FIPS and non-FIPS modes.
7811 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7812 ENGINE is used then we cannot handle that in the FIPS module so we
7813 keep original code iff non-FIPS operations are allowed.
7817 * Add -attime option to openssl utilities.
7819 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7821 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7825 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7826 FIPS EC methods unconditionally for now.
7830 * New build option no-ec2m to disable characteristic 2 code.
7834 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7835 all cases can be covered as some introduce binary incompatibilities.
7839 * Redirect RSA operations to FIPS module including keygen,
7840 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7844 * Add similar low-level API blocking to ciphers.
7848 * low-level digest APIs are not approved in FIPS mode: any attempt
7849 to use these will cause a fatal error. Applications that *really* want
7850 to use them can use the `private_*` version instead.
7854 * Redirect cipher operations to FIPS module for FIPS builds.
7858 * Redirect digest operations to FIPS module for FIPS builds.
7862 * Update build system to add "fips" flag which will link in fipscanister.o
7863 for static and shared library builds embedding a signature if needed.
7867 * Output TLS supported curves in preference order instead of numerical
7868 order. This is currently hardcoded for the highest order curves first.
7869 This should be configurable so applications can judge speed vs strength.
7873 * Add TLS v1.2 server support for client authentication.
7877 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7882 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7883 FIPS modules versions.
7887 * Add TLS v1.2 client side support for client authentication. Keep cache
7888 of handshake records longer as we don't know the hash algorithm to use
7889 until after the certificate request message is received.
7893 * Initial TLS v1.2 client support. Add a default signature algorithms
7894 extension including all the algorithms we support. Parse new signature
7895 format in client key exchange. Relax some ECC signing restrictions for
7896 TLS v1.2 as indicated in RFC5246.
7900 * Add server support for TLS v1.2 signature algorithms extension. Switch
7901 to new signature format when needed using client digest preference.
7902 All server ciphersuites should now work correctly in TLS v1.2. No client
7903 support yet and no support for client certificates.
7907 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7908 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7909 ciphersuites. At present only RSA key exchange ciphersuites work with
7910 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7911 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7912 and version checking.
7916 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7917 with this defined it will not be affected by any changes to ssl internal
7918 structures. Add several utility functions to allow openssl application
7919 to work with OPENSSL_NO_SSL_INTERN defined.
7923 * A long standing patch to add support for SRP from EdelWeb (Peter
7924 Sylvester and Christophe Renou) was integrated.
7925 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7926 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7929 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7933 * Permit abbreviated handshakes when renegotiating using the function
7934 SSL_renegotiate_abbreviated().
7936 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7938 * Add call to ENGINE_register_all_complete() to
7939 ENGINE_load_builtin_engines(), so some implementations get used
7940 automatically instead of needing explicit application support.
7944 * Add support for TLS key exporter as described in RFC5705.
7946 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7948 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7949 a few changes are required:
7951 Add SSL_OP_NO_TLSv1_1 flag.
7952 Add TLSv1_1 methods.
7953 Update version checking logic to handle version 1.1.
7954 Add explicit IV handling (ported from DTLS code).
7955 Add command line options to s_client/s_server.
7962 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7964 * X509_ATTRIBUTE memory leak
7966 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7967 memory. This structure is used by the PKCS#7 and CMS routines so any
7968 application which reads PKCS#7 or CMS data from untrusted sources is
7969 affected. SSL/TLS is not affected.
7971 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7977 * Race condition handling PSK identify hint
7979 If PSK identity hints are received by a multi-threaded client then
7980 the values are wrongly updated in the parent SSL_CTX structure. This can
7981 result in a race condition potentially leading to a double free of the
7987 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7989 * Malformed ECParameters causes infinite loop
7991 When processing an ECParameters structure OpenSSL enters an infinite loop
7992 if the curve specified is over a specially malformed binary polynomial
7995 This can be used to perform denial of service against any
7996 system which processes public keys, certificate requests or
7997 certificates. This includes TLS clients and TLS servers with
7998 client authentication enabled.
8000 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8005 * Exploitable out-of-bounds read in X509_cmp_time
8007 X509_cmp_time does not properly check the length of the ASN1_TIME
8008 string and can read a few bytes out of bounds. In addition,
8009 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8012 An attacker can use this to craft malformed certificates and CRLs of
8013 various sizes and potentially cause a segmentation fault, resulting in
8014 a DoS on applications that verify certificates or CRLs. TLS clients
8015 that verify CRLs are affected. TLS clients and servers with client
8016 authentication enabled may be affected if they use custom verification
8019 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8020 independently by Hanno Böck.
8025 * PKCS7 crash with missing EnvelopedContent
8027 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8028 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8029 with missing content and trigger a NULL pointer dereference on parsing.
8031 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8032 structures from untrusted sources are affected. OpenSSL clients and
8033 servers are not affected.
8035 This issue was reported to OpenSSL by Michal Zalewski (Google).
8040 * CMS verify infinite loop with unknown hash function
8042 When verifying a signedData message the CMS code can enter an infinite loop
8043 if presented with an unknown hash function OID. This can be used to perform
8044 denial of service against any system which verifies signedData messages using
8046 This issue was reported to OpenSSL by Johannes Bauer.
8051 * Race condition handling NewSessionTicket
8053 If a NewSessionTicket is received by a multi-threaded client when attempting to
8054 reuse a previous ticket then a race condition can occur potentially leading to
8055 a double free of the ticket data.
8060 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8062 * Segmentation fault in ASN1_TYPE_cmp fix
8064 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8065 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8066 certificate signature algorithm consistency this can be used to crash any
8067 certificate verification operation and exploited in a DoS attack. Any
8068 application which performs certificate verification is vulnerable including
8069 OpenSSL clients and servers which enable client authentication.
8074 * ASN.1 structure reuse memory corruption fix
8076 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8077 memory corruption via an invalid write. Such reuse is and has been
8078 strongly discouraged and is believed to be rare.
8080 Applications that parse structures containing CHOICE or ANY DEFINED BY
8081 components may be affected. Certificate parsing (d2i_X509 and related
8082 functions) are however not affected. OpenSSL clients and servers are
8088 * PKCS7 NULL pointer dereferences fix
8090 The PKCS#7 parsing code does not handle missing outer ContentInfo
8091 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8092 missing content and trigger a NULL pointer dereference on parsing.
8094 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8095 otherwise parse PKCS#7 structures from untrusted sources are
8096 affected. OpenSSL clients and servers are not affected.
8098 This issue was reported to OpenSSL by Michal Zalewski (Google).
8103 * DoS via reachable assert in SSLv2 servers fix
8105 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8106 servers that both support SSLv2 and enable export cipher suites by sending
8107 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8109 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8110 (OpenSSL development team).
8115 * Use After Free following d2i_ECPrivatekey error fix
8117 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8118 could cause a use after free condition. This, in turn, could cause a double
8119 free in several private key parsing functions (such as d2i_PrivateKey
8120 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8121 for applications that receive EC private keys from untrusted
8122 sources. This scenario is considered rare.
8124 This issue was discovered by the BoringSSL project and fixed in their
8130 * X509_to_X509_REQ NULL pointer deref fix
8132 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8133 the certificate key is invalid. This function is rarely used in practice.
8135 This issue was discovered by Brian Carpenter.
8140 * Removed the export ciphers from the DEFAULT ciphers
8144 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8146 * Build fixes for the Windows and OpenVMS platforms
8148 *Matt Caswell and Richard Levitte*
8150 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8152 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8153 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8154 dereference. This could lead to a Denial Of Service attack. Thanks to
8155 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8160 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8161 dtls1_buffer_record function under certain conditions. In particular this
8162 could occur if an attacker sent repeated DTLS records with the same
8163 sequence number but for the next epoch. The memory leak could be exploited
8164 by an attacker in a Denial of Service attack through memory exhaustion.
8165 Thanks to Chris Mueller for reporting this issue.
8170 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8171 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8172 method would be set to NULL which could later result in a NULL pointer
8173 dereference. Thanks to Frank Schmirler for reporting this issue.
8178 * Abort handshake if server key exchange message is omitted for ephemeral
8181 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8182 reporting this issue.
8187 * Remove non-export ephemeral RSA code on client and server. This code
8188 violated the TLS standard by allowing the use of temporary RSA keys in
8189 non-export ciphersuites and could be used by a server to effectively
8190 downgrade the RSA key length used to a value smaller than the server
8191 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8192 INRIA or reporting this issue.
8197 * Fixed issue where DH client certificates are accepted without verification.
8198 An OpenSSL server will accept a DH certificate for client authentication
8199 without the certificate verify message. This effectively allows a client to
8200 authenticate without the use of a private key. This only affects servers
8201 which trust a client certificate authority which issues certificates
8202 containing DH keys: these are extremely rare and hardly ever encountered.
8203 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8209 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8210 results on some platforms, including x86_64. This bug occurs at random
8211 with a very low probability, and is not known to be exploitable in any
8212 way, though its exact impact is difficult to determine. Thanks to Pieter
8213 Wuille (Blockstream) who reported this issue and also suggested an initial
8214 fix. Further analysis was conducted by the OpenSSL development team and
8215 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8216 the OpenSSL core team.
8221 * Fix various certificate fingerprint issues.
8223 By using non-DER or invalid encodings outside the signed portion of a
8224 certificate the fingerprint can be changed without breaking the signature.
8225 Although no details of the signed portion of the certificate can be changed
8226 this can cause problems with some applications: e.g. those using the
8227 certificate fingerprint for blacklists.
8229 1. Reject signatures with non zero unused bits.
8231 If the BIT STRING containing the signature has non zero unused bits reject
8232 the signature. All current signature algorithms require zero unused bits.
8234 2. Check certificate algorithm consistency.
8236 Check the AlgorithmIdentifier inside TBS matches the one in the
8237 certificate signature. NB: this will result in signature failure
8238 errors for some broken certificates.
8240 Thanks to Konrad Kraszewski from Google for reporting this issue.
8242 3. Check DSA/ECDSA signatures use DER.
8244 Reencode DSA/ECDSA signatures and compare with the original received
8245 signature. Return an error if there is a mismatch.
8247 This will reject various cases including garbage after signature
8248 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8249 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8250 (negative or with leading zeroes).
8252 Further analysis was conducted and fixes were developed by Stephen Henson
8253 of the OpenSSL core team.
8259 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8261 * Session Ticket Memory Leak.
8263 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8264 integrity of that ticket is first verified. In the event of a session
8265 ticket integrity check failing, OpenSSL will fail to free memory
8266 causing a memory leak. By sending a large number of invalid session
8267 tickets an attacker could exploit this issue in a Denial Of Service
8273 * Build option no-ssl3 is incomplete.
8275 When OpenSSL is configured with "no-ssl3" as a build option, servers
8276 could accept and complete a SSL 3.0 handshake, and clients could be
8277 configured to send them.
8280 *Akamai and the OpenSSL team*
8282 * Add support for TLS_FALLBACK_SCSV.
8283 Client applications doing fallback retries should call
8284 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8287 *Adam Langley, Bodo Moeller*
8289 * Add additional DigestInfo checks.
8291 Reencode DigestInto in DER and check against the original when
8292 verifying RSA signature: this will reject any improperly encoded
8293 DigestInfo structures.
8295 Note: this is a precautionary measure and no attacks are currently known.
8299 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8301 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8302 to a denial of service attack. A malicious server can crash the client
8303 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8304 ciphersuite and sending carefully crafted handshake messages.
8306 Thanks to Felix Gröbert (Google) for discovering and researching this
8312 * By sending carefully crafted DTLS packets an attacker could cause openssl
8313 to leak memory. This can be exploited through a Denial of Service attack.
8314 Thanks to Adam Langley for discovering and researching this issue.
8319 * An attacker can force openssl to consume large amounts of memory whilst
8320 processing DTLS handshake messages. This can be exploited through a
8321 Denial of Service attack.
8322 Thanks to Adam Langley for discovering and researching this issue.
8327 * An attacker can force an error condition which causes openssl to crash
8328 whilst processing DTLS packets due to memory being freed twice. This
8329 can be exploited through a Denial of Service attack.
8330 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8336 * If a multithreaded client connects to a malicious server using a resumed
8337 session and the server sends an ec point format extension it could write
8338 up to 255 bytes to freed memory.
8340 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8346 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8347 X509_name_oneline, X509_name_print_ex et al. to leak some information
8348 from the stack. Applications may be affected if they echo pretty printing
8349 output to the attacker.
8351 Thanks to Ivan Fratric (Google) for discovering this issue.
8354 *Emilia Käsper, and Steve Henson*
8356 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8357 for corner cases. (Certain input points at infinity could lead to
8358 bogus results, with non-infinity inputs mapped to infinity too.)
8362 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8364 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8365 handshake can force the use of weak keying material in OpenSSL
8366 SSL/TLS clients and servers.
8368 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8369 researching this issue. ([CVE-2014-0224])
8371 *KIKUCHI Masashi, Steve Henson*
8373 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8374 OpenSSL DTLS client the code can be made to recurse eventually crashing
8377 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8380 *Imre Rad, Steve Henson*
8382 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8383 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8384 client or server. This is potentially exploitable to run arbitrary
8385 code on a vulnerable client or server.
8387 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8389 *Jüri Aedla, Steve Henson*
8391 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8392 are subject to a denial of service attack.
8394 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8395 this issue. ([CVE-2014-3470])
8397 *Felix Gröbert, Ivan Fratric, Steve Henson*
8399 * Harmonize version and its documentation. -f flag is used to display
8402 *mancha <mancha1@zoho.com>*
8404 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8405 in i2d_ECPrivateKey.
8407 *mancha <mancha1@zoho.com>*
8409 * Fix some double frees. These are not thought to be exploitable.
8411 *mancha <mancha1@zoho.com>*
8413 * Fix for the attack described in the paper "Recovering OpenSSL
8414 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8415 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8416 <http://eprint.iacr.org/2014/140>
8418 Thanks to Yuval Yarom and Naomi Benger for discovering this
8419 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8421 *Yuval Yarom and Naomi Benger*
8423 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8425 * Keep original DTLS digest and encryption contexts in retransmission
8426 structures so we can use the previous session parameters if they need
8427 to be resent. ([CVE-2013-6450])
8431 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8432 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8433 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8434 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8435 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8436 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8438 *Rob Stradling, Adam Langley*
8440 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8442 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8444 This addresses the flaw in CBC record processing discovered by
8445 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8446 at: <http://www.isg.rhul.ac.uk/tls/>
8448 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8449 Security Group at Royal Holloway, University of London
8450 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8451 Emilia Käsper for the initial patch.
8454 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8456 * Return an error when checking OCSP signatures when key is NULL.
8457 This fixes a DoS attack. ([CVE-2013-0166])
8461 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8462 the right response is stapled. Also change SSL_get_certificate()
8463 so it returns the certificate actually sent.
8464 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8465 (This is a backport)
8467 *Rob Stradling <rob.stradling@comodo.com>*
8469 * Fix possible deadlock when decoding public keys.
8473 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8475 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8478 * Sanity check record length before skipping explicit IV in DTLS
8481 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8482 fuzzing as a service testing platform.
8487 * Initialise tkeylen properly when encrypting CMS messages.
8488 Thanks to Solar Designer of Openwall for reporting this issue.
8492 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8494 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8495 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8496 in CRYPTO_realloc_clean.
8498 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8499 issue and to Adam Langley <agl@chromium.org> for fixing it.
8502 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8504 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8506 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8507 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8508 content decryption and always return the same error. Note: this attack
8509 needs on average 2^20 messages so it only affects automated senders. The
8510 old behaviour can be re-enabled in the CMS code by setting the
8511 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8512 an MMA defence is not necessary.
8513 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8514 this issue. ([CVE-2012-0884])
8518 * Fix CVE-2011-4619: make sure we really are receiving a
8519 client hello before rejecting multiple SGC restarts. Thanks to
8520 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8524 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8526 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8527 Thanks to Antonio Martin, Enterprise Secure Access Research and
8528 Development, Cisco Systems, Inc. for discovering this bug and
8529 preparing a fix. ([CVE-2012-0050])
8533 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8535 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8536 of the Vaudenay padding oracle attack on CBC mode encryption
8537 which enables an efficient plaintext recovery attack against
8538 the OpenSSL implementation of DTLS. Their attack exploits timing
8539 differences arising during decryption processing. A research
8540 paper describing this attack can be found at:
8541 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8542 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8543 Security Group at Royal Holloway, University of London
8544 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8545 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8546 for preparing the fix. ([CVE-2011-4108])
8548 *Robin Seggelmann, Michael Tuexen*
8550 * Clear bytes used for block padding of SSL 3.0 records.
8553 *Adam Langley (Google)*
8555 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8556 Kadianakis <desnacked@gmail.com> for discovering this issue and
8557 Adam Langley for preparing the fix. ([CVE-2011-4619])
8559 *Adam Langley (Google)*
8561 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8563 *Andrey Kulikov <amdeich@gmail.com>*
8565 * Prevent malformed RFC3779 data triggering an assertion failure.
8566 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8567 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8569 *Rob Austein <sra@hactrn.net>*
8571 * Improved PRNG seeding for VOS.
8573 *Paul Green <Paul.Green@stratus.com>*
8575 * Fix ssl_ciph.c set-up race.
8577 *Adam Langley (Google)*
8579 * Fix spurious failures in ecdsatest.c.
8581 *Emilia Käsper (Google)*
8583 * Fix the BIO_f_buffer() implementation (which was mixing different
8584 interpretations of the `..._len` fields).
8586 *Adam Langley (Google)*
8588 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8589 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8590 threads won't reuse the same blinding coefficients.
8592 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8593 lock to call BN_BLINDING_invert_ex, and avoids one use of
8594 BN_BLINDING_update for each BN_BLINDING structure (previously,
8595 the last update always remained unused).
8597 *Emilia Käsper (Google)*
8599 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8601 *Bob Buckholz (Google)*
8603 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8605 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8606 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8608 *Kaspar Brand <ossl@velox.ch>*
8610 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8611 for multi-threaded use of ECDH. ([CVE-2011-3210])
8613 *Adam Langley (Google)*
8615 * Fix x509_name_ex_d2i memory leak on bad inputs.
8619 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8620 signature public key algorithm by using OID xref utilities instead.
8621 Before this you could only use some ECC ciphersuites with SHA1 only.
8625 * Add protection against ECDSA timing attacks as mentioned in the paper
8626 by Billy Bob Brumley and Nicola Tuveri, see:
8627 <http://eprint.iacr.org/2011/232.pdf>
8629 *Billy Bob Brumley and Nicola Tuveri*
8631 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8633 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8635 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8637 * Fix bug in string printing code: if *any* escaping is enabled we must
8638 escape the escape character (backslash) or the resulting string is
8643 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8645 * Disable code workaround for ancient and obsolete Netscape browsers
8646 and servers: an attacker can use it in a ciphersuite downgrade attack.
8647 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8651 * Fixed J-PAKE implementation error, originally discovered by
8652 Sebastien Martini, further info and confirmation from Stefan
8653 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8657 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8659 * Fix extension code to avoid race conditions which can result in a buffer
8660 overrun vulnerability: resumed sessions must not be modified as they can
8661 be shared by multiple threads. CVE-2010-3864
8665 * Fix WIN32 build system to correctly link an ENGINE directory into
8670 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8672 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8675 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8677 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8679 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8680 context. The operation can be customised via the ctrl mechanism in
8681 case ENGINEs want to include additional functionality.
8685 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8689 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8690 output hashes compatible with older versions of OpenSSL.
8692 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8694 * Fix compression algorithm handling: if resuming a session use the
8695 compression algorithm of the resumed session instead of determining
8696 it from client hello again. Don't allow server to change algorithm.
8700 * Add load_crls() function to commands tidying load_certs() too. Add option
8701 to verify utility to allow additional CRLs to be included.
8705 * Update OCSP request code to permit adding custom headers to the request:
8706 some responders need this.
8710 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8713 *Julia Lawall <julia@diku.dk>*
8715 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8716 needlessly dereferenced structures, used obsolete functions and
8717 didn't handle all updated verify codes correctly.
8721 * Disable MD2 in the default configuration.
8725 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8726 indicate the initial BIO being pushed or popped. This makes it possible
8727 to determine whether the BIO is the one explicitly called or as a result
8728 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8729 it handles reference counts correctly and doesn't zero out the I/O bio
8730 when it is not being explicitly popped. WARNING: applications which
8731 included workarounds for the old buggy behaviour will need to be modified
8732 or they could free up already freed BIOs.
8736 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8737 renaming to all platforms (within the 0.9.8 branch, this was
8738 done conditionally on Netware platforms to avoid a name clash).
8740 *Guenter <lists@gknw.net>*
8742 * Add ECDHE and PSK support to DTLS.
8744 *Michael Tuexen <tuexen@fh-muenster.de>*
8746 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8751 * Add "missing" function EVP_MD_flags() (without this the only way to
8752 retrieve a digest flags is by accessing the structure directly. Update
8753 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8754 or cipher is registered as in the "from" argument. Print out all
8755 registered digests in the dgst usage message instead of manually
8756 attempting to work them out.
8760 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8761 this allows the use of compression and extensions. Change default cipher
8762 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8763 by default unless an application cipher string requests it.
8767 * Alter match criteria in PKCS12_parse(). It used to try to use local
8768 key ids to find matching certificates and keys but some PKCS#12 files
8769 don't follow the (somewhat unwritten) rules and this strategy fails.
8770 Now just gather all certificates together and the first private key
8771 then look for the first certificate that matches the key.
8775 * Support use of registered digest and cipher names for dgst and cipher
8776 commands instead of having to add each one as a special case. So now
8783 openssl dgst -sha256 foo
8785 and this works for ENGINE based algorithms too.
8789 * Update Gost ENGINE to support parameter files.
8791 *Victor B. Wagner <vitus@cryptocom.ru>*
8793 * Support GeneralizedTime in ca utility.
8795 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8797 * Enhance the hash format used for certificate directory links. The new
8798 form uses the canonical encoding (meaning equivalent names will work
8799 even if they aren't identical) and uses SHA1 instead of MD5. This form
8800 is incompatible with the older format and as a result c_rehash should
8801 be used to rebuild symbolic links.
8805 * Make PKCS#8 the default write format for private keys, replacing the
8806 traditional format. This form is standardised, more secure and doesn't
8807 include an implicit MD5 dependency.
8811 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8812 committed to OpenSSL should pass this lot as a minimum.
8816 * Add session ticket override functionality for use by EAP-FAST.
8818 *Jouni Malinen <j@w1.fi>*
8820 * Modify HMAC functions to return a value. Since these can be implemented
8821 in an ENGINE errors can occur.
8825 * Type-checked OBJ_bsearch_ex.
8829 * Type-checked OBJ_bsearch. Also some constification necessitated
8830 by type-checking. Still to come: TXT_DB, bsearch(?),
8831 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8836 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8837 seconds to a tm structure directly, instead of going through OS
8838 specific date routines. This avoids any issues with OS routines such
8839 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8840 and X509_time_adj_ex() to cover the extended range. The existing
8841 X509_time_adj() is still usable and will no longer have any date issues.
8845 * Delta CRL support. New use deltas option which will attempt to locate
8846 and search any appropriate delta CRLs available.
8848 This work was sponsored by Google.
8852 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8853 code and add additional score elements. Validate alternate CRL paths
8854 as part of the CRL checking and indicate a new error "CRL path validation
8855 error" in this case. Applications wanting additional details can use
8856 the verify callback and check the new "parent" field. If this is not
8857 NULL CRL path validation is taking place. Existing applications won't
8858 see this because it requires extended CRL support which is off by
8861 This work was sponsored by Google.
8865 * Support for freshest CRL extension.
8867 This work was sponsored by Google.
8871 * Initial indirect CRL support. Currently only supported in the CRLs
8872 passed directly and not via lookup. Process certificate issuer
8873 CRL entry extension and lookup CRL entries by bother issuer name
8874 and serial number. Check and process CRL issuer entry in IDP extension.
8876 This work was sponsored by Google.
8880 * Add support for distinct certificate and CRL paths. The CRL issuer
8881 certificate is validated separately in this case. Only enabled if
8882 an extended CRL support flag is set: this flag will enable additional
8883 CRL functionality in future.
8885 This work was sponsored by Google.
8889 * Add support for policy mappings extension.
8891 This work was sponsored by Google.
8895 * Fixes to pathlength constraint, self issued certificate handling,
8896 policy processing to align with RFC3280 and PKITS tests.
8898 This work was sponsored by Google.
8902 * Support for name constraints certificate extension. DN, email, DNS
8903 and URI types are currently supported.
8905 This work was sponsored by Google.
8909 * To cater for systems that provide a pointer-based thread ID rather
8910 than numeric, deprecate the current numeric thread ID mechanism and
8911 replace it with a structure and associated callback type. This
8912 mechanism allows a numeric "hash" to be extracted from a thread ID in
8913 either case, and on platforms where pointers are larger than 'long',
8914 mixing is done to help ensure the numeric 'hash' is usable even if it
8915 can't be guaranteed unique. The default mechanism is to use "&errno"
8916 as a pointer-based thread ID to distinguish between threads.
8918 Applications that want to provide their own thread IDs should now use
8919 CRYPTO_THREADID_set_callback() to register a callback that will call
8920 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8922 Note that ERR_remove_state() is now deprecated, because it is tied
8923 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8924 to free the current thread's error state should be replaced by
8925 ERR_remove_thread_state(NULL).
8927 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8928 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8929 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8930 application was previously providing a numeric thread callback that
8931 was inappropriate for distinguishing threads, then uniqueness might
8932 have been obtained with &errno that happened immediately in the
8933 intermediate development versions of OpenSSL; this is no longer the
8934 case, the numeric thread callback will now override the automatic use
8937 *Geoff Thorpe, with help from Bodo Moeller*
8939 * Initial support for different CRL issuing certificates. This covers a
8940 simple case where the self issued certificates in the chain exist and
8941 the real CRL issuer is higher in the existing chain.
8943 This work was sponsored by Google.
8947 * Removed effectively defunct crypto/store from the build.
8951 * Revamp of STACK to provide stronger type-checking. Still to come:
8952 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8953 ASN1_STRING, CONF_VALUE.
8957 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8958 RAM on SSL connections. This option can save about 34k per idle SSL.
8962 * Revamp of LHASH to provide stronger type-checking. Still to come:
8963 STACK, TXT_DB, bsearch, qsort.
8967 * Initial support for Cryptographic Message Syntax (aka CMS) based
8968 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8969 support for data, signedData, compressedData, digestedData and
8970 encryptedData, envelopedData types included. Scripts to check against
8971 RFC4134 examples draft and interop and consistency checks of many
8972 content types and variants.
8976 * Add options to enc utility to support use of zlib compression BIO.
8980 * Extend mk1mf to support importing of options and assembly language
8981 files from Configure script, currently only included in VC-WIN32.
8982 The assembly language rules can now optionally generate the source
8983 files from the associated perl scripts.
8987 * Implement remaining functionality needed to support GOST ciphersuites.
8988 Interop testing has been performed using CryptoPro implementations.
8990 *Victor B. Wagner <vitus@cryptocom.ru>*
8992 * s390x assembler pack.
8996 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9001 * Implement Opaque PRF Input TLS extension as specified in
9002 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9003 official specification yet and no extension type assignment by
9004 IANA exists, this extension (for now) will have to be explicitly
9005 enabled when building OpenSSL by providing the extension number
9006 to use. For example, specify an option
9008 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9010 to the "config" or "Configure" script to enable the extension,
9011 assuming extension number 0x9527 (which is a completely arbitrary
9012 and unofficial assignment based on the MD5 hash of the Internet
9013 Draft). Note that by doing so, you potentially lose
9014 interoperability with other TLS implementations since these might
9015 be using the same extension number for other purposes.
9017 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9018 opaque PRF input value to use in the handshake. This will create
9019 an internal copy of the length-'len' string at 'src', and will
9020 return non-zero for success.
9022 To get more control and flexibility, provide a callback function
9025 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9026 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9030 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9033 Callback function 'cb' will be called in handshakes, and is
9034 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9035 Argument 'arg' is for application purposes (the value as given to
9036 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9037 be provided to the callback function). The callback function
9038 has to return non-zero to report success: usually 1 to use opaque
9039 PRF input just if possible, or 2 to enforce use of the opaque PRF
9040 input. In the latter case, the library will abort the handshake
9041 if opaque PRF input is not successfully negotiated.
9043 Arguments 'peerinput' and 'len' given to the callback function
9044 will always be NULL and 0 in the case of a client. A server will
9045 see the client's opaque PRF input through these variables if
9046 available (NULL and 0 otherwise). Note that if the server
9047 provides an opaque PRF input, the length must be the same as the
9048 length of the client's opaque PRF input.
9050 Note that the callback function will only be called when creating
9051 a new session (session resumption can resume whatever was
9052 previously negotiated), and will not be called in SSL 2.0
9053 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9054 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9055 for applications that need to enforce opaque PRF input.
9059 * Update ssl code to support digests other than SHA1+MD5 for handshake
9062 *Victor B. Wagner <vitus@cryptocom.ru>*
9064 * Add RFC4507 support to OpenSSL. This includes the corrections in
9065 RFC4507bis. The encrypted ticket format is an encrypted encoded
9066 SSL_SESSION structure, that way new session features are automatically
9069 If a client application caches session in an SSL_SESSION structure
9070 support is transparent because tickets are now stored in the encoded
9073 The SSL_CTX structure automatically generates keys for ticket
9074 protection in servers so again support should be possible
9075 with no application modification.
9077 If a client or server wishes to disable RFC4507 support then the option
9078 SSL_OP_NO_TICKET can be set.
9080 Add a TLS extension debugging callback to allow the contents of any client
9081 or server extensions to be examined.
9083 This work was sponsored by Google.
9087 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9088 OpenSSL should now compile cleanly on gcc 4.2
9090 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9092 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9093 support including streaming MAC support: this is required for GOST
9094 ciphersuite support.
9096 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9098 * Add option -stream to use PKCS#7 streaming in smime utility. New
9099 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9100 to output in BER and PEM format.
9104 * Experimental support for use of HMAC via EVP_PKEY interface. This
9105 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9106 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9107 ENGINE support for HMAC keys which are unextractable. New -mac and
9108 -macopt options to dgst utility.
9112 * New option -sigopt to dgst utility. Update dgst to use
9113 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9114 alternative signing parameters such as X9.31 or PSS in the dgst
9119 * Change ssl_cipher_apply_rule(), the internal function that does
9120 the work each time a ciphersuite string requests enabling
9121 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9122 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9123 the order of disabled ciphersuites such that those ciphersuites
9124 that most recently went from enabled to disabled not only stay
9125 in order with respect to each other, but also have higher priority
9126 than other disabled ciphersuites the next time ciphersuites are
9129 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9130 the same ciphersuites as with "HIGH" alone, but in a specific
9131 order where the PSK ciphersuites come first (since they are the
9132 most recently disabled ciphersuites when "HIGH" is parsed).
9134 Also, change ssl_create_cipher_list() (using this new
9135 functionality) such that between otherwise identical
9136 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9141 * Change ssl_create_cipher_list() so that it automatically
9142 arranges the ciphersuites in reasonable order before starting
9143 to process the rule string. Thus, the definition for "DEFAULT"
9144 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9145 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9146 This makes it much easier to arrive at a reasonable default order
9147 in applications for which anonymous ciphers are OK (meaning
9148 that you can't actually use DEFAULT).
9150 *Bodo Moeller; suggested by Victor Duchovni*
9152 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9153 processing) into multiple integers instead of setting
9154 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9155 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9156 (These masks as well as the individual bit definitions are hidden
9157 away into the non-exported interface ssl/ssl_locl.h, so this
9158 change to the definition of the SSL_CIPHER structure shouldn't
9159 affect applications.) This give us more bits for each of these
9160 categories, so there is no longer a need to coagulate AES128 and
9161 AES256 into a single algorithm bit, and to coagulate Camellia128
9162 and Camellia256 into a single algorithm bit, which has led to all
9165 Thus, among other things, the kludge introduced in 0.9.7m and
9166 0.9.8e for masking out AES256 independently of AES128 or masking
9167 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9169 With the change, we also introduce new ciphersuite aliases that
9170 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9175 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9176 Use the leftmost N bytes of the signature input if the input is
9177 larger than the prime q (with N being the size in bytes of q).
9181 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9182 it yet and it is largely untested.
9186 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9190 * Initial incomplete changes to avoid need for function casts in OpenSSL
9191 some compilers (gcc 4.2 and later) reject their use. Safestack is
9192 reimplemented. Update ASN1 to avoid use of legacy functions.
9196 * Win32/64 targets are linked with Winsock2.
9200 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9201 to external functions. This can be used to increase CRL handling
9202 efficiency especially when CRLs are very large by (for example) storing
9203 the CRL revoked certificates in a database.
9207 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9208 new CRLs added to a directory can be used. New command line option
9209 -verify_return_error to s_client and s_server. This causes real errors
9210 to be returned by the verify callback instead of carrying on no matter
9211 what. This reflects the way a "real world" verify callback would behave.
9215 * GOST engine, supporting several GOST algorithms and public key formats.
9216 Kindly donated by Cryptocom.
9220 * Partial support for Issuing Distribution Point CRL extension. CRLs
9221 partitioned by DP are handled but no indirect CRL or reason partitioning
9222 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9223 selected via a scoring technique which handles IDP and AKID in CRLs.
9227 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9228 will ultimately be used for all verify operations: this will remove the
9229 X509_STORE dependency on certificate verification and allow alternative
9230 lookup methods. X509_STORE based implementations of these two callbacks.
9234 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9235 Modify get_crl() to find a valid (unexpired) CRL if possible.
9239 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9240 this would be called X509_CRL_cmp() but that name is already used by
9241 a function that just compares CRL issuer names. Cache several CRL
9242 extensions in X509_CRL structure and cache CRLDP in X509.
9246 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9247 this maps equivalent X509_NAME structures into a consistent structure.
9248 Name comparison can then be performed rapidly using memcmp().
9252 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9257 * Allow digests to supply their own micalg string for S/MIME type using
9258 the ctrl EVP_MD_CTRL_MICALG.
9262 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9263 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9264 ctrl. It can then customise the structure before and/or after signing
9269 * New function OBJ_add_sigid() to allow application defined signature OIDs
9270 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9271 to free up any added signature OIDs.
9275 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9276 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9277 digest and cipher tables. New options added to openssl utility:
9278 list-message-digest-algorithms and list-cipher-algorithms.
9282 * Change the array representation of binary polynomials: the list
9283 of degrees of non-zero coefficients is now terminated with -1.
9284 Previously it was terminated with 0, which was also part of the
9285 value; thus, the array representation was not applicable to
9286 polynomials where t^0 has coefficient zero. This change makes
9287 the array representation useful in a more general context.
9291 * Various modifications and fixes to SSL/TLS cipher string
9292 handling. For ECC, the code now distinguishes between fixed ECDH
9293 with RSA certificates on the one hand and with ECDSA certificates
9294 on the other hand, since these are separate ciphersuites. The
9295 unused code for Fortezza ciphersuites has been removed.
9297 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9298 (not "ECDHE"). For consistency with the code for DH
9299 certificates, use of ECDH certificates is now considered ECDH
9300 authentication, not RSA or ECDSA authentication (the latter is
9301 merely the CA's signing algorithm and not actively used in the
9304 The temporary ciphersuite alias "ECCdraft" is no longer
9305 available, and ECC ciphersuites are no longer excluded from "ALL"
9306 and "DEFAULT". The following aliases now exist for RFC 4492
9307 ciphersuites, most of these by analogy with the DH case:
9309 kECDHr - ECDH cert, signed with RSA
9310 kECDHe - ECDH cert, signed with ECDSA
9311 kECDH - ECDH cert (signed with either RSA or ECDSA)
9312 kEECDH - ephemeral ECDH
9313 ECDH - ECDH cert or ephemeral ECDH
9319 AECDH - anonymous ECDH
9320 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9324 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9325 Use correct micalg parameters depending on digest(s) in signed message.
9329 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9330 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9334 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9335 an engine to register a method. Add ENGINE lookups for methods and
9336 functional reference processing.
9340 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9341 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9346 * New -resign option to smime utility. This adds one or more signers
9347 to an existing PKCS#7 signedData structure. Also -md option to use an
9348 alternative message digest algorithm for signing.
9352 * Tidy up PKCS#7 routines and add new functions to make it easier to
9353 create PKCS7 structures containing multiple signers. Update smime
9354 application to support multiple signers.
9358 * New -macalg option to pkcs12 utility to allow setting of an alternative
9363 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9364 Reorganize PBE internals to lookup from a static table using NIDs,
9365 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9366 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9367 PRF which will be automatically used with PBES2.
9371 * Replace the algorithm specific calls to generate keys in "req" with the
9376 * Update PKCS#7 enveloped data routines to use new API. This is now
9377 supported by any public key method supporting the encrypt operation. A
9378 ctrl is added to allow the public key algorithm to examine or modify
9379 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9384 * Add a ctrl to asn1 method to allow a public key algorithm to express
9385 a default digest type to use. In most cases this will be SHA1 but some
9386 algorithms (such as GOST) need to specify an alternative digest. The
9387 return value indicates how strong the preference is 1 means optional and
9388 2 is mandatory (that is it is the only supported type). Modify
9389 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9390 use the default md. Update openssl utilities to use the default digest
9391 type for signing if it is not explicitly indicated.
9395 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9396 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9397 signing method from the key type. This effectively removes the link
9398 between digests and public key types.
9402 * Add an OID cross reference table and utility functions. Its purpose is to
9403 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9404 rsaEncryption. This will allow some of the algorithm specific hackery
9405 needed to use the correct OID to be removed.
9409 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9410 structures for PKCS7_sign(). They are now set up by the relevant public
9415 * Add provisional EC pkey method with support for ECDSA and ECDH.
9419 * Add support for key derivation (agreement) in the API, DH method and
9424 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9425 public and private key formats. As a side effect these add additional
9426 command line functionality not previously available: DSA signatures can be
9427 generated and verified using pkeyutl and DH key support and generation in
9434 *Oliver Tappe <zooey@hirschkaefer.de>*
9436 * New make target "install_html_docs" installs HTML renditions of the
9439 *Oliver Tappe <zooey@hirschkaefer.de>*
9441 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9442 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9443 support key and parameter generation and add initial key generation
9444 functionality for RSA.
9448 * Add functions for main EVP_PKEY_method operations. The undocumented
9449 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9450 `EVP_PKEY_{encrypt,decrypt}_old`.
9454 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9455 key API, doesn't do much yet.
9459 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9460 public key algorithms. New option to openssl utility:
9461 "list-public-key-algorithms" to print out info.
9465 * Implement the Supported Elliptic Curves Extension for
9466 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9470 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9471 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9475 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9476 utilities such as rsa, dsa, dsaparam etc except they process any key
9481 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9482 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9483 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9488 * Initial support for pluggable public key ASN1.
9489 De-spaghettify the public key ASN1 handling. Move public and private
9490 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9491 algorithm specific handling to a single module within the relevant
9492 algorithm directory. Add functions to allow (near) opaque processing
9493 of public and private key structures.
9497 * Implement the Supported Point Formats Extension for
9498 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9502 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9503 for the psk identity [hint] and the psk callback functions to the
9504 SSL_SESSION, SSL and SSL_CTX structure.
9507 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9511 SSL_CTX_use_psk_identity_hint
9512 SSL_get_psk_identity_hint
9513 SSL_get_psk_identity
9514 SSL_use_psk_identity_hint
9516 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9518 * Add RFC 3161 compliant time stamp request creation, response generation
9519 and response verification functionality.
9521 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9523 * Add initial support for TLS extensions, specifically for the server_name
9524 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9525 have new members for a host name. The SSL data structure has an
9526 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9527 stored in that context to allow for session resumption, even after the
9528 SSL has been switched to a new SSL_CTX in reaction to a client's
9529 server_name extension.
9531 New functions (subject to change):
9533 SSL_get_servername()
9534 SSL_get_servername_type()
9537 New CTRL codes and macros (subject to change):
9539 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9540 - SSL_CTX_set_tlsext_servername_callback()
9541 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9542 - SSL_CTX_set_tlsext_servername_arg()
9543 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9545 openssl s_client has a new '-servername ...' option.
9547 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9548 '-key2 ...', '-servername_fatal' (subject to change). This allows
9549 testing the HostName extension for a specific single host name ('-cert'
9550 and '-key' remain fallbacks for handshakes without HostName
9551 negotiation). If the unrecognized_name alert has to be sent, this by
9552 default is a warning; it becomes fatal with the '-servername_fatal'
9555 *Peter Sylvester, Remy Allais, Christophe Renou*
9557 * Whirlpool hash implementation is added.
9561 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9562 bn(64,32). Because of instruction set limitations it doesn't have
9563 any negative impact on performance. This was done mostly in order
9564 to make it possible to share assembler modules, such as bn_mul_mont
9565 implementations, between 32- and 64-bit builds without hassle.
9569 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9570 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9575 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9576 dedicated Montgomery multiplication procedure, is introduced.
9577 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9578 "64-bit" performance on certain 32-bit targets.
9582 * New option SSL_OP_NO_COMP to disable use of compression selectively
9583 in SSL structures. New SSL ctrl to set maximum send fragment size.
9584 Save memory by setting the I/O buffer sizes dynamically instead of
9585 using the maximum available value.
9589 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9590 in addition to the text details.
9594 * Very, very preliminary EXPERIMENTAL support for printing of general
9595 ASN1 structures. This currently produces rather ugly output and doesn't
9596 handle several customised structures at all.
9600 * Integrated support for PVK file format and some related formats such
9601 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9602 these in the 'rsa' and 'dsa' utilities.
9606 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9610 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9611 place for the (very old) "NETSCAPE" format certificates which are now
9612 handled using new ASN1 code equivalents.
9616 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9617 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9618 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9622 * Modify CRL distribution points extension code to print out previously
9623 unsupported fields. Enhance extension setting code to allow setting of
9628 * Add print and set support for Issuing Distribution Point CRL extension.
9632 * Change 'Configure' script to enable Camellia by default.
9639 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9641 * When rejecting SSL/TLS records due to an incorrect version number, never
9642 update s->server with a new major version number. As of
9643 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9644 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9645 the previous behavior could result in a read attempt at NULL when
9646 receiving specific incorrect SSL/TLS records once record payload
9647 protection is active. ([CVE-2010-0740])
9649 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9651 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9652 could be crashed if the relevant tables were not present (e.g. chrooted).
9654 *Tomas Hoger <thoger@redhat.com>*
9656 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9658 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9660 *Martin Olsson, Neel Mehta*
9662 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9663 accommodate for stack sorting, always a write lock!).
9667 * On some versions of WIN32 Heap32Next is very slow. This can cause
9668 excessive delays in the RAND_poll(): over a minute. As a workaround
9669 include a time check in the inner Heap32Next loop too.
9673 * The code that handled flushing of data in SSL/TLS originally used the
9674 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9675 the problem outlined in PR#1949. The fix suggested there however can
9676 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9677 of Apache). So instead simplify the code to flush unconditionally.
9678 This should be fine since flushing with no data to flush is a no op.
9682 * Handle TLS versions 2.0 and later properly and correctly use the
9683 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9684 off ancient servers have a habit of sticking around for a while...
9688 * Modify compression code so it frees up structures without using the
9689 ex_data callbacks. This works around a problem where some applications
9690 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9691 restarting) then use compression (e.g. SSL with compression) later.
9692 This results in significant per-connection memory leaks and
9693 has caused some security issues including CVE-2008-1678 and
9698 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9699 change when encrypting or decrypting.
9703 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9704 connect and renegotiate with servers which do not support RI.
9705 Until RI is more widely deployed this option is enabled by default.
9709 * Add "missing" ssl ctrls to clear options and mode.
9713 * If client attempts to renegotiate and doesn't support RI respond with
9714 a no_renegotiation alert as required by RFC5746. Some renegotiating
9715 TLS clients will continue a connection gracefully when they receive
9716 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9717 waiting for a server hello which it will never receive. Now we treat a
9718 received no_renegotiation alert as a fatal error. This is because
9719 applications requesting a renegotiation might well expect it to succeed
9720 and would have no code in place to handle the server denying it so the
9721 only safe thing to do is to terminate the connection.
9725 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9726 peer supports secure renegotiation and 0 otherwise. Print out peer
9727 renegotiation support in s_client/s_server.
9731 * Replace the highly broken and deprecated SPKAC certification method with
9732 the updated NID creation version. This should correctly handle UTF8.
9736 * Implement RFC5746. Re-enable renegotiation but require the extension
9737 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9738 turns out to be a bad idea. It has been replaced by
9739 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9740 SSL_CTX_set_options(). This is really not recommended unless you
9741 know what you are doing.
9743 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9745 * Fixes to stateless session resumption handling. Use initial_ctx when
9746 issuing and attempting to decrypt tickets in case it has changed during
9747 servername handling. Use a non-zero length session ID when attempting
9748 stateless session resumption: this makes it possible to determine if
9749 a resumption has occurred immediately after receiving server hello
9750 (several places in OpenSSL subtly assume this) instead of later in
9755 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9756 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9757 fixes for a few places where the return code is not checked
9760 *Julia Lawall <julia@diku.dk>*
9762 * Add --strict-warnings option to Configure script to include devteam
9763 warnings in other configurations.
9767 * Add support for --libdir option and LIBDIR variable in makefiles. This
9768 makes it possible to install openssl libraries in locations which
9769 have names other than "lib", for example "/usr/lib64" which some
9772 *Steve Henson, based on patch from Jeremy Utley*
9774 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9775 X690 8.9.12 and can produce some misleading textual output of OIDs.
9777 *Steve Henson, reported by Dan Kaminsky*
9779 * Delete MD2 from algorithm tables. This follows the recommendation in
9780 several standards that it is not used in new applications due to
9781 several cryptographic weaknesses. For binary compatibility reasons
9782 the MD2 API is still compiled in by default.
9786 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9791 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9792 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9795 *Guenter <lists@gknw.net>*
9797 * Fix the server certificate chain building code to use X509_verify_cert(),
9798 it used to have an ad-hoc builder which was unable to cope with anything
9799 other than a simple chain.
9801 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9803 * Don't check self signed certificate signatures in X509_verify_cert()
9804 by default (a flag can override this): it just wastes time without
9805 adding any security. As a useful side effect self signed root CAs
9806 with non-FIPS digests are now usable in FIPS mode.
9810 * In dtls1_process_out_of_seq_message() the check if the current message
9811 is already buffered was missing. For every new message was memory
9812 allocated, allowing an attacker to perform an denial of service attack
9813 with sending out of seq handshake messages until there is no memory
9814 left. Additionally every future message was buffered, even if the
9815 sequence number made no sense and would be part of another handshake.
9816 So only messages with sequence numbers less than 10 in advance will be
9817 buffered. ([CVE-2009-1378])
9819 *Robin Seggelmann, discovered by Daniel Mentz*
9821 * Records are buffered if they arrive with a future epoch to be
9822 processed after finishing the corresponding handshake. There is
9823 currently no limitation to this buffer allowing an attacker to perform
9824 a DOS attack with sending records with future epochs until there is no
9825 memory left. This patch adds the pqueue_size() function to determine
9826 the size of a buffer and limits the record buffer to 100 entries.
9829 *Robin Seggelmann, discovered by Daniel Mentz*
9831 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9832 parent structure is freed. ([CVE-2009-1379])
9836 * Handle non-blocking I/O properly in SSL_shutdown() call.
9838 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9840 * Add `2.5.4.*` OIDs
9842 *Ilya O. <vrghost@gmail.com>*
9844 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9846 * Disable renegotiation completely - this fixes a severe security
9847 problem ([CVE-2009-3555]) at the cost of breaking all
9848 renegotiation. Renegotiation can be re-enabled by setting
9849 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9850 run-time. This is really not recommended unless you know what
9855 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9857 * Don't set val to NULL when freeing up structures, it is freed up by
9858 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9859 zeroing past the valid field. ([CVE-2009-0789])
9861 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9863 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9864 checked correctly. This would allow some invalid signed attributes to
9865 appear to verify correctly. ([CVE-2009-0591])
9867 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9869 * Reject UniversalString and BMPString types with invalid lengths. This
9870 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9871 a legal length. ([CVE-2009-0590])
9875 * Set S/MIME signing as the default purpose rather than setting it
9876 unconditionally. This allows applications to override it at the store
9881 * Permit restricted recursion of ASN1 strings. This is needed in practice
9882 to handle some structures.
9886 * Improve efficiency of mem_gets: don't search whole buffer each time
9889 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9891 * New -hex option for openssl rand.
9895 * Print out UTF8String and NumericString when parsing ASN1.
9899 * Support NumericString type for name components.
9903 * Allow CC in the environment to override the automatically chosen
9904 compiler. Note that nothing is done to ensure flags work with the
9909 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9911 * Properly check EVP_VerifyFinal() and similar return values
9914 *Ben Laurie, Bodo Moeller, Google Security Team*
9916 * Enable TLS extensions by default.
9920 * Allow the CHIL engine to be loaded, whether the application is
9921 multithreaded or not. (This does not release the developer from the
9922 obligation to set up the dynamic locking callbacks.)
9924 *Sander Temme <sander@temme.net>*
9926 * Use correct exit code if there is an error in dgst command.
9928 *Steve Henson; problem pointed out by Roland Dirlewanger*
9930 * Tweak Configure so that you need to say "experimental-jpake" to enable
9931 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9935 * Add experimental JPAKE support, including demo authentication in
9936 s_client and s_server.
9940 * Set the comparison function in v3_addr_canonize().
9942 *Rob Austein <sra@hactrn.net>*
9944 * Add support for XMPP STARTTLS in s_client.
9946 *Philip Paeps <philip@freebsd.org>*
9948 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9949 to ensure that even with this option, only ciphersuites in the
9950 server's preference list will be accepted. (Note that the option
9951 applies only when resuming a session, so the earlier behavior was
9952 just about the algorithm choice for symmetric cryptography.)
9956 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9958 * Fix NULL pointer dereference if a DTLS server received
9959 ChangeCipherSpec as first record ([CVE-2009-1386]).
9963 * Fix a state transition in s3_srvr.c and d1_srvr.c
9964 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9968 * The fix in 0.9.8c that supposedly got rid of unsafe
9969 double-checked locking was incomplete for RSA blinding,
9970 addressing just one layer of what turns out to have been
9971 doubly unsafe triple-checked locking.
9973 So now fix this for real by retiring the MONT_HELPER macro
9974 in crypto/rsa/rsa_eay.c.
9976 *Bodo Moeller; problem pointed out by Marius Schilder*
9978 * Various precautionary measures:
9980 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9982 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9983 (NB: This would require knowledge of the secret session ticket key
9984 to exploit, in which case you'd be SOL either way.)
9986 - Change bn_nist.c so that it will properly handle input BIGNUMs
9987 outside the expected range.
9989 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9992 *Neel Mehta, Bodo Moeller*
9994 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9995 the load fails. Useful for distros.
9997 *Ben Laurie and the FreeBSD team*
9999 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10003 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10007 * Expand ENGINE to support engine supplied SSL client certificate functions.
10009 This work was sponsored by Logica.
10013 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10014 keystores. Support for SSL/TLS client authentication too.
10015 Not compiled unless enable-capieng specified to Configure.
10017 This work was sponsored by Logica.
10021 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10022 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10023 attribute creation routines such as certificate requests and PKCS#12
10028 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10030 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10031 handshake which could lead to a client crash as found using the
10032 Codenomicon TLS test suite ([CVE-2008-1672])
10034 *Steve Henson, Mark Cox*
10036 * Fix double free in TLS server name extensions which could lead to
10037 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10041 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10043 Clear the error queue to ensure that error entries left from
10044 older function calls do not interfere with the correct operation.
10046 *Lutz Jaenicke, Erik de Castro Lopo*
10048 * Remove root CA certificates of commercial CAs:
10050 The OpenSSL project does not recommend any specific CA and does not
10051 have any policy with respect to including or excluding any CA.
10052 Therefore it does not make any sense to ship an arbitrary selection
10053 of root CA certificates with the OpenSSL software.
10057 * RSA OAEP patches to fix two separate invalid memory reads.
10058 The first one involves inputs when 'lzero' is greater than
10059 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10060 before the beginning of from). The second one involves inputs where
10061 the 'db' section contains nothing but zeroes (there is a one-byte
10062 invalid read after the end of 'db').
10064 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10066 * Partial backport from 0.9.9-dev:
10068 Introduce bn_mul_mont (dedicated Montgomery multiplication
10069 procedure) as a candidate for BIGNUM assembler implementation.
10070 While 0.9.9-dev uses assembler for various architectures, only
10071 x86_64 is available by default here in the 0.9.8 branch, and
10072 32-bit x86 is available through a compile-time setting.
10074 To try the 32-bit x86 assembler implementation, use Configure
10075 option "enable-montasm" (which exists only for this backport).
10077 As "enable-montasm" for 32-bit x86 disclaims code stability
10078 anyway, in this constellation we activate additional code
10079 backported from 0.9.9-dev for further performance improvements,
10080 namely BN_from_montgomery_word. (To enable this otherwise,
10081 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10083 *Andy Polyakov (backport partially by Bodo Moeller)*
10085 * Add TLS session ticket callback. This allows an application to set
10086 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10087 values. This is useful for key rollover for example where several key
10088 sets may exist with different names.
10092 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10093 This was broken until now in 0.9.8 releases, such that the only way
10094 a registered ENGINE could be used (assuming it initialises
10095 successfully on the host) was to explicitly set it as the default
10096 for the relevant algorithms. This is in contradiction with 0.9.7
10097 behaviour and the documentation. With this fix, when an ENGINE is
10098 registered into a given algorithm's table of implementations, the
10099 'uptodate' flag is reset so that auto-discovery will be used next
10100 time a new context for that algorithm attempts to select an
10103 *Ian Lister (tweaked by Geoff Thorpe)*
10105 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10106 implementation in the following ways:
10108 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10111 Lack of BER streaming support means one pass streaming processing is
10112 only supported if data is detached: setting the streaming flag is
10113 ignored for embedded content.
10115 CMS support is disabled by default and must be explicitly enabled
10116 with the enable-cms configuration option.
10120 * Update the GMP engine glue to do direct copies between BIGNUM and
10121 mpz_t when openssl and GMP use the same limb size. Otherwise the
10122 existing "conversion via a text string export" trick is still used.
10124 *Paul Sheer <paulsheer@gmail.com>*
10126 * Zlib compression BIO. This is a filter BIO which compressed and
10127 uncompresses any data passed through it.
10131 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10132 RFC3394 compatible AES key wrapping.
10136 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10137 sets string data without copying. X509_ALGOR_set0() and
10138 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10139 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10140 from an X509_ATTRIBUTE structure optionally checking it occurs only
10141 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10146 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10147 to get the expected BN_FLG_CONSTTIME behavior.
10149 *Bodo Moeller (Google)*
10153 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10154 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10155 - added some more tests to do_tests.pl
10156 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10157 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10158 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10159 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10160 - various changes to netware.pl to enable gcc-cross builds on Win32
10162 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10163 - various changes to fix missing prototype warnings
10164 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10165 - added AES, WHIRLPOOL and CPUID assembler code to build files
10166 - added missing AES assembler make rules to mk1mf.pl
10167 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10169 *Guenter Knauf <eflash@gmx.net>*
10171 * Implement certificate status request TLS extension defined in RFC3546.
10172 A client can set the appropriate parameters and receive the encoded
10173 OCSP response via a callback. A server can query the supplied parameters
10174 and set the encoded OCSP response in the callback. Add simplified examples
10175 to s_client and s_server.
10179 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10181 * Fix various bugs:
10182 + Binary incompatibility of ssl_ctx_st structure
10183 + DTLS interoperation with non-compliant servers
10184 + Don't call get_session_cb() without proposed session
10185 + Fix ia64 assembler code
10187 *Andy Polyakov, Steve Henson*
10189 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10191 * DTLS Handshake overhaul. There were longstanding issues with
10192 OpenSSL DTLS implementation, which were making it impossible for
10193 RFC 4347 compliant client to communicate with OpenSSL server.
10194 Unfortunately just fixing these incompatibilities would "cut off"
10195 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10196 server keeps tolerating non RFC compliant syntax. The opposite is
10197 not true, 0.9.8f client can not communicate with earlier server.
10198 This update even addresses CVE-2007-4995.
10202 * Changes to avoid need for function casts in OpenSSL: some compilers
10203 (gcc 4.2 and later) reject their use.
10204 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10207 * Add RFC4507 support to OpenSSL. This includes the corrections in
10208 RFC4507bis. The encrypted ticket format is an encrypted encoded
10209 SSL_SESSION structure, that way new session features are automatically
10212 If a client application caches session in an SSL_SESSION structure
10213 support is transparent because tickets are now stored in the encoded
10216 The SSL_CTX structure automatically generates keys for ticket
10217 protection in servers so again support should be possible
10218 with no application modification.
10220 If a client or server wishes to disable RFC4507 support then the option
10221 SSL_OP_NO_TICKET can be set.
10223 Add a TLS extension debugging callback to allow the contents of any client
10224 or server extensions to be examined.
10226 This work was sponsored by Google.
10230 * Add initial support for TLS extensions, specifically for the server_name
10231 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10232 have new members for a host name. The SSL data structure has an
10233 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10234 stored in that context to allow for session resumption, even after the
10235 SSL has been switched to a new SSL_CTX in reaction to a client's
10236 server_name extension.
10238 New functions (subject to change):
10240 SSL_get_servername()
10241 SSL_get_servername_type()
10244 New CTRL codes and macros (subject to change):
10246 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10247 - SSL_CTX_set_tlsext_servername_callback()
10248 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10249 - SSL_CTX_set_tlsext_servername_arg()
10250 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10252 openssl s_client has a new '-servername ...' option.
10254 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10255 '-key2 ...', '-servername_fatal' (subject to change). This allows
10256 testing the HostName extension for a specific single host name ('-cert'
10257 and '-key' remain fallbacks for handshakes without HostName
10258 negotiation). If the unrecognized_name alert has to be sent, this by
10259 default is a warning; it becomes fatal with the '-servername_fatal'
10262 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10264 * Add AES and SSE2 assembly language support to VC++ build.
10268 * Mitigate attack on final subtraction in Montgomery reduction.
10272 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10273 (which previously caused an internal error).
10277 * Squeeze another 10% out of IGE mode when in != out.
10281 * AES IGE mode speedup.
10283 *Dean Gaudet (Google)*
10285 * Add the Korean symmetric 128-bit cipher SEED (see
10286 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10287 add SEED ciphersuites from RFC 4162:
10289 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10290 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10291 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10292 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10294 To minimize changes between patchlevels in the OpenSSL 0.9.8
10295 series, SEED remains excluded from compilation unless OpenSSL
10296 is configured with 'enable-seed'.
10298 *KISA, Bodo Moeller*
10300 * Mitigate branch prediction attacks, which can be practical if a
10301 single processor is shared, allowing a spy process to extract
10302 information. For detailed background information, see
10303 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10304 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10305 and Necessary Software Countermeasures"). The core of the change
10306 are new versions BN_div_no_branch() and
10307 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10308 respectively, which are slower, but avoid the security-relevant
10309 conditional branches. These are automatically called by BN_div()
10310 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10311 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10312 remove a conditional branch.
10314 BN_FLG_CONSTTIME is the new name for the previous
10315 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10316 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10317 in the exponent causes BN_mod_exp_mont() to use the alternative
10318 implementation in BN_mod_exp_mont_consttime().) The old name
10319 remains as a deprecated alias.
10321 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10322 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10323 constant-time implementations for more than just exponentiation.
10324 Here too the old name is kept as a deprecated alias.
10326 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10327 the BN_BLINDING structure gets an independent copy of the
10328 modulus. This means that the previous `BIGNUM *m` argument to
10329 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10330 essentially becomes `const BIGNUM *m`, although we can't actually
10331 change this in the header file before 0.9.9. It allows
10332 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10333 enable BN_FLG_CONSTTIME.
10335 *Matthew D Wood (Intel Corp)*
10337 * In the SSL/TLS server implementation, be strict about session ID
10338 context matching (which matters if an application uses a single
10339 external cache for different purposes). Previously,
10340 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10341 set. This did ensure strict client verification, but meant that,
10342 with applications using a single external cache for quite
10343 different requirements, clients could circumvent ciphersuite
10344 restrictions for a given session ID context by starting a session
10345 in a different context.
10349 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10350 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10351 authentication-only ciphersuites.
10355 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10356 not complete and could lead to a possible single byte overflow
10357 ([CVE-2007-5135]) [Ben Laurie]
10359 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10361 * Since AES128 and AES256 (and similarly Camellia128 and
10362 Camellia256) share a single mask bit in the logic of
10363 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10364 kludge to work properly if AES128 is available and AES256 isn't
10365 (or if Camellia128 is available and Camellia256 isn't).
10369 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10370 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10371 When a point or a seed is encoded in a BIT STRING, we need to
10372 prevent the removal of trailing zero bits to get the proper DER
10373 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10374 of a NamedBitList, for which trailing 0 bits need to be removed.)
10378 * Have SSL/TLS server implementation tolerate "mismatched" record
10379 protocol version while receiving ClientHello even if the
10380 ClientHello is fragmented. (The server can't insist on the
10381 particular protocol version it has chosen before the ServerHello
10382 message has informed the client about his choice.)
10386 * Add RFC 3779 support.
10388 *Rob Austein for ARIN, Ben Laurie*
10390 * Load error codes if they are not already present instead of using a
10391 static variable. This allows them to be cleanly unloaded and reloaded.
10392 Improve header file function name parsing.
10396 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10397 or CAPABILITY handshake as required by RFCs.
10399 *Goetz Babin-Ebell*
10401 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10403 * Introduce limits to prevent malicious keys being able to
10404 cause a denial of service. ([CVE-2006-2940])
10406 *Steve Henson, Bodo Moeller*
10408 * Fix ASN.1 parsing of certain invalid structures that can result
10409 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10411 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10412 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10414 * Fix SSL client code which could crash if connecting to a
10415 malicious SSLv2 server. ([CVE-2006-4343])
10417 *Tavis Ormandy and Will Drewry, Google Security Team*
10419 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10420 match only those. Before that, "AES256-SHA" would be interpreted
10421 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10422 the same strength classification in 0.9.7h) as we currently only
10423 have a single AES bit in the ciphersuite description bitmap.
10424 That change, however, also applied to ciphersuite strings such as
10425 "RC4-MD5" that intentionally matched multiple ciphersuites --
10426 namely, SSL 2.0 ciphersuites in addition to the more common ones
10427 from SSL 3.0/TLS 1.0.
10429 So we change the selection algorithm again: Naming an explicit
10430 ciphersuite selects this one ciphersuite, and any other similar
10431 ciphersuite (same bitmap) from *other* protocol versions.
10432 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10433 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10435 Since SSL 2.0 does not have any ciphersuites for which the
10436 128/256 bit distinction would be relevant, this works for now.
10437 The proper fix will be to use different bits for AES128 and
10438 AES256, which would have avoided the problems from the beginning;
10439 however, bits are scarce, so we can only do this in a new release
10440 (not just a patchlevel) when we can change the SSL_CIPHER
10441 definition to split the single 'unsigned long mask' bitmap into
10442 multiple values to extend the available space.
10446 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10448 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10449 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10451 * Add AES IGE and biIGE modes.
10455 * Change the Unix randomness entropy gathering to use poll() when
10456 possible instead of select(), since the latter has some
10457 undesirable limitations.
10459 *Darryl Miles via Richard Levitte and Bodo Moeller*
10461 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10462 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10463 cannot be implicitly activated as part of, e.g., the "AES" alias.
10464 However, please upgrade to OpenSSL 0.9.9[-dev] for
10465 non-experimental use of the ECC ciphersuites to get TLS extension
10466 support, which is required for curve and point format negotiation
10467 to avoid potential handshake problems.
10471 * Disable rogue ciphersuites:
10473 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10474 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10475 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10477 The latter two were purportedly from
10478 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10481 Also deactivate the remaining ciphersuites from
10482 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10483 unofficial, and the ID has long expired.
10487 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10488 dual-core machines) and other potential thread-safety issues.
10492 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10493 versions), which is now available for royalty-free use
10494 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10495 Also, add Camellia TLS ciphersuites from RFC 4132.
10497 To minimize changes between patchlevels in the OpenSSL 0.9.8
10498 series, Camellia remains excluded from compilation unless OpenSSL
10499 is configured with 'enable-camellia'.
10503 * Disable the padding bug check when compression is in use. The padding
10504 bug check assumes the first packet is of even length, this is not
10505 necessarily true if compression is enabled and can result in false
10506 positives causing handshake failure. The actual bug test is ancient
10507 code so it is hoped that implementations will either have fixed it by
10508 now or any which still have the bug do not support compression.
10512 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10514 * When applying a cipher rule check to see if string match is an explicit
10515 cipher suite and only match that one cipher suite if it is.
10519 * Link in manifests for VC++ if needed.
10521 *Austin Ziegler <halostatue@gmail.com>*
10523 * Update support for ECC-based TLS ciphersuites according to
10524 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10525 TLS extensions, which are supported starting with the 0.9.9
10526 branch, not in the OpenSSL 0.9.8 branch).
10530 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10531 opaque EVP_CIPHER_CTX handling.
10535 * Fixes and enhancements to zlib compression code. We now only use
10536 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10537 to conform with the standards mentioned here:
10538 <http://www.zlib.net/DLL_FAQ.txt>
10539 Static zlib linking now works on Windows and the new --with-zlib-include
10540 --with-zlib-lib options to Configure can be used to supply the location
10541 of the headers and library. Gracefully handle case where zlib library
10546 * Several fixes and enhancements to the OID generation code. The old code
10547 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10548 handle numbers larger than ULONG_MAX, truncated printing and had a
10549 non standard OBJ_obj2txt() behaviour.
10553 * Add support for building of engines under engine/ as shared libraries
10554 under VC++ build system.
10558 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10559 Hopefully, we will not see any false combination of paths any more.
10563 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10565 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10566 (part of SSL_OP_ALL). This option used to disable the
10567 countermeasure against man-in-the-middle protocol-version
10568 rollback in the SSL 2.0 server implementation, which is a bad
10569 idea. ([CVE-2005-2969])
10571 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10572 for Information Security, National Institute of Advanced Industrial
10573 Science and Technology [AIST], Japan)*
10575 * Add two function to clear and return the verify parameter flags.
10579 * Keep cipherlists sorted in the source instead of sorting them at
10580 runtime, thus removing the need for a lock.
10584 * Avoid some small subgroup attacks in Diffie-Hellman.
10586 *Nick Mathewson and Ben Laurie*
10588 * Add functions for well-known primes.
10592 * Extended Windows CE support.
10594 *Satoshi Nakamura and Andy Polyakov*
10596 * Initialize SSL_METHOD structures at compile time instead of during
10597 runtime, thus removing the need for a lock.
10601 * Make PKCS7_decrypt() work even if no certificate is supplied by
10602 attempting to decrypt each encrypted key in turn. Add support to
10607 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10609 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10612 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10616 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10617 key into the same file any more.
10621 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10625 * Add -utf8 command line and config file option to 'ca'.
10627 *Stefan <stf@udoma.org*
10629 * Removed the macro des_crypt(), as it seems to conflict with some
10630 libraries. Use DES_crypt().
10634 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10635 involves renaming the source and generated shared-libs for
10636 both. The engines will accept the corrected or legacy ids
10637 ('ncipher' and '4758_cca' respectively) when binding. NB,
10638 this only applies when building 'shared'.
10640 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10642 * Add attribute functions to EVP_PKEY structure. Modify
10643 PKCS12_create() to recognize a CSP name attribute and
10644 use it. Make -CSP option work again in pkcs12 utility.
10648 * Add new functionality to the bn blinding code:
10649 - automatic re-creation of the BN_BLINDING parameters after
10650 a fixed number of uses (currently 32)
10651 - add new function for parameter creation
10652 - introduce flags to control the update behaviour of the
10653 BN_BLINDING parameters
10654 - hide BN_BLINDING structure
10655 Add a second BN_BLINDING slot to the RSA structure to improve
10656 performance when a single RSA object is shared among several
10661 * Add support for DTLS.
10663 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10665 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10666 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10670 * Remove buggy and incomplete DH cert support from
10671 ssl/ssl_rsa.c and ssl/s3_both.c
10675 * Use SHA-1 instead of MD5 as the default digest algorithm for
10676 the `apps/openssl` commands.
10680 * Compile clean with "-Wall -Wmissing-prototypes
10681 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10682 DEBUG_SAFESTACK must also be set.
10686 * Change ./Configure so that certain algorithms can be disabled by default.
10687 The new counterpiece to "no-xxx" is "enable-xxx".
10689 The patented RC5 and MDC2 algorithms will now be disabled unless
10690 "enable-rc5" and "enable-mdc2", respectively, are specified.
10692 (IDEA remains enabled despite being patented. This is because IDEA
10693 is frequently required for interoperability, and there is no license
10694 fee for non-commercial use. As before, "no-idea" can be used to
10695 avoid this algorithm.)
10699 * Add processing of proxy certificates (see RFC 3820). This work was
10700 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10701 EGEE (Enabling Grids for E-science in Europe).
10705 * RC4 performance overhaul on modern architectures/implementations, such
10706 as Intel P4, IA-64 and AMD64.
10710 * New utility extract-section.pl. This can be used specify an alternative
10711 section number in a pod file instead of having to treat each file as
10712 a separate case in Makefile. This can be done by adding two lines to the
10715 =for comment openssl_section:XXX
10717 The blank line is mandatory.
10721 * New arguments -certform, -keyform and -pass for s_client and s_server
10722 to allow alternative format key and certificate files and passphrase
10727 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10728 update associated structures and add various utility functions.
10730 Add new policy related verify parameters, include policy checking in
10731 standard verify code. Enhance 'smime' application with extra parameters
10732 to support policy checking and print out.
10736 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10737 Nehemiah processors. These extensions support AES encryption in hardware
10738 as well as RNG (though RNG support is currently disabled).
10740 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10742 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10746 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10748 *Andy Polyakov and a number of other people*
10750 * Improved PowerPC platform support. Most notably BIGNUM assembler
10751 implementation contributed by IBM.
10753 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10755 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10756 exponent rather than 'unsigned long'. There is a corresponding change to
10757 the new 'rsa_keygen' element of the RSA_METHOD structure.
10759 *Jelte Jansen, Geoff Thorpe*
10761 * Functionality for creating the initial serial number file is now
10762 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10764 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10765 number file to 1, which is bound to cause problems. To avoid
10766 the problems while respecting compatibility between different 0.9.7
10767 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10768 CA.pl for serial number initialization. With the new release 0.9.8,
10769 we can fix the problem directly in the 'ca' utility.)
10773 * Reduced header interdependencies by declaring more opaque objects in
10774 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10775 give fewer recursive includes, which could break lazy source code - so
10776 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10777 developers should define this symbol when building and using openssl to
10778 ensure they track the recommended behaviour, interfaces, [etc], but
10779 backwards-compatible behaviour prevails when this isn't defined.
10783 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10787 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10788 This will generate a random key of the appropriate length based on the
10789 cipher context. The EVP_CIPHER can provide its own random key generation
10790 routine to support keys of a specific form. This is used in the des and
10791 3des routines to generate a key of the correct parity. Update S/MIME
10792 code to use new functions and hence generate correct parity DES keys.
10793 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10794 valid (weak or incorrect parity).
10798 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10799 as looking them up. This is useful when the verified structure may contain
10800 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10801 present unless the new PKCS7_NO_CRL flag is asserted.
10805 * Extend ASN1 oid configuration module. It now additionally accepts the
10808 shortName = some long name, 1.2.3.4
10812 * Reimplemented the BN_CTX implementation. There is now no more static
10813 limitation on the number of variables it can handle nor the depth of the
10814 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10815 information can now expand as required, and rather than having a single
10816 static array of bignums, BN_CTX now uses a linked-list of such arrays
10817 allowing it to expand on demand whilst maintaining the usefulness of
10818 BN_CTX's "bundling".
10822 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10823 to allow all RSA operations to function using a single BN_CTX.
10827 * Preliminary support for certificate policy evaluation and checking. This
10828 is initially intended to pass the tests outlined in "Conformance Testing
10829 of Relying Party Client Certificate Path Processing Logic" v1.07.
10833 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10834 remained unused and not that useful. A variety of other little bignum
10835 tweaks and fixes have also been made continuing on from the audit (see
10840 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10841 associated ASN1, EVP and SSL functions and old ASN1 macros.
10845 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10846 and this should never fail. So the return value from the use of
10847 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10848 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10852 * BN_CTX_get() should return zero-valued bignums, providing the same
10853 initialised value as BN_new().
10855 *Geoff Thorpe, suggested by Ulf Möller*
10857 * Support for inhibitAnyPolicy certificate extension.
10861 * An audit of the BIGNUM code is underway, for which debugging code is
10862 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10863 is considered valid when processing BIGNUMs, and causes execution to
10864 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10865 further steps are taken to deliberately pollute unused data in BIGNUM
10866 structures to try and expose faulty code further on. For now, openssl will
10867 (in its default mode of operation) continue to tolerate the inconsistent
10868 forms that it has tolerated in the past, but authors and packagers should
10869 consider trying openssl and their own applications when compiled with
10870 these debugging symbols defined. It will help highlight potential bugs in
10871 their own code, and will improve the test coverage for OpenSSL itself. At
10872 some point, these tighter rules will become openssl's default to improve
10873 maintainability, though the assert()s and other overheads will remain only
10874 in debugging configurations. See bn.h for more details.
10876 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10878 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10879 that can only be obtained through BN_CTX_new() (which implicitly
10880 initialises it). The presence of this function only made it possible
10881 to overwrite an existing structure (and cause memory leaks).
10885 * Because of the callback-based approach for implementing LHASH as a
10886 template type, lh_insert() adds opaque objects to hash-tables and
10887 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10888 to clean up those corresponding objects before destroying the hash table
10889 (and losing the object pointers). So some over-zealous constifications in
10890 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10891 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10892 prototyped to have "const" restrictions on the object pointers they are
10893 given (and so aren't required to cast them away any more).
10897 * The tmdiff.h API was so ugly and minimal that our own timing utility
10898 (speed) prefers to use its own implementation. The two implementations
10899 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10900 its object type properly exposed (MS_TM) instead of casting to/from
10901 `char *`. This may still change yet if someone realises MS_TM and
10903 aren't necessarily the greatest nomenclatures - but this is what was used
10904 internally to the implementation so I've used that for now.
10908 * Ensure that deprecated functions do not get compiled when
10909 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10910 the self-tests were still using deprecated key-generation functions so
10911 these have been updated also.
10915 * Reorganise PKCS#7 code to separate the digest location functionality
10916 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10917 New function PKCS7_set_digest() to set the digest type for PKCS#7
10918 digestedData type. Add additional code to correctly generate the
10919 digestedData type and add support for this type in PKCS7 initialization
10924 * New function PKCS7_set0_type_other() this initializes a PKCS7
10925 structure of type "other".
10929 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10930 sure the loop does correctly stop and breaking ("division by zero")
10931 modulus operations are not performed. The (pre-generated) prime
10932 table crypto/bn/bn_prime.h was already correct, but it could not be
10933 re-generated on some platforms because of the "division by zero"
10934 situation in the script.
10936 *Ralf S. Engelschall*
10938 * Update support for ECC-based TLS ciphersuites according to
10939 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10940 SHA-1 now is only used for "small" curves (where the
10941 representation of a field element takes up to 24 bytes); for
10942 larger curves, the field element resulting from ECDH is directly
10943 used as premaster secret.
10945 *Douglas Stebila (Sun Microsystems Laboratories)*
10947 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10948 curve secp160r1 to the tests.
10950 *Douglas Stebila (Sun Microsystems Laboratories)*
10952 * Add the possibility to load symbols globally with DSO.
10954 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10956 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10957 control of the error stack.
10961 * Add support for STORE in ENGINE.
10965 * Add the STORE type. The intention is to provide a common interface
10966 to certificate and key stores, be they simple file-based stores, or
10967 HSM-type store, or LDAP stores, or...
10968 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10972 * Add a generic structure called OPENSSL_ITEM. This can be used to
10973 pass a list of arguments to any function as well as provide a way
10974 for a function to pass data back to the caller.
10978 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10979 works like BUF_strdup() but can be used to duplicate a portion of
10980 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10985 * Add the function sk_find_ex() which works like sk_find(), but will
10986 return an index to an element even if an exact match couldn't be
10987 found. The index is guaranteed to point at the element where the
10988 searched-for key would be inserted to preserve sorting order.
10992 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10993 takes an extra flags argument for optional functionality. Currently,
10994 the following flags are defined:
10996 OBJ_BSEARCH_VALUE_ON_NOMATCH
10997 This one gets OBJ_bsearch_ex() to return a pointer to the first
10998 element where the comparing function returns a negative or zero
11001 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11002 This one gets OBJ_bsearch_ex() to return a pointer to the first
11003 element where the comparing function returns zero. This is useful
11004 if there are more than one element where the comparing function
11009 * Make it possible to create self-signed certificates with 'openssl ca'
11010 in such a way that the self-signed certificate becomes part of the
11011 CA database and uses the same mechanisms for serial number generation
11012 as all other certificate signing. The new flag '-selfsign' enables
11013 this functionality. Adapt CA.sh and CA.pl.in.
11017 * Add functionality to check the public key of a certificate request
11018 against a given private. This is useful to check that a certificate
11019 request can be signed by that key (self-signing).
11023 * Make it possible to have multiple active certificates with the same
11024 subject in the CA index file. This is done only if the keyword
11025 'unique_subject' is set to 'no' in the main CA section (default
11026 if 'CA_default') of the configuration file. The value is saved
11027 with the database itself in a separate index attribute file,
11028 named like the index file with '.attr' appended to the name.
11032 * Generate multi-valued AVAs using '+' notation in config files for
11037 * Support for nameConstraints certificate extension.
11041 * Support for policyConstraints certificate extension.
11045 * Support for policyMappings certificate extension.
11049 * Make sure the default DSA_METHOD implementation only uses its
11050 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11051 and change its own handlers to be NULL so as to remove unnecessary
11052 indirection. This lets alternative implementations fallback to the
11053 default implementation more easily.
11057 * Support for directoryName in GeneralName related extensions
11062 * Make it possible to link applications using Makefile.shared.
11063 Make that possible even when linking against static libraries!
11067 * Support for single pass processing for S/MIME signing. This now
11068 means that S/MIME signing can be done from a pipe, in addition
11069 cleartext signing (multipart/signed type) is effectively streaming
11070 and the signed data does not need to be all held in memory.
11072 This is done with a new flag PKCS7_STREAM. When this flag is set
11073 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11074 is done after the data is output (and digests calculated) in
11075 SMIME_write_PKCS7().
11079 * Add full support for -rpath/-R, both in shared libraries and
11080 applications, at least on the platforms where it's known how
11085 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11086 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11087 will now compute a table of multiples of the generator that
11088 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11089 faster (notably in the case of a single point multiplication,
11090 scalar * generator).
11092 *Nils Larsch, Bodo Moeller*
11094 * IPv6 support for certificate extensions. The various extensions
11095 which use the IP:a.b.c.d can now take IPv6 addresses using the
11096 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11101 * Added an ENGINE that implements RSA by performing private key
11102 exponentiations with the GMP library. The conversions to and from
11103 GMP's mpz_t format aren't optimised nor are any montgomery forms
11104 cached, and on x86 it appears OpenSSL's own performance has caught up.
11105 However there are likely to be other architectures where GMP could
11106 provide a boost. This ENGINE is not built in by default, but it can be
11107 specified at Configure time and should be accompanied by the necessary
11108 linker additions, eg;
11109 ./config -DOPENSSL_USE_GMP -lgmp
11113 * "openssl engine" will not display ENGINE/DSO load failure errors when
11114 testing availability of engines with "-t" - the old behaviour is
11115 produced by increasing the feature's verbosity with "-tt".
11119 * ECDSA routines: under certain error conditions uninitialized BN objects
11120 could be freed. Solution: make sure initialization is performed early
11121 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11126 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11127 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11128 software implementations. For DSA and DH, parameter generation can
11129 also be overridden by providing the appropriate method callbacks.
11133 * Change the "progress" mechanism used in key-generation and
11134 primality testing to functions that take a new BN_GENCB pointer in
11135 place of callback/argument pairs. The new API functions have `_ex`
11136 postfixes and the older functions are reimplemented as wrappers for
11137 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11138 declarations of the old functions to help (graceful) attempts to
11139 migrate to the new functions. Also, the new key-generation API
11140 functions operate on a caller-supplied key-structure and return
11141 success/failure rather than returning a key or NULL - this is to
11142 help make "keygen" another member function of RSA_METHOD etc.
11144 Example for using the new callback interface:
11146 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11147 void *my_arg = ...;
11150 BN_GENCB_set(&my_cb, my_callback, my_arg);
11152 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11153 /* For the meaning of a, b in calls to my_callback(), see the
11154 * documentation of the function that calls the callback.
11155 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11156 * my_callback should return 1 if it wants BN_is_prime_ex()
11157 * to continue, or 0 to stop.
11162 * Change the ZLIB compression method to be stateful, and make it
11163 available to TLS with the number defined in
11164 draft-ietf-tls-compression-04.txt.
11168 * Add the ASN.1 structures and functions for CertificatePair, which
11169 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11171 CertificatePair ::= SEQUENCE {
11172 forward [0] Certificate OPTIONAL,
11173 reverse [1] Certificate OPTIONAL,
11174 -- at least one of the pair shall be present -- }
11176 Also implement the PEM functions to read and write certificate
11177 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11179 This needed to be defined, mostly for the sake of the LDAP
11180 attribute crossCertificatePair, but may prove useful elsewhere as
11185 * Make it possible to inhibit symlinking of shared libraries in
11186 Makefile.shared, for Cygwin's sake.
11190 * Extend the BIGNUM API by creating a function
11191 void BN_set_negative(BIGNUM *a, int neg);
11192 and a macro that behave like
11193 int BN_is_negative(const BIGNUM *a);
11195 to avoid the need to access 'a->neg' directly in applications.
11199 * Implement fast modular reduction for pseudo-Mersenne primes
11200 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11201 EC_GROUP_new_curve_GFp() will now automatically use this
11204 *Nils Larsch <nla@trustcenter.de>*
11206 * Add new lock type (CRYPTO_LOCK_BN).
11210 * Change the ENGINE framework to automatically load engines
11211 dynamically from specific directories unless they could be
11212 found to already be built in or loaded. Move all the
11213 current engines except for the cryptodev one to a new
11214 directory engines/.
11215 The engines in engines/ are built as shared libraries if
11216 the "shared" options was given to ./Configure or ./config.
11217 Otherwise, they are inserted in libcrypto.a.
11218 /usr/local/ssl/engines is the default directory for dynamic
11219 engines, but that can be overridden at configure time through
11220 the usual use of --prefix and/or --openssldir, and at run
11221 time with the environment variable OPENSSL_ENGINES.
11223 *Geoff Thorpe and Richard Levitte*
11225 * Add Makefile.shared, a helper makefile to build shared
11226 libraries. Adapt Makefile.org.
11230 * Add version info to Win32 DLLs.
11232 *Peter 'Luna' Runestig" <peter@runestig.com>*
11234 * Add new 'medium level' PKCS#12 API. Certificates and keys
11235 can be added using this API to created arbitrary PKCS#12
11236 files while avoiding the low-level API.
11238 New options to PKCS12_create(), key or cert can be NULL and
11239 will then be omitted from the output file. The encryption
11240 algorithm NIDs can be set to -1 for no encryption, the mac
11241 iteration count can be set to 0 to omit the mac.
11243 Enhance pkcs12 utility by making the -nokeys and -nocerts
11244 options work when creating a PKCS#12 file. New option -nomac
11245 to omit the mac, NONE can be set for an encryption algorithm.
11246 New code is modified to use the enhanced PKCS12_create()
11247 instead of the low-level API.
11251 * Extend ASN1 encoder to support indefinite length constructed
11252 encoding. This can output sequences tags and octet strings in
11253 this form. Modify pk7_asn1.c to support indefinite length
11254 encoding. This is experimental and needs additional code to
11255 be useful, such as an ASN1 bio and some enhanced streaming
11258 Extend template encode functionality so that tagging is passed
11259 down to the template encoder.
11263 * Let 'openssl req' fail if an argument to '-newkey' is not
11264 recognized instead of using RSA as a default.
11268 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11269 As these are not official, they are not included in "ALL";
11270 the "ECCdraft" ciphersuite group alias can be used to select them.
11272 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11274 * Add ECDH engine support.
11276 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11278 * Add ECDH in new directory crypto/ecdh/.
11280 *Douglas Stebila (Sun Microsystems Laboratories)*
11282 * Let BN_rand_range() abort with an error after 100 iterations
11283 without success (which indicates a broken PRNG).
11287 * Change BN_mod_sqrt() so that it verifies that the input value
11288 is really the square of the return value. (Previously,
11289 BN_mod_sqrt would show GIGO behaviour.)
11293 * Add named elliptic curves over binary fields from X9.62, SECG,
11294 and WAP/WTLS; add OIDs that were still missing.
11296 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11298 * Extend the EC library for elliptic curves over binary fields
11299 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11302 EC_GF2m_simple_method
11306 EC_GROUP_new_curve_GF2m
11307 EC_GROUP_set_curve_GF2m
11308 EC_GROUP_get_curve_GF2m
11309 EC_POINT_set_affine_coordinates_GF2m
11310 EC_POINT_get_affine_coordinates_GF2m
11311 EC_POINT_set_compressed_coordinates_GF2m
11313 Point compression for binary fields is disabled by default for
11314 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11317 As binary polynomials are represented as BIGNUMs, various members
11318 of the EC_GROUP and EC_POINT data structures can be shared
11319 between the implementations for prime fields and binary fields;
11320 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11321 are essentially identical to their `..._GFp` counterparts.
11322 (For simplicity, the `..._GFp` prefix has been dropped from
11323 various internal method names.)
11325 An internal 'field_div' method (similar to 'field_mul' and
11326 'field_sqr') has been added; this is used only for binary fields.
11328 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11330 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11331 through methods ('mul', 'precompute_mult').
11333 The generic implementations (now internally called 'ec_wNAF_mul'
11334 and 'ec_wNAF_precomputed_mult') remain the default if these
11335 methods are undefined.
11337 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11339 * New function EC_GROUP_get_degree, which is defined through
11340 EC_METHOD. For curves over prime fields, this returns the bit
11341 length of the modulus.
11343 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11345 * New functions EC_GROUP_dup, EC_POINT_dup.
11346 (These simply call ..._new and ..._copy).
11348 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11350 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11351 Polynomials are represented as BIGNUMs (where the sign bit is not
11352 used) in the following functions [macros]:
11355 BN_GF2m_sub [= BN_GF2m_add]
11356 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11357 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11358 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11360 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11361 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11362 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11363 BN_GF2m_cmp [= BN_ucmp]
11365 (Note that only the 'mod' functions are actually for fields GF(2^m).
11366 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11368 For some functions, an the irreducible polynomial defining a
11369 field can be given as an 'unsigned int[]' with strictly
11370 decreasing elements giving the indices of those bits that are set;
11371 i.e., p[] represents the polynomial
11372 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11374 p[0] > p[1] > ... > p[k] = 0.
11375 This applies to the following functions:
11378 BN_GF2m_mod_mul_arr
11379 BN_GF2m_mod_sqr_arr
11380 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11381 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11382 BN_GF2m_mod_exp_arr
11383 BN_GF2m_mod_sqrt_arr
11384 BN_GF2m_mod_solve_quad_arr
11388 Conversion can be performed by the following functions:
11393 bntest.c has additional tests for binary polynomial arithmetic.
11395 Two implementations for BN_GF2m_mod_div() are available.
11396 The default algorithm simply uses BN_GF2m_mod_inv() and
11397 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11398 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11399 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11401 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11403 * Add new error code 'ERR_R_DISABLED' that can be used when some
11404 functionality is disabled at compile-time.
11406 *Douglas Stebila <douglas.stebila@sun.com>*
11408 * Change default behaviour of 'openssl asn1parse' so that more
11409 information is visible when viewing, e.g., a certificate:
11411 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11412 mode the content of non-printable OCTET STRINGs is output in a
11413 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11414 avoid the appearance of a printable string.
11416 *Nils Larsch <nla@trustcenter.de>*
11418 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11420 EC_GROUP_set_asn1_flag()
11421 EC_GROUP_get_asn1_flag()
11422 EC_GROUP_set_point_conversion_form()
11423 EC_GROUP_get_point_conversion_form()
11424 These control ASN1 encoding details:
11425 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11426 has been set to OPENSSL_EC_NAMED_CURVE.
11427 - Points are encoded in uncompressed form by default; options for
11428 asn1_for are as for point2oct, namely
11429 POINT_CONVERSION_COMPRESSED
11430 POINT_CONVERSION_UNCOMPRESSED
11431 POINT_CONVERSION_HYBRID
11433 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11435 EC_GROUP_set_seed()
11436 EC_GROUP_get0_seed()
11437 EC_GROUP_get_seed_len()
11438 This is used only for ASN1 purposes (so far).
11440 *Nils Larsch <nla@trustcenter.de>*
11442 * Add 'field_type' member to EC_METHOD, which holds the NID
11443 of the appropriate field type OID. The new function
11444 EC_METHOD_get_field_type() returns this value.
11446 *Nils Larsch <nla@trustcenter.de>*
11449 EC_POINT_point2bn()
11450 EC_POINT_bn2point()
11451 EC_POINT_point2hex()
11452 EC_POINT_hex2point()
11453 providing useful interfaces to EC_POINT_point2oct() and
11454 EC_POINT_oct2point().
11456 *Nils Larsch <nla@trustcenter.de>*
11458 * Change internals of the EC library so that the functions
11459 EC_GROUP_set_generator()
11460 EC_GROUP_get_generator()
11461 EC_GROUP_get_order()
11462 EC_GROUP_get_cofactor()
11463 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11464 to methods, which would lead to unnecessary code duplication when
11465 adding different types of curves.
11467 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11469 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11470 arithmetic, and such that modified wNAFs are generated
11471 (which avoid length expansion in many cases).
11475 * Add a function EC_GROUP_check_discriminant() (defined via
11476 EC_METHOD) that verifies that the curve discriminant is non-zero.
11478 Add a function EC_GROUP_check() that makes some sanity tests
11479 on a EC_GROUP, its generator and order. This includes
11480 EC_GROUP_check_discriminant().
11482 *Nils Larsch <nla@trustcenter.de>*
11484 * Add ECDSA in new directory crypto/ecdsa/.
11486 Add applications 'openssl ecparam' and 'openssl ecdsa'
11487 (these are based on 'openssl dsaparam' and 'openssl dsa').
11489 ECDSA support is also included in various other files across the
11490 library. Most notably,
11491 - 'openssl req' now has a '-newkey ecdsa:file' option;
11492 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11493 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11494 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11495 them suitable for ECDSA where domain parameters must be
11496 extracted before the specific public key;
11497 - ECDSA engine support has been added.
11499 *Nils Larsch <nla@trustcenter.de>*
11501 * Include some named elliptic curves, and add OIDs from X9.62,
11502 SECG, and WAP/WTLS. Each curve can be obtained from the new
11504 EC_GROUP_new_by_curve_name(),
11505 and the list of available named curves can be obtained with
11506 EC_get_builtin_curves().
11507 Also add a 'curve_name' member to EC_GROUP objects, which can be
11509 EC_GROUP_set_curve_name()
11510 EC_GROUP_get_curve_name()
11512 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11514 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11515 was actually never needed) and in BN_mul(). The removal in BN_mul()
11516 required a small change in bn_mul_part_recursive() and the addition
11517 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11518 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11519 bn_sub_words() and bn_add_words() except they take arrays with
11524 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11526 * Cleanse PEM buffers before freeing them since they may contain
11529 *Benjamin Bennett <ben@psc.edu>*
11531 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11532 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11533 authentication-only ciphersuites.
11537 * Since AES128 and AES256 share a single mask bit in the logic of
11538 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11539 kludge to work properly if AES128 is available and AES256 isn't.
11543 * Expand security boundary to match 1.1.1 module.
11547 * Remove redundant features: hash file source, editing of test vectors
11548 modify fipsld to use external fips_premain.c signature.
11552 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11553 run algorithm test programs.
11557 * Make algorithm test programs more tolerant of whitespace.
11561 * Have SSL/TLS server implementation tolerate "mismatched" record
11562 protocol version while receiving ClientHello even if the
11563 ClientHello is fragmented. (The server can't insist on the
11564 particular protocol version it has chosen before the ServerHello
11565 message has informed the client about his choice.)
11569 * Load error codes if they are not already present instead of using a
11570 static variable. This allows them to be cleanly unloaded and reloaded.
11574 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11576 * Introduce limits to prevent malicious keys being able to
11577 cause a denial of service. ([CVE-2006-2940])
11579 *Steve Henson, Bodo Moeller*
11581 * Fix ASN.1 parsing of certain invalid structures that can result
11582 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11584 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11585 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11587 * Fix SSL client code which could crash if connecting to a
11588 malicious SSLv2 server. ([CVE-2006-4343])
11590 *Tavis Ormandy and Will Drewry, Google Security Team*
11592 * Change ciphersuite string processing so that an explicit
11593 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11594 will no longer include "AES128-SHA"), and any other similar
11595 ciphersuite (same bitmap) from *other* protocol versions (so that
11596 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11597 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11598 changes from 0.9.8b and 0.9.8d.
11602 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11604 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11605 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11607 * Change the Unix randomness entropy gathering to use poll() when
11608 possible instead of select(), since the latter has some
11609 undesirable limitations.
11611 *Darryl Miles via Richard Levitte and Bodo Moeller*
11613 * Disable rogue ciphersuites:
11615 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11616 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11617 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11619 The latter two were purportedly from
11620 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11623 Also deactivate the remaining ciphersuites from
11624 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11625 unofficial, and the ID has long expired.
11629 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11630 dual-core machines) and other potential thread-safety issues.
11634 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11636 * Adapt fipsld and the build system to link against the validated FIPS
11637 module in FIPS mode.
11641 * Fixes for VC++ 2005 build under Windows.
11645 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11646 from a Windows bash shell such as MSYS. It is autodetected from the
11647 "config" script when run from a VC++ environment. Modify standard VC++
11648 build to use fipscanister.o from the GNU make build.
11652 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11654 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11655 The value now differs depending on if you build for FIPS or not.
11656 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11657 safely run with a non-FIPSed libcrypto, as it may crash because of
11658 the difference induced by this change.
11662 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11664 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11665 (part of SSL_OP_ALL). This option used to disable the
11666 countermeasure against man-in-the-middle protocol-version
11667 rollback in the SSL 2.0 server implementation, which is a bad
11668 idea. ([CVE-2005-2969])
11670 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11671 for Information Security, National Institute of Advanced Industrial
11672 Science and Technology [AIST, Japan)]*
11674 * Minimal support for X9.31 signatures and PSS padding modes. This is
11675 mainly for FIPS compliance and not fully integrated at this stage.
11679 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11680 the exponentiation using a fixed-length exponent. (Otherwise,
11681 the information leaked through timing could expose the secret key
11682 after many signatures; cf. Bleichenbacher's attack on DSA with
11687 * Make a new fixed-window mod_exp implementation the default for
11688 RSA, DSA, and DH private-key operations so that the sequence of
11689 squares and multiplies and the memory access pattern are
11690 independent of the particular secret key. This will mitigate
11691 cache-timing and potential related attacks.
11693 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11694 and this is automatically used by BN_mod_exp_mont() if the new flag
11695 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11696 will use this BN flag for private exponents unless the flag
11697 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11698 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11700 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11702 * Change the client implementation for SSLv23_method() and
11703 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11704 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11705 (Previously, the SSL 2.0 backwards compatible Client Hello
11706 message format would be used even with SSL_OP_NO_SSLv2.)
11710 * Add support for smime-type MIME parameter in S/MIME messages which some
11715 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11716 a threadsafe manner. Modify rsa code to use new function and add calls
11717 to dsa and dh code (which had race conditions before).
11721 * Include the fixed error library code in the C error file definitions
11722 instead of fixing them up at runtime. This keeps the error code
11723 structures constant.
11727 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11729 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11732 * Fixes for newer kerberos headers. NB: the casts are needed because
11733 the 'length' field is signed on one version and unsigned on another
11734 with no (?) obvious way to tell the difference, without these VC++
11735 complains. Also the "definition" of FAR (blank) is no longer included
11736 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11737 some needed definitions.
11741 * Undo Cygwin change.
11745 * Added support for proxy certificates according to RFC 3820.
11746 Because they may be a security thread to unaware applications,
11747 they must be explicitly allowed in run-time. See
11748 docs/HOWTO/proxy_certificates.txt for further information.
11752 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11754 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11755 server and client random values. Previously
11756 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11757 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11759 This change has negligible security impact because:
11761 1. Server and client random values still have 24 bytes of pseudo random
11764 2. Server and client random values are sent in the clear in the initial
11767 3. The master secret is derived using the premaster secret (48 bytes in
11768 size for static RSA ciphersuites) as well as client server and random
11771 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11774 *Stephen Henson, reported by UK NISCC*
11776 * Use Windows randomness collection on Cygwin.
11780 * Fix hang in EGD/PRNGD query when communication socket is closed
11781 prematurely by EGD/PRNGD.
11783 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11785 * Prompt for pass phrases when appropriate for PKCS12 input format.
11789 * Back-port of selected performance improvements from development
11790 branch, as well as improved support for PowerPC platforms.
11794 * Add lots of checks for memory allocation failure, error codes to indicate
11795 failure and freeing up memory if a failure occurs.
11797 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11799 * Add new -passin argument to dgst.
11803 * Perform some character comparisons of different types in X509_NAME_cmp:
11804 this is needed for some certificates that re-encode DNs into UTF8Strings
11805 (in violation of RFC3280) and can't or won't issue name rollover
11810 * Make an explicit check during certificate validation to see that
11811 the CA setting in each certificate on the chain is correct. As a
11812 side effect always do the following basic checks on extensions,
11813 not just when there's an associated purpose to the check:
11815 - if there is an unhandled critical extension (unless the user
11816 has chosen to ignore this fault)
11817 - if the path length has been exceeded (if one is set at all)
11818 - that certain extensions fit the associated purpose (if one has
11823 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11825 * Avoid a race condition when CRLs are checked in a multi threaded
11826 environment. This would happen due to the reordering of the revoked
11827 entries during signature checking and serial number lookup. Now the
11828 encoding is cached and the serial number sort performed under a lock.
11829 Add new STACK function sk_is_sorted().
11833 * Add Delta CRL to the extension code.
11837 * Various fixes to s3_pkt.c so alerts are sent properly.
11839 *David Holmes <d.holmes@f5.com>*
11841 * Reduce the chances of duplicate issuer name and serial numbers (in
11842 violation of RFC3280) using the OpenSSL certificate creation utilities.
11843 This is done by creating a random 64 bit value for the initial serial
11844 number when a serial number file is created or when a self signed
11845 certificate is created using 'openssl req -x509'. The initial serial
11846 number file is created using 'openssl x509 -next_serial' in CA.pl
11847 rather than being initialized to 1.
11851 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11853 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11854 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11856 *Joe Orton, Steve Henson*
11858 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11861 *Joe Orton, Steve Henson*
11863 * Make it possible to have multiple active certificates with the same
11864 subject in the CA index file. This is done only if the keyword
11865 'unique_subject' is set to 'no' in the main CA section (default
11866 if 'CA_default') of the configuration file. The value is saved
11867 with the database itself in a separate index attribute file,
11868 named like the index file with '.attr' appended to the name.
11872 * X509 verify fixes. Disable broken certificate workarounds when
11873 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11874 keyUsage extension present. Don't accept CRLs with unhandled critical
11875 extensions: since verify currently doesn't process CRL extensions this
11876 rejects a CRL with *any* critical extensions. Add new verify error codes
11881 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11882 A clarification of RFC2560 will require the use of OCTET STRINGs and
11883 some implementations cannot handle the current raw format. Since OpenSSL
11884 copies and compares OCSP nonces as opaque blobs without any attempt at
11885 parsing them this should not create any compatibility issues.
11889 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11890 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11891 this HMAC (and other) operations are several times slower than OpenSSL
11896 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11898 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11900 * Use the correct content when signing type "other".
11904 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11906 * Fix various bugs revealed by running the NISCC test suite:
11908 Stop out of bounds reads in the ASN1 code when presented with
11909 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11911 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11913 If verify callback ignores invalid public key errors don't try to check
11914 certificate signature with the NULL public key.
11918 * New -ignore_err option in ocsp application to stop the server
11919 exiting on the first error in a request.
11923 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11924 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11929 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11930 extra data after the compression methods not only for TLS 1.0
11931 but also for SSL 3.0 (as required by the specification).
11933 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11935 * Change X509_certificate_type() to mark the key as exported/exportable
11936 when it's 512 *bits* long, not 512 bytes.
11940 * Change AES_cbc_encrypt() so it outputs exact multiple of
11941 blocks during encryption.
11945 * Various fixes to base64 BIO and non blocking I/O. On write
11946 flushes were not handled properly if the BIO retried. On read
11947 data was not being buffered properly and had various logic bugs.
11948 This also affects blocking I/O when the data being decoded is a
11953 * Various S/MIME bugfixes and compatibility changes:
11954 output correct application/pkcs7 MIME type if
11955 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11956 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11957 of files as .eml work). Correctly handle very long lines in MIME
11962 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11964 * Countermeasure against the Klima-Pokorny-Rosa extension of
11965 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11966 a protocol version number mismatch like a decryption error
11967 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11971 * Turn on RSA blinding by default in the default implementation
11972 to avoid a timing attack. Applications that don't want it can call
11973 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11974 They would be ill-advised to do so in most cases.
11976 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11978 * Change RSA blinding code so that it works when the PRNG is not
11979 seeded (in this case, the secret RSA exponent is abused as
11980 an unpredictable seed -- if it is not unpredictable, there
11981 is no point in blinding anyway). Make RSA blinding thread-safe
11982 by remembering the creator's thread ID in rsa->blinding and
11983 having all other threads use local one-time blinding factors
11984 (this requires more computation than sharing rsa->blinding, but
11985 avoids excessive locking; and if an RSA object is not shared
11986 between threads, blinding will still be very fast).
11990 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11991 ENGINE as defaults for all supported algorithms irrespective of
11992 the 'flags' parameter. 'flags' is now honoured, so applications
11993 should make sure they are passing it correctly.
11997 * Target "mingw" now allows native Windows code to be generated in
11998 the Cygwin environment as well as with the MinGW compiler.
12002 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12004 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12005 via timing by performing a MAC computation even if incorrect
12006 block cipher padding has been found. This is a countermeasure
12007 against active attacks where the attacker has to distinguish
12008 between bad padding and a MAC verification error. ([CVE-2003-0078])
12010 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12011 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12012 Martin Vuagnoux (EPFL, Ilion)*
12014 * Make the no-err option work as intended. The intention with no-err
12015 is not to have the whole error stack handling routines removed from
12016 libcrypto, it's only intended to remove all the function name and
12017 reason texts, thereby removing some of the footprint that may not
12018 be interesting if those errors aren't displayed anyway.
12020 NOTE: it's still possible for any application or module to have its
12021 own set of error texts inserted. The routines are there, just not
12022 used by default when no-err is given.
12026 * Add support for FreeBSD on IA64.
12028 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12030 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12031 Kerberos function mit_des_cbc_cksum(). Before this change,
12032 the value returned by DES_cbc_cksum() was like the one from
12033 mit_des_cbc_cksum(), except the bytes were swapped.
12035 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12037 * Allow an application to disable the automatic SSL chain building.
12038 Before this a rather primitive chain build was always performed in
12039 ssl3_output_cert_chain(): an application had no way to send the
12040 correct chain if the automatic operation produced an incorrect result.
12042 Now the chain builder is disabled if either:
12044 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12046 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12048 The reasoning behind this is that an application would not want the
12049 auto chain building to take place if extra chain certificates are
12050 present and it might also want a means of sending no additional
12051 certificates (for example the chain has two certificates and the
12056 * Add the possibility to build without the ENGINE framework.
12058 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12060 * Under Win32 gmtime() can return NULL: check return value in
12061 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12065 * DSA routines: under certain error conditions uninitialized BN objects
12066 could be freed. Solution: make sure initialization is performed early
12067 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12068 Nils Larsch <nla@trustcenter.de> via PR#459)
12072 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12073 checked on reconnect on the client side, therefore session resumption
12074 could still fail with a "ssl session id is different" error. This
12075 behaviour is masked when SSL_OP_ALL is used due to
12076 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12077 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12078 followup to PR #377.
12082 * IA-32 assembler support enhancements: unified ELF targets, support
12083 for SCO/Caldera platforms, fix for Cygwin shared build.
12087 * Add support for FreeBSD on sparc64. As a consequence, support for
12088 FreeBSD on non-x86 processors is separate from x86 processors on
12089 the config script, much like the NetBSD support.
12091 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12093 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12095 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12098 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12099 code (06) was taken as the first octet of the session ID and the last
12100 octet was ignored consequently. As a result SSLv2 client side session
12101 caching could not have worked due to the session ID mismatch between
12103 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12108 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12109 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12114 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12115 seems that in spite of existing for more than a year, many application
12116 author have done nothing to provide the necessary callbacks, which
12117 means that this particular engine will not work properly anywhere.
12118 This is a very unfortunate situation which forces us, in the name
12119 of usability, to give the hw_ncipher.c a static lock, which is part
12121 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12122 appear in 0.9.8 or later. We EXPECT application authors to have
12123 dealt properly with this when 0.9.8 is released (unless we actually
12124 make such changes in the libcrypto locking code that changes will
12125 have to be made anyway).
12129 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12130 octets have been read, EOF or an error occurs. Without this change
12131 some truncated ASN1 structures will not produce an error.
12135 * Disable Heimdal support, since it hasn't been fully implemented.
12136 Still give the possibility to force the use of Heimdal, but with
12137 warnings and a request that patches get sent to openssl-dev.
12141 * Add the VC-CE target, introduce the WINCE sysname, and add
12142 INSTALL.WCE and appropriate conditionals to make it build.
12144 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12146 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12147 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12148 edit numbers of the version.
12150 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12152 * Introduce safe string copy and catenation functions
12153 (BUF_strlcpy() and BUF_strlcat()).
12155 *Ben Laurie (CHATS) and Richard Levitte*
12157 * Avoid using fixed-size buffers for one-line DNs.
12159 *Ben Laurie (CHATS)*
12161 * Add BUF_MEM_grow_clean() to avoid information leakage when
12162 resizing buffers containing secrets, and use where appropriate.
12164 *Ben Laurie (CHATS)*
12166 * Avoid using fixed size buffers for configuration file location.
12168 *Ben Laurie (CHATS)*
12170 * Avoid filename truncation for various CA files.
12172 *Ben Laurie (CHATS)*
12174 * Use sizeof in preference to magic numbers.
12176 *Ben Laurie (CHATS)*
12178 * Avoid filename truncation in cert requests.
12180 *Ben Laurie (CHATS)*
12182 * Add assertions to check for (supposedly impossible) buffer
12185 *Ben Laurie (CHATS)*
12187 * Don't cache truncated DNS entries in the local cache (this could
12188 potentially lead to a spoofing attack).
12190 *Ben Laurie (CHATS)*
12192 * Fix various buffers to be large enough for hex/decimal
12193 representations in a platform independent manner.
12195 *Ben Laurie (CHATS)*
12197 * Add CRYPTO_realloc_clean() to avoid information leakage when
12198 resizing buffers containing secrets, and use where appropriate.
12200 *Ben Laurie (CHATS)*
12202 * Add BIO_indent() to avoid much slightly worrying code to do
12205 *Ben Laurie (CHATS)*
12207 * Convert sprintf()/BIO_puts() to BIO_printf().
12209 *Ben Laurie (CHATS)*
12211 * buffer_gets() could terminate with the buffer only half
12214 *Ben Laurie (CHATS)*
12216 * Add assertions to prevent user-supplied crypto functions from
12217 overflowing internal buffers by having large block sizes, etc.
12219 *Ben Laurie (CHATS)*
12221 * New OPENSSL_assert() macro (similar to assert(), but enabled
12224 *Ben Laurie (CHATS)*
12226 * Eliminate unused copy of key in RC4.
12228 *Ben Laurie (CHATS)*
12230 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12232 *Ben Laurie (CHATS)*
12234 * Fix off-by-one error in EGD path.
12236 *Ben Laurie (CHATS)*
12238 * If RANDFILE path is too long, ignore instead of truncating.
12240 *Ben Laurie (CHATS)*
12242 * Eliminate unused and incorrectly sized X.509 structure
12245 *Ben Laurie (CHATS)*
12247 * Eliminate unused and dangerous function knumber().
12249 *Ben Laurie (CHATS)*
12251 * Eliminate unused and dangerous structure, KSSL_ERR.
12253 *Ben Laurie (CHATS)*
12255 * Protect against overlong session ID context length in an encoded
12256 session object. Since these are local, this does not appear to be
12259 *Ben Laurie (CHATS)*
12261 * Change from security patch (see 0.9.6e below) that did not affect
12262 the 0.9.6 release series:
12264 Remote buffer overflow in SSL3 protocol - an attacker could
12265 supply an oversized master key in Kerberos-enabled versions.
12268 *Ben Laurie (CHATS)*
12270 * Change the SSL kerb5 codes to match RFC 2712.
12274 * Make -nameopt work fully for req and add -reqopt switch.
12276 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12278 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12280 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12282 * Make sure tests can be performed even if the corresponding algorithms
12283 have been removed entirely. This was also the last step to make
12284 OpenSSL compilable with DJGPP under all reasonable conditions.
12286 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12288 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12289 to allow version independent disabling of normally unselected ciphers,
12290 which may be activated as a side-effect of selecting a single cipher.
12292 (E.g., cipher list string "RSA" enables ciphersuites that are left
12293 out of "ALL" because they do not provide symmetric encryption.
12294 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12296 *Lutz Jaenicke, Bodo Moeller*
12298 * Add appropriate support for separate platform-dependent build
12299 directories. The recommended way to make a platform-dependent
12300 build directory is the following (tested on Linux), maybe with
12303 # Place yourself outside of the OpenSSL source tree. In
12304 # this example, the environment variable OPENSSL_SOURCE
12305 # is assumed to contain the absolute OpenSSL source directory.
12306 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12307 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12308 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12309 mkdir -p `dirname $F`
12310 ln -s $OPENSSL_SOURCE/$F $F
12313 To be absolutely sure not to disturb the source tree, a "make clean"
12314 is a good thing. If it isn't successful, don't worry about it,
12315 it probably means the source directory is very clean.
12319 * Make sure any ENGINE control commands make local copies of string
12320 pointers passed to them whenever necessary. Otherwise it is possible
12321 the caller may have overwritten (or deallocated) the original string
12322 data when a later ENGINE operation tries to use the stored values.
12324 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12326 * Improve diagnostics in file reading and command-line digests.
12328 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12330 * Add AES modes CFB and OFB to the object database. Correct an
12331 error in AES-CFB decryption.
12335 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12336 allows existing EVP_CIPHER_CTX structures to be reused after
12337 calling `EVP_*Final()`. This behaviour is used by encryption
12338 BIOs and some applications. This has the side effect that
12339 applications must explicitly clean up cipher contexts with
12340 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12344 * Check the values of dna and dnb in bn_mul_recursive before calling
12345 bn_mul_comba (a non zero value means the a or b arrays do not contain
12346 n2 elements) and fallback to bn_mul_normal if either is not zero.
12350 * Fix escaping of non-ASCII characters when using the -subj option
12351 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12355 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12356 form for "surname", serialNumber has no short form.
12357 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12358 therefore remove "mail" short name for "internet 7".
12359 The OID for unique identifiers in X509 certificates is
12360 x500UniqueIdentifier, not uniqueIdentifier.
12361 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12365 * Add an "init" command to the ENGINE config module and auto initialize
12366 ENGINEs. Without any "init" command the ENGINE will be initialized
12367 after all ctrl commands have been executed on it. If init=1 the
12368 ENGINE is initialized at that point (ctrls before that point are run
12369 on the uninitialized ENGINE and after on the initialized one). If
12370 init=0 then the ENGINE will not be initialized at all.
12374 * Fix the 'app_verify_callback' interface so that the user-defined
12375 argument is actually passed to the callback: In the
12376 SSL_CTX_set_cert_verify_callback() prototype, the callback
12377 declaration has been changed from
12380 int (*cb)(X509_STORE_CTX *,void *);
12381 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12382 i=s->ctx->app_verify_callback(&ctx)
12383 has been changed into
12384 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12386 To update applications using SSL_CTX_set_cert_verify_callback(),
12387 a dummy argument can be added to their callback functions.
12389 *D. K. Smetters <smetters@parc.xerox.com>*
12391 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12393 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12395 * Add and OPENSSL_LOAD_CONF define which will cause
12396 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12397 This allows older applications to transparently support certain
12398 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12399 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12400 load the config file and OPENSSL_add_all_algorithms_conf() which will
12401 always load it have also been added.
12405 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12406 Adjust NIDs and EVP layer.
12408 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12410 * Config modules support in openssl utility.
12412 Most commands now load modules from the config file,
12413 though in a few (such as version) this isn't done
12414 because it couldn't be used for anything.
12416 In the case of ca and req the config file used is
12417 the same as the utility itself: that is the -config
12418 command line option can be used to specify an
12423 * Move default behaviour from OPENSSL_config(). If appname is NULL
12424 use "openssl_conf" if filename is NULL use default openssl config file.
12428 * Add an argument to OPENSSL_config() to allow the use of an alternative
12429 config section name. Add a new flag to tolerate a missing config file
12430 and move code to CONF_modules_load_file().
12434 * Support for crypto accelerator cards from Accelerated Encryption
12435 Processing, www.aep.ie. (Use engine 'aep')
12436 The support was copied from 0.9.6c [engine] and adapted/corrected
12437 to work with the new engine framework.
12439 *AEP Inc. and Richard Levitte*
12441 * Support for SureWare crypto accelerator cards from Baltimore
12442 Technologies. (Use engine 'sureware')
12443 The support was copied from 0.9.6c [engine] and adapted
12444 to work with the new engine framework.
12448 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12449 make the newer ENGINE framework commands for the CHIL engine work.
12451 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12453 * Make it possible to produce shared libraries on ReliantUNIX.
12455 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12457 * Add the configuration target debug-linux-ppro.
12458 Make 'openssl rsa' use the general key loading routines
12459 implemented in `apps.c`, and make those routines able to
12460 handle the key format FORMAT_NETSCAPE and the variant
12463 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12465 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12467 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12469 * Add -keyform to rsautl, and document -engine.
12471 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12473 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12474 BIO_R_NO_SUCH_FILE error code rather than the generic
12475 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12479 * Add new functions
12480 ERR_peek_last_error
12481 ERR_peek_last_error_line
12482 ERR_peek_last_error_line_data.
12483 These are similar to
12485 ERR_peek_error_line
12486 ERR_peek_error_line_data,
12487 but report on the latest error recorded rather than the first one
12488 still in the error queue.
12490 *Ben Laurie, Bodo Moeller*
12492 * default_algorithms option in ENGINE config module. This allows things
12494 default_algorithms = ALL
12495 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12499 * Preliminary ENGINE config module.
12503 * New experimental application configuration code.
12507 * Change the AES code to follow the same name structure as all other
12508 symmetric ciphers, and behave the same way. Move everything to
12509 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12511 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12513 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12515 *Ben Laurie and Theo de Raadt*
12517 * Add option to output public keys in req command.
12519 *Massimiliano Pala madwolf@openca.org*
12521 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12522 (up to about 10% better than before for P-192 and P-224).
12526 * New functions/macros
12528 SSL_CTX_set_msg_callback(ctx, cb)
12529 SSL_CTX_set_msg_callback_arg(ctx, arg)
12530 SSL_set_msg_callback(ssl, cb)
12531 SSL_set_msg_callback_arg(ssl, arg)
12533 to request calling a callback function
12535 void cb(int write_p, int version, int content_type,
12536 const void *buf, size_t len, SSL *ssl, void *arg)
12538 whenever a protocol message has been completely received
12539 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12540 protocol version according to which the SSL library interprets
12541 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12542 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12543 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12544 specification (change_cipher_spec(20), alert(21), handshake(22)).
12545 'buf' and 'len' point to the actual message, 'ssl' to the
12546 SSL object, and 'arg' is the application-defined value set by
12547 SSL[_CTX]_set_msg_callback_arg().
12549 'openssl s_client' and 'openssl s_server' have new '-msg' options
12550 to enable a callback that displays all protocol messages.
12554 * Change the shared library support so shared libraries are built as
12555 soon as the corresponding static library is finished, and thereby get
12556 openssl and the test programs linked against the shared library.
12557 This still only happens when the keyword "shard" has been given to
12558 the configuration scripts.
12560 NOTE: shared library support is still an experimental thing, and
12561 backward binary compatibility is still not guaranteed.
12563 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12565 * Add support for Subject Information Access extension.
12567 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12569 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12570 additional bytes when new memory had to be allocated, not just
12571 when reusing an existing buffer.
12575 * New command line and configuration option 'utf8' for the req command.
12576 This allows field values to be specified as UTF8 strings.
12580 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12581 runs for the former and machine-readable output for the latter.
12585 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12586 of the e-mail address in the DN (i.e., it will go into a certificate
12587 extension only). The new configuration file option 'email_in_dn = no'
12588 has the same effect.
12590 *Massimiliano Pala madwolf@openca.org*
12592 * Change all functions with names starting with `des_` to be starting
12593 with `DES_` instead. Add wrappers that are compatible with libdes,
12594 but are named `_ossl_old_des_*`. Finally, add macros that map the
12595 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12596 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12597 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12600 Since we provide two compatibility mappings, the user needs to
12601 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12602 compatibility is desired. The default (i.e., when that macro
12603 isn't defined) is OpenSSL 0.9.6c compatibility.
12605 There are also macros that enable and disable the support of old
12606 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12607 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12608 are defined, the default will apply: to support the old des routines.
12610 In either case, one must include openssl/des.h to get the correct
12611 definitions. Do not try to just include openssl/des_old.h, that
12614 NOTE: This is a major break of an old API into a new one. Software
12615 authors are encouraged to switch to the `DES_` style functions. Some
12616 time in the future, des_old.h and the libdes compatibility functions
12617 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12618 default), and then completely removed.
12622 * Test for certificates which contain unsupported critical extensions.
12623 If such a certificate is found during a verify operation it is
12624 rejected by default: this behaviour can be overridden by either
12625 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12626 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12627 X509_supported_extension() has also been added which returns 1 if a
12628 particular extension is supported.
12632 * Modify the behaviour of EVP cipher functions in similar way to digests
12633 to retain compatibility with existing code.
12637 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12638 compatibility with existing code. In particular the 'ctx' parameter does
12639 not have to be to be initialized before the call to EVP_DigestInit() and
12640 it is tidied up after a call to EVP_DigestFinal(). New function
12641 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12642 EVP_MD_CTX_copy() changed to not require the destination to be
12643 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12644 requires the destination to be valid.
12646 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12647 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12651 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12652 so that complete 'Handshake' protocol structures are kept in memory
12653 instead of overwriting 'msg_type' and 'length' with 'body' data.
12657 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12659 *Massimo Santin via Richard Levitte*
12661 * Major restructuring to the underlying ENGINE code. This includes
12662 reduction of linker bloat, separation of pure "ENGINE" manipulation
12663 (initialisation, etc) from functionality dealing with implementations
12664 of specific crypto interfaces. This change also introduces integrated
12665 support for symmetric ciphers and digest implementations - so ENGINEs
12666 can now accelerate these by providing EVP_CIPHER and EVP_MD
12667 implementations of their own. This is detailed in
12668 [crypto/engine/README.md](crypto/engine/README.md)
12669 as it couldn't be adequately described here. However, there are a few
12670 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12671 were changed in the original introduction of ENGINE code have now
12672 reverted back - the hooking from this code to ENGINE is now a good
12673 deal more passive and at run-time, operations deal directly with
12674 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12675 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12676 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12677 they were not being used by the framework as there is no concept of a
12678 BIGNUM_METHOD and they could not be generalised to the new
12679 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12680 ENGINE_cpy() has been removed as it cannot be consistently defined in
12685 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12689 * Change mkdef.pl to sort symbols that get the same entry number,
12690 and make sure the automatically generated functions `ERR_load_*`
12691 become part of libeay.num as well.
12695 * New function SSL_renegotiate_pending(). This returns true once
12696 renegotiation has been requested (either SSL_renegotiate() call
12697 or HelloRequest/ClientHello received from the peer) and becomes
12698 false once a handshake has been completed.
12699 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12700 sends a HelloRequest, but does not ensure that a handshake takes
12701 place. SSL_renegotiate_pending() is useful for checking if the
12702 client has followed the request.)
12706 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12707 By default, clients may request session resumption even during
12708 renegotiation (if session ID contexts permit); with this option,
12709 session resumption is possible only in the first handshake.
12711 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12712 more bits available for options that should not be part of
12713 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12717 * Add some demos for certificate and certificate request creation.
12721 * Make maximum certificate chain size accepted from the peer application
12722 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12723 "Douglas E. Engert" <deengert@anl.gov>.
12727 * Add support for shared libraries for Unixware-7
12728 (Boyd Lynn Gerber <gerberb@zenez.com>).
12732 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12733 be done prior to destruction. Use this to unload error strings from
12734 ENGINEs that load their own error strings. NB: This adds two new API
12735 functions to "get" and "set" this destroy handler in an ENGINE.
12739 * Alter all existing ENGINE implementations (except "openssl" and
12740 "openbsd") to dynamically instantiate their own error strings. This
12741 makes them more flexible to be built both as statically-linked ENGINEs
12742 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12743 Also, add stub code to each that makes building them as self-contained
12744 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12748 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12749 implementations into applications that are completely implemented in
12750 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12751 commands that can be used to configure what shared-library to load and
12752 to control aspects of the way it is handled. Also, made an update to
12753 the [README-Engine.md](README-Engine.md) file
12754 that brings its information up-to-date and
12755 provides some information and instructions on the "dynamic" ENGINE
12756 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12760 * Make it possible to unload ranges of ERR strings with a new
12761 "ERR_unload_strings" function.
12765 * Add a copy() function to EVP_MD.
12769 * Make EVP_MD routines take a context pointer instead of just the
12770 md_data void pointer.
12774 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12775 that the digest can only process a single chunk of data
12776 (typically because it is provided by a piece of
12777 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12778 is only going to provide a single chunk of data, and hence the
12779 framework needn't accumulate the data for oneshot drivers.
12783 * As with "ERR", make it possible to replace the underlying "ex_data"
12784 functions. This change also alters the storage and management of global
12785 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12786 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12787 index counters. The API functions that use this state have been changed
12788 to take a "class_index" rather than pointers to the class's local STACK
12789 and counter, and there is now an API function to dynamically create new
12790 classes. This centralisation allows us to (a) plug a lot of the
12791 thread-safety problems that existed, and (b) makes it possible to clean
12792 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12793 such data would previously have always leaked in application code and
12794 workarounds were in place to make the memory debugging turn a blind eye
12795 to it. Application code that doesn't use this new function will still
12796 leak as before, but their memory debugging output will announce it now
12797 rather than letting it slide.
12799 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12800 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12801 has a return value to indicate success or failure.
12805 * Make it possible to replace the underlying "ERR" functions such that the
12806 global state (2 LHASH tables and 2 locks) is only used by the "default"
12807 implementation. This change also adds two functions to "get" and "set"
12808 the implementation prior to it being automatically set the first time
12809 any other ERR function takes place. Ie. an application can call "get",
12810 pass the return value to a module it has just loaded, and that module
12811 can call its own "set" function using that value. This means the
12812 module's "ERR" operations will use (and modify) the error state in the
12813 application and not in its own statically linked copy of OpenSSL code.
12817 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12818 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12819 the operation, and provides a more encapsulated way for external code
12820 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12821 to use these functions rather than manually incrementing the counts.
12823 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12827 * Add EVP test program.
12831 * Add symmetric cipher support to ENGINE. Expect the API to change!
12835 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12836 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12837 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12838 These allow a CRL to be built without having to access X509_CRL fields
12839 directly. Modify 'ca' application to use new functions.
12843 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12844 bug workarounds. Rollback attack detection is a security feature.
12845 The problem will only arise on OpenSSL servers when TLSv1 is not
12846 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12847 Software authors not wanting to support TLSv1 will have special reasons
12848 for their choice and can explicitly enable this option.
12850 *Bodo Moeller, Lutz Jaenicke*
12852 * Rationalise EVP so it can be extended: don't include a union of
12853 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12854 (similar to those existing for EVP_CIPHER_CTX).
12859 EVP_MD_CTX_init(&md); /* new function call */
12860 EVP_DigestInit(&md, EVP_sha1());
12861 EVP_DigestUpdate(&md, in, len);
12862 EVP_DigestFinal(&md, out, NULL);
12863 EVP_MD_CTX_cleanup(&md); /* new function call */
12867 * Make DES key schedule conform to the usual scheme, as well as
12868 correcting its structure. This means that calls to DES functions
12869 now have to pass a pointer to a des_key_schedule instead of a
12870 plain des_key_schedule (which was actually always a pointer
12873 des_key_schedule ks;
12875 des_set_key_checked(..., &ks);
12876 des_ncbc_encrypt(..., &ks, ...);
12878 (Note that a later change renames 'des_...' into 'DES_...'.)
12882 * Initial reduction of linker bloat: the use of some functions, such as
12883 PEM causes large amounts of unused functions to be linked in due to
12884 poor organisation. For example pem_all.c contains every PEM function
12885 which has a knock on effect of linking in large amounts of (unused)
12886 ASN1 code. Grouping together similar functions and splitting unrelated
12887 functions prevents this.
12891 * Cleanup of EVP macros.
12895 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12896 correct `_ecb suffix`.
12900 * Add initial OCSP responder support to ocsp application. The
12901 revocation information is handled using the text based index
12902 use by the ca application. The responder can either handle
12903 requests generated internally, supplied in files (for example
12904 via a CGI script) or using an internal minimal server.
12908 * Add configuration choices to get zlib compression for TLS.
12912 * Changes to Kerberos SSL for RFC 2712 compliance:
12913 1. Implemented real KerberosWrapper, instead of just using
12914 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12915 2. Implemented optional authenticator field of KerberosWrapper.
12917 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12918 and authenticator structs; see crypto/krb5/.
12920 Generalized Kerberos calls to support multiple Kerberos libraries.
12921 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12922 via Richard Levitte*
12924 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12925 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12926 values for each of the key sizes rather than having just
12927 parameters (and 'speed' generating keys each time).
12931 * Speed up EVP routines.
12934 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12935 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12936 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12937 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12939 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12940 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12941 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12944 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12946 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12950 * Added the OS2-EMX target.
12952 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12954 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12955 New functions to support `NCONF` routines in extension code.
12956 New function `CONF_set_nconf()`
12957 to allow functions which take an `NCONF` to also handle the old `LHASH`
12958 structure: this means that the old `CONF` compatible routines can be
12959 retained (in particular w.rt. extensions) without having to duplicate the
12960 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12964 * Enhance the general user interface with mechanisms for inner control
12965 and with possibilities to have yes/no kind of prompts.
12969 * Change all calls to low-level digest routines in the library and
12970 applications to use EVP. Add missing calls to HMAC_cleanup() and
12971 don't assume HMAC_CTX can be copied using memcpy().
12973 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12975 * Add the possibility to control engines through control names but with
12976 arbitrary arguments instead of just a string.
12977 Change the key loaders to take a UI_METHOD instead of a callback
12978 function pointer. NOTE: this breaks binary compatibility with earlier
12979 versions of OpenSSL [engine].
12980 Adapt the nCipher code for these new conditions and add a card insertion
12985 * Enhance the general user interface with mechanisms to better support
12986 dialog box interfaces, application-defined prompts, the possibility
12987 to use defaults (for example default passwords from somewhere else)
12988 and interrupts/cancellations.
12992 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12993 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12997 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12998 tidy up some unnecessarily weird code in 'sk_new()').
13000 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13002 * Change the key loading routines for ENGINEs to use the same kind
13003 callback (pem_password_cb) as all other routines that need this
13008 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13009 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13010 than this minimum value is recommended.
13014 * New random seeder for OpenVMS, using the system process statistics
13015 that are easily reachable.
13019 * Windows apparently can't transparently handle global
13020 variables defined in DLLs. Initialisations such as:
13022 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13024 won't compile. This is used by the any applications that need to
13025 declare their own ASN1 modules. This was fixed by adding the option
13026 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13027 needed for static libraries under Win32.
13031 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13032 setting of purpose and trust fields. New X509_STORE trust and
13033 purpose functions and tidy up setting in other SSL functions.
13037 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13038 structure. These are inherited by X509_STORE_CTX when it is
13039 initialised. This allows various defaults to be set in the
13040 X509_STORE structure (such as flags for CRL checking and custom
13041 purpose or trust settings) for functions which only use X509_STORE_CTX
13042 internally such as S/MIME.
13044 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13045 trust settings if they are not set in X509_STORE. This allows X509_STORE
13046 purposes and trust (in S/MIME for example) to override any set by default.
13048 Add command line options for CRL checking to smime, s_client and s_server
13053 * Initial CRL based revocation checking. If the CRL checking flag(s)
13054 are set then the CRL is looked up in the X509_STORE structure and
13055 its validity and signature checked, then if the certificate is found
13056 in the CRL the verify fails with a revoked error.
13058 Various new CRL related callbacks added to X509_STORE_CTX structure.
13060 Command line options added to 'verify' application to support this.
13062 This needs some additional work, such as being able to handle multiple
13063 CRLs with different times, extension based lookup (rather than just
13064 by subject name) and ultimately more complete V2 CRL extension
13069 * Add a general user interface API (crypto/ui/). This is designed
13070 to replace things like des_read_password and friends (backward
13071 compatibility functions using this new API are provided).
13072 The purpose is to remove prompting functions from the DES code
13073 section as well as provide for prompting through dialog boxes in
13074 a window system and the like.
13078 * Add "ex_data" support to ENGINE so implementations can add state at a
13079 per-structure level rather than having to store it globally.
13083 * Make it possible for ENGINE structures to be copied when retrieved by
13084 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13085 This causes the "original" ENGINE structure to act like a template,
13086 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13087 operational state can be localised to each ENGINE structure, despite the
13088 fact they all share the same "methods". New ENGINE structures returned in
13089 this case have no functional references and the return value is the single
13090 structural reference. This matches the single structural reference returned
13091 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13096 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13097 needs to match any other type at all we need to manually clear the
13102 * Changes to the "openssl engine" utility to include;
13103 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13104 about an ENGINE's available control commands.
13105 - executing control commands from command line arguments using the
13106 '-pre' and '-post' switches. '-post' is only used if '-t' is
13107 specified and the ENGINE is successfully initialised. The syntax for
13108 the individual commands are colon-separated, for example;
13109 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13113 * New dynamic control command support for ENGINEs. ENGINEs can now
13114 declare their own commands (numbers), names (strings), descriptions,
13115 and input types for run-time discovery by calling applications. A
13116 subset of these commands are implicitly classed as "executable"
13117 depending on their input type, and only these can be invoked through
13118 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13119 can be based on user input, config files, etc). The distinction is
13120 that "executable" commands cannot return anything other than a boolean
13121 result and can only support numeric or string input, whereas some
13122 discoverable commands may only be for direct use through
13123 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13124 pointers, or other custom uses. The "executable" commands are to
13125 support parameterisations of ENGINE behaviour that can be
13126 unambiguously defined by ENGINEs and used consistently across any
13127 OpenSSL-based application. Commands have been added to all the
13128 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13129 control over shared-library paths without source code alterations.
13133 * Changed all ENGINE implementations to dynamically allocate their
13134 ENGINEs rather than declaring them statically. Apart from this being
13135 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13136 this also allows the implementations to compile without using the
13137 internal engine_int.h header.
13141 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13142 'const' value. Any code that should be able to modify a RAND_METHOD
13143 should already have non-const pointers to it (ie. they should only
13144 modify their own ones).
13148 * Made a variety of little tweaks to the ENGINE code.
13149 - "atalla" and "ubsec" string definitions were moved from header files
13150 to C code. "nuron" string definitions were placed in variables
13151 rather than hard-coded - allowing parameterisation of these values
13152 later on via ctrl() commands.
13153 - Removed unused "#if 0"'d code.
13154 - Fixed engine list iteration code so it uses ENGINE_free() to release
13155 structural references.
13156 - Constified the RAND_METHOD element of ENGINE structures.
13157 - Constified various get/set functions as appropriate and added
13158 missing functions (including a catch-all ENGINE_cpy that duplicates
13159 all ENGINE values onto a new ENGINE except reference counts/state).
13160 - Removed NULL parameter checks in get/set functions. Setting a method
13161 or function to NULL is a way of cancelling out a previously set
13162 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13163 and doesn't justify the extra error symbols and code.
13164 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13165 flags from engine_int.h to engine.h.
13166 - Changed prototypes for ENGINE handler functions (init(), finish(),
13167 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13171 * Implement binary inversion algorithm for BN_mod_inverse in addition
13172 to the algorithm using long division. The binary algorithm can be
13173 used only if the modulus is odd. On 32-bit systems, it is faster
13174 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13175 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13176 up to 450 bits. In 64-bit environments, the binary algorithm
13177 appears to be advantageous for much longer moduli; here we use it
13178 for moduli up to 2048 bits.
13182 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13183 could not support the combine flag in choice fields.
13187 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13188 extensions from a certificate request to the certificate.
13192 * Allow multiple 'certopt' and 'nameopt' options to be separated
13193 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13194 file: this allows the display of the certificate about to be
13195 signed to be customised, to allow certain fields to be included
13196 or excluded and extension details. The old system didn't display
13197 multicharacter strings properly, omitted fields not in the policy
13198 and couldn't display additional details such as extensions.
13202 * Function EC_POINTs_mul for multiple scalar multiplication
13203 of an arbitrary number of elliptic curve points
13204 \sum scalars[i]*points[i],
13205 optionally including the generator defined for the EC_GROUP:
13206 scalar*generator + \sum scalars[i]*points[i].
13208 EC_POINT_mul is a simple wrapper function for the typical case
13209 that the point list has just one item (besides the optional
13214 * First EC_METHODs for curves over GF(p):
13216 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13217 operations and provides various method functions that can also
13218 operate with faster implementations of modular arithmetic.
13220 EC_GFp_mont_method() reuses most functions that are part of
13221 EC_GFp_simple_method, but uses Montgomery arithmetic.
13223 *Bodo Moeller; point addition and point doubling
13224 implementation directly derived from source code provided by
13225 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13227 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13228 crypto/ec/ec_lib.c):
13230 Curves are EC_GROUP objects (with an optional group generator)
13231 based on EC_METHODs that are built into the library.
13233 Points are EC_POINT objects based on EC_GROUP objects.
13235 Most of the framework would be able to handle curves over arbitrary
13236 finite fields, but as there are no obvious types for fields other
13237 than GF(p), some functions are limited to that for now.
13241 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13242 that the file contains a complete HTTP response.
13246 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13247 change the def and num file printf format specifier from "%-40sXXX"
13248 to "%-39s XXX". The latter will always guarantee a space after the
13249 field while the former will cause them to run together if the field
13250 is 40 of more characters long.
13254 * Constify the cipher and digest 'method' functions and structures
13255 and modify related functions to take constant EVP_MD and EVP_CIPHER
13260 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13261 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13265 * Modify `EVP_Digest*()` routines so they now return values. Although the
13266 internal software routines can never fail additional hardware versions
13271 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13273 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13274 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13277 ERR_R_NESTED_ASN1_ERROR
13279 ERR_R_MISSING_ASN1_EOS
13280 were 4 .. 9, conflicting with
13281 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13283 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13284 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13286 Add new error code 'ERR_R_INTERNAL_ERROR'.
13290 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13295 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13296 sets the subject name for a new request or supersedes the
13297 subject name in a given request. Formats that can be parsed are
13298 'CN=Some Name, OU=myOU, C=IT'
13300 'CN=Some Name/OU=myOU/C=IT'.
13302 Add options '-batch' and '-verbose' to 'openssl req'.
13304 *Massimiliano Pala <madwolf@hackmasters.net>*
13306 * Introduce the possibility to access global variables through
13307 functions on platform were that's the best way to handle exporting
13308 global variables in shared libraries. To enable this functionality,
13309 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13310 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13311 is normally done by Configure or something similar).
13313 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13314 in the source file (foo.c) like this:
13316 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13317 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13319 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13320 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13322 OPENSSL_DECLARE_GLOBAL(int,foo);
13323 #define foo OPENSSL_GLOBAL_REF(foo)
13324 OPENSSL_DECLARE_GLOBAL(double,bar);
13325 #define bar OPENSSL_GLOBAL_REF(bar)
13327 The #defines are very important, and therefore so is including the
13328 header file everywhere where the defined globals are used.
13330 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13331 of ASN.1 items, but that structure is a bit different.
13333 The largest change is in util/mkdef.pl which has been enhanced with
13334 better and easier to understand logic to choose which symbols should
13335 go into the Windows .def files as well as a number of fixes and code
13336 cleanup (among others, algorithm keywords are now sorted
13337 lexicographically to avoid constant rewrites).
13341 * In BN_div() keep a copy of the sign of 'num' before writing the
13342 result to 'rm' because if rm==num the value will be overwritten
13343 and produce the wrong result if 'num' is negative: this caused
13344 problems with BN_mod() and BN_nnmod().
13348 * Function OCSP_request_verify(). This checks the signature on an
13349 OCSP request and verifies the signer certificate. The signer
13350 certificate is just checked for a generic purpose and OCSP request
13355 * Add OCSP_check_validity() function to check the validity of OCSP
13356 responses. OCSP responses are prepared in real time and may only
13357 be a few seconds old. Simply checking that the current time lies
13358 between thisUpdate and nextUpdate max reject otherwise valid responses
13359 caused by either OCSP responder or client clock inaccuracy. Instead
13360 we allow thisUpdate and nextUpdate to fall within a certain period of
13361 the current time. The age of the response can also optionally be
13362 checked. Two new options -validity_period and -status_age added to
13367 * If signature or public key algorithm is unrecognized print out its
13368 OID rather that just UNKNOWN.
13372 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13373 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13374 ID to be generated from the issuer certificate alone which can then be
13375 passed to OCSP_id_issuer_cmp().
13379 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13380 ASN1 modules to export functions returning ASN1_ITEM pointers
13381 instead of the ASN1_ITEM structures themselves. This adds several
13382 new macros which allow the underlying ASN1 function/structure to
13383 be accessed transparently. As a result code should not use ASN1_ITEM
13384 references directly (such as &X509_it) but instead use the relevant
13385 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13386 use of the new ASN1 code on platforms where exporting structures
13387 is problematical (for example in shared libraries) but exporting
13388 functions returning pointers to structures is not.
13392 * Add support for overriding the generation of SSL/TLS session IDs.
13393 These callbacks can be registered either in an SSL_CTX or per SSL.
13394 The purpose of this is to allow applications to control, if they wish,
13395 the arbitrary values chosen for use as session IDs, particularly as it
13396 can be useful for session caching in multiple-server environments. A
13397 command-line switch for testing this (and any client code that wishes
13398 to use such a feature) has been added to "s_server".
13400 *Geoff Thorpe, Lutz Jaenicke*
13402 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13403 of the form `#if defined(...) || defined(...) || ...` and
13404 `#if !defined(...) && !defined(...) && ...`. This also avoids
13405 the growing number of special cases it was previously handling.
13409 * Make all configuration macros available for application by making
13410 sure they are available in opensslconf.h, by giving them names starting
13411 with `OPENSSL_` to avoid conflicts with other packages and by making
13412 sure e_os2.h will cover all platform-specific cases together with
13414 Additionally, it is now possible to define configuration/platform-
13415 specific names (called "system identities"). In the C code, these
13416 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13417 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13418 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13423 * New option -set_serial to 'req' and 'x509' this allows the serial
13424 number to use to be specified on the command line. Previously self
13425 signed certificates were hard coded with serial number 0 and the
13426 CA options of 'x509' had to use a serial number in a file which was
13431 * New options to 'ca' utility to support V2 CRL entry extensions.
13432 Currently CRL reason, invalidity date and hold instruction are
13433 supported. Add new CRL extensions to V3 code and some new objects.
13437 * New function EVP_CIPHER_CTX_set_padding() this is used to
13438 disable standard block padding (aka PKCS#5 padding) in the EVP
13439 API, which was previously mandatory. This means that the data is
13440 not padded in any way and so the total length much be a multiple
13441 of the block size, otherwise an error occurs.
13445 * Initial (incomplete) OCSP SSL support.
13449 * New function OCSP_parse_url(). This splits up a URL into its host,
13450 port and path components: primarily to parse OCSP URLs. New -url
13451 option to ocsp utility.
13455 * New nonce behavior. The return value of OCSP_check_nonce() now
13456 reflects the various checks performed. Applications can decide
13457 whether to tolerate certain situations such as an absent nonce
13458 in a response when one was present in a request: the ocsp application
13459 just prints out a warning. New function OCSP_add1_basic_nonce()
13460 this is to allow responders to include a nonce in a response even if
13461 the request is nonce-less.
13465 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13466 skipped when using openssl x509 multiple times on a single input file,
13467 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13471 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13472 set string type: to handle setting ASN1_TIME structures. Fix ca
13473 utility to correctly initialize revocation date of CRLs.
13477 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13478 the clients preferred ciphersuites and rather use its own preferences.
13479 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13480 Internet Explorer by ensuring unchanged hash method during stepup.
13481 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13485 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13486 to aes and add a new 'exist' option to print out symbols that don't
13491 * Additional options to ocsp utility to allow flags to be set and
13492 additional certificates supplied.
13496 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13497 OCSP client a number of certificate to only verify the response
13502 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13503 handle the new API. Currently only ECB, CBC modes supported. Add new
13506 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13507 Encryption Standard (AES) Ciphersuites for Transport Layer
13508 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13509 not enabled by default and were not part of the "ALL" ciphersuite
13510 alias because they were not yet official; they could be
13511 explicitly requested by specifying the "AESdraft" ciphersuite
13512 group alias. In the final release of OpenSSL 0.9.7, the group
13513 alias is called "AES" and is part of "ALL".)
13515 *Ben Laurie, Steve Henson, Bodo Moeller*
13517 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13518 request to response.
13522 * Functions for OCSP responders. OCSP_request_onereq_count(),
13523 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13524 extract information from a certificate request. OCSP_response_create()
13525 creates a response and optionally adds a basic response structure.
13526 OCSP_basic_add1_status() adds a complete single response to a basic
13527 response and returns the OCSP_SINGLERESP structure just added (to allow
13528 extensions to be included for example). OCSP_basic_add1_cert() adds a
13529 certificate to a basic response and OCSP_basic_sign() signs a basic
13530 response with various flags. New helper functions ASN1_TIME_check()
13531 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13532 (converts ASN1_TIME to GeneralizedTime).
13536 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13537 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13538 structure from a certificate. X509_pubkey_digest() digests the public_key
13539 contents: this is used in various key identifiers.
13543 * Make sk_sort() tolerate a NULL argument.
13545 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13547 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13548 passed by the function are trusted implicitly. If any of them signed the
13549 response then it is assumed to be valid and is not verified.
13553 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13554 to data. This was previously part of the PKCS7 ASN1 code. This
13555 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13556 *Steve Henson, reported by Kenneth R. Robinette
13557 <support@securenetterm.com>*
13559 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13560 routines: without these tracing memory leaks is very painful.
13561 Fix leaks in PKCS12 and PKCS7 routines.
13565 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13566 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13567 effectively meant GeneralizedTime would never be used. Now it
13568 is initialised to -1 but X509_time_adj() now has to check the value
13569 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13570 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13571 *Steve Henson, reported by Kenneth R. Robinette
13572 <support@securenetterm.com>*
13574 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13575 result in a zero length in the ASN1_INTEGER structure which was
13576 not consistent with the structure when d2i_ASN1_INTEGER() was used
13577 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13578 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13579 where it did not print out a minus for negative ASN1_INTEGER.
13583 * Add summary printout to ocsp utility. The various functions which
13584 convert status values to strings have been renamed to:
13585 OCSP_response_status_str(), OCSP_cert_status_str() and
13586 OCSP_crl_reason_str() and are no longer static. New options
13587 to verify nonce values and to disable verification. OCSP response
13588 printout format cleaned up.
13592 * Add additional OCSP certificate checks. These are those specified
13593 in RFC2560. This consists of two separate checks: the CA of the
13594 certificate being checked must either be the OCSP signer certificate
13595 or the issuer of the OCSP signer certificate. In the latter case the
13596 OCSP signer certificate must contain the OCSP signing extended key
13597 usage. This check is performed by attempting to match the OCSP
13598 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13599 in the OCSP_CERTID structures of the response.
13603 * Initial OCSP certificate verification added to OCSP_basic_verify()
13604 and related routines. This uses the standard OpenSSL certificate
13605 verify routines to perform initial checks (just CA validity) and
13606 to obtain the certificate chain. Then additional checks will be
13607 performed on the chain. Currently the root CA is checked to see
13608 if it is explicitly trusted for OCSP signing. This is used to set
13609 a root CA as a global signing root: that is any certificate that
13610 chains to that CA is an acceptable OCSP signing certificate.
13614 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13615 extensions from a separate configuration file.
13616 As when reading extensions from the main configuration file,
13617 the '-extensions ...' option may be used for specifying the
13620 *Massimiliano Pala <madwolf@comune.modena.it>*
13622 * New OCSP utility. Allows OCSP requests to be generated or
13623 read. The request can be sent to a responder and the output
13624 parsed, outputted or printed in text form. Not complete yet:
13625 still needs to check the OCSP response validity.
13629 * New subcommands for 'openssl ca':
13630 `openssl ca -status <serial>` prints the status of the cert with
13631 the given serial number (according to the index file).
13632 `openssl ca -updatedb` updates the expiry status of certificates
13635 *Massimiliano Pala <madwolf@comune.modena.it>*
13637 * New '-newreq-nodes' command option to CA.pl. This is like
13638 '-newreq', but calls 'openssl req' with the '-nodes' option
13639 so that the resulting key is not encrypted.
13641 *Damien Miller <djm@mindrot.org>*
13643 * New configuration for the GNU Hurd.
13645 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13647 * Initial code to implement OCSP basic response verify. This
13648 is currently incomplete. Currently just finds the signer's
13649 certificate and verifies the signature on the response.
13653 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13654 value of OPENSSLDIR. This is available via the new '-d' option
13655 to 'openssl version', and is also included in 'openssl version -a'.
13659 * Allowing defining memory allocation callbacks that will be given
13660 file name and line number information in additional arguments
13661 (a `const char*` and an int). The basic functionality remains, as
13662 well as the original possibility to just replace malloc(),
13663 realloc() and free() by functions that do not know about these
13664 additional arguments. To register and find out the current
13665 settings for extended allocation functions, the following
13666 functions are provided:
13668 CRYPTO_set_mem_ex_functions
13669 CRYPTO_set_locked_mem_ex_functions
13670 CRYPTO_get_mem_ex_functions
13671 CRYPTO_get_locked_mem_ex_functions
13673 These work the same way as CRYPTO_set_mem_functions and friends.
13674 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13675 extended allocation function is enabled.
13676 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13677 a conventional allocation function is enabled.
13679 *Richard Levitte, Bodo Moeller*
13681 * Finish off removing the remaining LHASH function pointer casts.
13682 There should no longer be any prototype-casting required when using
13683 the LHASH abstraction, and any casts that remain are "bugs". See
13684 the callback types and macros at the head of lhash.h for details
13685 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13689 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13690 If /dev/[u]random devices are not available or do not return enough
13691 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13693 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13694 /etc/entropy will be queried once each in this sequence, querying stops
13695 when enough entropy was collected without querying more sockets.
13699 * Change the Unix RAND_poll() variant to be able to poll several
13700 random devices, as specified by DEVRANDOM, until a sufficient amount
13701 of data has been collected. We spend at most 10 ms on each file
13702 (select timeout) and read in non-blocking mode. DEVRANDOM now
13703 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13704 (previously it was just the string "/dev/urandom"), so on typical
13705 platforms the 10 ms delay will never occur.
13706 Also separate out the Unix variant to its own file, rand_unix.c.
13707 For VMS, there's a currently-empty rand_vms.c.
13711 * Move OCSP client related routines to ocsp_cl.c. These
13712 provide utility functions which an application needing
13713 to issue a request to an OCSP responder and analyse the
13714 response will typically need: as opposed to those which an
13715 OCSP responder itself would need which will be added later.
13717 OCSP_request_sign() signs an OCSP request with an API similar
13718 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13719 response. OCSP_response_get1_basic() extracts basic response
13720 from response. OCSP_resp_find_status(): finds and extracts status
13721 information from an OCSP_CERTID structure (which will be created
13722 when the request structure is built). These are built from lower
13723 level functions which work on OCSP_SINGLERESP structures but
13724 won't normally be used unless the application wishes to examine
13725 extensions in the OCSP response for example.
13727 Replace nonce routines with a pair of functions.
13728 OCSP_request_add1_nonce() adds a nonce value and optionally
13729 generates a random value. OCSP_check_nonce() checks the
13730 validity of the nonce in an OCSP response.
13734 * Change function OCSP_request_add() to OCSP_request_add0_id().
13735 This doesn't copy the supplied OCSP_CERTID and avoids the
13736 need to free up the newly created id. Change return type
13737 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13738 This can then be used to add extensions to the request.
13739 Deleted OCSP_request_new(), since most of its functionality
13740 is now in OCSP_REQUEST_new() (and the case insensitive name
13741 clash) apart from the ability to set the request name which
13742 will be added elsewhere.
13746 * Update OCSP API. Remove obsolete extensions argument from
13747 various functions. Extensions are now handled using the new
13748 OCSP extension code. New simple OCSP HTTP function which
13749 can be used to send requests and parse the response.
13753 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13754 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13755 uses the special reorder version of SET OF to sort the attributes
13756 and reorder them to match the encoded order. This resolves a long
13757 standing problem: a verify on a PKCS7 structure just after signing
13758 it used to fail because the attribute order did not match the
13759 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13760 it uses the received order. This is necessary to tolerate some broken
13761 software that does not order SET OF. This is handled by encoding
13762 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13763 to produce the required SET OF.
13767 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13768 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13769 files to get correct declarations of the ASN.1 item variables.
13773 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13774 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13775 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13776 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13777 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13778 ASN1_ITEM and no wrapper functions.
13782 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13783 replace the old function pointer based I/O routines. Change most of
13784 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13788 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13789 lines, recognize more "algorithms" that can be deselected, and make
13790 it complain about algorithm deselection that isn't recognised.
13794 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13795 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13796 to use new functions. Add NO_ASN1_OLD which can be set to remove
13797 some old style ASN1 functions: this can be used to determine if old
13798 code will still work when these eventually go away.
13802 * New extension functions for OCSP structures, these follow the
13803 same conventions as certificates and CRLs.
13807 * New function X509V3_add1_i2d(). This automatically encodes and
13808 adds an extension. Its behaviour can be customised with various
13809 flags to append, replace or delete. Various wrappers added for
13810 certificates and CRLs.
13814 * Fix to avoid calling the underlying ASN1 print routine when
13815 an extension cannot be parsed. Correct a typo in the
13816 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13820 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13821 entries for variables.
13825 * Add functionality to `apps/openssl.c` for detecting locking
13826 problems: As the program is single-threaded, all we have
13827 to do is register a locking callback using an array for
13828 storing which locks are currently held by the program.
13832 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13833 SSL_get_ex_data_X509_STORE_idx(), which is used in
13834 ssl_verify_cert_chain() and thus can be called at any time
13835 during TLS/SSL handshakes so that thread-safety is essential.
13836 Unfortunately, the ex_data design is not at all suited
13837 for multi-threaded use, so it probably should be abolished.
13841 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13843 *Broadcom, tweaked and integrated by Geoff Thorpe*
13845 * Move common extension printing code to new function
13846 X509V3_print_extensions(). Reorganise OCSP print routines and
13847 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13851 * New function X509_signature_print() to remove duplication in some
13856 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13857 set (this was treated exactly the same as SET OF previously). This
13858 is used to reorder the STACK representing the structure to match the
13859 encoding. This will be used to get round a problem where a PKCS7
13860 structure which was signed could not be verified because the STACK
13861 order did not reflect the encoded order.
13865 * Reimplement the OCSP ASN1 module using the new code.
13869 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13870 for its ASN1 operations. The old style function pointers still exist
13871 for now but they will eventually go away.
13875 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13876 completely replaces the old ASN1 functionality with a table driven
13877 encoder and decoder which interprets an ASN1_ITEM structure describing
13878 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13879 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13880 has also been converted to the new form.
13884 * Change BN_mod_exp_recp so that negative moduli are tolerated
13885 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13886 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13887 for negative moduli.
13891 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13892 of not touching the result's sign bit.
13896 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13901 * Changed the LHASH code to use prototypes for callbacks, and created
13902 macros to declare and implement thin (optionally static) functions
13903 that provide type-safety and avoid function pointer casting for the
13904 type-specific callbacks.
13908 * Added Kerberos Cipher Suites to be used with TLS, as written in
13910 *Veers Staats <staatsvr@asc.hpc.mil>,
13911 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13913 * Reformat the FAQ so the different questions and answers can be divided
13914 in sections depending on the subject.
13918 * Have the zlib compression code load ZLIB.DLL dynamically under
13923 * New function BN_mod_sqrt for computing square roots modulo a prime
13924 (using the probabilistic Tonelli-Shanks algorithm unless
13925 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13926 be handled deterministically).
13928 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13930 * Make BN_mod_inverse faster by explicitly handling small quotients
13931 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13932 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13936 * New function BN_kronecker.
13940 * Fix BN_gcd so that it works on negative inputs; the result is
13941 positive unless both parameters are zero.
13942 Previously something reasonably close to an infinite loop was
13943 possible because numbers could be growing instead of shrinking
13944 in the implementation of Euclid's algorithm.
13948 * Fix BN_is_word() and BN_is_one() macros to take into account the
13949 sign of the number in question.
13951 Fix BN_is_word(a,w) to work correctly for w == 0.
13953 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13954 because its test if the absolute value of 'a' equals 'w'.
13955 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13956 it exists mostly for use in the implementations of BN_is_zero(),
13957 BN_is_one(), and BN_is_word().
13961 * New function BN_swap.
13965 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13966 the exponentiation functions are more likely to produce reasonable
13967 results on negative inputs.
13971 * Change BN_mod_mul so that the result is always non-negative.
13972 Previously, it could be negative if one of the factors was negative;
13973 I don't think anyone really wanted that behaviour.
13977 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13978 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13979 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13980 and add new functions:
13989 BN_mod_lshift1_quick
13991 BN_mod_lshift_quick
13993 These functions always generate non-negative results.
13995 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13996 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13998 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13999 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14000 be reduced modulo `m`.
14002 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14005 The following entry accidentally appeared in the CHANGES file
14006 distributed with OpenSSL 0.9.7. The modifications described in
14007 it do *not* apply to OpenSSL 0.9.7.
14009 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14010 was actually never needed) and in BN_mul(). The removal in BN_mul()
14011 required a small change in bn_mul_part_recursive() and the addition
14012 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14013 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14014 bn_sub_words() and bn_add_words() except they take arrays with
14020 * In 'openssl passwd', verify passwords read from the terminal
14021 unless the '-salt' option is used (which usually means that
14022 verification would just waste user's time since the resulting
14023 hash is going to be compared with some given password hash)
14024 or the new '-noverify' option is used.
14026 This is an incompatible change, but it does not affect
14027 non-interactive use of 'openssl passwd' (passwords on the command
14028 line, '-stdin' option, '-in ...' option) and thus should not
14029 cause any problems.
14033 * Remove all references to RSAref, since there's no more need for it.
14037 * Make DSO load along a path given through an environment variable
14038 (SHLIB_PATH) with shl_load().
14042 * Constify the ENGINE code as a result of BIGNUM constification.
14043 Also constify the RSA code and most things related to it. In a
14044 few places, most notable in the depth of the ASN.1 code, ugly
14045 casts back to non-const were required (to be solved at a later
14050 * Make it so the openssl application has all engines loaded by default.
14054 * Constify the BIGNUM routines a little more.
14058 * Add the following functions:
14060 ENGINE_load_cswift()
14062 ENGINE_load_atalla()
14063 ENGINE_load_nuron()
14064 ENGINE_load_builtin_engines()
14066 That way, an application can itself choose if external engines that
14067 are built-in in OpenSSL shall ever be used or not. The benefit is
14068 that applications won't have to be linked with libdl or other dso
14069 libraries unless it's really needed.
14071 Changed 'openssl engine' to load all engines on demand.
14072 Changed the engine header files to avoid the duplication of some
14073 declarations (they differed!).
14077 * 'openssl engine' can now list capabilities.
14081 * Better error reporting in 'openssl engine'.
14085 * Never call load_dh_param(NULL) in s_server.
14089 * Add engine application. It can currently list engines by name and
14090 identity, and test if they are actually available.
14094 * Improve RPM specification file by forcing symbolic linking and making
14095 sure the installed documentation is also owned by root.root.
14097 *Damien Miller <djm@mindrot.org>*
14099 * Give the OpenSSL applications more possibilities to make use of
14100 keys (public as well as private) handled by engines.
14104 * Add OCSP code that comes from CertCo.
14108 * Add VMS support for the Rijndael code.
14112 * Added untested support for Nuron crypto accelerator.
14116 * Add support for external cryptographic devices. This code was
14117 previously distributed separately as the "engine" branch.
14119 *Geoff Thorpe, Richard Levitte*
14121 * Rework the filename-translation in the DSO code. It is now possible to
14122 have far greater control over how a "name" is turned into a filename
14123 depending on the operating environment and any oddities about the
14124 different shared library filenames on each system.
14128 * Support threads on FreeBSD-elf in Configure.
14132 * Fix for SHA1 assembly problem with MASM: it produces
14133 warnings about corrupt line number information when assembling
14134 with debugging information. This is caused by the overlapping
14137 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14140 NCONF_get_number() has no error checking at all. As a replacement,
14141 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14142 promoted strongly. The old NCONF_get_number is kept around for
14143 binary backward compatibility.
14144 Make it possible for methods to load from something other than a BIO,
14145 by providing a function pointer that is given a name instead of a BIO.
14146 For example, this could be used to load configuration data from an
14151 * Fix for non blocking accept BIOs. Added new I/O special reason
14152 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14153 with non blocking I/O was not possible because no retry code was
14154 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14159 * Added the beginnings of Rijndael support.
14163 * Fix for bug in DirectoryString mask setting. Add support for
14164 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14165 to allow certificate printing to more controllable, additional
14166 'certopt' option to 'x509' to allow new printing options to be
14171 * Clean old EAY MD5 hack from e_os.h.
14175 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14177 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14178 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14180 *Joe Orton, Steve Henson*
14182 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14184 * Fix additional bug revealed by the NISCC test suite:
14186 Stop bug triggering large recursion when presented with
14187 certain ASN.1 tags ([CVE-2003-0851])
14191 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14193 * Fix various bugs revealed by running the NISCC test suite:
14195 Stop out of bounds reads in the ASN1 code when presented with
14196 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14198 If verify callback ignores invalid public key errors don't try to check
14199 certificate signature with the NULL public key.
14203 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14204 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14209 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14210 extra data after the compression methods not only for TLS 1.0
14211 but also for SSL 3.0 (as required by the specification).
14213 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14215 * Change X509_certificate_type() to mark the key as exported/exportable
14216 when it's 512 *bits* long, not 512 bytes.
14220 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14222 * Countermeasure against the Klima-Pokorny-Rosa extension of
14223 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14224 a protocol version number mismatch like a decryption error
14225 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14229 * Turn on RSA blinding by default in the default implementation
14230 to avoid a timing attack. Applications that don't want it can call
14231 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14232 They would be ill-advised to do so in most cases.
14234 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14236 * Change RSA blinding code so that it works when the PRNG is not
14237 seeded (in this case, the secret RSA exponent is abused as
14238 an unpredictable seed -- if it is not unpredictable, there
14239 is no point in blinding anyway). Make RSA blinding thread-safe
14240 by remembering the creator's thread ID in rsa->blinding and
14241 having all other threads use local one-time blinding factors
14242 (this requires more computation than sharing rsa->blinding, but
14243 avoids excessive locking; and if an RSA object is not shared
14244 between threads, blinding will still be very fast).
14248 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14250 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14251 via timing by performing a MAC computation even if incorrect
14252 block cipher padding has been found. This is a countermeasure
14253 against active attacks where the attacker has to distinguish
14254 between bad padding and a MAC verification error. ([CVE-2003-0078])
14256 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14257 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14258 Martin Vuagnoux (EPFL, Ilion)*
14260 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14262 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14263 memory from its contents. This is done with a counter that will
14264 place alternating values in each byte. This can be used to solve
14265 two issues: 1) the removal of calls to memset() by highly optimizing
14266 compilers, and 2) cleansing with other values than 0, since those can
14267 be read through on certain media, for example a swap space on disk.
14271 * Bugfix: client side session caching did not work with external caching,
14272 because the session->cipher setting was not restored when reloading
14273 from the external cache. This problem was masked, when
14274 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14275 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14279 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14280 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14282 *Zeev Lieber <zeev-l@yahoo.com>*
14284 * Undo an undocumented change introduced in 0.9.6e which caused
14285 repeated calls to OpenSSL_add_all_ciphers() and
14286 OpenSSL_add_all_digests() to be ignored, even after calling
14291 * Change the default configuration reader to deal with last line not
14292 being properly terminated.
14296 * Change X509_NAME_cmp() so it applies the special rules on handling
14297 DN values that are of type PrintableString, as well as RDNs of type
14298 emailAddress where the value has the type ia5String.
14300 *stefank@valicert.com via Richard Levitte*
14302 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14303 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14304 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14305 the bitwise-OR of the two for use by the majority of applications
14306 wanting this behaviour, and update the docs. The documented
14307 behaviour and actual behaviour were inconsistent and had been
14308 changing anyway, so this is more a bug-fix than a behavioural
14311 *Geoff Thorpe, diagnosed by Nadav Har'El*
14313 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14314 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14318 * Fix initialization code race conditions in
14319 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14320 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14321 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14322 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14323 ssl2_get_cipher_by_char(),
14324 ssl3_get_cipher_by_char().
14326 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14328 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14329 the cached sessions are flushed, as the remove_cb() might use ex_data
14330 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14331 (see [openssl.org #212]).
14333 *Geoff Thorpe, Lutz Jaenicke*
14335 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14336 length, instead of the encoding length to d2i_ASN1_OBJECT.
14340 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14342 * [In 0.9.6g-engine release:]
14343 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14345 *Lynn Gazis <lgazis@rainbow.com>*
14347 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14349 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14350 and get fix the header length calculation.
14351 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14352 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14354 * Use proper error handling instead of 'assertions' in buffer
14355 overflow checks added in 0.9.6e. This prevents DoS (the
14356 assertions could call abort()).
14358 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14360 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14362 * Add various sanity checks to asn1_get_length() to reject
14363 the ASN1 length bytes if they exceed sizeof(long), will appear
14364 negative or the content length exceeds the length of the
14367 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14369 * Fix cipher selection routines: ciphers without encryption had no flags
14370 for the cipher strength set and where therefore not handled correctly
14371 by the selection routines (PR #130).
14375 * Fix EVP_dsa_sha macro.
14380 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14381 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14382 that was added in OpenSSL 0.9.6d.
14384 As the countermeasure turned out to be incompatible with some
14385 broken SSL implementations, the new option is part of SSL_OP_ALL.
14386 SSL_OP_ALL is usually employed when compatibility with weird SSL
14387 implementations is desired (e.g. '-bugs' option to 's_client' and
14388 's_server'), so the new option is automatically set in many
14393 * Changes in security patch:
14395 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14396 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14397 Air Force Materiel Command, USAF, under agreement number
14400 * Add various sanity checks to asn1_get_length() to reject
14401 the ASN1 length bytes if they exceed sizeof(long), will appear
14402 negative or the content length exceeds the length of the
14403 supplied buffer. ([CVE-2002-0659])
14405 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14407 * Assertions for various potential buffer overflows, not known to
14408 happen in practice.
14410 *Ben Laurie (CHATS)*
14412 * Various temporary buffers to hold ASCII versions of integers were
14413 too small for 64 bit platforms. ([CVE-2002-0655])
14414 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14416 * Remote buffer overflow in SSL3 protocol - an attacker could
14417 supply an oversized session ID to a client. ([CVE-2002-0656])
14419 *Ben Laurie (CHATS)*
14421 * Remote buffer overflow in SSL2 protocol - an attacker could
14422 supply an oversized client master key. ([CVE-2002-0656])
14424 *Ben Laurie (CHATS)*
14426 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14428 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14429 encoded as NULL) with id-dsa-with-sha1.
14431 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14433 * Check various `X509_...()` return values in `apps/req.c`.
14435 *Nils Larsch <nla@trustcenter.de>*
14437 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14438 an end-of-file condition would erroneously be flagged, when the CRLF
14439 was just at the end of a processed block. The bug was discovered when
14440 processing data through a buffering memory BIO handing the data to a
14441 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14442 <ptsekov@syntrex.com> and Nedelcho Stanev.
14446 * Implement a countermeasure against a vulnerability recently found
14447 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14448 before application data chunks to avoid the use of known IVs
14449 with data potentially chosen by the attacker.
14453 * Fix length checks in ssl3_get_client_hello().
14457 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14458 to prevent ssl3_read_internal() from incorrectly assuming that
14459 ssl3_read_bytes() found application data while handshake
14460 processing was enabled when in fact s->s3->in_read_app_data was
14461 merely automatically cleared during the initial handshake.
14463 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14465 * Fix object definitions for Private and Enterprise: they were not
14466 recognized in their shortname (=lowercase) representation. Extend
14467 obj_dat.pl to issue an error when using undefined keywords instead
14468 of silently ignoring the problem (Svenning Sorensen
14469 <sss@sss.dnsalias.net>).
14473 * Fix DH_generate_parameters() so that it works for 'non-standard'
14474 generators, i.e. generators other than 2 and 5. (Previously, the
14475 code did not properly initialise the 'add' and 'rem' values to
14476 BN_generate_prime().)
14478 In the new general case, we do not insist that 'generator' is
14479 actually a primitive root: This requirement is rather pointless;
14480 a generator of the order-q subgroup is just as good, if not
14485 * Map new X509 verification errors to alerts. Discovered and submitted by
14486 Tom Wu <tom@arcot.com>.
14490 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14491 returning non-zero before the data has been completely received
14492 when using non-blocking I/O.
14494 *Bodo Moeller; problem pointed out by John Hughes*
14496 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14498 *Ben Laurie, Lutz Jaenicke*
14500 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14501 Yoram Zahavi <YoramZ@gilian.com>).
14505 * Add information about CygWin 1.3 and on, and preserve proper
14506 configuration for the versions before that.
14508 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14510 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14511 check whether we deal with a copy of a session and do not delete from
14512 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14513 <izhar@checkpoint.com>.
14517 * Do not store session data into the internal session cache, if it
14518 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14519 flag is set). Proposed by Aslam <aslam@funk.com>.
14523 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14528 * [In 0.9.6d-engine release:]
14529 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14531 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14533 * Add the configuration target linux-s390x.
14535 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14537 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14538 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14539 variable as an indication that a ClientHello message has been
14540 received. As the flag value will be lost between multiple
14541 invocations of ssl3_accept when using non-blocking I/O, the
14542 function may not be aware that a handshake has actually taken
14543 place, thus preventing a new session from being added to the
14546 To avoid this problem, we now set s->new_session to 2 instead of
14547 using a local variable.
14549 *Lutz Jaenicke, Bodo Moeller*
14551 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14552 if the SSL_R_LENGTH_MISMATCH error is detected.
14554 *Geoff Thorpe, Bodo Moeller*
14556 * New 'shared_ldflag' column in Configure platform table.
14560 * Fix EVP_CIPHER_mode macro.
14562 *"Dan S. Camper" <dan@bti.net>*
14564 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14565 type, we must throw them away by setting rr->length to 0.
14567 *D P Chang <dpc@qualys.com>*
14569 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14571 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14572 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14573 worked incorrectly for those cases where range = `10..._2` and
14574 `3*range` is two bits longer than range.)
14578 * Only add signing time to PKCS7 structures if it is not already
14583 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14584 OBJ_ld_ce should be OBJ_id_ce.
14585 Also some ip-pda OIDs in crypto/objects/objects.txt were
14586 incorrect (cf. RFC 3039).
14588 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14590 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14591 returns early because it has nothing to do.
14593 *Andy Schneider <andy.schneider@bjss.co.uk>*
14595 * [In 0.9.6c-engine release:]
14596 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14598 *Andy Schneider <andy.schneider@bjss.co.uk>*
14600 * [In 0.9.6c-engine release:]
14601 Add support for Cryptographic Appliance's keyserver technology.
14602 (Use engine 'keyclient')
14604 *Cryptographic Appliances and Geoff Thorpe*
14606 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14607 is called via tools/c89.sh because arguments have to be
14608 rearranged (all '-L' options must appear before the first object
14611 *Richard Shapiro <rshapiro@abinitio.com>*
14613 * [In 0.9.6c-engine release:]
14614 Add support for Broadcom crypto accelerator cards, backported
14617 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14619 * [In 0.9.6c-engine release:]
14620 Add support for SureWare crypto accelerator cards from
14621 Baltimore Technologies. (Use engine 'sureware')
14623 *Baltimore Technologies and Mark Cox*
14625 * [In 0.9.6c-engine release:]
14626 Add support for crypto accelerator cards from Accelerated
14627 Encryption Processing, www.aep.ie. (Use engine 'aep')
14629 *AEP Inc. and Mark Cox*
14631 * Add a configuration entry for gcc on UnixWare.
14633 *Gary Benson <gbenson@redhat.com>*
14635 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14636 messages are stored in a single piece (fixed-length part and
14637 variable-length part combined) and fix various bugs found on the way.
14641 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14642 instead. BIO_gethostbyname() does not know what timeouts are
14643 appropriate, so entries would stay in cache even when they have
14645 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14647 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14648 faced with a pathologically small ClientHello fragment that does
14649 not contain client_version: Instead of aborting with an error,
14650 simply choose the highest available protocol version (i.e.,
14651 TLS 1.0 unless it is disabled). In practice, ClientHello
14652 messages are never sent like this, but this change gives us
14653 strictly correct behaviour at least for TLS.
14657 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14658 never resets s->method to s->ctx->method when called from within
14659 one of the SSL handshake functions.
14661 *Bodo Moeller; problem pointed out by Niko Baric*
14663 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14664 (sent using the client's version number) if client_version is
14665 smaller than the protocol version in use. Also change
14666 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14667 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14668 the client will at least see that alert.
14672 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14677 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14678 client receives HelloRequest while in a handshake.
14680 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14682 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14683 should end in 'break', not 'goto end' which circumvents various
14684 cleanups done in state SSL_ST_OK. But session related stuff
14685 must be disabled for SSL_ST_OK in the case that we just sent a
14688 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14689 before just sending a HelloRequest.
14691 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14693 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14694 reveal whether illegal block cipher padding was found or a MAC
14695 verification error occurred. (Neither SSLerr() codes nor alerts
14696 are directly visible to potential attackers, but the information
14697 may leak via logfiles.)
14699 Similar changes are not required for the SSL 2.0 implementation
14700 because the number of padding bytes is sent in clear for SSL 2.0,
14701 and the extra bytes are just ignored. However ssl/s2_pkt.c
14702 failed to verify that the purported number of padding bytes is in
14707 * Add OpenUNIX-8 support including shared libraries
14708 (Boyd Lynn Gerber <gerberb@zenez.com>).
14712 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14713 'wristwatch attack' using huge encoding parameters (cf.
14714 James H. Manger's CRYPTO 2001 paper). Note that the
14715 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14716 encoding parameters and hence was not vulnerable.
14720 * BN_sqr() bug fix.
14722 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14724 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14725 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14726 followed by modular reduction.
14728 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14730 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14731 equivalent based on BN_pseudo_rand() instead of BN_rand().
14735 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14736 This function was broken, as the check for a new client hello message
14737 to handle SGC did not allow these large messages.
14738 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14742 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14746 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14747 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14751 * Rework the configuration and shared library support for Tru64 Unix.
14752 The configuration part makes use of modern compiler features and
14753 still retains old compiler behavior for those that run older versions
14754 of the OS. The shared library support part includes a variant that
14755 uses the RPATH feature, and is available through the special
14756 configuration target "alpha-cc-rpath", which will never be selected
14759 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14761 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14762 with the same message size as in ssl3_get_certificate_request().
14763 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14764 messages might inadvertently be reject as too long.
14766 *Petr Lampa <lampa@fee.vutbr.cz>*
14768 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14772 * Modified SSL library such that the verify_callback that has been set
14773 specifically for an SSL object with SSL_set_verify() is actually being
14774 used. Before the change, a verify_callback set with this function was
14775 ignored and the verify_callback() set in the SSL_CTX at the time of
14776 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14777 to allow the necessary settings.
14781 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14782 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14783 done automatically (in contradiction to the requirements of the C
14784 standard). This made problems when used from OpenSSH.
14788 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14789 dh->length and always used
14791 BN_rand_range(priv_key, dh->p).
14793 BN_rand_range() is not necessary for Diffie-Hellman, and this
14794 specific range makes Diffie-Hellman unnecessarily inefficient if
14795 dh->length (recommended exponent length) is much smaller than the
14796 length of dh->p. We could use BN_rand_range() if the order of
14797 the subgroup was stored in the DH structure, but we only have
14802 BN_rand(priv_key, l, ...)
14804 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14811 RSA_eay_public_encrypt
14812 RSA_eay_private_decrypt
14813 RSA_eay_private_encrypt (signing)
14814 RSA_eay_public_decrypt (signature verification)
14816 (default implementations for RSA_public_encrypt,
14817 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14818 always reject numbers >= n.
14822 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14823 to synchronize access to 'locking_thread'. This is necessary on
14824 systems where access to 'locking_thread' (an 'unsigned long'
14825 variable) is not atomic.
14829 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14830 *before* setting the 'crypto_lock_rand' flag. The previous code had
14831 a race condition if 0 is a valid thread ID.
14833 *Travis Vitek <vitek@roguewave.com>*
14835 * Add support for shared libraries under Irix.
14837 *Albert Chin-A-Young <china@thewrittenword.com>*
14839 * Add configuration option to build on Linux on both big-endian and
14840 little-endian MIPS.
14842 *Ralf Baechle <ralf@uni-koblenz.de>*
14844 * Add the possibility to create shared libraries on HP-UX.
14848 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14850 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14851 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14852 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14853 PRNG state recovery was possible based on the output of
14854 one PRNG request appropriately sized to gain knowledge on
14855 'md' followed by enough consecutive 1-byte PRNG requests
14856 to traverse all of 'state'.
14858 1. When updating 'md_local' (the current thread's copy of 'md')
14859 during PRNG output generation, hash all of the previous
14860 'md_local' value, not just the half used for PRNG output.
14862 2. Make the number of bytes from 'state' included into the hash
14863 independent from the number of PRNG bytes requested.
14865 The first measure alone would be sufficient to avoid
14866 Markku-Juhani's attack. (Actually it had never occurred
14867 to me that the half of 'md_local' used for chaining was the
14868 half from which PRNG output bytes were taken -- I had always
14869 assumed that the secret half would be used.) The second
14870 measure makes sure that additional data from 'state' is never
14871 mixed into 'md_local' in small portions; this heuristically
14872 further strengthens the PRNG.
14876 * Fix crypto/bn/asm/mips3.s.
14880 * When only the key is given to "enc", the IV is undefined. Print out
14881 an error message in this case.
14885 * Handle special case when X509_NAME is empty in X509 printing routines.
14889 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14890 positive and less than q.
14894 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14895 used: it isn't thread safe and the add_lock_callback should handle
14898 *Paul Rose <Paul.Rose@bridge.com>*
14900 * Verify that incoming data obeys the block size in
14901 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14907 *Ulf Möller, Bodo Möller*
14909 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14910 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14911 when fixing the server behaviour for backwards-compatible 'client
14912 hello' messages. (Note that the attack is impractical against
14913 SSL 3.0 and TLS 1.0 anyway because length and version checking
14914 means that the probability of guessing a valid ciphertext is
14915 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14918 Before 0.9.5, the countermeasure (hide the error by generating a
14919 random 'decryption result') did not work properly because
14920 ERR_clear_error() was missing, meaning that SSL_get_error() would
14921 detect the supposedly ignored error.
14923 Both problems are now fixed.
14927 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14928 (previously it was 1024).
14932 * Fix for compatibility mode trust settings: ignore trust settings
14933 unless some valid trust or reject settings are present.
14937 * Fix for blowfish EVP: its a variable length cipher.
14941 * Fix various bugs related to DSA S/MIME verification. Handle missing
14942 parameters in DSA public key structures and return an error in the
14943 DSA routines if parameters are absent.
14947 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14948 in the current directory if neither $RANDFILE nor $HOME was set.
14949 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14950 caused some confusion to Windows users who haven't defined $HOME.
14951 Thus RAND_file_name() is changed again: e_os.h can define a
14952 DEFAULT_HOME, which will be used if $HOME is not set.
14953 For Windows, we use "C:"; on other platforms, we still require
14954 environment variables.
14956 * Move 'if (!initialized) RAND_poll()' into regions protected by
14957 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14958 having multiple threads call RAND_poll() concurrently.
14962 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14963 combination of a flag and a thread ID variable.
14964 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14965 flag), *other* threads can enter ssleay_add_bytes without obeying
14966 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14967 that they do not hold after the first thread unsets add_do_not_lock).
14971 * Change bctest again: '-x' expressions are not available in all
14972 versions of 'test'.
14976 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14978 * Fix a couple of memory leaks in PKCS7_dataDecode()
14980 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14982 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14983 the default extension for executables, if any. Also, make the perl
14984 scripts that use symlink() to test if it really exists and use "cp"
14985 if it doesn't. All this made OpenSSL compilable and installable in
14990 * Fix for asn1_GetSequence() for indefinite length constructed data.
14991 If SEQUENCE is length is indefinite just set c->slen to the total
14992 amount of data available.
14994 *Steve Henson, reported by shige@FreeBSD.org*
14996 *This change does not apply to 0.9.7.*
14998 * Change bctest to avoid here-documents inside command substitution
14999 (workaround for FreeBSD /bin/sh bug).
15000 For compatibility with Ultrix, avoid shell functions (introduced
15001 in the bctest version that searches along $PATH).
15005 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15006 with des_encrypt() defined on some operating systems, like Solaris
15011 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15012 On the Importance of Eliminating Errors in Cryptographic
15013 Computations, J. Cryptology 14 (2001) 2, 101-119,
15014 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15018 * MIPS assembler BIGNUM division bug fix.
15022 * Disabled incorrect Alpha assembler code.
15026 * Fix PKCS#7 decode routines so they correctly update the length
15027 after reading an EOC for the EXPLICIT tag.
15031 *This change does not apply to 0.9.7.*
15033 * Fix bug in PKCS#12 key generation routines. This was triggered
15034 if a 3DES key was generated with a 0 initial byte. Include
15035 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15036 (but broken) behaviour.
15040 * Enhance bctest to search for a working bc along $PATH and print
15043 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15045 * Fix memory leaks in err.c: free err_data string if necessary;
15046 don't write to the wrong index in ERR_set_error_data.
15050 * Implement ssl23_peek (analogous to ssl23_read), which previously
15055 * Replace rdtsc with `_emit` statements for VC++ version 5.
15057 *Jeremy Cooper <jeremy@baymoo.org>*
15059 * Make it possible to reuse SSLv2 sessions.
15063 * In copy_email() check for >= 0 as a return value for
15064 X509_NAME_get_index_by_NID() since 0 is a valid index.
15066 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15068 * Avoid coredump with unsupported or invalid public keys by checking if
15069 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15070 PKCS7_verify() fails with non detached data.
15074 * Don't use getenv in library functions when run as setuid/setgid.
15075 New function OPENSSL_issetugid().
15079 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15080 due to incorrect handling of multi-threading:
15082 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15084 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15086 3. Count how many times MemCheck_off() has been called so that
15087 nested use can be treated correctly. This also avoids
15088 inband-signalling in the previous code (which relied on the
15089 assumption that thread ID 0 is impossible).
15093 * Add "-rand" option also to s_client and s_server.
15097 * Fix CPU detection on Irix 6.x.
15098 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15099 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15101 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15106 *This change does not apply to 0.9.7.*
15108 * Use the cached encoding of an X509_NAME structure rather than
15109 copying it. This is apparently the reason for the libsafe "errors"
15110 but the code is actually correct.
15114 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15115 Bleichenbacher's DSA attack.
15116 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15117 to be set and top=0 forces the highest bit to be set; top=-1 is new
15118 and leaves the highest bit random.
15120 *Ulf Moeller, Bodo Moeller*
15122 * In the `NCONF_...`-based implementations for `CONF_...` queries
15123 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15124 a temporary CONF structure with the data component set to NULL
15125 (which gives segmentation faults in lh_retrieve).
15126 Instead, use NULL for the CONF pointer in CONF_get_string and
15127 CONF_get_number (which may use environment variables) and directly
15128 return NULL from CONF_get_section.
15132 * Fix potential buffer overrun for EBCDIC.
15136 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15137 keyUsage if basicConstraints absent for a CA.
15141 * Make SMIME_write_PKCS7() write mail header values with a format that
15142 is more generally accepted (no spaces before the semicolon), since
15143 some programs can't parse those values properly otherwise. Also make
15144 sure BIO's that break lines after each write do not create invalid
15149 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15150 macros previously used would not encode an empty SEQUENCE OF
15151 and break the signature.
15155 *This change does not apply to 0.9.7.*
15157 * Zero the premaster secret after deriving the master secret in
15162 * Add some EVP_add_digest_alias registrations (as found in
15163 OpenSSL_add_all_digests()) to SSL_library_init()
15164 aka OpenSSL_add_ssl_algorithms(). This provides improved
15165 compatibility with peers using X.509 certificates
15166 with unconventional AlgorithmIdentifier OIDs.
15170 * Fix for Irix with NO_ASM.
15172 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15174 * ./config script fixes.
15176 *Ulf Moeller, Richard Levitte*
15178 * Fix 'openssl passwd -1'.
15182 * Change PKCS12_key_gen_asc() so it can cope with non null
15183 terminated strings whose length is passed in the passlen
15184 parameter, for example from PEM callbacks. This was done
15185 by adding an extra length parameter to asc2uni().
15187 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15189 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15190 call failed, free the DSA structure.
15194 * Fix to uni2asc() to cope with zero length Unicode strings.
15195 These are present in some PKCS#12 files.
15199 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15200 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15201 when writing a 32767 byte record.
15203 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15205 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15206 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15208 (RSA objects have a reference count access to which is protected
15209 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15210 so they are meant to be shared between threads.)
15211 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15212 "Reddie, Steven" <Steven.Reddie@ca.com>*
15214 * Fix a deadlock in CRYPTO_mem_leaks().
15218 * Use better test patterns in bntest.
15222 * rand_win.c fix for Borland C.
15226 * BN_rshift bugfix for n == 0.
15230 * Add a 'bctest' script that checks for some known 'bc' bugs
15231 so that 'make test' does not abort just because 'bc' is broken.
15235 * Store verify_result within SSL_SESSION also for client side to
15236 avoid potential security hole. (Re-used sessions on the client side
15237 always resulted in verify_result==X509_V_OK, not using the original
15238 result of the server certificate verification.)
15242 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15243 SSL3_RT_APPLICATION_DATA, return 0.
15244 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15249 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15250 releases, have been re-implemented by renaming the previous
15251 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15252 and ssl3_read_internal, respectively, and adding 'peek' parameters
15253 to them. The new ssl[23]_{read,peek} functions are calls to
15254 ssl[23]_read_internal with the 'peek' flag set appropriately.
15255 A 'peek' parameter has also been added to ssl3_read_bytes, which
15256 does the actual work for ssl3_read_internal.
15260 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15261 the method-specific "init()" handler. Also clean up ex_data after
15262 calling the method-specific "finish()" handler. Previously, this was
15263 happening the other way round.
15267 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15268 The previous value, 12, was not always sufficient for BN_mod_exp().
15272 * Make sure that shared libraries get the internal name engine with
15273 the full version number and not just 0. This should mark the
15274 shared libraries as not backward compatible. Of course, this should
15275 be changed again when we can guarantee backward binary compatibility.
15279 * Fix typo in get_cert_by_subject() in by_dir.c
15281 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15283 * Rework the system to generate shared libraries:
15285 - Make note of the expected extension for the shared libraries and
15286 if there is a need for symbolic links from for example libcrypto.so.0
15287 to libcrypto.so.0.9.7. There is extended info in Configure for
15290 - Make as few rebuilds of the shared libraries as possible.
15292 - Still avoid linking the OpenSSL programs with the shared libraries.
15294 - When installing, install the shared libraries separately from the
15299 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15301 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15302 and not in SSL_clear because the latter is also used by the
15303 accept/connect functions; previously, the settings made by
15304 SSL_set_read_ahead would be lost during the handshake.
15306 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15308 * Correct util/mkdef.pl to be selective about disabled algorithms.
15309 Previously, it would create entries for disabled algorithms no
15314 * Added several new manual pages for SSL_* function.
15318 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15320 * In ssl23_get_client_hello, generate an error message when faced
15321 with an initial SSL 3.0/TLS record that is too small to contain the
15322 first two bytes of the ClientHello message, i.e. client_version.
15323 (Note that this is a pathologic case that probably has never happened
15324 in real life.) The previous approach was to use the version number
15325 from the record header as a substitute; but our protocol choice
15326 should not depend on that one because it is not authenticated
15327 by the Finished messages.
15331 * More robust randomness gathering functions for Windows.
15333 *Jeffrey Altman <jaltman@columbia.edu>*
15335 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15336 not set then we don't setup the error code for issuer check errors
15337 to avoid possibly overwriting other errors which the callback does
15338 handle. If an application does set the flag then we assume it knows
15339 what it is doing and can handle the new informational codes
15344 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15345 a general "ANY" type, as such it should be able to decode anything
15346 including tagged types. However it didn't check the class so it would
15347 wrongly interpret tagged types in the same way as their universal
15348 counterpart and unknown types were just rejected. Changed so that the
15349 tagged and unknown types are handled in the same way as a SEQUENCE:
15350 that is the encoding is stored intact. There is also a new type
15351 "V_ASN1_OTHER" which is used when the class is not universal, in this
15352 case we have no idea what the actual type is so we just lump them all
15357 * On VMS, stdout may very well lead to a file that is written to
15358 in a record-oriented fashion. That means that every write() will
15359 write a separate record, which will be read separately by the
15360 programs trying to read from it. This can be very confusing.
15362 The solution is to put a BIO filter in the way that will buffer
15363 text until a linefeed is reached, and then write everything a
15364 line at a time, so every record written will be an actual line,
15365 not chunks of lines and not (usually doesn't happen, but I've
15366 seen it once) several lines in one record. BIO_f_linebuffer() is
15369 Currently, it's a VMS-only method, because that's where it has
15370 been tested well enough.
15374 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15375 it can return incorrect results.
15376 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15377 but it was in 0.9.6-beta[12].)
15381 * Disable the check for content being present when verifying detached
15382 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15383 include zero length content when signing messages.
15387 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15388 BIO_ctrl (for BIO pairs).
15392 * Add DSO method for VMS.
15396 * Bug fix: Montgomery multiplication could produce results with the
15401 * Add RPM specification openssl.spec and modify it to build three
15402 packages. The default package contains applications, application
15403 documentation and run-time libraries. The devel package contains
15404 include files, static libraries and function documentation. The
15405 doc package contains the contents of the doc directory. The original
15406 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15410 * Add a large number of documentation files for many SSL routines.
15412 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15414 * Add a configuration entry for Sony News 4.
15416 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15418 * Don't set the two most significant bits to one when generating a
15419 random number < q in the DSA library.
15423 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15424 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15425 the underlying transport is blocking) if a handshake took place.
15426 (The default behaviour is needed by applications such as s_client
15427 and s_server that use select() to determine when to use SSL_read;
15428 but for applications that know in advance when to expect data, it
15429 just makes things more complicated.)
15433 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15438 * Add a few more EBCDIC conditionals that make `req` and `x509`
15439 work better on such systems.
15441 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15443 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15444 Update PKCS12_parse() so it copies the friendlyName and the
15445 keyid to the certificates aux info.
15449 * Fix bug in PKCS7_verify() which caused an infinite loop
15450 if there was more than one signature.
15452 *Sven Uszpelkat <su@celocom.de>*
15454 * Major change in util/mkdef.pl to include extra information
15455 about each symbol, as well as presenting variables as well
15456 as functions. This change means that there's n more need
15457 to rebuild the .num files when some algorithms are excluded.
15461 * Allow the verify time to be set by an application,
15462 rather than always using the current time.
15466 * Phase 2 verify code reorganisation. The certificate
15467 verify code now looks up an issuer certificate by a
15468 number of criteria: subject name, authority key id
15469 and key usage. It also verifies self signed certificates
15470 by the same criteria. The main comparison function is
15471 X509_check_issued() which performs these checks.
15473 Lot of changes were necessary in order to support this
15474 without completely rewriting the lookup code.
15476 Authority and subject key identifier are now cached.
15478 The LHASH 'certs' is X509_STORE has now been replaced
15479 by a STACK_OF(X509_OBJECT). This is mainly because an
15480 LHASH can't store or retrieve multiple objects with
15481 the same hash value.
15483 As a result various functions (which were all internal
15484 use only) have changed to handle the new X509_STORE
15485 structure. This will break anything that messed round
15486 with X509_STORE internally.
15488 The functions X509_STORE_add_cert() now checks for an
15489 exact match, rather than just subject name.
15491 The X509_STORE API doesn't directly support the retrieval
15492 of multiple certificates matching a given criteria, however
15493 this can be worked round by performing a lookup first
15494 (which will fill the cache with candidate certificates)
15495 and then examining the cache for matches. This is probably
15496 the best we can do without throwing out X509_LOOKUP
15497 entirely (maybe later...).
15499 The X509_VERIFY_CTX structure has been enhanced considerably.
15501 All certificate lookup operations now go via a get_issuer()
15502 callback. Although this currently uses an X509_STORE it
15503 can be replaced by custom lookups. This is a simple way
15504 to bypass the X509_STORE hackery necessary to make this
15505 work and makes it possible to use more efficient techniques
15506 in future. A very simple version which uses a simple
15507 STACK for its trusted certificate store is also provided
15508 using X509_STORE_CTX_trusted_stack().
15510 The verify_cb() and verify() callbacks now have equivalents
15511 in the X509_STORE_CTX structure.
15513 X509_STORE_CTX also has a 'flags' field which can be used
15514 to customise the verify behaviour.
15518 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15519 excludes S/MIME capabilities.
15523 * When a certificate request is read in keep a copy of the
15524 original encoding of the signed data and use it when outputting
15525 again. Signatures then use the original encoding rather than
15526 a decoded, encoded version which may cause problems if the
15527 request is improperly encoded.
15531 * For consistency with other BIO_puts implementations, call
15532 buffer_write(b, ...) directly in buffer_puts instead of calling
15535 In BIO_puts, increment b->num_write as in BIO_write.
15537 *Peter.Sylvester@EdelWeb.fr*
15539 * Fix BN_mul_word for the case where the word is 0. (We have to use
15540 BN_zero, we may not return a BIGNUM with an array consisting of
15541 words set to zero.)
15545 * Avoid calling abort() from within the library when problems are
15546 detected, except if preprocessor symbols have been defined
15547 (such as REF_CHECK, BN_DEBUG etc.).
15551 * New openssl application 'rsautl'. This utility can be
15552 used for low-level RSA operations. DER public key
15553 BIO/fp routines also added.
15557 * New Configure entry and patches for compiling on QNX 4.
15559 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15561 * A demo state-machine implementation was sponsored by
15562 Nuron (<http://www.nuron.com/>) and is now available in
15563 demos/state_machine.
15567 * New options added to the 'dgst' utility for signature
15568 generation and verification.
15572 * Unrecognized PKCS#7 content types are now handled via a
15573 catch all ASN1_TYPE structure. This allows unsupported
15574 types to be stored as a "blob" and an application can
15575 encode and decode it manually.
15579 * Fix various signed/unsigned issues to make a_strex.c
15580 compile under VC++.
15582 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15584 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15585 length if passed a buffer. ASN1_INTEGER_to_BN failed
15586 if passed a NULL BN and its argument was negative.
15588 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15590 * Modification to PKCS#7 encoding routines to output definite
15591 length encoding. Since currently the whole structures are in
15592 memory there's not real point in using indefinite length
15593 constructed encoding. However if OpenSSL is compiled with
15594 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15598 * Added BIO_vprintf() and BIO_vsnprintf().
15602 * Added more prefixes to parse for in the strings written
15603 through a logging bio, to cover all the levels that are available
15604 through syslog. The prefixes are now:
15606 PANIC, EMERG, EMR => LOG_EMERG
15607 ALERT, ALR => LOG_ALERT
15608 CRIT, CRI => LOG_CRIT
15609 ERROR, ERR => LOG_ERR
15610 WARNING, WARN, WAR => LOG_WARNING
15611 NOTICE, NOTE, NOT => LOG_NOTICE
15612 INFO, INF => LOG_INFO
15613 DEBUG, DBG => LOG_DEBUG
15615 and as before, if none of those prefixes are present at the
15616 beginning of the string, LOG_ERR is chosen.
15618 On Win32, the `LOG_*` levels are mapped according to this:
15620 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15621 LOG_WARNING => EVENTLOG_WARNING_TYPE
15622 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15626 * Made it possible to reconfigure with just the configuration
15627 argument "reconf" or "reconfigure". The command line arguments
15628 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15629 and are retrieved from there when reconfiguring.
15635 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15637 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15641 * The obj_dat.pl script was messing up the sorting of object
15642 names. The reason was that it compared the quoted version
15643 of strings as a result "OCSP" > "OCSP Signing" because
15644 " > SPACE. Changed script to store unquoted versions of
15645 names and add quotes on output. It was also omitting some
15646 names from the lookup table if they were given a default
15647 value (that is if SN is missing it is given the same
15648 value as LN and vice versa), these are now added on the
15649 grounds that if an object has a name we should be able to
15650 look it up. Finally added warning output when duplicate
15651 short or long names are found.
15655 * Changes needed for Tandem NSK.
15657 *Scott Uroff <scott@xypro.com>*
15659 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15660 RSA_padding_check_SSLv23(), special padding was never detected
15661 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15662 version rollback attacks was not effective.
15664 In s23_clnt.c, don't use special rollback-attack detection padding
15665 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15666 client; similarly, in s23_srvr.c, don't do the rollback check if
15667 SSL 2.0 is the only protocol enabled in the server.
15671 * Make it possible to get hexdumps of unprintable data with 'openssl
15672 asn1parse'. By implication, the functions ASN1_parse_dump() and
15673 BIO_dump_indent() are added.
15677 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15678 these print out strings and name structures based on various
15679 flags including RFC2253 support and proper handling of
15680 multibyte characters. Added options to the 'x509' utility
15681 to allow the various flags to be set.
15685 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15686 Also change the functions X509_cmp_current_time() and
15687 X509_gmtime_adj() work with an ASN1_TIME structure,
15688 this will enable certificates using GeneralizedTime in validity
15689 dates to be checked.
15693 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15694 negative public key encodings) on by default,
15695 NO_NEG_PUBKEY_BUG can be set to disable it.
15699 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15700 content octets. An i2c_ASN1_OBJECT is unnecessary because
15701 the encoding can be trivially obtained from the structure.
15705 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15706 not read locks (`CRYPTO_r_[un]lock`).
15710 * A first attempt at creating official support for shared
15711 libraries through configuration. I've kept it so the
15712 default is static libraries only, and the OpenSSL programs
15713 are always statically linked for now, but there are
15714 preparations for dynamic linking in place.
15715 This has been tested on Linux and Tru64.
15719 * Randomness polling function for Win9x, as described in:
15720 Peter Gutmann, Software Generation of Practically Strong
15725 * Fix so PRNG is seeded in req if using an already existing
15730 * New options to smime application. -inform and -outform
15731 allow alternative formats for the S/MIME message including
15732 PEM and DER. The -content option allows the content to be
15733 specified separately. This should allow things like Netscape
15734 form signing output easier to verify.
15738 * Fix the ASN1 encoding of tags using the 'long form'.
15742 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15743 STRING types. These convert content octets to and from the
15744 underlying type. The actual tag and length octets are
15745 already assumed to have been read in and checked. These
15746 are needed because all other string types have virtually
15747 identical handling apart from the tag. By having versions
15748 of the ASN1 functions that just operate on content octets
15749 IMPLICIT tagging can be handled properly. It also allows
15750 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15751 and ASN1_INTEGER are identical apart from the tag.
15755 * Change the handling of OID objects as follows:
15757 - New object identifiers are inserted in objects.txt, following
15758 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15759 - objects.pl is used to process obj_mac.num and create a new
15761 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15764 This is currently kind of a hack, and the perl code in objects.pl
15765 isn't very elegant, but it works as I intended. The simplest way
15766 to check that it worked correctly is to look in obj_dat.h and
15767 check the array nid_objs and make sure the objects haven't moved
15768 around (this is important!). Additions are OK, as well as
15769 consistent name changes.
15773 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15777 * Addition of the command line parameter '-rand file' to 'openssl req'.
15778 The given file adds to whatever has already been seeded into the
15779 random pool through the RANDFILE configuration file option or
15780 environment variable, or the default random state file.
15784 * mkstack.pl now sorts each macro group into lexical order.
15785 Previously the output order depended on the order the files
15786 appeared in the directory, resulting in needless rewriting
15791 * Patches to make OpenSSL compile under Win32 again. Mostly
15792 work arounds for the VC++ problem that it treats func() as
15793 func(void). Also stripped out the parts of mkdef.pl that
15794 added extra typesafe functions: these no longer exist.
15798 * Reorganisation of the stack code. The macros are now all
15799 collected in safestack.h . Each macro is defined in terms of
15800 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15801 DEBUG_SAFESTACK is now handled in terms of function casts,
15802 this has the advantage of retaining type safety without the
15803 use of additional functions. If DEBUG_SAFESTACK is not defined
15804 then the non typesafe macros are used instead. Also modified the
15805 mkstack.pl script to handle the new form. Needs testing to see
15806 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15807 the default if no major problems. Similar behaviour for ASN1_SET_OF
15808 and PKCS12_STACK_OF.
15812 * When some versions of IIS use the 'NET' form of private key the
15813 key derivation algorithm is different. Normally MD5(password) is
15814 used as a 128 bit RC4 key. In the modified case
15815 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15816 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15817 as the old Netscape_RSA functions except they have an additional
15818 'sgckey' parameter which uses the modified algorithm. Also added
15819 an -sgckey command line option to the rsa utility. Thanks to
15820 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15821 algorithm to openssl-dev.
15825 * The evp_local.h macros were using 'c.##kname' which resulted in
15826 invalid expansion on some systems (SCO 5.0.5 for example).
15827 Corrected to 'c.kname'.
15829 *Phillip Porch <root@theporch.com>*
15831 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15832 a STACK of email addresses from a certificate or request, these look
15833 in the subject name and the subject alternative name extensions and
15834 omit any duplicate addresses.
15838 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15839 This makes DSA verification about 2 % faster.
15843 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15844 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15845 plus overhead for 1024 bit moduli).
15846 This makes exponentiations about 0.5 % faster for 1024 bit
15847 exponents (as measured by "openssl speed rsa2048").
15851 * Rename memory handling macros to avoid conflicts with other
15853 Malloc => OPENSSL_malloc
15854 Malloc_locked => OPENSSL_malloc_locked
15855 Realloc => OPENSSL_realloc
15856 Free => OPENSSL_free
15860 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15861 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15865 * CygWin32 support.
15867 *John Jarvie <jjarvie@newsguy.com>*
15869 * The type-safe stack code has been rejigged. It is now only compiled
15870 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15871 by default all type-specific stack functions are "#define"d back to
15872 standard stack functions. This results in more streamlined output
15873 but retains the type-safety checking possibilities of the original
15878 * The STACK code has been cleaned up, and certain type declarations
15879 that didn't make a lot of sense have been brought in line. This has
15880 also involved a cleanup of sorts in safestack.h to more correctly
15881 map type-safe stack functions onto their plain stack counterparts.
15882 This work has also resulted in a variety of "const"ifications of
15883 lots of the code, especially `_cmp` operations which should normally
15884 be prototyped with "const" parameters anyway.
15888 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15889 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15890 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15891 where all of 'md' is used each time the PRNG is used, but 'state'
15892 is used only indexed by a cyclic counter. As entropy may not be
15893 well distributed from the beginning, 'md' is important as a
15894 chaining variable. However, the output function chains only half
15895 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15896 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15897 in all of 'state' being rewritten, with the new values depending
15898 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15902 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15903 the handshake is continued after ssl_verify_cert_chain();
15904 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15905 can lead to 'unexplainable' connection aborts later.
15907 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15909 * Major EVP API cipher revision.
15910 Add hooks for extra EVP features. This allows various cipher
15911 parameters to be set in the EVP interface. Support added for variable
15912 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15913 setting of RC2 and RC5 parameters.
15915 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15918 Remove lots of duplicated code from the EVP library. For example *every*
15919 cipher init() function handles the 'iv' in the same way according to the
15920 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15921 for CFB and OFB modes they zero ctx->num.
15923 New functionality allows removal of S/MIME code RC2 hack.
15925 Most of the routines have the same form and so can be declared in terms
15928 By shifting this to the top level EVP_CipherInit() it can be removed from
15929 all individual ciphers. If the cipher wants to handle IVs or keys
15930 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15933 Change lots of functions like EVP_EncryptUpdate() to now return a
15934 value: although software versions of the algorithms cannot fail
15935 any installed hardware versions can.
15939 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15940 this option is set, tolerate broken clients that send the negotiated
15941 protocol version number instead of the requested protocol version
15946 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15947 i.e. non-zero for export ciphersuites, zero otherwise.
15948 Previous versions had this flag inverted, inconsistent with
15949 rsa_tmp_cb (..._TMP_RSA_CB).
15951 *Bodo Moeller; problem reported by Amit Chopra*
15953 * Add missing DSA library text string. Work around for some IIS
15954 key files with invalid SEQUENCE encoding.
15958 * Add a document (doc/standards.txt) that list all kinds of standards
15959 and so on that are implemented in OpenSSL.
15963 * Enhance c_rehash script. Old version would mishandle certificates
15964 with the same subject name hash and wouldn't handle CRLs at all.
15965 Added -fingerprint option to crl utility, to support new c_rehash
15970 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15974 * Fix for SSL server purpose checking. Server checking was
15975 rejecting certificates which had extended key usage present
15976 but no ssl client purpose.
15978 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15980 * Make PKCS#12 code work with no password. The PKCS#12 spec
15981 is a little unclear about how a blank password is handled.
15982 Since the password in encoded as a BMPString with terminating
15983 double NULL a zero length password would end up as just the
15984 double NULL. However no password at all is different and is
15985 handled differently in the PKCS#12 key generation code. NS
15986 treats a blank password as zero length. MSIE treats it as no
15987 password on export: but it will try both on import. We now do
15988 the same: PKCS12_parse() tries zero length and no password if
15989 the password is set to "" or NULL (NULL is now a valid password:
15990 it wasn't before) as does the pkcs12 application.
15994 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15995 perror when PEM_read_bio_X509_REQ fails, the error message must
15996 be obtained from the error queue.
16000 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16001 it in ERR_remove_state if appropriate, and change ERR_get_state
16002 accordingly to avoid race conditions (this is necessary because
16003 thread_hash is no longer constant once set).
16007 * Bugfix for linux-elf makefile.one.
16011 * RSA_get_default_method() will now cause a default
16012 RSA_METHOD to be chosen if one doesn't exist already.
16013 Previously this was only set during a call to RSA_new()
16014 or RSA_new_method(NULL) meaning it was possible for
16015 RSA_get_default_method() to return NULL.
16019 * Added native name translation to the existing DSO code
16020 that will convert (if the flag to do so is set) filenames
16021 that are sufficiently small and have no path information
16022 into a canonical native form. Eg. "blah" converted to
16023 "libblah.so" or "blah.dll" etc.
16027 * New function ERR_error_string_n(e, buf, len) which is like
16028 ERR_error_string(e, buf), but writes at most 'len' bytes
16029 including the 0 terminator. For ERR_error_string_n, 'buf'
16032 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16034 * CONF library reworked to become more general. A new CONF
16035 configuration file reader "class" is implemented as well as a
16036 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16037 old `CONF_*` functions are still there, but are reimplemented to
16038 work in terms of the new functions. Also, a set of functions
16039 to handle the internal storage of the configuration data is
16040 provided to make it easier to write new configuration file
16041 reader "classes" (I can definitely see something reading a
16042 configuration file in XML format, for example), called `_CONF_*`,
16043 or "the configuration storage API"...
16045 The new configuration file reading functions are:
16047 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16048 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16050 NCONF_default, NCONF_WIN32
16052 NCONF_dump_fp, NCONF_dump_bio
16054 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16055 NCONF_new creates a new CONF object. This works in the same way
16056 as other interfaces in OpenSSL, like the BIO interface.
16057 `NCONF_dump_*` dump the internal storage of the configuration file,
16058 which is useful for debugging. All other functions take the same
16059 arguments as the old `CONF_*` functions with the exception of the
16060 first that must be a `CONF *` instead of a `LHASH *`.
16062 To make it easier to use the new classes with the old `CONF_*` functions,
16063 the function CONF_set_default_method is provided.
16067 * Add '-tls1' option to 'openssl ciphers', which was already
16068 mentioned in the documentation but had not been implemented.
16069 (This option is not yet really useful because even the additional
16070 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16074 * Initial DSO code added into libcrypto for letting OpenSSL (and
16075 OpenSSL-based applications) load shared libraries and bind to
16076 them in a portable way.
16078 *Geoff Thorpe, with contributions from Richard Levitte*
16080 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16082 * Make sure _lrotl and _lrotr are only used with MSVC.
16084 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16085 (the default implementation of RAND_status).
16087 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16088 to '-clrext' (= clear extensions), as intended and documented.
16089 *Bodo Moeller; inconsistency pointed out by Michael Attili
16090 <attili@amaxo.com>*
16092 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16093 was larger than the MD block size.
16095 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16097 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16098 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16099 using the passed key: if the passed key was a private key the result
16100 of X509_print(), for example, would be to print out all the private key
16105 * des_quad_cksum() byte order bug fix.
16106 *Ulf Möller, using the problem description in krb4-0.9.7, where
16107 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16109 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16112 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16114 * For easily testing in shell scripts whether some command
16115 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16116 returns with exit code 0 iff no command of the given name is available.
16117 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16118 the output goes to stdout and nothing is printed to stderr.
16119 Additional arguments are always ignored.
16121 Since for each cipher there is a command of the same name,
16122 the 'no-cipher' compilation switches can be tested this way.
16124 ('openssl no-XXX' is not able to detect pseudo-commands such
16125 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16129 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16133 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16134 is set; it will be thrown away anyway because each handshake creates
16136 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16137 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16138 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16139 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16143 * New s_client option -ign_eof: EOF at stdin is ignored, and
16144 'Q' and 'R' lose their special meanings (quit/renegotiate).
16145 This is part of what -quiet does; unlike -quiet, -ign_eof
16146 does not suppress any output.
16150 * Add compatibility options to the purpose and trust code. The
16151 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16152 accepts a certificate or CA, this was the previous behaviour,
16153 with all the associated security issues.
16155 X509_TRUST_COMPAT is the old trust behaviour: only and
16156 automatically trust self signed roots in certificate store. A
16157 new trust setting X509_TRUST_DEFAULT is used to specify that
16158 a purpose has no associated trust setting and it should instead
16159 use the value in the default purpose.
16163 * Fix the PKCS#8 DSA private key code so it decodes keys again
16164 and fix a memory leak.
16168 * In util/mkerr.pl (which implements 'make errors'), preserve
16169 reason strings from the previous version of the .c file, as
16170 the default to have only downcase letters (and digits) in
16171 automatically generated reasons codes is not always appropriate.
16175 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16176 using strerror. Previously, ERR_reason_error_string() returned
16177 library names as reason strings for SYSerr; but SYSerr is a special
16178 case where small numbers are errno values, not library numbers.
16182 * Add '-dsaparam' option to 'openssl dhparam' application. This
16183 converts DSA parameters into DH parameters. (When creating parameters,
16184 DSA_generate_parameters is used.)
16188 * Include 'length' (recommended exponent length) in C code generated
16189 by 'openssl dhparam -C'.
16193 * The second argument to set_label in perlasm was already being used
16194 so couldn't be used as a "file scope" flag. Moved to third argument
16199 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16200 instead of RAND_bytes for encryption IVs and salts.
16204 * Include RAND_status() into RAND_METHOD instead of implementing
16205 it only for md_rand.c Otherwise replacing the PRNG by calling
16206 RAND_set_rand_method would be impossible.
16210 * Don't let DSA_generate_key() enter an infinite loop if the random
16211 number generation fails.
16215 * New 'rand' application for creating pseudo-random output.
16219 * Added configuration support for Linux/IA64
16221 *Rolf Haberrecker <rolf@suse.de>*
16223 * Assembler module support for Mingw32.
16227 * Shared library support for HPUX (in shlib/).
16229 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16231 * Shared library support for Solaris gcc.
16233 *Lutz Behnke <behnke@trustcenter.de>*
16235 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16237 * PKCS7_encrypt() was adding text MIME headers twice because they
16238 were added manually and by SMIME_crlf_copy().
16242 * In bntest.c don't call BN_rand with zero bits argument.
16244 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16246 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16247 case was implemented. This caused BN_div_recp() to fail occasionally.
16251 * Add an optional second argument to the set_label() in the perl
16252 assembly language builder. If this argument exists and is set
16253 to 1 it signals that the assembler should use a symbol whose
16254 scope is the entire file, not just the current function. This
16255 is needed with MASM which uses the format label:: for this scope.
16257 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16259 * Change the ASN1 types so they are typedefs by default. Before
16260 almost all types were #define'd to ASN1_STRING which was causing
16261 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16266 * Change names of new functions to the new get1/get0 naming
16267 convention: After 'get1', the caller owns a reference count
16268 and has to call `..._free`; 'get0' returns a pointer to some
16269 data structure without incrementing reference counters.
16270 (Some of the existing 'get' functions increment a reference
16271 counter, some don't.)
16272 Similarly, 'set1' and 'add1' functions increase reference
16273 counters or duplicate objects.
16277 * Allow for the possibility of temp RSA key generation failure:
16278 the code used to assume it always worked and crashed on failure.
16282 * Fix potential buffer overrun problem in BIO_printf().
16283 *Ulf Möller, using public domain code by Patrick Powell; problem
16284 pointed out by David Sacerdote <das33@cornell.edu>*
16286 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16287 RAND_egd() and RAND_status(). In the command line application,
16288 the EGD socket can be specified like a seed file using RANDFILE
16293 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16294 Some CAs (e.g. Verisign) distribute certificates in this form.
16298 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16299 list to exclude them. This means that no special compilation option
16300 is needed to use anonymous DH: it just needs to be included in the
16305 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16306 EVP_MD_type. The old functionality is available in a new macro called
16307 EVP_MD_md(). Change code that uses it and update docs.
16311 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16312 where the `void *` argument is replaced by a function pointer argument.
16313 Previously `void *` was abused to point to functions, which works on
16314 many platforms, but is not correct. As these functions are usually
16315 called by macros defined in OpenSSL header files, most source code
16316 should work without changes.
16320 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16321 sections with information on -D... compiler switches used for
16322 compiling the library so that applications can see them. To enable
16323 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16324 must be defined. E.g.,
16325 #define OPENSSL_ALGORITHM_DEFINES
16326 #include <openssl/opensslconf.h>
16327 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16329 *Richard Levitte, Ulf and Bodo Möller*
16331 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16336 * Change the 'other' type in certificate aux info to a STACK_OF
16337 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16338 the required ASN1 format: arbitrary types determined by an OID.
16342 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16343 argument to 'req'. This is not because the function is newer or
16344 better than others it just uses the work 'NEW' in the certificate
16345 request header lines. Some software needs this.
16349 * Reorganise password command line arguments: now passwords can be
16350 obtained from various sources. Delete the PEM_cb function and make
16351 it the default behaviour: i.e. if the callback is NULL and the
16352 usrdata argument is not NULL interpret it as a null terminated pass
16353 phrase. If usrdata and the callback are NULL then the pass phrase
16354 is prompted for as usual.
16358 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16359 the support is automatically enabled. The resulting binaries will
16360 autodetect the card and use it if present.
16362 *Ben Laurie and Compaq Inc.*
16364 * Work around for Netscape hang bug. This sends certificate request
16365 and server done in one record. Since this is perfectly legal in the
16366 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16367 the bugs/SSLv3 entry for more info.
16371 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16375 * Add -rand argument to smime and pkcs12 applications and read/write
16380 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16384 * Add command line password options to the remaining applications.
16388 * Bug fix for BN_div_recp() for numerators with an even number of
16393 * More tests in bntest.c, and changed test_bn output.
16397 * ./config recognizes MacOS X now.
16401 * Bug fix for BN_div() when the first words of num and divisor are
16402 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16406 * Add support for various broken PKCS#8 formats, and command line
16407 options to produce them.
16411 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16412 get temporary BIGNUMs from a BN_CTX.
16416 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16421 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16422 include a #define from the old name to the new. The original intent
16423 was that statically linked binaries could for example just call
16424 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16425 link with digests. This never worked because SSLeay_add_all_digests()
16426 and SSLeay_add_all_ciphers() were in the same source file so calling
16427 one would link with the other. They are now in separate source files.
16431 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16435 * Use a less unusual form of the Miller-Rabin primality test (it used
16436 a binary algorithm for exponentiation integrated into the Miller-Rabin
16437 loop, our standard modexp algorithms are faster).
16441 * Support for the EBCDIC character set completed.
16443 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16445 * Source code cleanups: use const where appropriate, eliminate casts,
16446 use `void *` instead of `char *` in lhash.
16450 * Bugfix: ssl3_send_server_key_exchange was not restartable
16451 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16452 this the server could overwrite ephemeral keys that the client
16457 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16458 using 50 iterations of the Rabin-Miller test.
16460 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16461 iterations of the Rabin-Miller test as required by the appendix
16462 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16463 As BN_is_prime_fasttest includes trial division, DSA parameter
16464 generation becomes much faster.
16466 This implies a change for the callback functions in DSA_is_prime
16467 and DSA_generate_parameters: The callback function is called once
16468 for each positive witness in the Rabin-Miller test, not just
16469 occasionally in the inner loop; and the parameters to the
16470 callback function now provide an iteration count for the outer
16471 loop rather than for the current invocation of the inner loop.
16472 DSA_generate_parameters additionally can call the callback
16473 function with an 'iteration count' of -1, meaning that a
16474 candidate has passed the trial division test (when q is generated
16475 from an application-provided seed, trial division is skipped).
16479 * New function BN_is_prime_fasttest that optionally does trial
16480 division before starting the Rabin-Miller test and has
16481 an additional BN_CTX * argument (whereas BN_is_prime always
16482 has to allocate at least one BN_CTX).
16483 'callback(1, -1, cb_arg)' is called when a number has passed the
16484 trial division stage.
16488 * Fix for bug in CRL encoding. The validity dates weren't being handled
16493 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16497 * New function BN_pseudo_rand().
16501 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16502 bignum version of BN_from_montgomery() with the working code from
16503 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16508 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16509 made it impossible to use the same SSL_SESSION data structure in
16510 SSL2 clients in multiple threads.
16514 * The return value of RAND_load_file() no longer counts bytes obtained
16515 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16516 to seed the PRNG (previously an explicit byte count was required).
16518 *Ulf Möller, Bodo Möller*
16520 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16521 used `char *` instead of `void *` and had casts all over the place.
16525 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16529 * Retain source code compatibility for BN_prime_checks macro:
16530 BN_is_prime(..., BN_prime_checks, ...) now uses
16531 BN_prime_checks_for_size to determine the appropriate number of
16532 Rabin-Miller iterations.
16536 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16537 DH_CHECK_P_NOT_SAFE_PRIME.
16538 (Check if this is true? OpenPGP calls them "strong".)
16542 * Merge the functionality of "dh" and "gendh" programs into a new program
16543 "dhparam". The old programs are retained for now but will handle DH keys
16544 (instead of parameters) in future.
16548 * Make the ciphers, s_server and s_client programs check the return values
16549 when a new cipher list is set.
16553 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16554 ciphers. Before when the 56bit ciphers were enabled the sorting was
16557 The syntax for the cipher sorting has been extended to support sorting by
16558 cipher-strength (using the strength_bits hard coded in the tables).
16559 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16561 Fix a bug in the cipher-command parser: when supplying a cipher command
16562 string with an "undefined" symbol (neither command nor alphanumeric
16563 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16564 an error is flagged.
16566 Due to the strength-sorting extension, the code of the
16567 ssl_create_cipher_list() function was completely rearranged. I hope that
16568 the readability was also increased :-)
16570 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16572 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16573 for the first serial number and places 2 in the serial number file. This
16574 avoids problems when the root CA is created with serial number zero and
16575 the first user certificate has the same issuer name and serial number
16580 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16581 the new code. Add documentation for this stuff.
16585 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16586 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16587 structures and behave in an analogous way to the X509v3 functions:
16588 they shouldn't be called directly but wrapper functions should be used
16591 So we also now have some wrapper functions that call the X509at functions
16592 when passed certificate requests. (TO DO: similar things can be done with
16593 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16594 things. Some of these need some d2i or i2d and print functionality
16595 because they handle more complex structures.)
16599 * Add missing #ifndefs that caused missing symbols when building libssl
16600 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16601 NO_RSA in `ssl/s2*.c`.
16603 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16605 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16606 has a return value which indicates the quality of the random data
16607 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16608 error queue. New function RAND_pseudo_bytes() generates output that is
16609 guaranteed to be unique but not unpredictable. RAND_add is like
16610 RAND_seed, but takes an extra argument for an entropy estimate
16611 (RAND_seed always assumes full entropy).
16615 * Do more iterations of Rabin-Miller probable prime test (specifically,
16616 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16617 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16618 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16619 false-positive rate of at most 2^-80 for random input.
16623 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16627 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16628 in the 0.9.5 release), this returns the chain
16629 from an X509_CTX structure with a dup of the stack and all
16630 the X509 reference counts upped: so the stack will exist
16631 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16634 Also make SSL_SESSION_print() print out the verify return
16639 * Add manpage for the pkcs12 command. Also change the default
16640 behaviour so MAC iteration counts are used unless the new
16641 -nomaciter option is used. This improves file security and
16642 only older versions of MSIE (4.0 for example) need it.
16646 * Honor the no-xxx Configure options when creating .DEF files.
16650 * Add PKCS#10 attributes to field table: challengePassword,
16651 unstructuredName and unstructuredAddress. These are taken from
16652 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16653 international characters are used.
16655 More changes to X509_ATTRIBUTE code: allow the setting of types
16656 based on strings. Remove the 'loc' parameter when adding
16657 attributes because these will be a SET OF encoding which is sorted
16662 * Initial changes to the 'req' utility to allow request generation
16663 automation. This will allow an application to just generate a template
16664 file containing all the field values and have req construct the
16667 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16668 used all over the place including certificate requests and PKCS#7
16669 structures. They are currently handled manually where necessary with
16670 some primitive wrappers for PKCS#7. The new functions behave in a
16671 manner analogous to the X509 extension functions: they allow
16672 attributes to be looked up by NID and added.
16674 Later something similar to the X509V3 code would be desirable to
16675 automatically handle the encoding, decoding and printing of the
16676 more complex types. The string types like challengePassword can
16677 be handled by the string table functions.
16679 Also modified the multi byte string table handling. Now there is
16680 a 'global mask' which masks out certain types. The table itself
16681 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16682 is useful when for example there is only one permissible type
16683 (as in countryName) and using the mask might result in no valid
16688 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16689 SSL_get_peer_finished to allow applications to obtain the latest
16690 Finished messages sent to the peer or expected from the peer,
16691 respectively. (SSL_get_peer_finished is usually the Finished message
16692 actually received from the peer, otherwise the protocol will be aborted.)
16694 As the Finished message are message digests of the complete handshake
16695 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16696 be used for external authentication procedures when the authentication
16697 provided by SSL/TLS is not desired or is not enough.
16701 * Enhanced support for Alpha Linux is added. Now ./config checks if
16702 the host supports BWX extension and if Compaq C is present on the
16703 $PATH. Just exploiting of the BWX extension results in 20-30%
16704 performance kick for some algorithms, e.g. DES and RC4 to mention
16705 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16710 * Add support for MS "fast SGC". This is arguably a violation of the
16711 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16712 weak crypto and after checking the certificate is SGC a second one
16713 with strong crypto. MS SGC stops the first handshake after receiving
16714 the server certificate message and sends a second client hello. Since
16715 a server will typically do all the time consuming operations before
16716 expecting any further messages from the client (server key exchange
16717 is the most expensive) there is little difference between the two.
16719 To get OpenSSL to support MS SGC we have to permit a second client
16720 hello message after we have sent server done. In addition we have to
16721 reset the MAC if we do get this second client hello.
16725 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16726 if a DER encoded private key is RSA or DSA traditional format. Changed
16727 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16728 format DER encoded private key. Newer code should use PKCS#8 format which
16729 has the key type encoded in the ASN1 structure. Added DER private key
16730 support to pkcs8 application.
16734 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16735 ciphersuites has been selected (as required by the SSL 3/TLS 1
16736 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16737 is set, we interpret this as a request to violate the specification
16738 (the worst that can happen is a handshake failure, and 'correct'
16739 behaviour would result in a handshake failure anyway).
16743 * In SSL_CTX_add_session, take into account that there might be multiple
16744 SSL_SESSION structures with the same session ID (e.g. when two threads
16745 concurrently obtain them from an external cache).
16746 The internal cache can handle only one SSL_SESSION with a given ID,
16747 so if there's a conflict, we now throw out the old one to achieve
16752 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16753 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16754 some routines that use cipher OIDs: some ciphers do not have OIDs
16755 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16760 * Simplify the trust setting structure and code. Now we just have
16761 two sequences of OIDs for trusted and rejected settings. These will
16762 typically have values the same as the extended key usage extension
16763 and any application specific purposes.
16765 The trust checking code now has a default behaviour: it will just
16766 check for an object with the same NID as the passed id. Functions can
16767 be provided to override either the default behaviour or the behaviour
16768 for a given id. SSL client, server and email already have functions
16769 in place for compatibility: they check the NID and also return "trusted"
16770 if the certificate is self signed.
16774 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16775 traditional format into an EVP_PKEY structure.
16779 * Add a password callback function PEM_cb() which either prompts for
16780 a password if usr_data is NULL or otherwise assumes it is a null
16781 terminated password. Allow passwords to be passed on command line
16782 environment or config files in a few more utilities.
16786 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16787 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16788 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16789 Update documentation.
16793 * Support for ASN1 "NULL" type. This could be handled before by using
16794 ASN1_TYPE but there wasn't any function that would try to read a NULL
16795 and produce an error if it couldn't. For compatibility we also have
16796 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16797 don't allocate anything because they don't need to.
16801 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16804 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16806 * Rebuild of the memory allocation routines used by OpenSSL code and
16807 possibly others as well. The purpose is to make an interface that
16808 provide hooks so anyone can build a separate set of allocation and
16809 deallocation routines to be used by OpenSSL, for example memory
16810 pool implementations, or something else, which was previously hard
16811 since Malloc(), Realloc() and Free() were defined as macros having
16812 the values malloc, realloc and free, respectively (except for Win32
16813 compilations). The same is provided for memory debugging code.
16814 OpenSSL already comes with functionality to find memory leaks, but
16815 this gives people a chance to debug other memory problems.
16817 With these changes, a new set of functions and macros have appeared:
16819 CRYPTO_set_mem_debug_functions() [F]
16820 CRYPTO_get_mem_debug_functions() [F]
16821 CRYPTO_dbg_set_options() [F]
16822 CRYPTO_dbg_get_options() [F]
16823 CRYPTO_malloc_debug_init() [M]
16825 The memory debug functions are NULL by default, unless the library
16826 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16827 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16828 gives the standard debugging functions that come with OpenSSL) or
16829 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16830 provided by the library user) must be used. When the standard
16831 debugging functions are used, CRYPTO_dbg_set_options can be used to
16832 request additional information:
16833 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16834 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16836 Also, things like CRYPTO_set_mem_functions will always give the
16837 expected result (the new set of functions is used for allocation
16838 and deallocation) at all times, regardless of platform and compiler
16841 To finish it up, some functions that were never use in any other
16842 way than through macros have a new API and new semantic:
16844 CRYPTO_dbg_malloc()
16845 CRYPTO_dbg_realloc()
16848 All macros of value have retained their old syntax.
16850 *Richard Levitte and Bodo Moeller*
16852 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16853 ordering of SMIMECapabilities wasn't in "strength order" and there
16854 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16859 * Some ASN1 types with illegal zero length encoding (INTEGER,
16860 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16862 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16864 * Merge in my S/MIME library for OpenSSL. This provides a simple
16865 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16866 functionality to handle multipart/signed properly) and a utility
16867 called 'smime' to call all this stuff. This is based on code I
16868 originally wrote for Celo who have kindly allowed it to be
16869 included in OpenSSL.
16873 * Add variants des_set_key_checked and des_set_key_unchecked of
16874 des_set_key (aka des_key_sched). Global variable des_check_key
16875 decides which of these is called by des_set_key; this way
16876 des_check_key behaves as it always did, but applications and
16877 the library itself, which was buggy for des_check_key == 1,
16878 have a cleaner way to pick the version they need.
16882 * New function PKCS12_newpass() which changes the password of a
16887 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16888 dynamic mix. In both cases the ids can be used as an index into the
16889 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16890 functions so they accept a list of the field values and the
16891 application doesn't need to directly manipulate the X509_TRUST
16896 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16901 * Modify the way the V3 extension code looks up extensions. This now
16902 works in a similar way to the object code: we have some "standard"
16903 extensions in a static table which is searched with OBJ_bsearch()
16904 and the application can add dynamic ones if needed. The file
16905 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16906 updated whenever a new extension is added to the core code and kept
16907 in ext_nid order. There is a simple program 'tabtest.c' which checks
16908 this. New extensions are not added too often so this file can readily
16909 be maintained manually.
16911 There are two big advantages in doing things this way. The extensions
16912 can be looked up immediately and no longer need to be "added" using
16913 X509V3_add_standard_extensions(): this function now does nothing.
16914 Side note: I get *lots* of email saying the extension code doesn't
16915 work because people forget to call this function.
16916 Also no dynamic allocation is done unless new extensions are added:
16917 so if we don't add custom extensions there is no need to call
16918 X509V3_EXT_cleanup().
16922 * Modify enc utility's salting as follows: make salting the default. Add a
16923 magic header, so unsalted files fail gracefully instead of just decrypting
16924 to garbage. This is because not salting is a big security hole, so people
16925 should be discouraged from doing it.
16929 * Fixes and enhancements to the 'x509' utility. It allowed a message
16930 digest to be passed on the command line but it only used this
16931 parameter when signing a certificate. Modified so all relevant
16932 operations are affected by the digest parameter including the
16933 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16934 DSA key was used because it didn't fix the digest.
16938 * Initial certificate chain verify code. Currently tests the untrusted
16939 certificates for consistency with the verify purpose (which is set
16940 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16942 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16943 this is because it will reject chains with invalid extensions whereas
16944 every previous version of OpenSSL and SSLeay made no checks at all.
16946 Trust code: checks the root CA for the relevant trust settings. Trust
16947 settings have an initial value consistent with the verify purpose: e.g.
16948 if the verify purpose is for SSL client use it expects the CA to be
16949 trusted for SSL client use. However the default value can be changed to
16950 permit custom trust settings: one example of this would be to only trust
16951 certificates from a specific "secure" set of CAs.
16953 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16954 which should be used for version portability: especially since the
16955 verify structure is likely to change more often now.
16957 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16958 to set them. If not set then assume SSL clients will verify SSL servers
16961 Two new options to the verify program: -untrusted allows a set of
16962 untrusted certificates to be passed in and -purpose which sets the
16963 intended purpose of the certificate. If a purpose is set then the
16964 new chain verify code is used to check extension consistency.
16968 * Support for the authority information access extension.
16972 * Modify RSA and DSA PEM read routines to transparently handle
16973 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16974 public keys in a format compatible with certificate
16975 SubjectPublicKeyInfo structures. Unfortunately there were already
16976 functions called *_PublicKey_* which used various odd formats so
16977 these are retained for compatibility: however the DSA variants were
16978 never in a public release so they have been deleted. Changed dsa/rsa
16979 utilities to handle the new format: note no releases ever handled public
16980 keys so we should be OK.
16982 The primary motivation for this change is to avoid the same fiasco
16983 that dogs private keys: there are several incompatible private key
16984 formats some of which are standard and some OpenSSL specific and
16985 require various evil hacks to allow partial transparent handling and
16986 even then it doesn't work with DER formats. Given the option anything
16987 other than PKCS#8 should be dumped: but the other formats have to
16988 stay in the name of compatibility.
16990 With public keys and the benefit of hindsight one standard format
16991 is used which works with EVP_PKEY, RSA or DSA structures: though
16992 it clearly returns an error if you try to read the wrong kind of key.
16994 Added a -pubkey option to the 'x509' utility to output the public key.
16995 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16996 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16997 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16998 that do the same as the `EVP_PKEY_assign_*()` except they up the
16999 reference count of the added key (they don't "swallow" the
17004 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17005 CRLs would fail if the file contained no certificates or no CRLs:
17006 added a new function to read in both types and return the number
17007 read: this means that if none are read it will be an error. The
17008 DER versions of the certificate and CRL reader would always fail
17009 because it isn't possible to mix certificates and CRLs in DER format
17010 without choking one or the other routine. Changed this to just read
17011 a certificate: this is the best we can do. Also modified the code
17012 in `apps/verify.c` to take notice of return codes: it was previously
17013 attempting to read in certificates from NULL pointers and ignoring
17014 any errors: this is one reason why the cert and CRL reader seemed
17015 to work. It doesn't check return codes from the default certificate
17016 routines: these may well fail if the certificates aren't installed.
17020 * Code to support otherName option in GeneralName.
17024 * First update to verify code. Change the verify utility
17025 so it warns if it is passed a self signed certificate:
17026 for consistency with the normal behaviour. X509_verify
17027 has been modified to it will now verify a self signed
17028 certificate if *exactly* the same certificate appears
17029 in the store: it was previously impossible to trust a
17030 single self signed certificate. This means that:
17031 openssl verify ss.pem
17032 now gives a warning about a self signed certificate but
17033 openssl verify -CAfile ss.pem ss.pem
17038 * For servers, store verify_result in SSL_SESSION data structure
17039 (and add it to external session representation).
17040 This is needed when client certificate verifications fails,
17041 but an application-provided verification callback (set by
17042 SSL_CTX_set_cert_verify_callback) allows accepting the session
17043 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17044 but returns 1): When the session is reused, we have to set
17045 ssl->verify_result to the appropriate error code to avoid
17048 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17050 * Fix a bug in the new PKCS#7 code: it didn't consider the
17051 case in PKCS7_dataInit() where the signed PKCS7 structure
17052 didn't contain any existing data because it was being created.
17054 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17056 * Add a salt to the key derivation routines in enc.c. This
17057 forms the first 8 bytes of the encrypted file. Also add a
17058 -S option to allow a salt to be input on the command line.
17062 * New function X509_cmp(). Oddly enough there wasn't a function
17063 to compare two certificates. We do this by working out the SHA1
17064 hash and comparing that. X509_cmp() will be needed by the trust
17069 * SSL_get1_session() is like SSL_get_session(), but increments
17070 the reference count in the SSL_SESSION returned.
17072 *Geoff Thorpe <geoff@eu.c2.net>*
17074 * Fix for 'req': it was adding a null to request attributes.
17075 Also change the X509_LOOKUP and X509_INFO code to handle
17076 certificate auxiliary information.
17080 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17085 * Add the possibility to add extra information to the memory leak
17086 detecting output, to form tracebacks, showing from where each
17087 allocation was originated: CRYPTO_push_info("constant string") adds
17088 the string plus current file name and line number to a per-thread
17089 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17090 is like calling CYRPTO_pop_info() until the stack is empty.
17091 Also updated memory leak detection code to be multi-thread-safe.
17095 * Add options -text and -noout to pkcs7 utility and delete the
17096 encryption options which never did anything. Update docs.
17100 * Add options to some of the utilities to allow the pass phrase
17101 to be included on either the command line (not recommended on
17102 OSes like Unix) or read from the environment. Update the
17103 manpages and fix a few bugs.
17107 * Add a few manpages for some of the openssl commands.
17111 * Fix the -revoke option in ca. It was freeing up memory twice,
17112 leaking and not finding already revoked certificates.
17116 * Extensive changes to support certificate auxiliary information.
17117 This involves the use of X509_CERT_AUX structure and X509_AUX
17118 functions. An X509_AUX function such as PEM_read_X509_AUX()
17119 can still read in a certificate file in the usual way but it
17120 will also read in any additional "auxiliary information". By
17121 doing things this way a fair degree of compatibility can be
17122 retained: existing certificates can have this information added
17123 using the new 'x509' options.
17125 Current auxiliary information includes an "alias" and some trust
17126 settings. The trust settings will ultimately be used in enhanced
17127 certificate chain verification routines: currently a certificate
17128 can only be trusted if it is self signed and then it is trusted
17133 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17134 The problem was that one of the replacement routines had not been working
17135 since SSLeay releases. For now the offending routine has been replaced
17136 with non-optimised assembler. Even so, this now gives around 95%
17137 performance improvement for 1024 bit RSA signs.
17141 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17142 handling. Most clients have the effective key size in bits equal to
17143 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17144 A few however don't do this and instead use the size of the decrypted key
17145 to determine the RC2 key length and the AlgorithmIdentifier to determine
17146 the effective key length. In this case the effective key length can still
17147 be 40 bits but the key length can be 168 bits for example. This is fixed
17148 by manually forcing an RC2 key into the EVP_PKEY structure because the
17149 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17150 the key length and effective key length are equal.
17154 * Add a bunch of functions that should simplify the creation of
17155 X509_NAME structures. Now you should be able to do:
17156 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17157 and have it automatically work out the correct field type and fill in
17158 the structures. The more adventurous can try:
17159 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17160 and it will (hopefully) work out the correct multibyte encoding.
17164 * Change the 'req' utility to use the new field handling and multibyte
17165 copy routines. Before the DN field creation was handled in an ad hoc
17166 way in req, ca, and x509 which was rather broken and didn't support
17167 BMPStrings or UTF8Strings. Since some software doesn't implement
17168 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17169 using the dirstring_type option. See the new comment in the default
17170 openssl.cnf for more info.
17174 * Make crypto/rand/md_rand.c more robust:
17175 - Assure unique random numbers after fork().
17176 - Make sure that concurrent threads access the global counter and
17177 md serializably so that we never lose entropy in them
17178 or use exactly the same state in multiple threads.
17179 Access to the large state is not always serializable because
17180 the additional locking could be a performance killer, and
17181 md should be large enough anyway.
17185 * New file `apps/app_rand.c` with commonly needed functionality
17186 for handling the random seed file.
17188 Use the random seed file in some applications that previously did not:
17190 dsaparam -genkey (which also ignored its '-rand' option),
17193 x509 (when signing).
17194 Except on systems with /dev/urandom, it is crucial to have a random
17195 seed file at least for key creation, DSA signing, and for DH exchanges;
17196 for RSA signatures we could do without one.
17198 gendh and gendsa (unlike genrsa) used to read only the first byte
17199 of each file listed in the '-rand' option. The function as previously
17200 found in genrsa is now in app_rand.c and is used by all programs
17201 that support '-rand'.
17205 * In RAND_write_file, use mode 0600 for creating files;
17206 don't just chmod when it may be too late.
17210 * Report an error from X509_STORE_load_locations
17211 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17215 * New function ASN1_mbstring_copy() this copies a string in either
17216 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17217 into an ASN1_STRING type. A mask of permissible types is passed
17218 and it chooses the "minimal" type to use or an error if not type
17223 * Add function equivalents to the various macros in asn1.h. The old
17224 macros are retained with an `M_` prefix. Code inside the library can
17225 use the `M_` macros. External code (including the openssl utility)
17226 should *NOT* in order to be "shared library friendly".
17230 * Add various functions that can check a certificate's extensions
17231 to see if it usable for various purposes such as SSL client,
17232 server or S/MIME and CAs of these types. This is currently
17233 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17234 verification. Also added a -purpose flag to x509 utility to
17235 print out all the purposes.
17239 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17244 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17245 for, obtain and decode and extension and obtain its critical flag.
17246 This allows all the necessary extension code to be handled in a
17247 single function call.
17251 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17252 platforms. See crypto/rc4/rc4_enc.c for further details.
17256 * New -noout option to asn1parse. This causes no output to be produced
17257 its main use is when combined with -strparse and -out to extract data
17258 from a file (which may not be in ASN.1 format).
17262 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17263 when producing the local key id.
17265 *Richard Levitte <levitte@stacken.kth.se>*
17267 * New option -dhparam in s_server. This allows a DH parameter file to be
17268 stated explicitly. If it is not stated then it tries the first server
17269 certificate file. The previous behaviour hard coded the filename
17274 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17275 a public key to be input or output. For example:
17276 openssl rsa -in key.pem -pubout -out pubkey.pem
17277 Also added necessary DSA public key functions to handle this.
17281 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17282 in the message. This was handled by allowing
17283 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17285 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17287 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17288 to the end of the strings whereas this didn't. This would cause problems
17289 if strings read with d2i_ASN1_bytes() were later modified.
17291 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17293 * Fix for base64 decode bug. When a base64 bio reads only one line of
17294 data and it contains EOF it will end up returning an error. This is
17295 caused by input 46 bytes long. The cause is due to the way base64
17296 BIOs find the start of base64 encoded data. They do this by trying a
17297 trial decode on each line until they find one that works. When they
17298 do a flag is set and it starts again knowing it can pass all the
17299 data directly through the decoder. Unfortunately it doesn't reset
17300 the context it uses. This means that if EOF is reached an attempt
17301 is made to pass two EOFs through the context and this causes the
17302 resulting error. This can also cause other problems as well. As is
17303 usual with these problems it takes *ages* to find and the fix is
17304 trivial: move one line.
17306 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17308 * Ugly workaround to get s_client and s_server working under Windows. The
17309 old code wouldn't work because it needed to select() on sockets and the
17310 tty (for keypresses and to see if data could be written). Win32 only
17311 supports select() on sockets so we select() with a 1s timeout on the
17312 sockets and then see if any characters are waiting to be read, if none
17313 are present then we retry, we also assume we can always write data to
17314 the tty. This isn't nice because the code then blocks until we've
17315 received a complete line of data and it is effectively polling the
17316 keyboard at 1s intervals: however it's quite a bit better than not
17317 working at all :-) A dedicated Windows application might handle this
17318 with an event loop for example.
17322 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17323 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17324 will be called when RSA_sign() and RSA_verify() are used. This is useful
17325 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17326 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17327 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17328 This necessitated the support of an extra signature type NID_md5_sha1
17329 for SSL signatures and modifications to the SSL library to use it instead
17330 of calling RSA_public_decrypt() and RSA_private_encrypt().
17334 * Add new -verify -CAfile and -CApath options to the crl program, these
17335 will lookup a CRL issuers certificate and verify the signature in a
17336 similar way to the verify program. Tidy up the crl program so it
17337 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17338 less strict. It will now permit CRL extensions even if it is not
17339 a V2 CRL: this will allow it to tolerate some broken CRLs.
17343 * Initialize all non-automatic variables each time one of the openssl
17344 sub-programs is started (this is necessary as they may be started
17345 multiple times from the "OpenSSL>" prompt).
17347 *Lennart Bang, Bodo Moeller*
17349 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17350 removing all other RSA functionality (this is what NO_RSA does). This
17351 is so (for example) those in the US can disable those operations covered
17352 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17357 * Non-copying interface to BIO pairs.
17358 (still largely untested)
17362 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17363 ASCII string. This was handled independently in various places before.
17367 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17368 UTF8 strings a character at a time.
17372 * Use client_version from client hello to select the protocol
17373 (s23_srvr.c) and for RSA client key exchange verification
17374 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17378 * Add various utility functions to handle SPKACs, these were previously
17379 handled by poking round in the structure internals. Added new function
17380 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17381 print, verify and generate SPKACs. Based on an original idea from
17382 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17386 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17390 * Allow the config file extension section to be overwritten on the
17391 command line. Based on an original idea from Massimiliano Pala
17392 <madwolf@comune.modena.it>. The new option is called -extensions
17393 and can be applied to ca, req and x509. Also -reqexts to override
17394 the request extensions in req and -crlexts to override the crl extensions
17399 * Add new feature to the SPKAC handling in ca. Now you can include
17400 the same field multiple times by preceding it by "XXXX." for example:
17403 this is the same syntax as used in the req config file.
17407 * Allow certificate extensions to be added to certificate requests. These
17408 are specified in a 'req_extensions' option of the req section of the
17409 config file. They can be printed out with the -text option to req but
17410 are otherwise ignored at present.
17414 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17415 data read consists of only the final block it would not decrypted because
17416 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17417 A misplaced 'break' also meant the decrypted final block might not be
17418 copied until the next read.
17422 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17423 a few extra parameters to the DH structure: these will be useful if
17424 for example we want the value of 'q' or implement X9.42 DH.
17428 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17429 provides hooks that allow the default DSA functions or functions on a
17430 "per key" basis to be replaced. This allows hardware acceleration and
17431 hardware key storage to be handled without major modification to the
17432 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17433 associated functions.
17437 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17438 as "read only": it can't be written to and the buffer it points to will
17439 not be freed. Reading from a read only BIO is much more efficient than
17440 a normal memory BIO. This was added because there are several times when
17441 an area of memory needs to be read from a BIO. The previous method was
17442 to create a memory BIO and write the data to it, this results in two
17443 copies of the data and an O(n^2) reading algorithm. There is a new
17444 function BIO_new_mem_buf() which creates a read only memory BIO from
17445 an area of memory. Also modified the PKCS#7 routines to use read only
17450 * Bugfix: ssl23_get_client_hello did not work properly when called in
17451 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17452 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17453 but a retry condition occurred while trying to read the rest.
17457 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17458 NID_pkcs7_encrypted by default: this was wrong since this should almost
17459 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17460 the encrypted data type: this is a more sensible place to put it and it
17461 allows the PKCS#12 code to be tidied up that duplicated this
17466 * Changed obj_dat.pl script so it takes its input and output files on
17467 the command line. This should avoid shell escape redirection problems
17472 * Initial support for certificate extension requests, these are included
17473 in things like Xenroll certificate requests. Included functions to allow
17474 extensions to be obtained and added.
17478 * -crlf option to s_client and s_server for sending newlines as
17479 CRLF (as required by many protocols).
17483 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17485 * Install libRSAglue.a when OpenSSL is built with RSAref.
17487 *Ralf S. Engelschall*
17489 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17491 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17493 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17498 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17499 DH parameters/keys (q is lost during that conversion, but the resulting
17500 DH parameters contain its length).
17502 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17503 much faster than DH_generate_parameters (which creates parameters
17504 where `p = 2*q + 1`), and also the smaller q makes DH computations
17505 much more efficient (160-bit exponentiation instead of 1024-bit
17506 exponentiation); so this provides a convenient way to support DHE
17507 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17508 utter importance to use
17509 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17511 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17512 when such DH parameters are used, because otherwise small subgroup
17513 attacks may become possible!
17517 * Avoid memory leak in i2d_DHparams.
17521 * Allow the -k option to be used more than once in the enc program:
17522 this allows the same encrypted message to be read by multiple recipients.
17526 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17527 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17528 it will always use the numerical form of the OID, even if it has a short
17533 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17534 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17535 otherwise bn_mod_exp was called. In the case of hardware keys for example
17536 no private key components need be present and it might store extra data
17537 in the RSA structure, which cannot be accessed from bn_mod_exp.
17538 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17539 private key operations.
17543 * Added support for SPARC Linux.
17547 * pem_password_cb function type incompatibly changed from
17548 typedef int pem_password_cb(char *buf, int size, int rwflag);
17550 ....(char *buf, int size, int rwflag, void *userdata);
17551 so that applications can pass data to their callbacks:
17552 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17553 additional void * argument, which is just handed through whenever
17554 the password callback is called.
17556 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17558 New function SSL_CTX_set_default_passwd_cb_userdata.
17560 Compatibility note: As many C implementations push function arguments
17561 onto the stack in reverse order, the new library version is likely to
17562 interoperate with programs that have been compiled with the old
17563 pem_password_cb definition (PEM_whatever takes some data that
17564 happens to be on the stack as its last argument, and the callback
17565 just ignores this garbage); but there is no guarantee whatsoever that
17568 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17569 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17570 problems not only on Windows, but also on some Unix platforms.
17571 To avoid problematic command lines, these definitions are now in an
17572 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17573 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17577 * MIPS III/IV assembler module is reimplemented.
17581 * More DES library cleanups: remove references to srand/rand and
17582 delete an unused file.
17586 * Add support for the free Netwide assembler (NASM) under Win32,
17587 since not many people have MASM (ml) and it can be hard to obtain.
17588 This is currently experimental but it seems to work OK and pass all
17589 the tests. Check out INSTALL.W32 for info.
17593 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17594 without temporary keys kept an extra copy of the server key,
17595 and connections with temporary keys did not free everything in case
17600 * New function RSA_check_key and new openssl rsa option -check
17601 for verifying the consistency of RSA keys.
17603 *Ulf Moeller, Bodo Moeller*
17605 * Various changes to make Win32 compile work:
17606 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17607 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17608 comparison" warnings.
17609 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17613 * Add a debugging option to PKCS#5 v2 key generation function: when
17614 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17615 derived keys are printed to stderr.
17619 * Copy the flags in ASN1_STRING_dup().
17621 *Roman E. Pavlov <pre@mo.msk.ru>*
17623 * The x509 application mishandled signing requests containing DSA
17624 keys when the signing key was also DSA and the parameters didn't match.
17626 It was supposed to omit the parameters when they matched the signing key:
17627 the verifying software was then supposed to automatically use the CA's
17628 parameters if they were absent from the end user certificate.
17630 Omitting parameters is no longer recommended. The test was also
17631 the wrong way round! This was probably due to unusual behaviour in
17632 EVP_cmp_parameters() which returns 1 if the parameters match.
17633 This meant that parameters were omitted when they *didn't* match and
17634 the certificate was useless. Certificates signed with 'ca' didn't have
17637 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17639 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17640 The interface is as follows:
17641 Applications can use
17642 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17643 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17644 "off" is now the default.
17645 The library internally uses
17646 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17647 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17648 to disable memory-checking temporarily.
17650 Some inconsistent states that previously were possible (and were
17651 even the default) are now avoided.
17653 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17654 with each memory chunk allocated; this is occasionally more helpful
17655 than just having a counter.
17657 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17659 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17664 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17665 which largely parallels "options", but is for changing API behaviour,
17666 whereas "options" are about protocol behaviour.
17667 Initial "mode" flags are:
17669 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17670 a single record has been written.
17671 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17672 retries use the same buffer location.
17673 (But all of the contents must be
17678 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17681 * Fix problems with no-hmac etc.
17683 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17685 * New functions RSA_get_default_method(), RSA_set_method() and
17686 RSA_get_method(). These allows replacement of RSA_METHODs without having
17687 to mess around with the internals of an RSA structure.
17691 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17692 Also really enable memory leak checks in openssl.c and in some
17695 *Chad C. Mulligan, Bodo Moeller*
17697 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17698 up the length of negative integers. This has now been simplified to just
17699 store the length when it is first determined and use it later, rather
17700 than trying to keep track of where data is copied and updating it to
17702 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17704 * Add a new function PKCS7_signatureVerify. This allows the verification
17705 of a PKCS#7 signature but with the signing certificate passed to the
17706 function itself. This contrasts with PKCS7_dataVerify which assumes the
17707 certificate is present in the PKCS#7 structure. This isn't always the
17708 case: certificates can be omitted from a PKCS#7 structure and be
17709 distributed by "out of band" means (such as a certificate database).
17713 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17714 function prototypes in pem.h, also change util/mkdef.pl to add the
17715 necessary function names.
17719 * mk1mf.pl (used by Windows builds) did not properly read the
17720 options set by Configure in the top level Makefile, and Configure
17721 was not even able to write more than one option correctly.
17722 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17726 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17727 file to be loaded from a BIO or FILE pointer. The BIO version will
17728 for example allow memory BIOs to contain config info.
17732 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17733 Whoever hopes to achieve shared-library compatibility across versions
17734 must use this, not the compile-time macro.
17735 (Exercise 0.9.4: Which is the minimum library version required by
17737 Note: All this applies only to multi-threaded programs, others don't
17742 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17743 through a BIO pair triggered the default case, i.e.
17744 SSLerr(...,SSL_R_UNKNOWN_STATE).
17748 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17749 can use the SSL library even if none of the specific BIOs is
17754 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17755 for the encoded length.
17757 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17759 * Add initial documentation of the X509V3 functions.
17763 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17764 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17765 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17766 secure PKCS#8 private key format with a high iteration count.
17770 * Fix determination of Perl interpreter: A perl or perl5
17771 *directory* in $PATH was also accepted as the interpreter.
17773 *Ralf S. Engelschall*
17775 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17776 wrong with it but it was very old and did things like calling
17777 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17778 unusual formatting.
17782 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17783 to use the new extension code.
17787 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17788 with macros. This should make it easier to change their form, add extra
17789 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17794 * Add to configuration table a new entry that can specify an alternative
17795 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17796 according to Mark Crispin <MRC@Panda.COM>.
17800 * DES CBC did not update the IV. Weird.
17804 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17805 Changing the behaviour of the former might break existing programs --
17806 where IV updating is needed, des_ncbc_encrypt can be used.
17809 * When bntest is run from "make test" it drives bc to check its
17810 calculations, as well as internally checking them. If an internal check
17811 fails, it needs to cause bc to give a non-zero result or make test carries
17812 on without noticing the failure. Fixed.
17816 * DES library cleanups.
17820 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17821 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17822 ciphers. NOTE: although the key derivation function has been verified
17823 against some published test vectors it has not been extensively tested
17824 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17829 * Instead of "mkdir -p", which is not fully portable, use new
17830 Perl script "util/mkdir-p.pl".
17834 * Rewrite the way password based encryption (PBE) is handled. It used to
17835 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17836 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17837 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17838 the 'parameter' field of the AlgorithmIdentifier is passed to the
17839 underlying key generation function so it must do its own ASN1 parsing.
17840 This has also changed the EVP_PBE_CipherInit() function which now has a
17841 'parameter' argument instead of literal salt and iteration count values
17842 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17846 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17847 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17848 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17849 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17850 value was just used as a "magic string" and not used directly its
17851 value doesn't matter.
17855 * Introduce some semblance of const correctness to BN. Shame C doesn't
17860 * "linux-sparc64" configuration (ultrapenguin).
17862 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17863 "linux-sparc" configuration.
17865 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17867 * config now generates no-xxx options for missing ciphers.
17871 * Support the EBCDIC character set (work in progress).
17872 File ebcdic.c not yet included because it has a different license.
17874 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17876 * Support BS2000/OSD-POSIX.
17878 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17880 * Make callbacks for key generation use `void *` instead of `char *`.
17884 * Make S/MIME samples compile (not yet tested).
17888 * Additional typesafe stacks.
17892 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17896 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17898 * New configuration variant "sco5-gcc".
17900 * Updated some demos.
17902 *Sean O Riordain, Wade Scholine*
17904 * Add missing BIO_free at exit of pkcs12 application.
17908 * Fix memory leak in conf.c.
17912 * Updates for Win32 to assembler version of MD5.
17916 * Set #! path to perl in `apps/der_chop` to where we found it
17917 instead of using a fixed path.
17921 * SHA library changes for irix64-mips4-cc.
17925 * Improvements for VMS support.
17929 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17931 * Bignum library bug fix. IRIX 6 passes "make test" now!
17932 This also avoids the problems with SC4.2 and unpatched SC5.
17934 *Andy Polyakov <appro@fy.chalmers.se>*
17936 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17937 These are required because of the typesafe stack would otherwise break
17938 existing code. If old code used a structure member which used to be STACK
17939 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17940 sk_num or sk_value it would produce an error because the num, data members
17941 are not present in STACK_OF. Now it just produces a warning. sk_set
17942 replaces the old method of assigning a value to sk_value
17943 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17944 that does this will no longer work (and should use sk_set instead) but
17945 this could be regarded as a "questionable" behaviour anyway.
17949 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17950 correctly handle encrypted S/MIME data.
17954 * Change type of various DES function arguments from des_cblock
17955 (which means, in function argument declarations, pointer to char)
17956 to des_cblock * (meaning pointer to array with 8 char elements),
17957 which allows the compiler to do more typechecking; it was like
17958 that back in SSLeay, but with lots of ugly casts.
17960 Introduce new type const_des_cblock.
17964 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17965 problems: find RecipientInfo structure that matches recipient certificate
17966 and initialise the ASN1 structures properly based on passed cipher.
17970 * Belatedly make the BN tests actually check the results.
17974 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17975 to and from BNs: it was completely broken. New compilation option
17976 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17977 key elements as negative integers.
17981 * Reorganize and speed up MD5.
17983 *Andy Polyakov <appro@fy.chalmers.se>*
17987 *Richard Levitte <richard@levitte.org>*
17989 * New option -out to asn1parse to allow the parsed structure to be
17990 output to a file. This is most useful when combined with the -strparse
17991 option to examine the output of things like OCTET STRINGS.
17995 * Make SSL library a little more fool-proof by not requiring any longer
17996 that `SSL_set_{accept,connect}_state` be called before
17997 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17998 in many applications because usually everything *appeared* to work as
17999 intended anyway -- now it really works as intended).
18003 * Move openssl.cnf out of lib/.
18007 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18008 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18009 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18011 *Ralf S. Engelschall*
18013 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18014 handle PKCS#7 enveloped data properly.
18016 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18018 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18019 copying pointers. The cert_st handling is changed by this in
18020 various ways (and thus what used to be known as ctx->default_cert
18021 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18022 any longer when s->cert does not give us what we need).
18023 ssl_cert_instantiate becomes obsolete by this change.
18024 As soon as we've got the new code right (possibly it already is?),
18025 we have solved a couple of bugs of the earlier code where s->cert
18026 was used as if it could not have been shared with other SSL structures.
18028 Note that using the SSL API in certain dirty ways now will result
18029 in different behaviour than observed with earlier library versions:
18030 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18031 does not influence s as it used to.
18033 In order to clean up things more thoroughly, inside SSL_SESSION
18034 we don't use CERT any longer, but a new structure SESS_CERT
18035 that holds per-session data (if available); currently, this is
18036 the peer's certificate chain and, for clients, the server's certificate
18037 and temporary key. CERT holds only those values that can have
18038 meaningful defaults in an SSL_CTX.
18042 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18043 from the internal representation. Various PKCS#7 fixes: remove some
18044 evil casts and set the enc_dig_alg field properly based on the signing
18049 * Allow PKCS#12 password to be set from the command line or the
18050 environment. Let 'ca' get its config file name from the environment
18051 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18056 * Allow certificate policies extension to use an IA5STRING for the
18057 organization field. This is contrary to the PKIX definition but
18058 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18063 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18064 without disallowing inline assembler and the like for non-pedantic builds.
18068 * Support Borland C++ builder.
18070 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18076 * SHA-1 cleanups and performance enhancements.
18078 *Andy Polyakov <appro@fy.chalmers.se>*
18080 * Sparc v8plus assembler for the bignum library.
18082 *Andy Polyakov <appro@fy.chalmers.se>*
18084 * Accept any -xxx and +xxx compiler options in Configure.
18088 * Update HPUX configuration.
18092 * Add missing `sk_<type>_unshift()` function to safestack.h
18094 *Ralf S. Engelschall*
18096 * New function SSL_CTX_use_certificate_chain_file that sets the
18097 "extra_cert"s in addition to the certificate. (This makes sense
18098 only for "PEM" format files, as chains as a whole are not
18103 * Support verify_depth from the SSL API.
18104 x509_vfy.c had what can be considered an off-by-one-error:
18105 Its depth (which was not part of the external interface)
18106 was actually counting the number of certificates in a chain;
18107 now it really counts the depth.
18111 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18112 instead of X509err, which often resulted in confusing error
18113 messages since the error codes are not globally unique
18114 (e.g. an alleged error in ssl3_accept when a certificate
18115 didn't match the private key).
18117 * New function SSL_CTX_set_session_id_context that allows to set a default
18118 value (so that you don't need SSL_set_session_id_context for each
18119 connection using the SSL_CTX).
18123 * OAEP decoding bug fix.
18127 * Support INSTALL_PREFIX for package builders, as proposed by
18132 * New Configure options "threads" and "no-threads". For systems
18133 where the proper compiler options are known (currently Solaris
18134 and Linux), "threads" is the default.
18138 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18142 * Install various scripts to $(OPENSSLDIR)/misc, not to
18143 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18144 such as /usr/local/bin.
18148 * "make linux-shared" to build shared libraries.
18150 *Niels Poppe <niels@netbox.org>*
18152 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18156 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18157 extension adding in x509 utility.
18161 * Remove NOPROTO sections and error code comments.
18165 * Partial rewrite of the DEF file generator to now parse the ANSI
18170 * New Configure options --prefix=DIR and --openssldir=DIR.
18174 * Complete rewrite of the error code script(s). It is all now handled
18175 by one script at the top level which handles error code gathering,
18176 header rewriting and C source file generation. It should be much better
18177 than the old method: it now uses a modified version of Ulf's parser to
18178 read the ANSI prototypes in all header files (thus the old K&R definitions
18179 aren't needed for error creation any more) and do a better job of
18180 translating function codes into names. The old 'ASN1 error code embedded
18181 in a comment' is no longer necessary and it doesn't use .err files which
18182 have now been deleted. Also the error code call doesn't have to appear all
18183 on one line (which resulted in some large lines...).
18187 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18191 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18192 0 (which usually indicates a closed connection), but continue reading.
18196 * Fix some race conditions.
18200 * Add support for CRL distribution points extension. Add Certificate
18201 Policies and CRL distribution points documentation.
18205 * Move the autogenerated header file parts to crypto/opensslconf.h.
18209 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18210 8 of keying material. Merlin has also confirmed interop with this fix
18211 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18213 *Merlin Hughes <merlin@baltimore.ie>*
18215 * Fix lots of warnings.
18217 *Richard Levitte <levitte@stacken.kth.se>*
18219 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18220 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18222 *Richard Levitte <levitte@stacken.kth.se>*
18224 * Fix problems with sizeof(long) == 8.
18226 *Andy Polyakov <appro@fy.chalmers.se>*
18228 * Change functions to ANSI C.
18232 * Fix typos in error codes.
18234 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18236 * Remove defunct assembler files from Configure.
18240 * SPARC v8 assembler BIGNUM implementation.
18242 *Andy Polyakov <appro@fy.chalmers.se>*
18244 * Support for Certificate Policies extension: both print and set.
18245 Various additions to support the r2i method this uses.
18249 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18250 return a const string when you are expecting an allocated buffer.
18254 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18255 types DirectoryString and DisplayText.
18259 * Add code to allow r2i extensions to access the configuration database,
18260 add an LHASH database driver and add several ctx helper functions.
18264 * Fix an evil bug in bn_expand2() which caused various BN functions to
18265 fail when they extended the size of a BIGNUM.
18269 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18270 support typesafe stack.
18274 * Fix typo in SSL_[gs]et_options().
18276 *Nils Frostberg <nils@medcom.se>*
18278 * Delete various functions and files that belonged to the (now obsolete)
18279 old X509V3 handling code.
18283 * New Configure option "rsaref".
18287 * Don't auto-generate pem.h.
18291 * Introduce type-safe ASN.1 SETs.
18295 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18297 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18299 * Introduce type-safe STACKs. This will almost certainly break lots of code
18300 that links with OpenSSL (well at least cause lots of warnings), but fear
18301 not: the conversion is trivial, and it eliminates loads of evil casts. A
18302 few STACKed things have been converted already. Feel free to convert more.
18303 In the fullness of time, I'll do away with the STACK type altogether.
18307 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18308 specified in `<certfile>` by updating the entry in the index.txt file.
18309 This way one no longer has to edit the index.txt file manually for
18310 revoking a certificate. The -revoke option does the gory details now.
18312 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18314 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18315 `-text` option at all and this way the `-noout -text` combination was
18316 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18318 *Ralf S. Engelschall*
18320 * Make sure a corresponding plain text error message exists for the
18321 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18322 verify callback function determined that a certificate was revoked.
18324 *Ralf S. Engelschall*
18326 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18327 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18328 all available ciphers including rc5, which was forgotten until now.
18329 In order to let the testing shell script know which algorithms
18330 are available, a new (up to now undocumented) command
18331 `openssl list-cipher-commands` is used.
18335 * Bugfix: s_client occasionally would sleep in select() when
18336 it should have checked SSL_pending() first.
18340 * New functions DSA_do_sign and DSA_do_verify to provide access to
18341 the raw DSA values prior to ASN.1 encoding.
18345 * Tweaks to Configure
18347 *Niels Poppe <niels@netbox.org>*
18349 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18354 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18358 * New config option to avoid instructions that are illegal on the 80386.
18359 The default code is faster, but requires at least a 486.
18363 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18364 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18365 same as SSL2_VERSION anyway.
18369 * New "-showcerts" option for s_client.
18373 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18374 application. Various cleanups and fixes.
18378 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18379 modify error routines to work internally. Add error codes and PBE init
18380 to library startup routines.
18384 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18385 packing functions to asn1 and evp. Changed function names and error
18386 codes along the way.
18390 * PKCS12 integration: and so it begins... First of several patches to
18391 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18392 objects to objects.h
18396 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18397 and display support for Thawte strong extranet extension.
18401 * Add LinuxPPC support.
18403 *Jeff Dubrule <igor@pobox.org>*
18405 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18406 bn_div_words in alpha.s.
18408 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18410 * Make sure the RSA OAEP test is skipped under -DRSAref because
18411 OAEP isn't supported when OpenSSL is built with RSAref.
18413 *Ulf Moeller <ulf@fitug.de>*
18415 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18416 so they no longer are missing under -DNOPROTO.
18418 *Soren S. Jorvang <soren@t.dk>*
18420 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18422 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18423 doesn't work when the session is reused. Coming soon!
18427 * Fix a security hole, that allows sessions to be reused in the wrong
18428 context thus bypassing client cert protection! All software that uses
18429 client certs and session caches in multiple contexts NEEDS PATCHING to
18430 allow session reuse! A fuller solution is in the works.
18432 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18434 * Some more source tree cleanups (removed obsolete files
18435 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18436 permission on "config" script to be executable) and a fix for the INSTALL
18439 *Ulf Moeller <ulf@fitug.de>*
18441 * Remove some legacy and erroneous uses of malloc, free instead of
18444 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18446 * Make rsa_oaep_test return non-zero on error.
18448 *Ulf Moeller <ulf@fitug.de>*
18450 * Add support for native Solaris shared libraries. Configure
18451 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18452 if someone would make that last step automatic.
18454 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18456 * ctx_size was not built with the right compiler during "make links". Fixed.
18460 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18461 except NULL ciphers". This means the default cipher list will no longer
18462 enable NULL ciphers. They need to be specifically enabled e.g. with
18463 the string "DEFAULT:eNULL".
18467 * Fix to RSA private encryption routines: if p < q then it would
18468 occasionally produce an invalid result. This will only happen with
18469 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18473 * Be less restrictive and allow also `perl util/perlpath.pl
18474 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18475 because this way one can also use an interpreter named `perl5` (which is
18476 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18477 installed as `perl`).
18479 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18481 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18483 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18485 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18486 advapi32.lib to Win32 build and change the pem test comparison
18487 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18488 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18489 and crypto/des/ede_cbcm_enc.c.
18493 * DES quad checksum was broken on big-endian architectures. Fixed.
18497 * Comment out two functions in bio.h that aren't implemented. Fix up the
18498 Win32 test batch file so it (might) work again. The Win32 test batch file
18499 is horrible: I feel ill....
18503 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18504 in e_os.h. Audit of header files to check ANSI and non ANSI
18505 sections: 10 functions were absent from non ANSI section and not exported
18506 from Windows DLLs. Fixed up libeay.num for new functions.
18510 * Make `openssl version` output lines consistent.
18512 *Ralf S. Engelschall*
18514 * Fix Win32 symbol export lists for BIO functions: Added
18515 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18516 to ms/libeay{16,32}.def.
18518 *Ralf S. Engelschall*
18520 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18521 fine under Unix and passes some trivial tests I've now added. But the
18522 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18523 added to make sure no one expects that this stuff really works in the
18524 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18525 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18528 *Ralf S. Engelschall*
18530 * Fix the generation of two part addresses in perl.
18532 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18534 * Add config entry for Linux on MIPS.
18536 *John Tobey <jtobey@channel1.com>*
18538 * Make links whenever Configure is run, unless we are on Windoze.
18542 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18543 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18548 * Add a useful kludge to allow package maintainers to specify compiler and
18549 other platforms details on the command line without having to patch the
18550 Configure script every time: One now can use
18551 `perl Configure <id>:<details>`,
18552 i.e. platform ids are allowed to have details appended
18553 to them (separated by colons). This is treated as there would be a static
18554 pre-configured entry in Configure's %table under key `<id>` with value
18555 `<details>` and `perl Configure <id>` is called. So, when you want to
18556 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18557 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18558 now, which overrides the FreeBSD-elf entry on-the-fly.
18560 *Ralf S. Engelschall*
18562 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18566 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18567 on the `perl Configure ...` command line. This way one can compile
18568 OpenSSL libraries with Position Independent Code (PIC) which is needed
18569 for linking it into DSOs.
18571 *Ralf S. Engelschall*
18573 * Remarkably, export ciphers were totally broken and no-one had noticed!
18578 * Cleaned up the LICENSE document: The official contact for any license
18579 questions now is the OpenSSL core team under openssl-core@openssl.org.
18580 And add a paragraph about the dual-license situation to make sure people
18581 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18582 to the OpenSSL toolkit.
18584 *Ralf S. Engelschall*
18586 * General source tree makefile cleanups: Made `making xxx in yyy...`
18587 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18588 Additionally cleaned up the `make links` target: Remove unnecessary
18589 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18590 to speed processing and no longer clutter the display with confusing
18591 stuff. Instead only the actually done links are displayed.
18593 *Ralf S. Engelschall*
18595 * Permit null encryption ciphersuites, used for authentication only. It used
18596 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18597 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18602 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18603 signed attributes when verifying signatures (this would break them),
18604 the detached data encoding was wrong and public keys obtained using
18605 X509_get_pubkey() weren't freed.
18609 * Add text documentation for the BUFFER functions. Also added a work around
18610 to a Win95 console bug. This was triggered by the password read stuff: the
18611 last character typed gets carried over to the next fread(). If you were
18612 generating a new cert request using 'req' for example then the last
18613 character of the passphrase would be CR which would then enter the first
18618 * Added the new 'Includes OpenSSL Cryptography Software' button as
18619 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18620 button and can be used by applications based on OpenSSL to show the
18621 relationship to the OpenSSL project.
18623 *Ralf S. Engelschall*
18625 * Remove confusing variables in function signatures in files
18626 ssl/ssl_lib.c and ssl/ssl.h.
18628 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18630 * Don't install bss_file.c under PREFIX/include/
18632 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18634 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18635 functions that return function pointers and has support for NT specific
18636 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18637 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18638 unsigned to signed types: this was killing the Win32 compile.
18642 * Add new certificate file to stack functions,
18643 SSL_add_dir_cert_subjects_to_stack() and
18644 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18645 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18646 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18647 This means that Apache-SSL and similar packages don't have to mess around
18648 to add as many CAs as they want to the preferred list.
18652 * Experiment with doxygen documentation. Currently only partially applied to
18654 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18655 openssl.doxy as the configuration file.
18659 * Get rid of remaining C++-style comments which strict C compilers hate.
18661 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18663 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18664 compiled in by default: it has problems with large keys.
18668 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18669 DH private keys and/or callback functions which directly correspond to
18670 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18671 is needed for applications which have to configure certificates on a
18672 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18674 For the RSA certificate situation is makes no difference, but
18675 for the DSA certificate situation this fixes the "no shared cipher"
18676 problem where the OpenSSL cipher selection procedure failed because the
18677 temporary keys were not overtaken from the context and the API provided
18678 no way to reconfigure them.
18679 The new functions now let applications reconfigure the stuff and they
18680 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18681 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18682 non-public-API function ssl_cert_instantiate() is used as a helper
18683 function and also to reduce code redundancy inside ssl_rsa.c.
18685 *Ralf S. Engelschall*
18687 * Move s_server -dcert and -dkey options out of the undocumented feature
18688 area because they are useful for the DSA situation and should be
18689 recognized by the users.
18691 *Ralf S. Engelschall*
18693 * Fix the cipher decision scheme for export ciphers: the export bits are
18694 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18695 SSL_EXP_MASK. So, the original variable has to be used instead of the
18696 already masked variable.
18698 *Richard Levitte <levitte@stacken.kth.se>*
18700 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18702 *Richard Levitte <levitte@stacken.kth.se>*
18704 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18705 from `int` to `unsigned int` because it is a length and initialized by
18706 EVP_DigestFinal() which expects an `unsigned int *`.
18708 *Richard Levitte <levitte@stacken.kth.se>*
18710 * Don't hard-code path to Perl interpreter on shebang line of Configure
18711 script. Instead use the usual Shell->Perl transition trick.
18713 *Ralf S. Engelschall*
18715 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18716 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18717 -noout -modulus` as it's already the case for `openssl rsa -noout
18718 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18719 currently the public key is printed (a decision which was already done by
18720 `openssl dsa -modulus` in the past) which serves a similar purpose.
18721 Additionally the NO_RSA no longer completely removes the whole -modulus
18722 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18725 *Ralf S. Engelschall*
18727 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18728 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18730 *Arne Ansper <arne@ats.cyber.ee>*
18732 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18733 to be added. Now both 'req' and 'ca' can use new objects defined in the
18738 * Add cool BIO that does syslog (or event log on NT).
18740 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18742 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18743 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18744 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18745 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18749 * Add preliminary config info for new extension code.
18753 * Make RSA_NO_PADDING really use no padding.
18755 *Ulf Moeller <ulf@fitug.de>*
18757 * Generate errors when private/public key check is done.
18761 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18762 for some CRL extensions and new objects added.
18766 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18767 key usage extension and fuller support for authority key id.
18771 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18772 padding method for RSA, which is recommended for new applications in PKCS
18773 #1 v2.0 (RFC 2437, October 1998).
18774 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18775 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18776 against Bleichbacher's attack on RSA.
18777 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18780 * Updates to the new SSL compression code
18782 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18784 * Fix so that the version number in the master secret, when passed
18785 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18786 (because the server will not accept higher), that the version number
18787 is 0x03,0x01, not 0x03,0x00
18789 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18791 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18792 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18793 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18797 * Support for RAW extensions where an arbitrary extension can be
18798 created by including its DER encoding. See `apps/openssl.cnf` for
18803 * Make sure latest Perl versions don't interpret some generated C array
18804 code as Perl array code in the crypto/err/err_genc.pl script.
18806 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18808 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18809 not many people have the assembler. Various Win32 compilation fixes and
18810 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18811 build instructions.
18815 * Modify configure script 'Configure' to automatically create crypto/date.h
18816 file under Win32 and also build pem.h from pem.org. New script
18817 util/mkfiles.pl to create the MINFO file on environments that can't do a
18818 'make files': perl util/mkfiles.pl >MINFO should work.
18822 * Major rework of DES function declarations, in the pursuit of correctness
18823 and purity. As a result, many evil casts evaporated, and some weirdness,
18824 too. You may find this causes warnings in your code. Zapping your evil
18825 casts will probably fix them. Mostly.
18829 * Fix for a typo in asn1.h. Bug fix to object creation script
18830 obj_dat.pl. It considered a zero in an object definition to mean
18831 "end of object": none of the objects in objects.h have any zeros
18832 so it wasn't spotted.
18834 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18836 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18837 Masking (CBCM). In the absence of test vectors, the best I have been able
18838 to do is check that the decrypt undoes the encrypt, so far. Send me test
18839 vectors if you have them.
18843 * Correct calculation of key length for export ciphers (too much space was
18844 allocated for null ciphers). This has not been tested!
18848 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18849 message is now correct (it understands "crypto" and "ssl" on its
18850 command line). There is also now an "update" option. This will update
18851 the util/ssleay.num and util/libeay.num files with any new functions.
18853 perl util/mkdef.pl crypto ssl update
18854 it will update them.
18858 * Overhauled the Perl interface:
18859 - ported BN stuff to OpenSSL's different BN library
18860 - made the perl/ source tree CVS-aware
18861 - renamed the package from SSLeay to OpenSSL (the files still contain
18862 their history because I've copied them in the repository)
18863 - removed obsolete files (the test scripts will be replaced
18864 by better Test::Harness variants in the future)
18866 *Ralf S. Engelschall*
18868 * First cut for a very conservative source tree cleanup:
18869 1. merge various obsolete readme texts into doc/ssleay.txt
18870 where we collect the old documents and readme texts.
18871 2. remove the first part of files where I'm already sure that we no
18872 longer need them because of three reasons: either they are just temporary
18873 files which were left by Eric or they are preserved original files where
18874 I've verified that the diff is also available in the CVS via "cvs diff
18875 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18876 the crypto/md/ stuff).
18878 *Ralf S. Engelschall*
18880 * More extension code. Incomplete support for subject and issuer alt
18881 name, issuer and authority key id. Change the i2v function parameters
18882 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18883 what that's for :-) Fix to ASN1 macro which messed up
18884 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18888 * Preliminary support for ENUMERATED type. This is largely copied from the
18893 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18895 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18897 * Make sure `make rehash` target really finds the `openssl` program.
18899 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18901 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18902 like to hear about it if this slows down other processors.
18906 * Add CygWin32 platform information to Configure script.
18908 *Alan Batie <batie@aahz.jf.intel.com>*
18910 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18912 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18914 * New program nseq to manipulate netscape certificate sequences
18918 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18923 * Fixes to BN code. Previously the default was to define BN_RECURSION
18924 but the BN code had some problems that would cause failures when
18925 doing certificate verification and some other functions.
18927 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18929 * Add ASN1 and PEM code to support netscape certificate sequences.
18933 * Add ASN1 and PEM code to support netscape certificate sequences.
18937 * Add several PKIX and private extended key usage OIDs.
18941 * Modify the 'ca' program to handle the new extension code. Modify
18942 openssl.cnf for new extension format, add comments.
18946 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18947 and add a sample to openssl.cnf so req -x509 now adds appropriate
18952 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18953 error code, add initial support to X509_print() and x509 application.
18957 * Takes a deep breath and start adding X509 V3 extension support code. Add
18958 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18959 stuff is currently isolated and isn't even compiled yet.
18963 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18964 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18965 Removed the versions check from X509 routines when loading extensions:
18966 this allows certain broken certificates that don't set the version
18967 properly to be processed.
18971 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18972 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18973 can still be regenerated with "make depend".
18977 * Spelling mistake in C version of CAST-128.
18979 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18981 * Changes to the error generation code. The perl script err-code.pl
18982 now reads in the old error codes and retains the old numbers, only
18983 adding new ones if necessary. It also only changes the .err files if new
18984 codes are added. The makefiles have been modified to only insert errors
18985 when needed (to avoid needlessly modifying header files). This is done
18986 by only inserting errors if the .err file is newer than the auto generated
18987 C file. To rebuild all the error codes from scratch (the old behaviour)
18988 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18989 or delete all the .err files.
18993 * CAST-128 was incorrectly implemented for short keys. The C version has
18994 been fixed, but is untested. The assembler versions are also fixed, but
18995 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18996 to regenerate it if needed.
18997 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18998 Hagino <itojun@kame.net>*
19000 * File was opened incorrectly in randfile.c.
19002 *Ulf Möller <ulf@fitug.de>*
19004 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19005 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19006 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19007 al: it's just almost always a UTCTime. Note this patch adds new error
19008 codes so do a "make errors" if there are problems.
19012 * Correct Linux 1 recognition in config.
19014 *Ulf Möller <ulf@fitug.de>*
19016 * Remove pointless MD5 hash when using DSA keys in ca.
19018 *Anonymous <nobody@replay.com>*
19020 * Generate an error if given an empty string as a cert directory. Also
19021 generate an error if handed NULL (previously returned 0 to indicate an
19022 error, but didn't set one).
19024 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19026 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19030 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19031 parameters. This was causing a warning which killed off the Win32 compile.
19035 * Remove C++ style comments from crypto/bn/bn_local.h.
19037 *Neil Costigan <neil.costigan@celocom.com>*
19039 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19040 based on a text string, looking up short and long names and finally
19041 "dot" format. The "dot" format stuff didn't work. Added new function
19042 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19043 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19044 OID is not part of the table.
19048 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19049 X509_LOOKUP_by_alias().
19053 * Sort openssl functions by name.
19057 * Get the `gendsa` command working and add it to the `list` command. Remove
19058 encryption from sample DSA keys (in case anyone is interested the password
19063 * Make *all* `*_free` functions accept a NULL pointer.
19065 *Frans Heymans <fheymans@isaserver.be>*
19067 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19070 *Anonymous <nobody@replay.com>*
19072 * s_server should send the CAfile as acceptable CAs, not its own cert.
19074 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19076 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19078 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19080 * Temp key "for export" tests were wrong in s3_srvr.c.
19082 *Anonymous <nobody@replay.com>*
19084 * Add prototype for temp key callback functions
19085 SSL_CTX_set_tmp_{rsa,dh}_callback().
19089 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19090 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19094 * X509_name_add_entry() freed the wrong thing after an error.
19096 *Arne Ansper <arne@ats.cyber.ee>*
19098 * rsa_eay.c would attempt to free a NULL context.
19100 *Arne Ansper <arne@ats.cyber.ee>*
19102 * BIO_s_socket() had a broken should_retry() on Windoze.
19104 *Arne Ansper <arne@ats.cyber.ee>*
19106 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19108 *Arne Ansper <arne@ats.cyber.ee>*
19110 * Make sure the already existing X509_STORE->depth variable is initialized
19111 in X509_STORE_new(), but document the fact that this variable is still
19112 unused in the certificate verification process.
19114 *Ralf S. Engelschall*
19116 * Fix the various library and `apps/` files to free up pkeys obtained from
19117 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19121 * Fix reference counting in X509_PUBKEY_get(). This makes
19122 demos/maurice/example2.c work, amongst others, probably.
19124 *Steve Henson and Ben Laurie*
19126 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19127 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19128 are no longer created. This way we have a single and consistent command
19129 line interface `openssl <command>`, similar to `cvs <command>`.
19131 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19133 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19134 BIT STRING wrapper always have zero unused bits.
19138 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19142 * Make the top-level INSTALL documentation easier to understand.
19146 * Makefiles updated to exit if an error occurs in a sub-directory
19147 make (including if user presses ^C) [Paul Sutton]
19149 * Make Montgomery context stuff explicit in RSA data structure.
19153 * Fix build order of pem and err to allow for generated pem.h.
19157 * Fix renumbering bug in X509_NAME_delete_entry().
19161 * Enhanced the err-ins.pl script so it makes the error library number
19162 global and can add a library name. This is needed for external ASN1 and
19163 other error libraries.
19167 * Fixed sk_insert which never worked properly.
19171 * Fix ASN1 macros so they can handle indefinite length constructed
19172 EXPLICIT tags. Some non standard certificates use these: they can now
19177 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19178 into a single doc/ssleay.txt bundle. This way the information is still
19179 preserved but no longer messes up this directory. Now it's new room for
19180 the new set of documentation files.
19182 *Ralf S. Engelschall*
19184 * SETs were incorrectly DER encoded. This was a major pain, because they
19185 shared code with SEQUENCEs, which aren't coded the same. This means that
19186 almost everything to do with SETs or SEQUENCEs has either changed name or
19187 number of arguments.
19189 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19191 * Fix test data to work with the above.
19195 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19196 was already fixed by Eric for 0.9.1 it seems.
19198 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19200 * Autodetect FreeBSD3.
19204 * Fix various bugs in Configure. This affects the following platforms:
19208 unixware-2.0-pentium
19213 * Eliminate generated files from CVS. Reorder tests to regenerate files
19214 before they are needed.
19218 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19222 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19224 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19225 changed SSLeay to OpenSSL in version strings.
19227 *Ralf S. Engelschall*
19229 * Some fixups to the top-level documents.
19233 * Fixed the nasty bug where rsaref.h was not found under compile-time
19234 because the symlink to include/ was missing.
19236 *Ralf S. Engelschall*
19238 * Incorporated the popular no-RSA/DSA-only patches
19239 which allow to compile a RSA-free SSLeay.
19241 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19243 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19244 when "ssleay" is still not found.
19246 *Ralf S. Engelschall*
19248 * Added more platforms to Configure: Cray T3E, HPUX 11,
19250 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19252 * Updated the README file.
19254 *Ralf S. Engelschall*
19256 * Added various .cvsignore files in the CVS repository subdirs
19257 to make a "cvs update" really silent.
19259 *Ralf S. Engelschall*
19261 * Recompiled the error-definition header files and added
19262 missing symbols to the Win32 linker tables.
19264 *Ralf S. Engelschall*
19266 * Cleaned up the top-level documents;
19267 o new files: CHANGES and LICENSE
19268 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19269 o merged COPYRIGHT into LICENSE
19270 o removed obsolete TODO file
19271 o renamed MICROSOFT to INSTALL.W32
19273 *Ralf S. Engelschall*
19275 * Removed dummy files from the 0.9.1b source tree:
19276 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19277 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19278 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19279 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19280 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19282 *Ralf S. Engelschall*
19284 * Added various platform portability fixes.
19288 * The Genesis of the OpenSSL rpject:
19289 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19290 Young and Tim J. Hudson created while they were working for C2Net until
19293 *The OpenSSL Project*
19295 ### Changes between 0.9.0b and 0.9.1b [not released]
19297 * Updated a few CA certificates under certs/
19301 * Changed some BIGNUM api stuff.
19305 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19306 DGUX x86, Linux Alpha, etc.
19310 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19311 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19316 * Add -strparse option to asn1pars program which parses nested
19319 *Dr Stephen Henson <shenson@bigfoot.com>*
19321 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19325 * DSA fix for "ca" program.
19329 * Added "-genkey" option to "dsaparam" program.
19333 * Added RIPE MD160 (rmd160) message digest.
19337 * Added -a (all) option to "ssleay version" command.
19341 * Added PLATFORM define which is the id given to Configure.
19345 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19349 * Extended the ASN.1 parser routines.
19353 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19357 * Added a BN_CTX to the BN library.
19361 * Fixed the weak key values in DES library
19365 * Changed API in EVP library for cipher aliases.
19369 * Added support for RC2/64bit cipher.
19373 * Converted the lhash library to the crypto/mem.c functions.
19377 * Added more recognized ASN.1 object ids.
19381 * Added more RSA padding checks for SSL/TLS.
19385 * Added BIO proxy/filter functionality.
19389 * Added extra_certs to SSL_CTX which can be used
19390 send extra CA certificates to the client in the CA cert chain sending
19391 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19395 * Now Fortezza is denied in the authentication phase because
19396 this is key exchange mechanism is not supported by SSLeay at all.
19400 * Additional PKCS1 checks.
19404 * Support the string "TLSv1" for all TLS v1 ciphers.
19408 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19409 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19413 * Fixed a few memory leaks.
19417 * Fixed various code and comment typos.
19421 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19422 bytes sent in the client random.
19424 *Edward Bishop <ebishop@spyglass.com>*
19428 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19429 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19430 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19431 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19432 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19433 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19434 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19435 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19436 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19437 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19438 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19439 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19440 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19441 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19442 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19443 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19444 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19445 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19446 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19447 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19448 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19449 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19450 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19451 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19452 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19453 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19454 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19455 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19456 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19457 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19458 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19459 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19460 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19461 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19462 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19463 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19464 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19465 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19466 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19467 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19468 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19469 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19470 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19471 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19472 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19473 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19474 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19475 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19476 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19477 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19478 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19479 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19480 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19481 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19482 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19483 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19484 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19485 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19486 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19487 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19488 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19489 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19490 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19491 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19492 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19493 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19494 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19495 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19496 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19497 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19498 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19499 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19500 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19501 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19502 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19503 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19504 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19505 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19506 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19507 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19508 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19509 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19510 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19511 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19512 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19513 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19514 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19515 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19516 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19517 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19518 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19519 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19520 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19521 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19522 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19523 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19524 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19525 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19526 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19527 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19528 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19529 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19530 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19531 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19532 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19533 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19534 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19535 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19536 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19537 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19538 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19539 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19540 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19541 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19542 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19543 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19544 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19545 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19546 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19547 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19548 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19549 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19550 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19551 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19552 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19553 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19554 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19555 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19556 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19557 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19558 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19559 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19560 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19561 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19562 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19563 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19564 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19565 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19566 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19567 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19568 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19569 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19570 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19571 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19572 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19573 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19574 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19575 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19576 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19577 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19578 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19579 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19580 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19581 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19582 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19583 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19584 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19585 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19586 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19587 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19588 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19589 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19590 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19591 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655