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 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.8 and 3.0.9 [xx XXX xxxx]
33 * Fixed an issue where invalid certificate policies in leaf certificates are
34 silently ignored by OpenSSL and other certificate policy checks are skipped
35 for that certificate. A malicious CA could use this to deliberately assert
36 invalid certificate policies in order to circumvent policy checking on the
37 certificate altogether.
42 * Limited the number of nodes created in a policy tree to mitigate
43 against CVE-2023-0464. The default limit is set to 1000 nodes, which
44 should be sufficient for most installations. If required, the limit
45 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
46 time define to a desired maximum number of nodes or zero to allow
52 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
54 * Fixed NULL dereference during PKCS7 data verification.
56 A NULL pointer can be dereferenced when signatures are being
57 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
58 algorithm used for the signature is known to the OpenSSL library but
59 the implementation of the hash algorithm is not available the digest
60 initialization will fail. There is a missing check for the return
61 value from the initialization function which later leads to invalid
62 usage of the digest API most likely leading to a crash.
65 PKCS7 data is processed by the SMIME library calls and also by the
66 time stamp (TS) library calls. The TLS implementation in OpenSSL does
67 not call these functions however third party applications would be
68 affected if they call these functions to verify signatures on untrusted
73 * Fixed X.400 address type confusion in X.509 GeneralName.
75 There is a type confusion vulnerability relating to X.400 address processing
76 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
77 but the public structure definition for GENERAL_NAME incorrectly specified
78 the type of the x400Address field as ASN1_TYPE. This field is subsequently
79 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
82 When CRL checking is enabled (i.e. the application sets the
83 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
84 pass arbitrary pointers to a memcmp call, enabling them to read memory
85 contents or enact a denial of service.
90 * Fixed NULL dereference validating DSA public key.
92 An invalid pointer dereference on read can be triggered when an
93 application tries to check a malformed DSA public key by the
94 EVP_PKEY_public_check() function. This will most likely lead
95 to an application crash. This function can be called on public
96 keys supplied from untrusted sources which could allow an attacker
97 to cause a denial of service attack.
99 The TLS implementation in OpenSSL does not call this function
100 but applications might call the function if there are additional
101 security requirements imposed by standards such as FIPS 140-3.
104 *Shane Lontis, Tomáš Mráz*
106 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
108 An invalid pointer dereference on read can be triggered when an
109 application tries to load malformed PKCS7 data with the
110 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
112 The result of the dereference is an application crash which could
113 lead to a denial of service attack. The TLS implementation in OpenSSL
114 does not call this function however third party applications might
115 call these functions on untrusted data.
120 * Fixed Use-after-free following BIO_new_NDEF.
122 The public API function BIO_new_NDEF is a helper function used for
123 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
124 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
125 be called directly by end user applications.
127 The function receives a BIO from the caller, prepends a new BIO_f_asn1
128 filter BIO onto the front of it to form a BIO chain, and then returns
129 the new head of the BIO chain to the caller. Under certain conditions,
130 for example if a CMS recipient public key is invalid, the new filter BIO
131 is freed and the function returns a NULL result indicating a failure.
132 However, in this case, the BIO chain is not properly cleaned up and the
133 BIO passed by the caller still retains internal pointers to the previously
134 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
135 then a use-after-free will occur. This will most likely result in a crash.
138 *Viktor Dukhovni, Matt Caswell*
140 * Fixed Double free after calling PEM_read_bio_ex.
142 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
143 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
144 data. If the function succeeds then the "name_out", "header" and "data"
145 arguments are populated with pointers to buffers containing the relevant
146 decoded data. The caller is responsible for freeing those buffers. It is
147 possible to construct a PEM file that results in 0 bytes of payload data.
148 In this case PEM_read_bio_ex() will return a failure code but will populate
149 the header argument with a pointer to a buffer that has already been freed.
150 If the caller also frees this buffer then a double free will occur. This
151 will most likely lead to a crash.
153 The functions PEM_read_bio() and PEM_read() are simple wrappers around
154 PEM_read_bio_ex() and therefore these functions are also directly affected.
156 These functions are also called indirectly by a number of other OpenSSL
157 functions including PEM_X509_INFO_read_bio_ex() and
158 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
159 internal uses of these functions are not vulnerable because the caller does
160 not free the header argument if PEM_read_bio_ex() returns a failure code.
163 *Kurt Roeckx, Matt Caswell*
165 * Fixed Timing Oracle in RSA Decryption.
167 A timing based side channel exists in the OpenSSL RSA Decryption
168 implementation which could be sufficient to recover a plaintext across
169 a network in a Bleichenbacher style attack. To achieve a successful
170 decryption an attacker would have to be able to send a very large number
171 of trial messages for decryption. The vulnerability affects all RSA padding
172 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
175 *Dmitry Belyavsky, Hubert Kario*
177 * Fixed X.509 Name Constraints Read Buffer Overflow.
179 A read buffer overrun can be triggered in X.509 certificate verification,
180 specifically in name constraint checking. The read buffer overrun might
181 result in a crash which could lead to a denial of service attack.
182 In a TLS client, this can be triggered by connecting to a malicious
183 server. In a TLS server, this can be triggered if the server requests
184 client authentication and a malicious client connects.
189 * Fixed X.509 Policy Constraints Double Locking security issue.
191 If an X.509 certificate contains a malformed policy constraint and
192 policy processing is enabled, then a write lock will be taken twice
193 recursively. On some operating systems (most widely: Windows) this
194 results in a denial of service when the affected process hangs. Policy
195 processing being enabled on a publicly facing server is not considered
196 to be a common setup.
201 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
202 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
203 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
204 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
205 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
206 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
207 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
208 for legacy EC and SM2 keys is also changed similarly to honor the
209 equivalent conversion format flag as specified in the underlying
210 `EC_KEY` object being exported to a provider, when this function is
211 called through `EVP_PKEY_export()`.
215 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
217 * Fixed two buffer overflows in punycode decoding functions.
219 A buffer overrun can be triggered in X.509 certificate verification,
220 specifically in name constraint checking. Note that this occurs after
221 certificate chain signature verification and requires either a CA to
222 have signed the malicious certificate or for the application to continue
223 certificate verification despite failure to construct a path to a trusted
226 In a TLS client, this can be triggered by connecting to a malicious
227 server. In a TLS server, this can be triggered if the server requests
228 client authentication and a malicious client connects.
230 An attacker can craft a malicious email address to overflow
231 an arbitrary number of bytes containing the `.` character (decimal 46)
232 on the stack. This buffer overflow could result in a crash (causing a
236 An attacker can craft a malicious email address to overflow four
237 attacker-controlled bytes on the stack. This buffer overflow could
238 result in a crash (causing a denial of service) or potentially remote code
239 execution depending on stack layout for any given platform/compiler.
244 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
245 parameters in OpenSSL code.
246 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
247 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
248 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
249 Using these invalid names may cause algorithms to use slower methods
250 that ignore the CRT parameters.
254 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
259 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
260 data to be signed before signing the certificate.
264 * Added RIPEMD160 to the default provider.
268 * Ensured that the key share group sent or accepted for the key exchange
269 is allowed for the protocol version.
273 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
275 * OpenSSL supports creating a custom cipher via the legacy
276 EVP_CIPHER_meth_new() function and associated function calls. This function
277 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
278 to use the new provider mechanism in order to implement custom ciphers.
280 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
281 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
282 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
283 and decryption initialisation functions). Instead of using the custom cipher
284 directly it incorrectly tries to fetch an equivalent cipher from the
285 available providers. An equivalent cipher is found based on the NID passed to
286 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
287 given cipher. However it is possible for an application to incorrectly pass
288 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
289 is used in this way the OpenSSL encryption/decryption initialisation function
290 will match the NULL cipher as being equivalent and will fetch this from the
291 available providers. This will succeed if the default provider has been
292 loaded (or if a third party provider has been loaded that offers this
293 cipher). Using the NULL cipher means that the plaintext is emitted as the
296 Applications are only affected by this issue if they call
297 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
298 encryption/decryption initialisation function. Applications that only use
299 SSL/TLS are not impacted by this issue.
304 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
309 * Fixed the linux-mips64 Configure target which was missing the
310 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
315 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
320 * Correctly handle a retransmitted ClientHello in DTLS
324 * Fixed detection of ktls support in cross-compile environment on Linux
328 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
333 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
334 report correct results in some cases
338 * Fix UWP builds by defining VirtualLock
342 * For known safe primes use the minimum key length according to RFC 7919.
343 Longer private key sizes unnecessarily raise the cycles needed to compute the
344 shared secret without any increase of the real security. This fixes a
345 regression from 1.1.1 where these shorter keys were generated for the known
350 * Added the loongarch64 target
354 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
355 only passed to the FIPS provider and not to the default or legacy provider.
359 * Fixed reported performance degradation on aarch64. Restored the
360 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
361 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
362 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
363 The new algorithm is still used for 32 bit targets.
367 * Added a missing header for memcmp that caused compilation failure on some
372 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
374 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
375 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
376 This issue makes the RSA implementation with 2048 bit private keys
377 incorrect on such machines and memory corruption will happen during
378 the computation. As a consequence of the memory corruption an attacker
379 may be able to trigger a remote code execution on the machine performing
382 SSL/TLS servers or other servers using 2048 bit RSA private keys running
383 on machines supporting AVX512IFMA instructions of the X86_64 architecture
384 are affected by this issue.
389 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
390 implementation would not encrypt the entirety of the data under some
391 circumstances. This could reveal sixteen bytes of data that was
392 preexisting in the memory that wasn't written. In the special case of
393 "in place" encryption, sixteen bytes of the plaintext would be revealed.
395 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
396 they are both unaffected.
399 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
401 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
403 * In addition to the c_rehash shell command injection identified in
404 CVE-2022-1292, further bugs where the c_rehash script does not
405 properly sanitise shell metacharacters to prevent command injection have been
408 When the CVE-2022-1292 was fixed it was not discovered that there
409 are other places in the script where the file names of certificates
410 being hashed were possibly passed to a command executed through the shell.
412 This script is distributed by some operating systems in a manner where
413 it is automatically executed. On such operating systems, an attacker
414 could execute arbitrary commands with the privileges of the script.
416 Use of the c_rehash script is considered obsolete and should be replaced
417 by the OpenSSL rehash command line tool.
420 *Daniel Fiala, Tomáš Mráz*
422 * Case insensitive string comparison no longer uses locales. It has instead
423 been directly implemented.
427 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
429 * Case insensitive string comparison is reimplemented via new locale-agnostic
430 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
431 comparison. The previous implementation had problems when the Turkish locale
436 * Fixed a bug in the c_rehash script which was not properly sanitising shell
437 metacharacters to prevent command injection. This script is distributed by
438 some operating systems in a manner where it is automatically executed. On
439 such operating systems, an attacker could execute arbitrary commands with the
440 privileges of the script.
442 Use of the c_rehash script is considered obsolete and should be replaced
443 by the OpenSSL rehash command line tool.
448 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
449 certificate on an OCSP response. The bug caused the function in the case
450 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
451 response (meaning a successful verification) even in the case where the
452 response signing certificate fails to verify.
454 It is anticipated that most users of `OCSP_basic_verify` will not use the
455 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
456 a negative value (indicating a fatal error) in the case of a certificate
457 verification failure. The normal expected return value in this case would be
460 This issue also impacts the command line OpenSSL "ocsp" application. When
461 verifying an ocsp response with the "-no_cert_checks" option the command line
462 application will report that the verification is successful even though it
463 has in fact failed. In this case the incorrect successful response will also
464 be accompanied by error messages showing the failure and contradicting the
465 apparently successful result.
470 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
471 AAD data as the MAC key. This made the MAC key trivially predictable.
473 An attacker could exploit this issue by performing a man-in-the-middle attack
474 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
475 that the modified data would still pass the MAC integrity check.
477 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
478 endpoint will always be rejected by the recipient and the connection will
479 fail at that point. Many application protocols require data to be sent from
480 the client to the server first. Therefore, in such a case, only an OpenSSL
481 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
483 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
484 sent in both directions. In this case both clients and servers could be
485 affected, regardless of the application protocol.
487 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
488 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
489 the handshake when using this ciphersuite.
491 The confidentiality of data is not impacted by this issue, i.e. an attacker
492 cannot decrypt data that has been encrypted using this ciphersuite - they can
495 In order for this attack to work both endpoints must legitimately negotiate
496 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
497 OpenSSL 3.0, and is not available within the default provider or the default
498 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
499 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
500 following must have occurred:
502 1) OpenSSL must have been compiled with the (non-default) compile time option
503 enable-weak-ssl-ciphers
505 2) OpenSSL must have had the legacy provider explicitly loaded (either
506 through application code or via configuration)
508 3) The ciphersuite must have been explicitly added to the ciphersuite list
510 4) The libssl security level must have been set to 0 (default is 1)
512 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
514 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
515 others that both endpoints have in common
520 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
521 occuppied by the removed hash table entries.
523 This function is used when decoding certificates or keys. If a long lived
524 process periodically decodes certificates or keys its memory usage will
525 expand without bounds and the process might be terminated by the operating
526 system causing a denial of service. Also traversing the empty hash table
527 entries will take increasingly more time.
529 Typically such long lived processes might be TLS clients or TLS servers
530 configured to accept client certificate authentication.
533 *Hugo Landau, Aliaksei Levin*
535 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
536 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
537 statistics are no longer supported. For compatibility, these statistics are
538 still listed in the output but are now always reported as zero.
542 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
544 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
545 for non-prime moduli.
547 Internally this function is used when parsing certificates that contain
548 elliptic curve public keys in compressed form or explicit elliptic curve
549 parameters with a base point encoded in compressed form.
551 It is possible to trigger the infinite loop by crafting a certificate that
552 has invalid explicit curve parameters.
554 Since certificate parsing happens prior to verification of the certificate
555 signature, any process that parses an externally supplied certificate may thus
556 be subject to a denial of service attack. The infinite loop can also be
557 reached when parsing crafted private keys as they can contain explicit
558 elliptic curve parameters.
560 Thus vulnerable situations include:
562 - TLS clients consuming server certificates
563 - TLS servers consuming client certificates
564 - Hosting providers taking certificates or private keys from customers
565 - Certificate authorities parsing certification requests from subscribers
566 - Anything else which parses ASN.1 elliptic curve parameters
568 Also any other applications that use the BN_mod_sqrt() where the attacker
569 can control the parameter values are vulnerable to this DoS issue.
574 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
575 to the list of ciphersuites providing Perfect Forward Secrecy as
576 required by SECLEVEL >= 3.
578 *Dmitry Belyavskiy, Nicola Tuveri*
580 * Made the AES constant time code for no-asm configurations
581 optional due to the resulting 95% performance degradation.
582 The AES constant time code can be enabled, for no assembly
583 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
587 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
592 * The negative return value handling of the certificate verification callback
593 was reverted. The replacement is to set the verification retry state with
594 the SSL_set_retry_verify() function.
598 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
600 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
601 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
602 verify a certificate supplied by a server. That function may return a
603 negative return value to indicate an internal error (for example out of
604 memory). Such a negative return value is mishandled by OpenSSL and will cause
605 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
606 success and a subsequent call to SSL_get_error() to return the value
607 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
608 returned by OpenSSL if the application has previously called
609 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
610 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
611 totally unexpected and applications may not behave correctly as a result. The
612 exact behaviour will depend on the application but it could result in
613 crashes, infinite loops or other similar incorrect responses.
615 This issue is made more serious in combination with a separate bug in OpenSSL
616 3.0 that will cause X509_verify_cert() to indicate an internal error when
617 processing a certificate chain. This will occur where a certificate does not
618 include the Subject Alternative Name extension but where a Certificate
619 Authority has enforced name constraints. This issue can occur even with valid
625 * Corrected a few file name and file reference bugs in the build,
626 installation and setup scripts, which lead to installation verification
627 failures. Slightly enhanced the installation verification script.
631 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
636 * Fixed PVK encoder to properly query for the passphrase.
640 * Multiple fixes in the OSSL_HTTP API functions.
644 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
645 OSSL_PARAM_INTEGER data type and return error on negative numbers
646 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
647 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
651 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
655 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
659 * Multiple threading fixes.
663 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
667 * Allow fetching an operation from the provider that owns an unexportable key
668 as a fallback if that is still allowed by the property query.
672 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
674 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
679 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
680 S390X capability vector to zero. This simplifies testing of different code
681 paths on S390X architecture.
685 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
686 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
687 SP 800-38D". The communication will fail at this point.
691 * The EC_GROUP_clear_free() function is deprecated as there is nothing
692 confidential in EC_GROUP data.
696 * The byte order mark (BOM) character is ignored if encountered at the
697 beginning of a PEM-formatted file.
701 * Added CMS support for the Russian GOST algorithms.
705 * Due to move of the implementation of cryptographic operations
706 to the providers, validation of various operation parameters can
707 be postponed until the actual operation is executed where previously
708 it happened immediately when an operation parameter was set.
710 For example when setting an unsupported curve with
711 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
712 fail but later keygen operations with the EVP_PKEY_CTX will fail.
714 *OpenSSL team members and many third party contributors*
716 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
717 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
718 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
719 instead to retrieve these algorithms from a provider.
723 * On build targets where the multilib postfix is set in the build
724 configuration the libdir directory was changing based on whether
725 the lib directory with the multilib postfix exists on the system
726 or not. This unpredictable behavior was removed and eventual
727 multilib postfix is now always added to the default libdir. Use
728 `--libdir=lib` to override the libdir if adding the postfix is
733 * The triple DES key wrap functionality now conforms to RFC 3217 but is
734 no longer interoperable with OpenSSL 1.1.1.
738 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
739 function codes, this function can only cause problems for calling
744 * Add a configurable flag to output date formats as ISO 8601. Does not
745 change the default date format.
749 * Version of MSVC earlier than 1300 could get link warnings, which could
750 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
751 Support for this flag has been removed.
755 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
756 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
757 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
758 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
759 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
763 * The signatures of the functions to get and set options on SSL and
764 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
765 Some source code changes may be required.
769 * The public definitions of conf_method_st and conf_st have been
770 deprecated. They will be made opaque in a future release.
772 *Rich Salz and Tomáš Mráz*
774 * Client-initiated renegotiation is disabled by default. To allow it, use
775 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
776 flag, or the "ClientRenegotiation" config parameter as appropriate.
780 * Add "abspath" and "includedir" pragma's to config files, to prevent,
781 or modify relative pathname inclusion.
785 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
786 validated. Please consult the README-FIPS and
787 README-PROVIDERS files, as well as the migration guide.
789 *OpenSSL team members and many third party contributors*
791 * For the key types DH and DHX the allowed settable parameters are now different.
795 * The openssl commands that read keys, certificates, and CRLs now
796 automatically detect the PEM or DER format of the input files.
798 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
800 * Added enhanced PKCS#12 APIs which accept a library context.
804 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
808 * Added support for Kernel TLS (KTLS).
810 *Boris Pismenny, John Baldwin and Andrew Gallatin*
812 * Support for RFC 5746 secure renegotiation is now required by default for
813 SSL or TLS connections to succeed.
817 * The signature of the `copy` functional parameter of the
818 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
819 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
820 the signature of the `pub_decode` functional parameter of the
821 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
822 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
826 * The error return values from some control calls (ctrl) have changed.
830 * A public key check is now performed during EVP_PKEY_derive_set_peer().
834 * Many functions in the EVP_ namespace that are getters of values from
835 implementations or contexts were renamed to include get or get0 in their
836 names. Old names are provided as macro aliases for compatibility and
841 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
842 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
843 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
848 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
851 * The output from the command line applications may have minor
856 * The output from numerous "printing" may have minor changes.
860 * Windows thread synchronization uses read/write primitives (SRWLock) when
861 supported by the OS, otherwise CriticalSection continues to be used.
865 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
866 work on read only BIO source/sinks that do not support these functions.
867 This allows piping or redirection of a file BIO using stdin to be buffered
868 into memory. This is used internally in OSSL_DECODER_from_bio().
872 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
873 this function would return one of the values OSSL_STORE_INFO_NAME,
874 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
875 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
876 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
877 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
878 using this function should be amended to handle the changed return value.
882 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
883 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
884 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
885 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
886 contains more than one certificate identifier: This means that all
887 certificates referenced there MUST be part of the validation chain.
891 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
892 RC5, DESX and DES have been moved to the legacy provider.
896 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
897 RIPEMD-160 have been moved to the legacy provider.
901 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
906 * The deprecated functions EVP_PKEY_get0_RSA(),
907 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
908 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
909 well as the similarly named "get1" functions behave differently in
914 * A number of functions handling low-level keys or engines were deprecated
915 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
916 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
917 EVP_PKEY_get0_siphash().
921 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
922 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
923 will need to load the legacy crypto provider. This includes these PBE
924 algorithms which use this KDF:
925 - NID_pbeWithMD2AndDES_CBC
926 - NID_pbeWithMD5AndDES_CBC
927 - NID_pbeWithSHA1AndRC2_CBC
928 - NID_pbeWithMD2AndRC2_CBC
929 - NID_pbeWithMD5AndRC2_CBC
930 - NID_pbeWithSHA1AndDES_CBC
934 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
935 BIO_debug_callback() functions.
939 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
940 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
944 * The RAND_METHOD APIs have been deprecated.
948 * The SRP APIs have been deprecated.
952 * Add a compile time option to prevent the caching of provider fetched
953 algorithms. This is enabled by including the no-cached-fetch option
954 at configuration time.
958 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
959 count of PKCS12_DEFAULT_ITER.
961 *Tomáš Mráz and Sahana Prasad*
963 * The openssl speed command does not use low-level API calls anymore.
967 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
970 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
972 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
976 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
977 providers may supply their own group implementations (using either the "key
978 exchange" or the "key encapsulation" methods) which will automatically be
979 detected and used by libssl.
981 *Matt Caswell, Nicola Tuveri*
983 * The undocumented function X509_certificate_type() has been deprecated;
987 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
991 * Removed RSA padding mode for SSLv23 (which was only used for
992 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
993 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
998 * Deprecated the obsolete X9.31 RSA key generation related functions.
1000 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1001 is not allowed to return a value > 1, this is no more taken as failure.
1003 *Viktor Dukhovni and David von Oheimb*
1005 * Deprecated the obsolete X9.31 RSA key generation related functions
1006 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1007 BN_X931_generate_prime_ex().
1011 * The default key generation method for the regular 2-prime RSA keys was
1012 changed to the FIPS 186-4 B.3.6 method.
1016 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1020 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1024 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1025 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1027 *Rich Salz, Richard Levitte, and David von Oheimb*
1029 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1033 * Deprecated `OCSP_parse_url()`.
1037 * Validation of SM2 keys has been separated from the validation of regular EC
1042 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1043 switches: a validation failure triggers an early exit, returning a failure
1044 exit status to the parent process.
1048 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1049 to ignore unknown ciphers.
1053 * The `-cipher-commands` and `-digest-commands` options
1054 of the command line utility `list` have been deprecated.
1055 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1059 * Added convenience functions for generating asymmetric key pairs:
1060 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1061 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1065 * All of the low level EC_KEY functions have been deprecated.
1067 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1069 * Deprecated all the libcrypto and libssl error string loading
1074 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1075 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1080 * The `-crypt` option to the `passwd` command line tool has been removed.
1084 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1089 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1093 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1094 EVP_PKEY_get1_tls_encodedpoint().
1098 * The security callback, which can be customised by application code, supports
1099 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1100 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1104 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1105 interface. Their functionality remains unchanged.
1109 * Added new option for 'openssl list', '-providers', which will display the
1110 list of loaded providers, their names, version and status. It optionally
1111 displays their gettable parameters.
1115 * Removed EVP_PKEY_set_alias_type().
1119 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1120 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1124 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1125 parameters are still checked for type safety at compile time via helper
1130 * Remove the RAND_DRBG API
1132 *Paul Dale and Matthias St. Pierre*
1134 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1135 as well as actual hostnames.
1139 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1140 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1141 conversely, silently ignore DTLS protocol version bounds when configuring
1142 TLS-based contexts. The commands can be repeated to set bounds of both
1143 types. The same applies with the corresponding "min_protocol" and
1144 "max_protocol" command-line switches, in case some application uses both TLS
1147 SSL_CTX instances that are created for a fixed protocol version (e.g.
1148 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1149 attempts to apply bounds to these protocol versions would result in an
1150 error. Now only the "version-flexible" SSL_CTX instances are subject to
1151 limits in configuration files in command-line options.
1155 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1160 * Reworked the recorded ERR codes to make better space for system errors.
1161 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1162 given code is a system error (true) or an OpenSSL error (false).
1166 * Reworked the test perl framework to better allow parallel testing.
1168 *Nicola Tuveri and David von Oheimb*
1170 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1171 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1175 * 'Configure' has been changed to figure out the configuration target if
1176 none is given on the command line. Consequently, the 'config' script is
1177 now only a mere wrapper. All documentation is changed to only mention
1180 *Rich Salz and Richard Levitte*
1182 * Added a library context `OSSL_LIB_CTX` that applications as well as
1183 other libraries can use to form a separate context within which
1184 libcrypto operations are performed.
1188 * Added various `_ex` functions to the OpenSSL API that support using
1189 a non-default `OSSL_LIB_CTX`.
1193 * Handshake now fails if Extended Master Secret extension is dropped
1198 * Dropped interactive mode from the `openssl` program.
1202 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1204 *David von Oheimb and Shane Lontis*
1206 * Deprecated `EC_METHOD_get_field_type()`.
1210 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1211 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1212 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1216 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1220 * Add CAdES-BES signature verification support, mostly derived
1221 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1223 *Filipe Raimundo da Silva*
1225 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1229 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1230 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1234 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1238 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1239 EC_KEY_precompute_mult().
1243 * Deprecated EC_POINTs_mul().
1247 * Removed FIPS_mode() and FIPS_mode_set().
1251 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1255 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1256 EC_POINT_get_Jprojective_coordinates_GFp().
1260 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1261 arrays to be more easily constructed via a series of utility functions.
1262 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1263 the various push functions and finally convert to a passable OSSL_PARAM
1264 array using OSSL_PARAM_BLD_to_param().
1268 * The security strength of SHA1 and MD5 based signatures in TLS has been
1273 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1274 contain a provider side internal key.
1278 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1282 * Project text documents not yet having a proper file name extension
1283 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1284 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1288 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1289 have been converted to Markdown with the goal to produce documents
1290 which not only look pretty when viewed online in the browser, but
1291 remain well readable inside a plain text editor.
1293 To achieve this goal, a 'minimalistic' Markdown style has been applied
1294 which avoids formatting elements that interfere too much with the
1295 reading flow in the text file. For example, it
1297 * avoids [ATX headings][] and uses [setext headings][] instead
1298 (which works for `<h1>` and `<h2>` headings only).
1299 * avoids [inline links][] and uses [reference links][] instead.
1300 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1302 [ATX headings]: https://github.github.com/gfm/#atx-headings
1303 [setext headings]: https://github.github.com/gfm/#setext-headings
1304 [inline links]: https://github.github.com/gfm/#inline-link
1305 [reference links]: https://github.github.com/gfm/#reference-link
1306 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1307 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1309 *Matthias St. Pierre*
1311 * The test suite is changed to preserve results of each test recipe.
1312 A new directory test-runs/ with subdirectories named like the
1313 test recipes are created in the build tree for this purpose.
1317 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1318 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1319 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1321 *David von Oheimb, Martin Peylo*
1323 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1324 It supports arbitrary request and response content types, GET redirection,
1325 TLS, connections via HTTP(S) proxies, connections and exchange via
1326 user-defined BIOs (allowing implicit connections), persistent connections,
1327 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1328 The legacy OCSP-focused (and only partly documented) API
1329 is retained for backward compatibility, while most of it is deprecated.
1333 * Added `util/check-format.pl`, a tool for checking adherence to the
1334 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1335 The checks performed are incomplete and yield some false positives.
1336 Still the tool should be useful for detecting most typical glitches.
1340 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1341 If domain name resolution yields multiple IP addresses all of them are tried
1342 after `connect()` failures.
1346 * All of the low level RSA functions have been deprecated.
1350 * X509 certificates signed using SHA1 are no longer allowed at security
1355 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1356 modified to use PKEY APIs. These commands are now in maintenance mode
1357 and no new features will be added to them.
1361 * The command line utility rsautl has been deprecated.
1365 * The command line utilities genrsa and rsa have been modified to use PKEY
1366 APIs. They now write PKCS#8 keys by default. These commands are now in
1367 maintenance mode and no new features will be added to them.
1371 * All of the low level DH functions have been deprecated.
1373 *Paul Dale and Matt Caswell*
1375 * All of the low level DSA functions have been deprecated.
1379 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1380 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1384 * Deprecated low level ECDH and ECDSA functions.
1388 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1392 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1393 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1394 a new formulation to include all the things it can be used for,
1395 as well as words of caution.
1399 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1403 * All of the low level HMAC functions have been deprecated.
1405 *Paul Dale and David von Oheimb*
1407 * Over two thousand fixes were made to the documentation, including:
1408 - Common options (such as -rand/-writerand, TLS version control, etc)
1409 were refactored and point to newly-enhanced descriptions in openssl.pod.
1410 - Added style conformance for all options (with help from Richard Levitte),
1411 documented all reported missing options, added a CI build to check
1412 that all options are documented and that no unimplemented options
1414 - Documented some internals, such as all use of environment variables.
1415 - Addressed all internal broken L<> references.
1419 * All of the low level CMAC functions have been deprecated.
1423 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1424 functions have been deprecated.
1426 *Paul Dale and David von Oheimb*
1428 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1429 set of functions. The documentation mentioned negative values for some
1430 errors, but this was never the case, so the mention of negative values
1433 Code that followed the documentation and thereby check with something
1434 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1438 * All of the low level cipher functions have been deprecated.
1440 *Matt Caswell and Paul Dale*
1442 * Removed include/openssl/opensslconf.h.in and replaced it with
1443 include/openssl/configuration.h.in, which differs in not including
1444 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1445 was added to include both.
1447 This allows internal hacks where one might need to modify the set
1448 of configured macros, for example this if deprecated symbols are
1449 still supposed to be available internally:
1451 #include <openssl/configuration.h>
1453 #undef OPENSSL_NO_DEPRECATED
1454 #define OPENSSL_SUPPRESS_DEPRECATED
1456 #include <openssl/macros.h>
1458 This should not be used by applications that use the exported
1459 symbols, as that will lead to linking errors.
1463 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1464 used in exponentiation with 512-bit moduli. No EC algorithms are
1465 affected. Analysis suggests that attacks against 2-prime RSA1024,
1466 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1467 difficult to perform and are not believed likely. Attacks against DH512
1468 are considered just feasible. However, for an attack the target would
1469 have to re-use the DH512 private key, which is not recommended anyway.
1470 Also applications directly using the low-level API BN_mod_exp may be
1471 affected if they use BN_FLG_CONSTTIME.
1476 * Most memory-debug features have been deprecated, and the functionality
1477 replaced with no-ops.
1481 * Added documentation for the STACK API.
1485 * Introduced a new method type and API, OSSL_ENCODER, to represent
1486 generic encoders. These do the same sort of job that PEM writers
1487 and d2i functions do, but with support for methods supplied by
1488 providers, and the possibility for providers to support other
1493 * Introduced a new method type and API, OSSL_DECODER, to represent
1494 generic decoders. These do the same sort of job that PEM readers
1495 and i2d functions do, but with support for methods supplied by
1496 providers, and the possibility for providers to support other
1501 * Added a .pragma directive to the syntax of configuration files, to
1502 allow varying behavior in a supported and predictable manner.
1503 Currently added pragma:
1507 This allows dollar signs to be a keyword character unless it's
1508 followed by a opening brace or parenthesis. This is useful for
1509 platforms where dollar signs are commonly used in names, such as
1510 volume names and system directory names on VMS.
1514 * Added functionality to create an EVP_PKEY from user data.
1518 * Change the interpretation of the '--api' configuration option to
1519 mean that this is a desired API compatibility level with no
1520 further meaning. The previous interpretation, that this would
1521 also mean to remove all deprecated symbols up to and including
1522 the given version, no requires that 'no-deprecated' is also used
1523 in the configuration.
1525 When building applications, the desired API compatibility level
1526 can be set with the OPENSSL_API_COMPAT macro like before. For
1527 API compatibility version below 3.0, the old style numerical
1528 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1529 For version 3.0 and on, the value is expected to be the decimal
1530 value calculated from the major and minor version like this:
1532 MAJOR * 10000 + MINOR * 100
1536 -DOPENSSL_API_COMPAT=30000 For 3.0
1537 -DOPENSSL_API_COMPAT=30200 For 3.2
1539 To hide declarations that are deprecated up to and including the
1540 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1541 given when building the application as well.
1545 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1546 access to certificate and CRL stores via URIs and OSSL_STORE
1549 This adds the following functions:
1551 - X509_LOOKUP_store()
1552 - X509_STORE_load_file()
1553 - X509_STORE_load_path()
1554 - X509_STORE_load_store()
1555 - SSL_add_store_cert_subjects_to_stack()
1556 - SSL_CTX_set_default_verify_store()
1557 - SSL_CTX_load_verify_file()
1558 - SSL_CTX_load_verify_dir()
1559 - SSL_CTX_load_verify_store()
1563 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1564 The presence of this system service is determined at run-time.
1568 * Added functionality to create an EVP_PKEY context based on data
1569 for methods from providers. This takes an algorithm name and a
1570 property query string and simply stores them, with the intent
1571 that any operation that uses this context will use those strings
1572 to fetch the needed methods implicitly, thereby making the port
1573 of application written for pre-3.0 OpenSSL easier.
1577 * The undocumented function NCONF_WIN32() has been deprecated; for
1578 conversion details see the HISTORY section of doc/man5/config.pod
1582 * Introduced the new functions EVP_DigestSignInit_ex() and
1583 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1584 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1585 pages for further details.
1589 * Over two thousand fixes were made to the documentation, including:
1590 adding missing command flags, better style conformance, documentation
1593 *Rich Salz, Richard Levitte*
1595 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1596 X25519, X448, Ed25519 and Ed448.
1600 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1605 * Deprecated the public definition of `ERR_STATE` as well as the function
1606 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1611 * Added ERR functionality to give callers access to the stored function
1612 names that have replaced the older function code based functions.
1614 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1615 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1616 ERR_peek_error_all() and ERR_peek_last_error_all().
1618 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1619 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1620 ERR_func_error_string().
1624 * Extended testing to be verbose for failing tests only. The make variables
1625 VERBOSE_FAILURE or VF can be used to enable this:
1627 $ make VF=1 test # Unix
1628 $ mms /macro=(VF=1) test ! OpenVMS
1629 $ nmake VF=1 test # Windows
1633 * Added the `-copy_extensions` option to the `x509` command for use with
1634 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1635 all extensions in the request are copied to the certificate or vice versa.
1637 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1639 * Added the `-copy_extensions` option to the `req` command for use with
1640 `-x509`. When given with the `copy` or `copyall` argument,
1641 all extensions in the certification request are copied to the certificate.
1645 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1646 they generate are by default RFC 5280 compliant in the following sense:
1647 There is a subjectKeyIdentifier extension with a hash value of the public key
1648 and for not self-signed certs there is an authorityKeyIdentifier extension
1649 with a keyIdentifier field or issuer information identifying the signing key.
1650 This is done unless some configuration overrides the new default behavior,
1651 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1655 * Added several checks to `X509_verify_cert()` according to requirements in
1656 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1657 (which may be done by using the CLI option `-x509_strict`):
1658 * The basicConstraints of CA certificates must be marked critical.
1659 * CA certificates must explicitly include the keyUsage extension.
1660 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1661 * The issuer name of any certificate must not be empty.
1662 * The subject name of CA certs, certs with keyUsage crlSign,
1663 and certs without subjectAlternativeName must not be empty.
1664 * If a subjectAlternativeName extension is given it must not be empty.
1665 * The signatureAlgorithm field and the cert signature must be consistent.
1666 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1667 must not be marked critical.
1668 * The authorityKeyIdentifier must be given for X.509v3 certs
1669 unless they are self-signed.
1670 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1674 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1675 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1679 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1680 used even when parsing explicit parameters, when loading a encoded key
1681 or calling `EC_GROUP_new_from_ecpkparameters()`/
1682 `EC_GROUP_new_from_ecparameters()`.
1683 This prevents bypass of security hardening and performance gains,
1684 especially for curves with specialized EC_METHODs.
1685 By default, if a key encoded with explicit parameters is loaded and later
1686 encoded, the output is still encoded with explicit parameters, even if
1687 internally a "named" EC_GROUP is used for computation.
1691 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1692 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1693 NULL. After this change, only the cofactor parameter can be NULL. It also
1694 does some minimal sanity checks on the passed order.
1699 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1700 An attack is simple, if the first CMS_recipientInfo is valid but the
1701 second CMS_recipientInfo is chosen ciphertext. If the second
1702 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1703 encryption key will be replaced by garbage, and the message cannot be
1704 decoded, but if the RSA decryption fails, the correct encryption key is
1705 used and the recipient will not notice the attack.
1706 As a work around for this potential attack the length of the decrypted
1707 key must be equal to the cipher default key length, in case the
1708 certifiate is not given and all recipientInfo are tried out.
1709 The old behaviour can be re-enabled in the CMS code by setting the
1710 CMS_DEBUG_DECRYPT flag.
1714 * Early start up entropy quality from the DEVRANDOM seed source has been
1715 improved for older Linux systems. The RAND subsystem will wait for
1716 /dev/random to be producing output before seeding from /dev/urandom.
1717 The seeded state is stored for future library initialisations using
1718 a system global shared memory segment. The shared memory identifier
1719 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1720 the desired value. The default identifier is 114.
1724 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1725 when primes for RSA keys are computed.
1726 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1727 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1728 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1729 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1730 This avoids possible fingerprinting of newly generated RSA modules.
1734 * Correct the extended master secret constant on EBCDIC systems. Without this
1735 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1736 negotiate EMS will fail. Unfortunately this also means that TLS connections
1737 between EBCDIC systems with this fix, and EBCDIC systems without this
1738 fix will fail if they negotiate EMS.
1742 * Changed the library initialisation so that the config file is now loaded
1743 by default. This was already the case for libssl. It now occurs for both
1744 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1745 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1749 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1750 where the former acts as a replacement for `ERR_put_error()`, and the
1751 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1752 `ERR_raise_data()` adds more flexibility by taking a format string and
1753 an arbitrary number of arguments following it, to be processed with
1758 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1759 to check if a named provider is loaded and available. When called, it
1760 will also activate all fallback providers if such are still present.
1764 * Enforce a minimum DH modulus size of 512 bits.
1768 * Changed DH parameters to generate the order q subgroup instead of 2q.
1769 Previously generated DH parameters are still accepted by DH_check
1770 but DH_generate_key works around that by clearing bit 0 of the
1771 private key for those. This avoids leaking bit 0 of the private key.
1775 * Significantly reduce secure memory usage by the randomness pools.
1779 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1784 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1785 algorithms. An implementation of a key exchange algorithm can be obtained
1786 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1787 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1788 the older EVP_PKEY_derive_init() function. See the man pages for the new
1789 functions for further details.
1793 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1797 * Removed the function names from error messages and deprecated the
1802 * Removed NextStep support and the macro OPENSSL_UNISTD
1806 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1807 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1808 Also removed "export var as function" capability; we do not export
1809 variables, only functions.
1813 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1814 an error and 1 indicating success. In previous versions of OpenSSL this
1815 was a void type. If a key was set longer than the maximum possible this
1820 * Support SM2 signing and verification schemes with X509 certificate.
1824 * Use SHA256 as the default digest for TS query in the `ts` app.
1828 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1832 * Default cipher lists/suites are now available via a function, the
1833 #defines are deprecated.
1837 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1838 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1839 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1843 * Join the directories crypto/x509 and crypto/x509v3
1847 * Added command 'openssl kdf' that uses the EVP_KDF API.
1851 * Added command 'openssl mac' that uses the EVP_MAC API.
1855 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1856 as default directories. Also added the command 'openssl info'
1857 for scripting purposes.
1861 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1866 * Add prediction resistance to the DRBG reseeding process.
1870 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1871 mandated by IEEE Std 1619-2018.
1875 * Added newline escaping functionality to a filename when using openssl dgst.
1876 This output format is to replicate the output format found in the `*sum`
1877 checksum programs. This aims to preserve backward compatibility.
1879 *Matt Eaton, Richard Levitte, and Paul Dale*
1881 * Removed the heartbeat message in DTLS feature, as it has very
1882 little usage and doesn't seem to fulfill a valuable purpose.
1883 The configuration option is now deprecated.
1887 * Changed the output of 'openssl {digestname} < file' to display the
1888 digest name in its output.
1892 * Added a new generic trace API which provides support for enabling
1893 instrumentation through trace output.
1895 *Richard Levitte & Matthias St. Pierre*
1897 * Added build tests for C++. These are generated files that only do one
1898 thing, to include one public OpenSSL head file each. This tests that
1899 the public header files can be usefully included in a C++ application.
1901 This test isn't enabled by default. It can be enabled with the option
1902 'enable-buildtest-c++'.
1906 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1910 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1914 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1918 * Added KMAC to EVP_MAC.
1922 * Added property based algorithm implementation selection framework to
1927 * Added SCA hardening for modular field inversion in EC_GROUP through
1928 a new dedicated field_inv() pointer in EC_METHOD.
1929 This also addresses a leakage affecting conversions from projective
1930 to affine coordinates.
1932 *Billy Bob Brumley, Nicola Tuveri*
1934 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1935 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1936 those algorithms that were already supported through the EVP_PKEY API
1937 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1938 and scrypt are now wrappers that call EVP_KDF.
1942 * Build devcrypto engine as a dynamic engine.
1944 *Eneas U de Queiroz*
1946 * Add keyed BLAKE2 to EVP_MAC.
1950 * Fix a bug in the computation of the endpoint-pair shared secret used
1951 by DTLS over SCTP. This breaks interoperability with older versions
1952 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1953 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1954 interoperability with such broken implementations. However, enabling
1955 this switch breaks interoperability with correct implementations.
1957 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1958 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1962 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1966 * Changed the license to the Apache License v2.0.
1970 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1972 - Major releases (indicated by incrementing the MAJOR release number)
1973 may introduce incompatible API/ABI changes.
1974 - Minor releases (indicated by incrementing the MINOR release number)
1975 may introduce new features but retain API/ABI compatibility.
1976 - Patch releases (indicated by incrementing the PATCH number)
1977 are intended for bug fixes and other improvements of existing
1978 features only (like improving performance or adding documentation)
1979 and retain API/ABI compatibility.
1983 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1987 * Remove the 'dist' target and add a tarball building script. The
1988 'dist' target has fallen out of use, and it shouldn't be
1989 necessary to configure just to create a source distribution.
1993 * Recreate the OS390-Unix config target. It no longer relies on a
1994 special script like it did for OpenSSL pre-1.1.0.
1998 * Instead of having the source directories listed in Configure, add
1999 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2004 * Add GMAC to EVP_MAC.
2008 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2012 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2013 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2014 to facilitate the continued use of MACs through raw private keys in
2015 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2019 * Deprecate ECDH_KDF_X9_62().
2023 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2024 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2025 are retained for backwards compatibility.
2029 * AES-XTS mode now enforces that its two keys are different to mitigate
2030 the attacked described in "Efficient Instantiations of Tweakable
2031 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2032 Details of this attack can be obtained from:
2033 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2037 * Rename the object files, i.e. give them other names than in previous
2038 versions. Their names now include the name of the final product, as
2039 well as its type mnemonic (bin, lib, shlib).
2043 * Added new option for 'openssl list', '-objects', which will display the
2044 list of built in objects, i.e. OIDs with names.
2048 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2049 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2054 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2055 improves application performance by removing data copies and providing
2056 applications with zero-copy system calls such as sendfile and splice.
2060 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2064 * `PKCS12_parse` now maintains the order of the parsed certificates
2065 when outputting them via `*ca` (rather than reversing it).
2069 * Deprecated pthread fork support methods.
2073 * Added support for FFDHE key exchange in TLS 1.3.
2077 * Added a new concept for OpenSSL plugability: providers. This
2078 functionality is designed to replace the ENGINE API and ENGINE
2079 implementations, and to be much more dynamic, allowing provider
2080 authors to introduce new algorithms among other things, as long as
2081 there's an API that supports the algorithm type.
2083 With this concept comes a new core API for interaction between
2084 libcrypto and provider implementations. Public libcrypto functions
2085 that want to use providers do so through this core API.
2087 The main documentation for this core API is found in
2088 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2089 refer to other manuals describing the API specific for supported
2090 algorithm types (also called operations).
2097 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2099 * Avoid loading of a dynamic engine twice.
2103 * Prioritise DANE TLSA issuer certs over peer certs
2107 * Fixed random API for MacOS prior to 10.12
2109 These MacOS versions don't support the CommonCrypto APIs
2113 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2115 * Fixed an SM2 Decryption Buffer Overflow.
2117 In order to decrypt SM2 encrypted data an application is expected to
2118 call the API function EVP_PKEY_decrypt(). Typically an application will
2119 call this function twice. The first time, on entry, the "out" parameter
2120 can be NULL and, on exit, the "outlen" parameter is populated with the
2121 buffer size required to hold the decrypted plaintext. The application
2122 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2123 again, but this time passing a non-NULL value for the "out" parameter.
2125 A bug in the implementation of the SM2 decryption code means that the
2126 calculation of the buffer size required to hold the plaintext returned
2127 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2128 size required by the second call. This can lead to a buffer overflow
2129 when EVP_PKEY_decrypt() is called by the application a second time with
2130 a buffer that is too small.
2132 A malicious attacker who is able present SM2 content for decryption to
2133 an application could cause attacker chosen data to overflow the buffer
2134 by up to a maximum of 62 bytes altering the contents of other data held
2135 after the buffer, possibly changing application behaviour or causing
2136 the application to crash. The location of the buffer is application
2137 dependent but is typically heap allocated.
2142 * Fixed various read buffer overruns processing ASN.1 strings
2144 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2145 structure which contains a buffer holding the string data and a field
2146 holding the buffer length. This contrasts with normal C strings which
2147 are repesented as a buffer for the string data which is terminated
2148 with a NUL (0) byte.
2150 Although not a strict requirement, ASN.1 strings that are parsed using
2151 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2152 well as any string whose value has been set with the ASN1_STRING_set()
2153 function will additionally NUL terminate the byte array in the
2154 ASN1_STRING structure.
2156 However, it is possible for applications to directly construct valid
2157 ASN1_STRING structures which do not NUL terminate the byte array by
2158 directly setting the "data" and "length" fields in the ASN1_STRING
2159 array. This can also happen by using the ASN1_STRING_set0() function.
2161 Numerous OpenSSL functions that print ASN.1 data have been found to
2162 assume that the ASN1_STRING byte array will be NUL terminated, even
2163 though this is not guaranteed for strings that have been directly
2164 constructed. Where an application requests an ASN.1 structure to be
2165 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2166 been directly constructed by the application without NUL terminating
2167 the "data" field, then a read buffer overrun can occur.
2169 The same thing can also occur during name constraints processing
2170 of certificates (for example if a certificate has been directly
2171 constructed by the application instead of loading it via the OpenSSL
2172 parsing functions, and the certificate contains non NUL terminated
2173 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2174 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2176 If a malicious actor can cause an application to directly construct an
2177 ASN1_STRING and then process it through one of the affected OpenSSL
2178 functions then this issue could be hit. This might result in a crash
2179 (causing a Denial of Service attack). It could also result in the
2180 disclosure of private memory contents (such as private keys, or
2181 sensitive plaintext).
2186 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2188 * Fixed a problem with verifying a certificate chain when using the
2189 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2190 the certificates present in a certificate chain. It is not set by default.
2192 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2193 the chain that have explicitly encoded elliptic curve parameters was added
2194 as an additional strict check.
2196 An error in the implementation of this check meant that the result of a
2197 previous check to confirm that certificates in the chain are valid CA
2198 certificates was overwritten. This effectively bypasses the check
2199 that non-CA certificates must not be able to issue other certificates.
2201 If a "purpose" has been configured then there is a subsequent opportunity
2202 for checks that the certificate is a valid CA. All of the named "purpose"
2203 values implemented in libcrypto perform this check. Therefore, where
2204 a purpose is set the certificate chain will still be rejected even when the
2205 strict flag has been used. A purpose is set by default in libssl client and
2206 server certificate verification routines, but it can be overridden or
2207 removed by an application.
2209 In order to be affected, an application must explicitly set the
2210 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2211 for the certificate verification or, in the case of TLS client or server
2212 applications, override the default purpose.
2217 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2218 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2219 renegotiation ClientHello omits the signature_algorithms extension (where it
2220 was present in the initial ClientHello), but includes a
2221 signature_algorithms_cert extension then a NULL pointer dereference will
2222 result, leading to a crash and a denial of service attack.
2224 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2225 (which is the default configuration). OpenSSL TLS clients are not impacted by
2229 *Peter Kästle and Samuel Sapalski*
2231 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2233 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2234 create a unique hash value based on the issuer and serial number data
2235 contained within an X509 certificate. However it was failing to correctly
2236 handle any errors that may occur while parsing the issuer field (which might
2237 occur if the issuer field is maliciously constructed). This may subsequently
2238 result in a NULL pointer deref and a crash leading to a potential denial of
2244 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2245 padding mode to correctly check for rollback attacks. This is considered a
2246 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2251 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2252 functions. Previously they could overflow the output length argument in some
2253 cases where the input length is close to the maximum permissable length for
2254 an integer on the platform. In such cases the return value from the function
2255 call would be 1 (indicating success), but the output length value would be
2256 negative. This could cause applications to behave incorrectly or crash.
2261 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2262 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2263 could be exploited in a side channel attack to recover the password. Since
2264 the attack is local host only this is outside of the current OpenSSL
2265 threat model and therefore no CVE is assigned.
2267 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2272 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2274 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2275 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2276 If an attacker can control both items being compared then this could lead
2277 to a possible denial of service attack. OpenSSL itself uses the
2278 GENERAL_NAME_cmp function for two purposes:
2279 1) Comparing CRL distribution point names between an available CRL and a
2280 CRL distribution point embedded in an X509 certificate
2281 2) When verifying that a timestamp response token signer matches the
2282 timestamp authority name (exposed via the API functions
2283 TS_RESP_verify_response and TS_RESP_verify_token)
2288 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2290 * Certificates with explicit curve parameters are now disallowed in
2291 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2295 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2296 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2297 conversely, silently ignore DTLS protocol version bounds when configuring
2298 TLS-based contexts. The commands can be repeated to set bounds of both
2299 types. The same applies with the corresponding "min_protocol" and
2300 "max_protocol" command-line switches, in case some application uses both TLS
2303 SSL_CTX instances that are created for a fixed protocol version (e.g.
2304 TLSv1_server_method()) also silently ignore version bounds. Previously
2305 attempts to apply bounds to these protocol versions would result in an
2306 error. Now only the "version-flexible" SSL_CTX instances are subject to
2307 limits in configuration files in command-line options.
2311 * Handshake now fails if Extended Master Secret extension is dropped
2316 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2318 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2320 * Fixed segmentation fault in SSL_check_chain()
2321 Server or client applications that call the SSL_check_chain() function
2322 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2323 dereference as a result of incorrect handling of the
2324 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2325 or unrecognised signature algorithm is received from the peer. This could
2326 be exploited by a malicious peer in a Denial of Service attack.
2331 * Added AES consttime code for no-asm configurations
2332 an optional constant time support for AES was added
2333 when building openssl for no-asm.
2334 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2335 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2336 At this time this feature is by default disabled.
2337 It will be enabled by default in 3.0.
2341 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2343 * Revert the change of EOF detection while reading in libssl to avoid
2344 regressions in applications depending on the current way of reporting
2345 the EOF. As the existing method is not fully accurate the change to
2346 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2347 branch and will be present in the 3.0 release.
2351 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2352 when primes for RSA keys are computed.
2353 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2354 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2355 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2356 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2357 This avoids possible fingerprinting of newly generated RSA modules.
2361 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2363 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2364 while reading in libssl then we would report an error back to the
2365 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2366 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2367 therefore give a hint as to what went wrong.
2371 * Check that ed25519 and ed448 are allowed by the security level. Previously
2372 signature algorithms not using an MD were not being checked that they were
2373 allowed by the security level.
2377 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2378 was not quite right. The behaviour was not consistent between resumption
2379 and normal handshakes, and also not quite consistent with historical
2380 behaviour. The behaviour in various scenarios has been clarified and
2381 it has been updated to make it match historical behaviour as closely as
2386 * *[VMS only]* The header files that the VMS compilers include automatically,
2387 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2388 that the C++ compiler doesn't understand. This is a shortcoming in the
2389 compiler, but can be worked around with `__cplusplus` guards.
2391 C++ applications that use OpenSSL libraries must be compiled using the
2392 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2393 functions. Otherwise, only functions with symbols of less than 31
2394 characters can be used, as the linker will not be able to successfully
2395 resolve symbols with longer names.
2399 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2400 The presence of this system service is determined at run-time.
2404 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2409 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2411 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2412 number generator (RNG). This was intended to include protection in the
2413 event of a fork() system call in order to ensure that the parent and child
2414 processes did not share the same RNG state. However this protection was not
2415 being used in the default case.
2417 A partial mitigation for this issue is that the output from a high
2418 precision timer is mixed into the RNG state so the likelihood of a parent
2419 and child process sharing state is significantly reduced.
2421 If an application already calls OPENSSL_init_crypto() explicitly using
2422 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2425 *Matthias St. Pierre*
2427 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2428 used even when parsing explicit parameters, when loading a encoded key
2429 or calling `EC_GROUP_new_from_ecpkparameters()`/
2430 `EC_GROUP_new_from_ecparameters()`.
2431 This prevents bypass of security hardening and performance gains,
2432 especially for curves with specialized EC_METHODs.
2433 By default, if a key encoded with explicit parameters is loaded and later
2434 encoded, the output is still encoded with explicit parameters, even if
2435 internally a "named" EC_GROUP is used for computation.
2439 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2440 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2441 NULL. After this change, only the cofactor parameter can be NULL. It also
2442 does some minimal sanity checks on the passed order.
2447 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2448 An attack is simple, if the first CMS_recipientInfo is valid but the
2449 second CMS_recipientInfo is chosen ciphertext. If the second
2450 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2451 encryption key will be replaced by garbage, and the message cannot be
2452 decoded, but if the RSA decryption fails, the correct encryption key is
2453 used and the recipient will not notice the attack.
2454 As a work around for this potential attack the length of the decrypted
2455 key must be equal to the cipher default key length, in case the
2456 certifiate is not given and all recipientInfo are tried out.
2457 The old behaviour can be re-enabled in the CMS code by setting the
2458 CMS_DEBUG_DECRYPT flag.
2463 * Early start up entropy quality from the DEVRANDOM seed source has been
2464 improved for older Linux systems. The RAND subsystem will wait for
2465 /dev/random to be producing output before seeding from /dev/urandom.
2466 The seeded state is stored for future library initialisations using
2467 a system global shared memory segment. The shared memory identifier
2468 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2469 the desired value. The default identifier is 114.
2473 * Correct the extended master secret constant on EBCDIC systems. Without this
2474 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2475 negotiate EMS will fail. Unfortunately this also means that TLS connections
2476 between EBCDIC systems with this fix, and EBCDIC systems without this
2477 fix will fail if they negotiate EMS.
2481 * Use Windows installation paths in the mingw builds
2483 Mingw isn't a POSIX environment per se, which means that Windows
2484 paths should be used for installation.
2489 * Changed DH_check to accept parameters with order q and 2q subgroups.
2490 With order 2q subgroups the bit 0 of the private key is not secret
2491 but DH_generate_key works around that by clearing bit 0 of the
2492 private key for those. This avoids leaking bit 0 of the private key.
2496 * Significantly reduce secure memory usage by the randomness pools.
2500 * Revert the DEVRANDOM_WAIT feature for Linux systems
2502 The DEVRANDOM_WAIT feature added a select() call to wait for the
2503 /dev/random device to become readable before reading from the
2504 /dev/urandom device.
2506 It turned out that this change had negative side effects on
2507 performance which were not acceptable. After some discussion it
2508 was decided to revert this feature and leave it up to the OS
2509 resp. the platform maintainer to ensure a proper initialization
2510 during early boot time.
2512 *Matthias St. Pierre*
2514 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2516 * Add build tests for C++. These are generated files that only do one
2517 thing, to include one public OpenSSL head file each. This tests that
2518 the public header files can be usefully included in a C++ application.
2520 This test isn't enabled by default. It can be enabled with the option
2521 'enable-buildtest-c++'.
2525 * Enable SHA3 pre-hashing for ECDSA and DSA.
2529 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2530 This changes the size when using the `genpkey` command when no size is given.
2531 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2532 generation commands to use 2048 bits by default.
2536 * Reorganize the manual pages to consistently have RETURN VALUES,
2537 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2538 util/fix-doc-nits accordingly.
2540 *Paul Yang, Joshua Lock*
2542 * Add the missing accessor EVP_PKEY_get0_engine()
2546 * Have commands like `s_client` and `s_server` output the signature scheme
2547 along with other cipher suite parameters when debugging.
2551 * Make OPENSSL_config() error agnostic again.
2555 * Do the error handling in RSA decryption constant time.
2559 * Prevent over long nonces in ChaCha20-Poly1305.
2561 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2562 for every encryption operation. RFC 7539 specifies that the nonce value
2563 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2564 and front pads the nonce with 0 bytes if it is less than 12
2565 bytes. However it also incorrectly allows a nonce to be set of up to 16
2566 bytes. In this case only the last 12 bytes are significant and any
2567 additional leading bytes are ignored.
2569 It is a requirement of using this cipher that nonce values are
2570 unique. Messages encrypted using a reused nonce value are susceptible to
2571 serious confidentiality and integrity attacks. If an application changes
2572 the default nonce length to be longer than 12 bytes and then makes a
2573 change to the leading bytes of the nonce expecting the new value to be a
2574 new unique nonce then such an application could inadvertently encrypt
2575 messages with a reused nonce.
2577 Additionally the ignored bytes in a long nonce are not covered by the
2578 integrity guarantee of this cipher. Any application that relies on the
2579 integrity of these ignored leading bytes of a long nonce may be further
2580 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2581 is safe because no such use sets such a long nonce value. However user
2582 applications that use this cipher directly and set a non-default nonce
2583 length to be longer than 12 bytes may be vulnerable.
2585 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2591 * Add DEVRANDOM_WAIT feature for Linux systems
2593 On older Linux systems where the getrandom() system call is not available,
2594 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2595 Contrary to getrandom(), the /dev/urandom device will not block during
2596 early boot when the kernel CSPRNG has not been seeded yet.
2598 To mitigate this known weakness, use select() to wait for /dev/random to
2599 become readable before reading from /dev/urandom.
2601 * Ensure that SM2 only uses SM3 as digest algorithm
2605 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2607 * Change the info callback signals for the start and end of a post-handshake
2608 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2609 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2610 confused by this and assume that a TLSv1.2 renegotiation has started. This
2611 can break KeyUpdate handling. Instead we no longer signal the start and end
2612 of a post handshake message exchange (although the messages themselves are
2613 still signalled). This could break some applications that were expecting
2614 the old signals. However without this KeyUpdate is not usable for many
2619 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2621 * Timing vulnerability in DSA signature generation
2623 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2624 timing side channel attack. An attacker could use variations in the signing
2625 algorithm to recover the private key.
2627 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2632 * Timing vulnerability in ECDSA signature generation
2634 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2635 timing side channel attack. An attacker could use variations in the signing
2636 algorithm to recover the private key.
2638 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2643 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2644 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2645 of two gigabytes and the error handling improved.
2647 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2648 categorized as a normal bug, not a security issue, because the DRBG reseeds
2649 automatically and is fully functional even without additional randomness
2650 provided by the application.
2652 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2654 * Add a new ClientHello callback. Provides a callback interface that gives
2655 the application the ability to adjust the nascent SSL object at the
2656 earliest stage of ClientHello processing, immediately after extensions have
2657 been collected but before they have been processed. In particular, this
2658 callback can adjust the supported TLS versions in response to the contents
2663 * Add SM2 base algorithm support.
2667 * s390x assembly pack: add (improved) hardware-support for the following
2668 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2669 aes-cfb/cfb8, aes-ecb.
2673 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2674 parameter is no longer accepted, as it leads to a corrupt table. NULL
2675 pem_str is reserved for alias entries only.
2679 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2680 step for prime curves. The new implementation is based on formulae from
2681 differential addition-and-doubling in homogeneous projective coordinates
2682 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2683 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2684 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2685 to work in projective coordinates.
2687 *Billy Bob Brumley, Nicola Tuveri*
2689 * Change generating and checking of primes so that the error rate of not
2690 being prime depends on the intended use based on the size of the input.
2691 For larger primes this will result in more rounds of Miller-Rabin.
2692 The maximal error rate for primes with more than 1080 bits is lowered
2695 *Kurt Roeckx, Annie Yousar*
2697 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2701 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2702 moving between systems, and to avoid confusion when a Windows build is
2703 done with mingw vs with MSVC. For POSIX installs, there's still a
2704 symlink or copy named 'tsget' to avoid that confusion as well.
2708 * Revert blinding in ECDSA sign and instead make problematic addition
2709 length-invariant. Switch even to fixed-length Montgomery multiplication.
2713 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2714 step for binary curves. The new implementation is based on formulae from
2715 differential addition-and-doubling in mixed Lopez-Dahab projective
2716 coordinates, modified to independently blind the operands.
2718 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2720 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2721 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2722 EC_METHODs to implement their own specialized "ladder step", to take
2723 advantage of more favorable coordinate systems or more efficient
2724 differential addition-and-doubling algorithms.
2726 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2728 * Modified the random device based seed sources to keep the relevant
2729 file descriptors open rather than reopening them on each access.
2730 This allows such sources to operate in a chroot() jail without
2731 the associated device nodes being available. This behaviour can be
2732 controlled using RAND_keep_random_devices_open().
2736 * Numerous side-channel attack mitigations have been applied. This may have
2737 performance impacts for some algorithms for the benefit of improved
2738 security. Specific changes are noted in this change log by their respective
2743 * AIX shared library support overhaul. Switch to AIX "natural" way of
2744 handling shared libraries, which means collecting shared objects of
2745 different versions and bitnesses in one common archive. This allows to
2746 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2747 doesn't affect the way 3rd party applications are linked, only how
2748 multi-version installation is managed.
2752 * Make ec_group_do_inverse_ord() more robust and available to other
2753 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2754 mitigations are applied to the fallback BN_mod_inverse().
2755 When using this function rather than BN_mod_inverse() directly, new
2756 EC cryptosystem implementations are then safer-by-default.
2760 * Add coordinate blinding for EC_POINT and implement projective
2761 coordinate blinding for generic prime curves as a countermeasure to
2762 chosen point SCA attacks.
2764 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2766 * Add blinding to ECDSA and DSA signatures to protect against side channel
2767 attacks discovered by Keegan Ryan (NCC Group).
2771 * Enforce checking in the `pkeyutl` command to ensure that the input
2772 length does not exceed the maximum supported digest length when performing
2773 a sign, verify or verifyrecover operation.
2777 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2778 I/O in combination with something like select() or poll() will hang. This
2779 can be turned off again using SSL_CTX_clear_mode().
2780 Many applications do not properly handle non-application data records, and
2781 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2782 around the problems in those applications, but can also break some.
2783 It's recommended to read the manpages about SSL_read(), SSL_write(),
2784 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2785 SSL_CTX_set_read_ahead() again.
2789 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2790 now allow empty (zero character) pass phrases.
2794 * Apply blinding to binary field modular inversion and remove patent
2795 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2799 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2800 binary and prime elliptic curves.
2804 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2805 constant time fixed point multiplication.
2809 * Revise elliptic curve scalar multiplication with timing attack
2810 defenses: ec_wNAF_mul redirects to a constant time implementation
2811 when computing fixed point and variable point multiplication (which
2812 in OpenSSL are mostly used with secret scalars in keygen, sign,
2813 ECDH derive operations).
2814 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2817 * Updated CONTRIBUTING
2821 * Updated DRBG / RAND to request nonce and additional low entropy
2822 randomness from the system.
2824 *Matthias St. Pierre*
2826 * Updated 'openssl rehash' to use OpenSSL consistent default.
2830 * Moved the load of the ssl_conf module to libcrypto, which helps
2831 loading engines that libssl uses before libssl is initialised.
2835 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2839 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2841 *Ingo Schwarze, Rich Salz*
2843 * Added output of accepting IP address and port for 'openssl s_server'
2847 * Added a new API for TLSv1.3 ciphersuites:
2848 SSL_CTX_set_ciphersuites()
2849 SSL_set_ciphersuites()
2853 * Memory allocation failures consistently add an error to the error
2858 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2859 in libcrypto when run as setuid/setgid.
2863 * Load any config file by default when libssl is used.
2867 * Added new public header file <openssl/rand_drbg.h> and documentation
2868 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2870 *Matthias St. Pierre*
2872 * QNX support removed (cannot find contributors to get their approval
2873 for the license change).
2877 * TLSv1.3 replay protection for early data has been implemented. See the
2878 SSL_read_early_data() man page for further details.
2882 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2883 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2884 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2885 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2886 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2887 configuration has been separated out. See the ciphers man page or the
2888 SSL_CTX_set_ciphersuites() man page for more information.
2892 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2893 in responder mode now supports the new "-multi" option, which
2894 spawns the specified number of child processes to handle OCSP
2895 requests. The "-timeout" option now also limits the OCSP
2896 responder's patience to wait to receive the full client request
2897 on a newly accepted connection. Child processes are respawned
2898 as needed, and the CA index file is automatically reloaded
2899 when changed. This makes it possible to run the "ocsp" responder
2900 as a long-running service, making the OpenSSL CA somewhat more
2901 feature-complete. In this mode, most diagnostic messages logged
2902 after entering the event loop are logged via syslog(3) rather than
2907 * Added support for X448 and Ed448. Heavily based on original work by
2912 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2913 objects loaded. This adds the functions OSSL_STORE_expect() and
2914 OSSL_STORE_find() as well as needed tools to construct searches and
2915 get the search data out of them.
2919 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2920 version of OpenSSL should review their configuration settings to ensure
2921 that they are still appropriate for TLSv1.3. For further information see:
2922 <https://wiki.openssl.org/index.php/TLS1.3>
2926 * Grand redesign of the OpenSSL random generator
2928 The default RAND method now utilizes an AES-CTR DRBG according to
2929 NIST standard SP 800-90Ar1. The new random generator is essentially
2930 a port of the default random generator from the OpenSSL FIPS 2.0
2931 object module. It is a hybrid deterministic random bit generator
2932 using an AES-CTR bit stream and which seeds and reseeds itself
2933 automatically using trusted system entropy sources.
2935 Some of its new features are:
2936 - Support for multiple DRBG instances with seed chaining.
2937 - The default RAND method makes use of a DRBG.
2938 - There is a public and private DRBG instance.
2939 - The DRBG instances are fork-safe.
2940 - Keep all global DRBG instances on the secure heap if it is enabled.
2941 - The public and private DRBG instance are per thread for lock free
2944 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2946 * Changed Configure so it only says what it does and doesn't dump
2947 so much data. Instead, ./configdata.pm should be used as a script
2948 to display all sorts of configuration data.
2952 * Added processing of "make variables" to Configure.
2956 * Added SHA512/224 and SHA512/256 algorithm support.
2960 * The last traces of Netware support, first removed in 1.1.0, have
2965 * Get rid of Makefile.shared, and in the process, make the processing
2966 of certain files (rc.obj, or the .def/.map/.opt files produced from
2967 the ordinal files) more visible and hopefully easier to trace and
2968 debug (or make silent).
2972 * Make it possible to have environment variable assignments as
2973 arguments to config / Configure.
2977 * Add multi-prime RSA (RFC 8017) support.
2981 * Add SM3 implemented according to GB/T 32905-2016
2982 *Jack Lloyd <jack.lloyd@ribose.com>,*
2983 *Ronald Tse <ronald.tse@ribose.com>,*
2984 *Erick Borsboom <erick.borsboom@ribose.com>*
2986 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2987 as documented in RFC6066.
2988 Based on a patch from Tomasz Moń
2990 *Filipe Raimundo da Silva*
2992 * Add SM4 implemented according to GB/T 32907-2016.
2993 *Jack Lloyd <jack.lloyd@ribose.com>,*
2994 *Ronald Tse <ronald.tse@ribose.com>,*
2995 *Erick Borsboom <erick.borsboom@ribose.com>*
2997 * Reimplement -newreq-nodes and ERR_error_string_n; the
2998 original author does not agree with the license change.
3002 * Add ARIA AEAD TLS support.
3006 * Some macro definitions to support VS6 have been removed. Visual
3007 Studio 6 has not worked since 1.1.0
3011 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3012 without clearing the errors.
3016 * Add "atfork" functions. If building on a system that without
3017 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3018 requirements. The RAND facility now uses/requires this.
3026 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3027 not possible to disable entirely. However, it's still possible to
3028 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3031 To disable, configure with 'no-ui-console'. 'no-ui' is still
3032 possible to use as an alias. Check at compile time with the
3033 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3034 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3038 * Add a STORE module, which implements a uniform and URI based reader of
3039 stores that can contain keys, certificates, CRLs and numerous other
3040 objects. The main API is loosely based on a few stdio functions,
3041 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3042 OSSL_STORE_error and OSSL_STORE_close.
3043 The implementation uses backends called "loaders" to implement arbitrary
3044 URI schemes. There is one built in "loader" for the 'file' scheme.
3048 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3049 then adjusted to work on FreeBSD 8.4 as well.
3050 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3051 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3055 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3056 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3057 error code calls like this:
3059 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3061 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3062 that can be encoded in C. For the foreseeable future, this will only
3065 *Richard Levitte and Tim Hudson*
3067 * Removed BSD cryptodev engine.
3071 * Add a build target 'build_all_generated', to build all generated files
3072 and only that. This can be used to prepare everything that requires
3073 things like perl for a system that lacks perl and then move everything
3074 to that system and do the rest of the build there.
3078 * In the UI interface, make it possible to duplicate the user data. This
3079 can be used by engines that need to retain the data for a longer time
3080 than just the call where this user data is passed.
3084 * Ignore the '-named_curve auto' value for compatibility of applications
3087 *Tomáš Mráz <tmraz@fedoraproject.org>*
3089 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3090 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3091 alerts across multiple records (some of which could be empty). In practice
3092 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3093 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3094 support this at all. Supporting it adds significant complexity to the
3095 record layer, and its removal is unlikely to cause interoperability
3100 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3101 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3102 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3107 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3108 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3110 *Richard Levitte, Andy Polyakov*
3112 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3117 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3118 platform rather than 'mingw'.
3122 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3123 success if they are asked to add an object which already exists
3124 in the store. This change cascades to other functions which load
3125 certificates and CRLs.
3129 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3130 facilitate stack unwinding even from assembly subroutines.
3134 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3135 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3139 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3140 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3141 which is the minimum version we support.
3145 * Certificate time validation (X509_cmp_time) enforces stricter
3146 compliance with RFC 5280. Fractional seconds and timezone offsets
3147 are no longer allowed.
3151 * Add support for ARIA
3155 * s_client will now send the Server Name Indication (SNI) extension by
3156 default unless the new "-noservername" option is used. The server name is
3157 based on the host provided to the "-connect" option unless overridden by
3158 using "-servername".
3162 * Add support for SipHash
3166 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3167 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3168 prevent issues where no progress is being made and the peer continually
3169 sends unrecognised record types, using up resources processing them.
3173 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3174 using the algorithm defined in
3175 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3179 * Heartbeat support has been removed; the ABI is changed for now.
3181 *Richard Levitte, Rich Salz*
3183 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3187 * The RSA "null" method, which was partially supported to avoid patent
3188 issues, has been replaced to always returns NULL.
3195 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3197 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3198 used even when parsing explicit parameters, when loading a encoded key
3199 or calling `EC_GROUP_new_from_ecpkparameters()`/
3200 `EC_GROUP_new_from_ecparameters()`.
3201 This prevents bypass of security hardening and performance gains,
3202 especially for curves with specialized EC_METHODs.
3203 By default, if a key encoded with explicit parameters is loaded and later
3204 encoded, the output is still encoded with explicit parameters, even if
3205 internally a "named" EC_GROUP is used for computation.
3209 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3210 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3211 NULL. After this change, only the cofactor parameter can be NULL. It also
3212 does some minimal sanity checks on the passed order.
3217 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3218 An attack is simple, if the first CMS_recipientInfo is valid but the
3219 second CMS_recipientInfo is chosen ciphertext. If the second
3220 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3221 encryption key will be replaced by garbage, and the message cannot be
3222 decoded, but if the RSA decryption fails, the correct encryption key is
3223 used and the recipient will not notice the attack.
3224 As a work around for this potential attack the length of the decrypted
3225 key must be equal to the cipher default key length, in case the
3226 certifiate is not given and all recipientInfo are tried out.
3227 The old behaviour can be re-enabled in the CMS code by setting the
3228 CMS_DEBUG_DECRYPT flag.
3233 * Use Windows installation paths in the mingw builds
3235 Mingw isn't a POSIX environment per se, which means that Windows
3236 paths should be used for installation.
3241 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3243 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3244 This changes the size when using the `genpkey` command when no size is given.
3245 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3246 generation commands to use 2048 bits by default.
3250 * Prevent over long nonces in ChaCha20-Poly1305.
3252 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3253 for every encryption operation. RFC 7539 specifies that the nonce value
3254 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3255 and front pads the nonce with 0 bytes if it is less than 12
3256 bytes. However it also incorrectly allows a nonce to be set of up to 16
3257 bytes. In this case only the last 12 bytes are significant and any
3258 additional leading bytes are ignored.
3260 It is a requirement of using this cipher that nonce values are
3261 unique. Messages encrypted using a reused nonce value are susceptible to
3262 serious confidentiality and integrity attacks. If an application changes
3263 the default nonce length to be longer than 12 bytes and then makes a
3264 change to the leading bytes of the nonce expecting the new value to be a
3265 new unique nonce then such an application could inadvertently encrypt
3266 messages with a reused nonce.
3268 Additionally the ignored bytes in a long nonce are not covered by the
3269 integrity guarantee of this cipher. Any application that relies on the
3270 integrity of these ignored leading bytes of a long nonce may be further
3271 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3272 is safe because no such use sets such a long nonce value. However user
3273 applications that use this cipher directly and set a non-default nonce
3274 length to be longer than 12 bytes may be vulnerable.
3276 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3282 * Added SCA hardening for modular field inversion in EC_GROUP through
3283 a new dedicated field_inv() pointer in EC_METHOD.
3284 This also addresses a leakage affecting conversions from projective
3285 to affine coordinates.
3287 *Billy Bob Brumley, Nicola Tuveri*
3289 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3290 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3294 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3298 * Remove the 'dist' target and add a tarball building script. The
3299 'dist' target has fallen out of use, and it shouldn't be
3300 necessary to configure just to create a source distribution.
3304 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3306 * Timing vulnerability in DSA signature generation
3308 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3309 timing side channel attack. An attacker could use variations in the signing
3310 algorithm to recover the private key.
3312 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3317 * Timing vulnerability in ECDSA signature generation
3319 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3320 timing side channel attack. An attacker could use variations in the signing
3321 algorithm to recover the private key.
3323 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3328 * Add coordinate blinding for EC_POINT and implement projective
3329 coordinate blinding for generic prime curves as a countermeasure to
3330 chosen point SCA attacks.
3332 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3334 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3336 * Client DoS due to large DH parameter
3338 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3339 malicious server can send a very large prime value to the client. This will
3340 cause the client to spend an unreasonably long period of time generating a
3341 key for this prime resulting in a hang until the client has finished. This
3342 could be exploited in a Denial Of Service attack.
3344 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3349 * Cache timing vulnerability in RSA Key Generation
3351 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3352 a cache timing side channel attack. An attacker with sufficient access to
3353 mount cache timing attacks during the RSA key generation process could
3354 recover the private key.
3356 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3357 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3362 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3363 parameter is no longer accepted, as it leads to a corrupt table. NULL
3364 pem_str is reserved for alias entries only.
3368 * Revert blinding in ECDSA sign and instead make problematic addition
3369 length-invariant. Switch even to fixed-length Montgomery multiplication.
3373 * Change generating and checking of primes so that the error rate of not
3374 being prime depends on the intended use based on the size of the input.
3375 For larger primes this will result in more rounds of Miller-Rabin.
3376 The maximal error rate for primes with more than 1080 bits is lowered
3379 *Kurt Roeckx, Annie Yousar*
3381 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3385 * Add blinding to ECDSA and DSA signatures to protect against side channel
3386 attacks discovered by Keegan Ryan (NCC Group).
3390 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3391 now allow empty (zero character) pass phrases.
3395 * Certificate time validation (X509_cmp_time) enforces stricter
3396 compliance with RFC 5280. Fractional seconds and timezone offsets
3397 are no longer allowed.
3401 * Fixed a text canonicalisation bug in CMS
3403 Where a CMS detached signature is used with text content the text goes
3404 through a canonicalisation process first prior to signing or verifying a
3405 signature. This process strips trailing space at the end of lines, converts
3406 line terminators to CRLF and removes additional trailing line terminators
3407 at the end of a file. A bug in the canonicalisation process meant that
3408 some characters, such as form-feed, were incorrectly treated as whitespace
3409 and removed. This is contrary to the specification (RFC5485). This fix
3410 could mean that detached text data signed with an earlier version of
3411 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3412 signed with a fixed OpenSSL may fail to verify with an earlier version of
3413 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3414 and use the "-binary" flag (for the "cms" command line application) or set
3415 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3419 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3421 * Constructed ASN.1 types with a recursive definition could exceed the stack
3423 Constructed ASN.1 types with a recursive definition (such as can be found
3424 in PKCS7) could eventually exceed the stack given malicious input with
3425 excessive recursion. This could result in a Denial Of Service attack. There
3426 are no such structures used within SSL/TLS that come from untrusted sources
3427 so this is considered safe.
3429 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3435 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3437 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3438 effectively reduced to only comparing the least significant bit of each
3439 byte. This allows an attacker to forge messages that would be considered as
3440 authenticated in an amount of tries lower than that guaranteed by the
3441 security claims of the scheme. The module can only be compiled by the
3442 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3444 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3450 * Add a build target 'build_all_generated', to build all generated files
3451 and only that. This can be used to prepare everything that requires
3452 things like perl for a system that lacks perl and then move everything
3453 to that system and do the rest of the build there.
3457 * Backport SSL_OP_NO_RENGOTIATION
3459 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3460 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3461 changes this is no longer possible in 1.1.0. Therefore the new
3462 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3463 1.1.0 to provide equivalent functionality.
3465 Note that if an application built against 1.1.0h headers (or above) is run
3466 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3467 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3471 * Removed the OS390-Unix config target. It relied on a script that doesn't
3476 * rsaz_1024_mul_avx2 overflow bug on x86_64
3478 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3479 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3480 Analysis suggests that attacks against RSA and DSA as a result of this
3481 defect would be very difficult to perform and are not believed likely.
3482 Attacks against DH1024 are considered just feasible, because most of the
3483 work necessary to deduce information about a private key may be performed
3484 offline. The amount of resources required for such an attack would be
3485 significant. However, for an attack on TLS to be meaningful, the server
3486 would have to share the DH1024 private key among multiple clients, which is
3487 no longer an option since CVE-2016-0701.
3489 This only affects processors that support the AVX2 but not ADX extensions
3490 like Intel Haswell (4th generation).
3492 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3493 was originally found via the OSS-Fuzz project.
3498 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3500 * bn_sqrx8x_internal carry bug on x86_64
3502 There is a carry propagating bug in the x86_64 Montgomery squaring
3503 procedure. No EC algorithms are affected. Analysis suggests that attacks
3504 against RSA and DSA as a result of this defect would be very difficult to
3505 perform and are not believed likely. Attacks against DH are considered just
3506 feasible (although very difficult) because most of the work necessary to
3507 deduce information about a private key may be performed offline. The amount
3508 of resources required for such an attack would be very significant and
3509 likely only accessible to a limited number of attackers. An attacker would
3510 additionally need online access to an unpatched system using the target
3511 private key in a scenario with persistent DH parameters and a private
3512 key that is shared between multiple clients.
3514 This only affects processors that support the BMI1, BMI2 and ADX extensions
3515 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3517 This issue was reported to OpenSSL by the OSS-Fuzz project.
3522 * Malformed X.509 IPAddressFamily could cause OOB read
3524 If an X.509 certificate has a malformed IPAddressFamily extension,
3525 OpenSSL could do a one-byte buffer overread. The most likely result
3526 would be an erroneous display of the certificate in text format.
3528 This issue was reported to OpenSSL by the OSS-Fuzz project.
3533 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3535 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3536 platform rather than 'mingw'.
3540 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3541 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3542 which is the minimum version we support.
3546 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3548 * Encrypt-Then-Mac renegotiation crash
3550 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3551 negotiated where it was not in the original handshake (or vice-versa) then
3552 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3553 and servers are affected.
3555 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3560 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3562 * Truncated packet could crash via OOB read
3564 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3565 cipher is being used, then a truncated packet can cause that host to
3566 perform an out-of-bounds read, usually resulting in a crash.
3568 This issue was reported to OpenSSL by Robert Święcki of Google.
3573 * Bad (EC)DHE parameters cause a client crash
3575 If a malicious server supplies bad parameters for a DHE or ECDHE key
3576 exchange then this can result in the client attempting to dereference a
3577 NULL pointer leading to a client crash. This could be exploited in a Denial
3580 This issue was reported to OpenSSL by Guido Vranken.
3585 * BN_mod_exp may produce incorrect results on x86_64
3587 There is a carry propagating bug in the x86_64 Montgomery squaring
3588 procedure. No EC algorithms are affected. Analysis suggests that attacks
3589 against RSA and DSA as a result of this defect would be very difficult to
3590 perform and are not believed likely. Attacks against DH are considered just
3591 feasible (although very difficult) because most of the work necessary to
3592 deduce information about a private key may be performed offline. The amount
3593 of resources required for such an attack would be very significant and
3594 likely only accessible to a limited number of attackers. An attacker would
3595 additionally need online access to an unpatched system using the target
3596 private key in a scenario with persistent DH parameters and a private
3597 key that is shared between multiple clients. For example this can occur by
3598 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3599 similar to CVE-2015-3193 but must be treated as a separate problem.
3601 This issue was reported to OpenSSL by the OSS-Fuzz project.
3606 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3608 * ChaCha20/Poly1305 heap-buffer-overflow
3610 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3611 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3612 crash. This issue is not considered to be exploitable beyond a DoS.
3614 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3619 * CMS Null dereference
3621 Applications parsing invalid CMS structures can crash with a NULL pointer
3622 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3623 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3624 structure callback if an attempt is made to free certain invalid encodings.
3625 Only CHOICE structures using a callback which do not handle NULL value are
3628 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3633 * Montgomery multiplication may produce incorrect results
3635 There is a carry propagating bug in the Broadwell-specific Montgomery
3636 multiplication procedure that handles input lengths divisible by, but
3637 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3638 and DH private keys are impossible. This is because the subroutine in
3639 question is not used in operations with the private key itself and an input
3640 of the attacker's direct choice. Otherwise the bug can manifest itself as
3641 transient authentication and key negotiation failures or reproducible
3642 erroneous outcome of public-key operations with specially crafted input.
3643 Among EC algorithms only Brainpool P-512 curves are affected and one
3644 presumably can attack ECDH key negotiation. Impact was not analyzed in
3645 detail, because pre-requisites for attack are considered unlikely. Namely
3646 multiple clients have to choose the curve in question and the server has to
3647 share the private key among them, neither of which is default behaviour.
3648 Even then only clients that chose the curve will be affected.
3650 This issue was publicly reported as transient failures and was not
3651 initially recognized as a security issue. Thanks to Richard Morgan for
3652 providing reproducible case.
3657 * Removed automatic addition of RPATH in shared libraries and executables,
3658 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3662 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3664 * Fix Use After Free for large message sizes
3666 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3667 message larger than approx 16k is received then the underlying buffer to
3668 store the incoming message is reallocated and moved. Unfortunately a
3669 dangling pointer to the old location is left which results in an attempt to
3670 write to the previously freed location. This is likely to result in a
3671 crash, however it could potentially lead to execution of arbitrary code.
3673 This issue only affects OpenSSL 1.1.0a.
3675 This issue was reported to OpenSSL by Robert Święcki.
3680 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3682 * OCSP Status Request extension unbounded memory growth
3684 A malicious client can send an excessively large OCSP Status Request
3685 extension. If that client continually requests renegotiation, sending a
3686 large OCSP Status Request extension each time, then there will be unbounded
3687 memory growth on the server. This will eventually lead to a Denial Of
3688 Service attack through memory exhaustion. Servers with a default
3689 configuration are vulnerable even if they do not support OCSP. Builds using
3690 the "no-ocsp" build time option are not affected.
3692 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3697 * SSL_peek() hang on empty record
3699 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3700 sends an empty record. This could be exploited by a malicious peer in a
3701 Denial Of Service attack.
3703 This issue was reported to OpenSSL by Alex Gaynor.
3708 * Excessive allocation of memory in tls_get_message_header() and
3709 dtls1_preprocess_fragment()
3711 A (D)TLS message includes 3 bytes for its length in the header for the
3712 message. This would allow for messages up to 16Mb in length. Messages of
3713 this length are excessive and OpenSSL includes a check to ensure that a
3714 peer is sending reasonably sized messages in order to avoid too much memory
3715 being consumed to service a connection. A flaw in the logic of version
3716 1.1.0 means that memory for the message is allocated too early, prior to
3717 the excessive message length check. Due to way memory is allocated in
3718 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3719 to service a connection. This could lead to a Denial of Service through
3720 memory exhaustion. However, the excessive message length check still takes
3721 place, and this would cause the connection to immediately fail. Assuming
3722 that the application calls SSL_free() on the failed connection in a timely
3723 manner then the 21Mb of allocated memory will then be immediately freed
3724 again. Therefore the excessive memory allocation will be transitory in
3725 nature. This then means that there is only a security impact if:
3727 1) The application does not call SSL_free() in a timely manner in the event
3728 that the connection fails
3730 2) The application is working in a constrained environment where there is
3731 very little free memory
3733 3) The attacker initiates multiple connection attempts such that there are
3734 multiple connections in a state where memory has been allocated for the
3735 connection; SSL_free() has not yet been called; and there is insufficient
3736 memory to service the multiple requests.
3738 Except in the instance of (1) above any Denial Of Service is likely to be
3739 transitory because as soon as the connection fails the memory is
3740 subsequently freed again in the SSL_free() call. However there is an
3741 increased risk during this period of application crashes due to the lack of
3742 memory - which would then mean a more serious Denial of Service.
3744 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3745 (CVE-2016-6307 and CVE-2016-6308)
3749 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3750 had to be removed. Primary reason is that vendor assembler can't
3751 assemble our modules with -KPIC flag. As result it, assembly
3752 support, was not even available as option. But its lack means
3753 lack of side-channel resistant code, which is incompatible with
3754 security by todays standards. Fortunately gcc is readily available
3755 prepackaged option, which we firmly point at...
3759 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3761 * Windows command-line tool supports UTF-8 opt-in option for arguments
3762 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3763 (to any value) allows Windows user to access PKCS#12 file generated
3764 with Windows CryptoAPI and protected with non-ASCII password, as well
3765 as files generated under UTF-8 locale on Linux also protected with
3770 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3771 have been disabled by default and removed from DEFAULT, just like RC4.
3772 See the RC4 item below to re-enable both.
3776 * The method for finding the storage location for the Windows RAND seed file
3777 has changed. First we check %RANDFILE%. If that is not set then we check
3778 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3779 all else fails we fall back to C:\.
3783 * The EVP_EncryptUpdate() function has had its return type changed from void
3784 to int. A return of 0 indicates and error while a return of 1 indicates
3789 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3790 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3791 off the constant time implementation for RSA, DSA and DH have been made
3792 no-ops and deprecated.
3796 * Windows RAND implementation was simplified to only get entropy by
3797 calling CryptGenRandom(). Various other RAND-related tickets
3800 *Joseph Wylie Yandle, Rich Salz*
3802 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3803 and `OPENSSL_LH_`, respectively. The old names are available
3804 with API compatibility. They new names are now completely documented.
3808 * Unify TYPE_up_ref(obj) methods signature.
3809 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3810 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3811 int (instead of void) like all others TYPE_up_ref() methods.
3812 So now these methods also check the return value of CRYPTO_atomic_add(),
3813 and the validity of object reference counter.
3815 *fdasilvayy@gmail.com*
3817 * With Windows Visual Studio builds, the .pdb files are installed
3818 alongside the installed libraries and executables. For a static
3819 library installation, ossl_static.pdb is the associate compiler
3820 generated .pdb file to be used when linking programs.
3824 * Remove openssl.spec. Packaging files belong with the packagers.
3828 * Automatic Darwin/OSX configuration has had a refresh, it will now
3829 recognise x86_64 architectures automatically. You can still decide
3830 to build for a different bitness with the environment variable
3831 KERNEL_BITS (can be 32 or 64), for example:
3833 KERNEL_BITS=32 ./config
3837 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3838 256 bit AES and HMAC with SHA256.
3842 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3846 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3850 * To enable users to have their own config files and build file templates,
3851 Configure looks in the directory indicated by the environment variable
3852 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3853 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3854 name and is used as is.
3858 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3859 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3860 X509_CERT_FILE_CTX was removed.
3864 * "shared" builds are now the default. To create only static libraries use
3865 the "no-shared" Configure option.
3869 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3870 All of these option have not worked for some while and are fundamental
3875 * Make various cleanup routines no-ops and mark them as deprecated. Most
3876 global cleanup functions are no longer required because they are handled
3877 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3878 Explicitly de-initing can cause problems (e.g. where a library that uses
3879 OpenSSL de-inits, but an application is still using it). The affected
3880 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3881 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3882 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3883 COMP_zlib_cleanup().
3887 * --strict-warnings no longer enables runtime debugging options
3888 such as REF_DEBUG. Instead, debug options are automatically
3889 enabled with '--debug' builds.
3891 *Andy Polyakov, Emilia Käsper*
3893 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3894 have been moved out of the public header files. New functions for managing
3895 these have been added.
3899 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3900 objects have been moved out of the public header files. New
3901 functions for managing these have been added.
3905 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3906 have been moved out of the public header files. New functions for managing
3907 these have been added.
3911 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3912 moved out of the public header files. New functions for managing these
3917 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3921 * Removed the mk1mf build scripts.
3925 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3926 it is always safe to #include a header now.
3930 * Removed the aged BC-32 config and all its supporting scripts
3934 * Removed support for Ultrix, Netware, and OS/2.
3938 * Add support for HKDF.
3940 *Alessandro Ghedini*
3942 * Add support for blake2b and blake2s
3946 * Added support for "pipelining". Ciphers that have the
3947 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3948 encryptions/decryptions simultaneously. There are currently no built-in
3949 ciphers with this property but the expectation is that engines will be able
3950 to offer it to significantly improve throughput. Support has been extended
3951 into libssl so that multiple records for a single connection can be
3952 processed in one go (for >=TLS 1.1).
3956 * Added the AFALG engine. This is an async capable engine which is able to
3957 offload work to the Linux kernel. In this initial version it only supports
3958 AES128-CBC. The kernel must be version 4.1.0 or greater.
3962 * OpenSSL now uses a new threading API. It is no longer necessary to
3963 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3964 are two supported threading models: pthreads and windows threads. It is
3965 also possible to configure OpenSSL at compile time for "no-threads". The
3966 old threading API should no longer be used. The functions have been
3967 replaced with "no-op" compatibility macros.
3969 *Alessandro Ghedini, Matt Caswell*
3971 * Modify behavior of ALPN to invoke callback after SNI/servername
3972 callback, such that updates to the SSL_CTX affect ALPN.
3976 * Add SSL_CIPHER queries for authentication and key-exchange.
3980 * Changes to the DEFAULT cipherlist:
3981 - Prefer (EC)DHE handshakes over plain RSA.
3982 - Prefer AEAD ciphers over legacy ciphers.
3983 - Prefer ECDSA over RSA when both certificates are available.
3984 - Prefer TLSv1.2 ciphers/PRF.
3985 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3990 * Change the ECC default curve list to be this, in order: x25519,
3991 secp256r1, secp521r1, secp384r1.
3995 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3996 disabled by default. They can be re-enabled using the
3997 enable-weak-ssl-ciphers option to Configure.
4001 * If the server has ALPN configured, but supports no protocols that the
4002 client advertises, send a fatal "no_application_protocol" alert.
4003 This behaviour is SHALL in RFC 7301, though it isn't universally
4004 implemented by other servers.
4008 * Add X25519 support.
4009 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4010 for public and private key encoding using the format documented in
4011 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4012 key generation and key derivation.
4014 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4019 * Deprecate SRP_VBASE_get_by_user.
4020 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4021 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4022 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4023 seed, even if the seed is configured.
4025 Users should use SRP_VBASE_get1_by_user instead. Note that in
4026 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4027 also that even though configuring the SRP seed attempts to hide
4028 invalid usernames by continuing the handshake with fake
4029 credentials, this behaviour is not constant time and no strong
4030 guarantees are made that the handshake is indistinguishable from
4031 that of a valid user.
4035 * Configuration change; it's now possible to build dynamic engines
4036 without having to build shared libraries and vice versa. This
4037 only applies to the engines in `engines/`, those in `crypto/engine/`
4038 will always be built into libcrypto (i.e. "static").
4040 Building dynamic engines is enabled by default; to disable, use
4041 the configuration option "disable-dynamic-engine".
4043 The only requirements for building dynamic engines are the
4044 presence of the DSO module and building with position independent
4045 code, so they will also automatically be disabled if configuring
4046 with "disable-dso" or "disable-pic".
4048 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4049 are also taken away from openssl/opensslconf.h, as they are
4054 * Configuration change; if there is a known flag to compile
4055 position independent code, it will always be applied on the
4056 libcrypto and libssl object files, and never on the application
4057 object files. This means other libraries that use routines from
4058 libcrypto / libssl can be made into shared libraries regardless
4059 of how OpenSSL was configured.
4061 If this isn't desirable, the configuration options "disable-pic"
4062 or "no-pic" can be used to disable the use of PIC. This will
4063 also disable building shared libraries and dynamic engines.
4067 * Removed JPAKE code. It was experimental and has no wide use.
4071 * The INSTALL_PREFIX Makefile variable has been renamed to
4072 DESTDIR. That makes for less confusion on what this variable
4073 is for. Also, the configuration option --install_prefix is
4078 * Heartbeat for TLS has been removed and is disabled by default
4079 for DTLS; configure with enable-heartbeats. Code that uses the
4080 old #define's might need to be updated.
4082 *Emilia Käsper, Rich Salz*
4084 * Rename REF_CHECK to REF_DEBUG.
4088 * New "unified" build system
4090 The "unified" build system is aimed to be a common system for all
4091 platforms we support. With it comes new support for VMS.
4093 This system builds supports building in a different directory tree
4094 than the source tree. It produces one Makefile (for unix family
4095 or lookalikes), or one descrip.mms (for VMS).
4097 The source of information to make the Makefile / descrip.mms is
4098 small files called 'build.info', holding the necessary
4099 information for each directory with source to compile, and a
4100 template in Configurations, like unix-Makefile.tmpl or
4103 With this change, the library names were also renamed on Windows
4104 and on VMS. They now have names that are closer to the standard
4105 on Unix, and include the major version number, and in certain
4106 cases, the architecture they are built for. See "Notes on shared
4107 libraries" in INSTALL.
4109 We rely heavily on the perl module Text::Template.
4113 * Added support for auto-initialisation and de-initialisation of the library.
4114 OpenSSL no longer requires explicit init or deinit routines to be called,
4115 except in certain circumstances. See the OPENSSL_init_crypto() and
4116 OPENSSL_init_ssl() man pages for further information.
4120 * The arguments to the DTLSv1_listen function have changed. Specifically the
4121 "peer" argument is now expected to be a BIO_ADDR object.
4123 * Rewrite of BIO networking library. The BIO library lacked consistent
4124 support of IPv6, and adding it required some more extensive
4125 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4126 which hold all types of addresses and chains of address information.
4127 It also introduces a new API, with functions like BIO_socket,
4128 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4129 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4130 have been adapted accordingly.
4134 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4139 * CRIME protection: disable compression by default, even if OpenSSL is
4140 compiled with zlib enabled. Applications can still enable compression
4141 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4142 using the SSL_CONF library to configure compression.
4146 * The signature of the session callback configured with
4147 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4148 was explicitly marked as `const unsigned char*` instead of
4153 * Always DPURIFY. Remove the use of uninitialized memory in the
4154 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4158 * Removed many obsolete configuration items, including
4159 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4160 MD2_CHAR, MD2_INT, MD2_LONG
4162 IDEA_SHORT, IDEA_LONG
4163 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4165 *Rich Salz, with advice from Andy Polyakov*
4167 * Many BN internals have been moved to an internal header file.
4169 *Rich Salz with help from Andy Polyakov*
4171 * Configuration and writing out the results from it has changed.
4172 Files such as Makefile include/openssl/opensslconf.h and are now
4173 produced through general templates, such as Makefile.in and
4174 crypto/opensslconf.h.in and some help from the perl module
4177 Also, the center of configuration information is no longer
4178 Makefile. Instead, Configure produces a perl module in
4179 configdata.pm which holds most of the config data (in the hash
4180 table %config), the target data that comes from the target
4181 configuration in one of the `Configurations/*.conf` files (in
4186 * To clarify their intended purposes, the Configure options
4187 --prefix and --openssldir change their semantics, and become more
4188 straightforward and less interdependent.
4190 --prefix shall be used exclusively to give the location INSTALLTOP
4191 where programs, scripts, libraries, include files and manuals are
4192 going to be installed. The default is now /usr/local.
4194 --openssldir shall be used exclusively to give the default
4195 location OPENSSLDIR where certificates, private keys, CRLs are
4196 managed. This is also where the default openssl.cnf gets
4198 If the directory given with this option is a relative path, the
4199 values of both the --prefix value and the --openssldir value will
4200 be combined to become OPENSSLDIR.
4201 The default for --openssldir is INSTALLTOP/ssl.
4203 Anyone who uses --openssldir to specify where OpenSSL is to be
4204 installed MUST change to use --prefix instead.
4208 * The GOST engine was out of date and therefore it has been removed. An up
4209 to date GOST engine is now being maintained in an external repository.
4210 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4211 support for GOST ciphersuites (these are only activated if a GOST engine
4216 * EGD is no longer supported by default; use enable-egd when
4219 *Ben Kaduk and Rich Salz*
4221 * The distribution now has Makefile.in files, which are used to
4222 create Makefile's when Configure is run. *Configure must be run
4223 before trying to build now.*
4227 * The return value for SSL_CIPHER_description() for error conditions
4232 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4234 Obtaining and performing DNSSEC validation of TLSA records is
4235 the application's responsibility. The application provides
4236 the TLSA records of its choice to OpenSSL, and these are then
4237 used to authenticate the peer.
4239 The TLSA records need not even come from DNS. They can, for
4240 example, be used to implement local end-entity certificate or
4241 trust-anchor "pinning", where the "pin" data takes the form
4242 of TLSA records, which can augment or replace verification
4243 based on the usual WebPKI public certification authorities.
4247 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4248 continues to support deprecated interfaces in default builds.
4249 However, applications are strongly advised to compile their
4250 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4251 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4252 or the 1.1.0 releases.
4254 In environments in which all applications have been ported to
4255 not use any deprecated interfaces OpenSSL's Configure script
4256 should be used with the --api=1.1.0 option to entirely remove
4257 support for the deprecated features from the library and
4258 unconditionally disable them in the installed headers.
4259 Essentially the same effect can be achieved with the "no-deprecated"
4260 argument to Configure, except that this will always restrict
4261 the build to just the latest API, rather than a fixed API
4264 As applications are ported to future revisions of the API,
4265 they should update their compile-time OPENSSL_API_COMPAT define
4266 accordingly, but in most cases should be able to continue to
4267 compile with later releases.
4269 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4270 0x10000000L and 0x00908000L, respectively. However those
4271 versions did not support the OPENSSL_API_COMPAT feature, and
4272 so applications are not typically tested for explicit support
4273 of just the undeprecated features of either release.
4277 * Add support for setting the minimum and maximum supported protocol.
4278 It can bet set via the SSL_set_min_proto_version() and
4279 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4280 MaxProtocol. It's recommended to use the new APIs to disable
4281 protocols instead of disabling individual protocols using
4282 SSL_set_options() or SSL_CONF's Protocol. This change also
4283 removes support for disabling TLS 1.2 in the OpenSSL TLS
4284 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4288 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4292 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4293 and integrates ECDSA and ECDH functionality into EC. Implementations can
4294 now redirect key generation and no longer need to convert to or from
4297 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4298 include the ec.h header file instead.
4302 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4303 ciphers who are no longer supported and drops support the ephemeral RSA key
4304 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4308 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4309 opaque. For HMAC_CTX, the following constructors and destructors
4312 HMAC_CTX *HMAC_CTX_new(void);
4313 void HMAC_CTX_free(HMAC_CTX *ctx);
4315 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4316 destroy such methods has been added. See EVP_MD_meth_new(3) and
4317 EVP_CIPHER_meth_new(3) for documentation.
4320 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4321 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4322 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4323 an already created structure.
4324 2) For consistency with the majority of our object creators and
4325 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4326 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4327 for deprecated builds.
4331 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4332 cryptographic operations to be performed asynchronously as long as an
4333 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4334 further details. Libssl has also had this capability integrated with the
4335 introduction of the new mode SSL_MODE_ASYNC and associated error
4336 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4337 pages. This work was developed in partnership with Intel Corp.
4341 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4342 always enabled now. If you want to disable the support you should
4343 exclude it using the list of supported ciphers. This also means that the
4344 "-no_ecdhe" option has been removed from s_server.
4348 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4349 SSL_{CTX_}set1_curves() which can set a list.
4353 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4354 curve you want to support using SSL_{CTX_}set1_curves().
4358 * State machine rewrite. The state machine code has been significantly
4359 refactored in order to remove much duplication of code and solve issues
4360 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4361 further details). This change does have some associated API changes.
4362 Notably the SSL_state() function has been removed and replaced by
4363 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4364 SSL_set_state() has been removed altogether. The previous handshake states
4365 defined in ssl.h and ssl3.h have also been removed.
4369 * All instances of the string "ssleay" in the public API were replaced
4370 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4371 Some error codes related to internal RSA_eay API's were renamed.
4375 * The demo files in crypto/threads were moved to demo/threads.
4379 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4382 *Matt Caswell, Rich Salz*
4384 * New ASN.1 embed macro.
4386 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4387 structure is not allocated: it is part of the parent. That is instead of
4395 This reduces memory fragmentation and make it impossible to accidentally
4396 set a mandatory field to NULL.
4398 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4399 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4400 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4405 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4409 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4410 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4411 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4412 DES and RC4 ciphersuites.
4416 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4417 This changes the decoding behaviour for some invalid messages,
4418 though the change is mostly in the more lenient direction, and
4419 legacy behaviour is preserved as much as possible.
4423 * Fix no-stdio build.
4424 *David Woodhouse <David.Woodhouse@intel.com> and also*
4425 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4427 * New testing framework
4428 The testing framework has been largely rewritten and is now using
4429 perl and the perl modules Test::Harness and an extended variant of
4430 Test::More called OpenSSL::Test to do its work. All test scripts in
4431 test/ have been rewritten into test recipes, and all direct calls to
4432 executables in test/Makefile have become individual recipes using the
4433 simplified testing OpenSSL::Test::Simple.
4435 For documentation on our testing modules, do:
4437 perldoc test/testlib/OpenSSL/Test/Simple.pm
4438 perldoc test/testlib/OpenSSL/Test.pm
4442 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4443 are used; the latter aborts on memory leaks (usually checked on exit).
4444 Some undocumented "set malloc, etc., hooks" functions were removed
4445 and others were changed. All are now documented.
4449 * In DSA_generate_parameters_ex, if the provided seed is too short,
4452 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4454 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4455 from RFC4279, RFC4785, RFC5487, RFC5489.
4457 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4458 original RSA_PSK patch.
4462 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4463 era flag was never set throughout the codebase (only read). Also removed
4464 SSL3_FLAGS_POP_BUFFER which was only used if
4465 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4469 * Changed the default name options in the "ca", "crl", "req" and "x509"
4470 to be "oneline" instead of "compat".
4474 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4475 not aware of clients that still exhibit this bug, and the workaround
4476 hasn't been working properly for a while.
4480 * The return type of BIO_number_read() and BIO_number_written() as well as
4481 the corresponding num_read and num_write members in the BIO structure has
4482 changed from unsigned long to uint64_t. On platforms where an unsigned
4483 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4488 * Given the pervasive nature of TLS extensions it is inadvisable to run
4489 OpenSSL without support for them. It also means that maintaining
4490 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4491 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4495 * Removed support for the two export grade static DH ciphersuites
4496 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4497 were newly added (along with a number of other static DH ciphersuites) to
4498 1.0.2. However the two export ones have *never* worked since they were
4499 introduced. It seems strange in any case to be adding new export
4500 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4504 * Version negotiation has been rewritten. In particular SSLv23_method(),
4505 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4506 and turned into macros which simply call the new preferred function names
4507 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4508 should use the new names instead. Also as part of this change the ssl23.h
4509 header file has been removed.
4513 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4514 code and the associated standard is no longer considered fit-for-purpose.
4518 * RT2547 was closed. When generating a private key, try to make the
4519 output file readable only by the owner. This behavior change might
4520 be noticeable when interacting with other software.
4522 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4527 * Added HTTP GET support to the ocsp command.
4531 * Changed default digest for the dgst and enc commands from MD5 to
4536 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4540 * Added support for TLS extended master secret from
4541 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4542 initial patch which was a great help during development.
4546 * All libssl internal structures have been removed from the public header
4547 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4548 now redundant). Users should not attempt to access internal structures
4549 directly. Instead they should use the provided API functions.
4553 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4554 Access to deprecated functions can be re-enabled by running config with
4555 "enable-deprecated". In addition applications wishing to use deprecated
4556 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4557 will, by default, disable some transitive includes that previously existed
4558 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4562 * Added support for OCB mode. OpenSSL has been granted a patent license
4563 compatible with the OpenSSL license for use of OCB. Details are available
4564 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4565 for OCB can be removed by calling config with no-ocb.
4569 * SSLv2 support has been removed. It still supports receiving a SSLv2
4570 compatible client hello.
4574 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4575 done while fixing the error code for the key-too-small case.
4577 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4579 * CA.sh has been removed; use CA.pl instead.
4583 * Removed old DES API.
4587 * Remove various unsupported platforms:
4593 Sinix/ReliantUNIX RM400
4598 16-bit platforms such as WIN16
4602 * Clean up OPENSSL_NO_xxx #define's
4603 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4604 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4605 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4606 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4607 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4608 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4609 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4610 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4611 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4612 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4616 * Cleaned up dead code
4617 Remove all but one '#ifdef undef' which is to be looked at.
4621 * Clean up calling of xxx_free routines.
4622 Just like free(), fix most of the xxx_free routines to accept
4623 NULL. Remove the non-null checks from callers. Save much code.
4627 * Add secure heap for storage of private keys (when possible).
4628 Add BIO_s_secmem(), CBIGNUM, etc.
4629 Contributed by Akamai Technologies under our Corporate CLA.
4633 * Experimental support for a new, fast, unbiased prime candidate generator,
4634 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4636 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4638 * New output format NSS in the sess_id command line tool. This allows
4639 exporting the session id and the master key in NSS keylog format.
4641 *Martin Kaiser <martin@kaiser.cx>*
4643 * Harmonize version and its documentation. -f flag is used to display
4646 *mancha <mancha1@zoho.com>*
4648 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4649 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4651 *mancha <mancha1@zoho.com>*
4653 * Fix some double frees. These are not thought to be exploitable.
4655 *mancha <mancha1@zoho.com>*
4657 * A missing bounds check in the handling of the TLS heartbeat extension
4658 can be used to reveal up to 64k of memory to a connected client or
4661 Thanks for Neel Mehta of Google Security for discovering this bug and to
4662 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4663 preparing the fix ([CVE-2014-0160])
4665 *Adam Langley, Bodo Moeller*
4667 * Fix for the attack described in the paper "Recovering OpenSSL
4668 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4669 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4670 <http://eprint.iacr.org/2014/140>
4672 Thanks to Yuval Yarom and Naomi Benger for discovering this
4673 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4675 *Yuval Yarom and Naomi Benger*
4677 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4678 this fixes a limitation in previous versions of OpenSSL.
4682 * Experimental encrypt-then-mac support.
4684 Experimental support for encrypt then mac from
4685 draft-gutmann-tls-encrypt-then-mac-02.txt
4687 To enable it set the appropriate extension number (0x42 for the test
4688 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4690 For non-compliant peers (i.e. just about everything) this should have no
4693 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4697 * Add EVP support for key wrapping algorithms, to avoid problems with
4698 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4699 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4700 algorithms and include tests cases.
4704 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4709 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4710 MGF1 digest and OAEP label.
4714 * Make openssl verify return errors.
4716 *Chris Palmer <palmer@google.com> and Ben Laurie*
4718 * New function ASN1_TIME_diff to calculate the difference between two
4719 ASN1_TIME structures or one structure and the current time.
4723 * Update fips_test_suite to support multiple command line options. New
4724 test to induce all self test errors in sequence and check expected
4729 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4730 sign or verify all in one operation.
4734 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4735 test programs and fips_test_suite. Includes functionality to parse
4736 the minimal script output of fipsalgest.pl directly.
4740 * Add authorisation parameter to FIPS_module_mode_set().
4744 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4748 * Use separate DRBG fields for internal and external flags. New function
4749 FIPS_drbg_health_check() to perform on demand health checking. Add
4750 generation tests to fips_test_suite with reduced health check interval to
4751 demonstrate periodic health checking. Add "nodh" option to
4752 fips_test_suite to skip very slow DH test.
4756 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4761 * More extensive health check for DRBG checking many more failure modes.
4762 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4763 combination: call this in fips_test_suite.
4767 * Add support for canonical generation of DSA parameter 'g'. See
4770 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4771 POST to handle HMAC cases.
4775 * Add functions FIPS_module_version() and FIPS_module_version_text()
4776 to return numerical and string versions of the FIPS module number.
4780 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4781 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4782 outside the validated module in the FIPS capable OpenSSL.
4786 * Minor change to DRBG entropy callback semantics. In some cases
4787 there is no multiple of the block length between min_len and
4788 max_len. Allow the callback to return more than max_len bytes
4789 of entropy but discard any extra: it is the callback's responsibility
4790 to ensure that the extra data discarded does not impact the
4791 requested amount of entropy.
4795 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4796 information in FIPS186-3, SP800-57 and SP800-131A.
4800 * CCM support via EVP. Interface is very similar to GCM case except we
4801 must supply all data in one chunk (i.e. no update, final) and the
4802 message length must be supplied if AAD is used. Add algorithm test
4807 * Initial version of POST overhaul. Add POST callback to allow the status
4808 of POST to be monitored and/or failures induced. Modify fips_test_suite
4809 to use callback. Always run all selftests even if one fails.
4813 * XTS support including algorithm test driver in the fips_gcmtest program.
4814 Note: this does increase the maximum key length from 32 to 64 bytes but
4815 there should be no binary compatibility issues as existing applications
4816 will never use XTS mode.
4820 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4821 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4822 performs algorithm blocking for unapproved PRNG types. Also do not
4823 set PRNG type in FIPS_mode_set(): leave this to the application.
4824 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4825 the standard OpenSSL PRNG: set additional data to a date time vector.
4829 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4830 This shouldn't present any incompatibility problems because applications
4831 shouldn't be using these directly and any that are will need to rethink
4832 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4836 * Extensive self tests and health checking required by SP800-90 DRBG.
4837 Remove strength parameter from FIPS_drbg_instantiate and always
4838 instantiate at maximum supported strength.
4842 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4846 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4850 * New function DH_compute_key_padded() to compute a DH key and pad with
4851 leading zeroes if needed: this complies with SP800-56A et al.
4855 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4856 anything, incomplete, subject to change and largely untested at present.
4860 * Modify fipscanisteronly build option to only build the necessary object
4861 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4865 * Add experimental option FIPSSYMS to give all symbols in
4866 fipscanister.o and FIPS or fips prefix. This will avoid
4867 conflicts with future versions of OpenSSL. Add perl script
4868 util/fipsas.pl to preprocess assembly language source files
4869 and rename any affected symbols.
4873 * Add selftest checks and algorithm block of non-fips algorithms in
4874 FIPS mode. Remove DES2 from selftests.
4878 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4879 return internal method without any ENGINE dependencies. Add new
4880 tiny fips sign and verify functions.
4884 * New build option no-ec2m to disable characteristic 2 code.
4888 * New build option "fipscanisteronly". This only builds fipscanister.o
4889 and (currently) associated fips utilities. Uses the file Makefile.fips
4890 instead of Makefile.org as the prototype.
4894 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4895 Update fips_gcmtest to use IV generator.
4899 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4900 setting output buffer to NULL. The `*Final` function must be
4901 called although it will not retrieve any additional data. The tag
4902 can be set or retrieved with a ctrl. The IV length is by default 12
4903 bytes (96 bits) but can be set to an alternative value. If the IV
4904 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4909 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4910 underlying do_cipher function handles all cipher semantics itself
4911 including padding and finalisation. This is useful if (for example)
4912 an ENGINE cipher handles block padding itself. The behaviour of
4913 do_cipher is subtly changed if this flag is set: the return value
4914 is the number of characters written to the output buffer (zero is
4915 no longer an error code) or a negative error code. Also if the
4916 input buffer is NULL and length 0 finalisation should be performed.
4920 * If a candidate issuer certificate is already part of the constructed
4921 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4925 * Improve forward-security support: add functions
4927 void SSL_CTX_set_not_resumable_session_callback(
4928 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4929 void SSL_set_not_resumable_session_callback(
4930 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4932 for use by SSL/TLS servers; the callback function will be called whenever a
4933 new session is created, and gets to decide whether the session may be
4934 cached to make it resumable (return 0) or not (return 1). (As by the
4935 SSL/TLS protocol specifications, the session_id sent by the server will be
4936 empty to indicate that the session is not resumable; also, the server will
4937 not generate RFC 4507 (RFC 5077) session tickets.)
4939 A simple reasonable callback implementation is to return is_forward_secure.
4940 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4941 by the SSL/TLS server library, indicating whether it can provide forward
4944 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4946 * New -verify_name option in command line utilities to set verification
4951 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4952 Add CMAC pkey methods.
4956 * Experimental renegotiation in s_server -www mode. If the client
4957 browses /reneg connection is renegotiated. If /renegcert it is
4958 renegotiated requesting a certificate.
4962 * Add an "external" session cache for debugging purposes to s_server. This
4963 should help trace issues which normally are only apparent in deployed
4964 multi-process servers.
4968 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4969 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4970 BIO_set_cipher() and some obscure PEM functions were changed so they
4971 can now return an error. The RAND changes required a change to the
4972 RAND_METHOD structure.
4976 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4977 a gcc attribute to warn if the result of a function is ignored. This
4978 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4979 whose return value is often ignored.
4983 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4984 These allow SCTs (signed certificate timestamps) to be requested and
4985 validated when establishing a connection.
4987 *Rob Percival <robpercival@google.com>*
4992 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4994 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4995 used even when parsing explicit parameters, when loading a encoded key
4996 or calling `EC_GROUP_new_from_ecpkparameters()`/
4997 `EC_GROUP_new_from_ecparameters()`.
4998 This prevents bypass of security hardening and performance gains,
4999 especially for curves with specialized EC_METHODs.
5000 By default, if a key encoded with explicit parameters is loaded and later
5001 encoded, the output is still encoded with explicit parameters, even if
5002 internally a "named" EC_GROUP is used for computation.
5006 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5007 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5008 NULL. After this change, only the cofactor parameter can be NULL. It also
5009 does some minimal sanity checks on the passed order.
5014 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5015 An attack is simple, if the first CMS_recipientInfo is valid but the
5016 second CMS_recipientInfo is chosen ciphertext. If the second
5017 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5018 encryption key will be replaced by garbage, and the message cannot be
5019 decoded, but if the RSA decryption fails, the correct encryption key is
5020 used and the recipient will not notice the attack.
5021 As a work around for this potential attack the length of the decrypted
5022 key must be equal to the cipher default key length, in case the
5023 certifiate is not given and all recipientInfo are tried out.
5024 The old behaviour can be re-enabled in the CMS code by setting the
5025 CMS_DEBUG_DECRYPT flag.
5030 * Document issue with installation paths in diverse Windows builds
5032 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5033 binaries and run-time config file.
5038 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5040 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5041 This changes the size when using the `genpkey` command when no size is given.
5042 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5043 generation commands to use 2048 bits by default.
5047 * Add FIPS support for Android Arm 64-bit
5049 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5050 Module in Version 2.0.10. For some reason, the corresponding target
5051 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5052 built with FIPS support on Android Arm 64-bit. This omission has been
5055 *Matthias St. Pierre*
5057 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5059 * 0-byte record padding oracle
5061 If an application encounters a fatal protocol error and then calls
5062 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5063 then OpenSSL can respond differently to the calling application if a 0 byte
5064 record is received with invalid padding compared to if a 0 byte record is
5065 received with an invalid MAC. If the application then behaves differently
5066 based on that in a way that is detectable to the remote peer, then this
5067 amounts to a padding oracle that could be used to decrypt data.
5069 In order for this to be exploitable "non-stitched" ciphersuites must be in
5070 use. Stitched ciphersuites are optimised implementations of certain
5071 commonly used ciphersuites. Also the application must call SSL_shutdown()
5072 twice even if a protocol error has occurred (applications should not do
5073 this but some do anyway).
5075 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5076 Aviram, with additional investigation by Steven Collison and Andrew
5077 Hourselt. It was reported to OpenSSL on 10th December 2018.
5082 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5086 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5088 * Microarchitecture timing vulnerability in ECC scalar multiplication
5090 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5091 shown to be vulnerable to a microarchitecture timing side channel attack.
5092 An attacker with sufficient access to mount local timing attacks during
5093 ECDSA signature generation could recover the private key.
5095 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5096 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5102 * Timing vulnerability in DSA signature generation
5104 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5105 timing side channel attack. An attacker could use variations in the signing
5106 algorithm to recover the private key.
5108 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5113 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5114 Module, accidentally introduced while backporting security fixes from the
5115 development branch and hindering the use of ECC in FIPS mode.
5119 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5121 * Client DoS due to large DH parameter
5123 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5124 malicious server can send a very large prime value to the client. This will
5125 cause the client to spend an unreasonably long period of time generating a
5126 key for this prime resulting in a hang until the client has finished. This
5127 could be exploited in a Denial Of Service attack.
5129 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5134 * Cache timing vulnerability in RSA Key Generation
5136 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5137 a cache timing side channel attack. An attacker with sufficient access to
5138 mount cache timing attacks during the RSA key generation process could
5139 recover the private key.
5141 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5142 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5147 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5148 parameter is no longer accepted, as it leads to a corrupt table. NULL
5149 pem_str is reserved for alias entries only.
5153 * Revert blinding in ECDSA sign and instead make problematic addition
5154 length-invariant. Switch even to fixed-length Montgomery multiplication.
5158 * Change generating and checking of primes so that the error rate of not
5159 being prime depends on the intended use based on the size of the input.
5160 For larger primes this will result in more rounds of Miller-Rabin.
5161 The maximal error rate for primes with more than 1080 bits is lowered
5164 *Kurt Roeckx, Annie Yousar*
5166 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5170 * Add blinding to ECDSA and DSA signatures to protect against side channel
5171 attacks discovered by Keegan Ryan (NCC Group).
5175 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5176 now allow empty (zero character) pass phrases.
5180 * Certificate time validation (X509_cmp_time) enforces stricter
5181 compliance with RFC 5280. Fractional seconds and timezone offsets
5182 are no longer allowed.
5186 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5188 * Constructed ASN.1 types with a recursive definition could exceed the stack
5190 Constructed ASN.1 types with a recursive definition (such as can be found
5191 in PKCS7) could eventually exceed the stack given malicious input with
5192 excessive recursion. This could result in a Denial Of Service attack. There
5193 are no such structures used within SSL/TLS that come from untrusted sources
5194 so this is considered safe.
5196 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5202 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5204 * Read/write after SSL object in error state
5206 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5207 mechanism. The intent was that if a fatal error occurred during a handshake
5208 then OpenSSL would move into the error state and would immediately fail if
5209 you attempted to continue the handshake. This works as designed for the
5210 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5211 SSL_connect()), however due to a bug it does not work correctly if
5212 SSL_read() or SSL_write() is called directly. In that scenario, if the
5213 handshake fails then a fatal error will be returned in the initial function
5214 call. If SSL_read()/SSL_write() is subsequently called by the application
5215 for the same SSL object then it will succeed and the data is passed without
5216 being decrypted/encrypted directly from the SSL/TLS record layer.
5218 In order to exploit this issue an application bug would have to be present
5219 that resulted in a call to SSL_read()/SSL_write() being issued after having
5220 already received a fatal error.
5222 This issue was reported to OpenSSL by David Benjamin (Google).
5227 * rsaz_1024_mul_avx2 overflow bug on x86_64
5229 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5230 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5231 Analysis suggests that attacks against RSA and DSA as a result of this
5232 defect would be very difficult to perform and are not believed likely.
5233 Attacks against DH1024 are considered just feasible, because most of the
5234 work necessary to deduce information about a private key may be performed
5235 offline. The amount of resources required for such an attack would be
5236 significant. However, for an attack on TLS to be meaningful, the server
5237 would have to share the DH1024 private key among multiple clients, which is
5238 no longer an option since CVE-2016-0701.
5240 This only affects processors that support the AVX2 but not ADX extensions
5241 like Intel Haswell (4th generation).
5243 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5244 was originally found via the OSS-Fuzz project.
5249 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5251 * bn_sqrx8x_internal carry bug on x86_64
5253 There is a carry propagating bug in the x86_64 Montgomery squaring
5254 procedure. No EC algorithms are affected. Analysis suggests that attacks
5255 against RSA and DSA as a result of this defect would be very difficult to
5256 perform and are not believed likely. Attacks against DH are considered just
5257 feasible (although very difficult) because most of the work necessary to
5258 deduce information about a private key may be performed offline. The amount
5259 of resources required for such an attack would be very significant and
5260 likely only accessible to a limited number of attackers. An attacker would
5261 additionally need online access to an unpatched system using the target
5262 private key in a scenario with persistent DH parameters and a private
5263 key that is shared between multiple clients.
5265 This only affects processors that support the BMI1, BMI2 and ADX extensions
5266 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5268 This issue was reported to OpenSSL by the OSS-Fuzz project.
5273 * Malformed X.509 IPAddressFamily could cause OOB read
5275 If an X.509 certificate has a malformed IPAddressFamily extension,
5276 OpenSSL could do a one-byte buffer overread. The most likely result
5277 would be an erroneous display of the certificate in text format.
5279 This issue was reported to OpenSSL by the OSS-Fuzz project.
5283 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5285 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5286 platform rather than 'mingw'.
5290 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5292 * Truncated packet could crash via OOB read
5294 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5295 cipher is being used, then a truncated packet can cause that host to
5296 perform an out-of-bounds read, usually resulting in a crash.
5298 This issue was reported to OpenSSL by Robert Święcki of Google.
5303 * BN_mod_exp may produce incorrect results on x86_64
5305 There is a carry propagating bug in the x86_64 Montgomery squaring
5306 procedure. No EC algorithms are affected. Analysis suggests that attacks
5307 against RSA and DSA as a result of this defect would be very difficult to
5308 perform and are not believed likely. Attacks against DH are considered just
5309 feasible (although very difficult) because most of the work necessary to
5310 deduce information about a private key may be performed offline. The amount
5311 of resources required for such an attack would be very significant and
5312 likely only accessible to a limited number of attackers. An attacker would
5313 additionally need online access to an unpatched system using the target
5314 private key in a scenario with persistent DH parameters and a private
5315 key that is shared between multiple clients. For example this can occur by
5316 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5317 similar to CVE-2015-3193 but must be treated as a separate problem.
5319 This issue was reported to OpenSSL by the OSS-Fuzz project.
5324 * Montgomery multiplication may produce incorrect results
5326 There is a carry propagating bug in the Broadwell-specific Montgomery
5327 multiplication procedure that handles input lengths divisible by, but
5328 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5329 and DH private keys are impossible. This is because the subroutine in
5330 question is not used in operations with the private key itself and an input
5331 of the attacker's direct choice. Otherwise the bug can manifest itself as
5332 transient authentication and key negotiation failures or reproducible
5333 erroneous outcome of public-key operations with specially crafted input.
5334 Among EC algorithms only Brainpool P-512 curves are affected and one
5335 presumably can attack ECDH key negotiation. Impact was not analyzed in
5336 detail, because pre-requisites for attack are considered unlikely. Namely
5337 multiple clients have to choose the curve in question and the server has to
5338 share the private key among them, neither of which is default behaviour.
5339 Even then only clients that chose the curve will be affected.
5341 This issue was publicly reported as transient failures and was not
5342 initially recognized as a security issue. Thanks to Richard Morgan for
5343 providing reproducible case.
5348 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5349 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5350 prevent issues where no progress is being made and the peer continually
5351 sends unrecognised record types, using up resources processing them.
5355 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5357 * Missing CRL sanity check
5359 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5360 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5361 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5363 This issue only affects the OpenSSL 1.0.2i
5368 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5370 * OCSP Status Request extension unbounded memory growth
5372 A malicious client can send an excessively large OCSP Status Request
5373 extension. If that client continually requests renegotiation, sending a
5374 large OCSP Status Request extension each time, then there will be unbounded
5375 memory growth on the server. This will eventually lead to a Denial Of
5376 Service attack through memory exhaustion. Servers with a default
5377 configuration are vulnerable even if they do not support OCSP. Builds using
5378 the "no-ocsp" build time option are not affected.
5380 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5385 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5388 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5394 * OOB write in MDC2_Update()
5396 An overflow can occur in MDC2_Update() either if called directly or
5397 through the EVP_DigestUpdate() function using MDC2. If an attacker
5398 is able to supply very large amounts of input data after a previous
5399 call to EVP_EncryptUpdate() with a partial block then a length check
5400 can overflow resulting in a heap corruption.
5402 The amount of data needed is comparable to SIZE_MAX which is impractical
5405 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5410 * Malformed SHA512 ticket DoS
5412 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5413 DoS attack where a malformed ticket will result in an OOB read which will
5416 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5417 a custom server callback and ticket lookup mechanism.
5419 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5424 * OOB write in BN_bn2dec()
5426 The function BN_bn2dec() does not check the return value of BN_div_word().
5427 This can cause an OOB write if an application uses this function with an
5428 overly large BIGNUM. This could be a problem if an overly large certificate
5429 or CRL is printed out from an untrusted source. TLS is not affected because
5430 record limits will reject an oversized certificate before it is parsed.
5432 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5437 * OOB read in TS_OBJ_print_bio()
5439 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5440 the total length the OID text representation would use and not the amount
5441 of data written. This will result in OOB reads when large OIDs are
5444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5449 * Pointer arithmetic undefined behaviour
5451 Avoid some undefined pointer arithmetic
5453 A common idiom in the codebase is to check limits in the following manner:
5456 Where "p" points to some malloc'd data of SIZE bytes and
5459 "len" here could be from some externally supplied data (e.g. from a TLS
5462 The rules of C pointer arithmetic are such that "p + len" is only well
5463 defined where len <= SIZE. Therefore the above idiom is actually
5464 undefined behaviour.
5466 For example this could cause problems if some malloc implementation
5467 provides an address for "p" such that "p + len" actually overflows for
5468 values of len that are too big and therefore p + len < limit.
5470 This issue was reported to OpenSSL by Guido Vranken
5475 * Constant time flag not preserved in DSA signing
5477 Operations in the DSA signing algorithm should run in constant time in
5478 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5479 implementation means that a non-constant time codepath is followed for
5480 certain operations. This has been demonstrated through a cache-timing
5481 attack to be sufficient for an attacker to recover the private DSA key.
5483 This issue was reported by César Pereida (Aalto University), Billy Brumley
5484 (Tampere University of Technology), and Yuval Yarom (The University of
5485 Adelaide and NICTA).
5490 * DTLS buffered message DoS
5492 In a DTLS connection where handshake messages are delivered out-of-order
5493 those messages that OpenSSL is not yet ready to process will be buffered
5494 for later use. Under certain circumstances, a flaw in the logic means that
5495 those messages do not get removed from the buffer even though the handshake
5496 has been completed. An attacker could force up to approx. 15 messages to
5497 remain in the buffer when they are no longer required. These messages will
5498 be cleared when the DTLS connection is closed. The default maximum size for
5499 a message is 100k. Therefore the attacker could force an additional 1500k
5500 to be consumed per connection. By opening many simulataneous connections an
5501 attacker could cause a DoS attack through memory exhaustion.
5503 This issue was reported to OpenSSL by Quan Luo.
5508 * DTLS replay protection DoS
5510 A flaw in the DTLS replay attack protection mechanism means that records
5511 that arrive for future epochs update the replay protection "window" before
5512 the MAC for the record has been validated. This could be exploited by an
5513 attacker by sending a record for the next epoch (which does not have to
5514 decrypt or have a valid MAC), with a very large sequence number. This means
5515 that all subsequent legitimate packets are dropped causing a denial of
5516 service for a specific DTLS connection.
5518 This issue was reported to OpenSSL by the OCAP audit team.
5523 * Certificate message OOB reads
5525 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5526 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5527 theoretical DoS risk but this has not been observed in practice on common
5530 The messages affected are client certificate, client certificate request
5531 and server certificate. As a result the attack can only be performed
5532 against a client or a server which enables client authentication.
5534 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5539 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5541 * Prevent padding oracle in AES-NI CBC MAC check
5543 A MITM attacker can use a padding oracle attack to decrypt traffic
5544 when the connection uses an AES CBC cipher and the server support
5547 This issue was introduced as part of the fix for Lucky 13 padding
5548 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5549 constant time by making sure that always the same bytes are read and
5550 compared against either the MAC or padding bytes. But it no longer
5551 checked that there was enough data to have both the MAC and padding
5554 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5558 * Fix EVP_EncodeUpdate overflow
5560 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5561 Base64 encoding of binary data. If an attacker is able to supply very large
5562 amounts of input data then a length check can overflow resulting in a heap
5565 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5566 the `PEM_write_bio*` family of functions. These are mainly used within the
5567 OpenSSL command line applications, so any application which processes data
5568 from an untrusted source and outputs it as a PEM file should be considered
5569 vulnerable to this issue. User applications that call these APIs directly
5570 with large amounts of untrusted data may also be vulnerable.
5572 This issue was reported by Guido Vranken.
5577 * Fix EVP_EncryptUpdate overflow
5579 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5580 is able to supply very large amounts of input data after a previous call to
5581 EVP_EncryptUpdate() with a partial block then a length check can overflow
5582 resulting in a heap corruption. Following an analysis of all OpenSSL
5583 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5584 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5585 the first called function after an EVP_EncryptInit(), and therefore that
5586 specific call must be safe. The second form is where the length passed to
5587 EVP_EncryptUpdate() can be seen from the code to be some small value and
5588 therefore there is no possibility of an overflow. Since all instances are
5589 one of these two forms, it is believed that there can be no overflows in
5590 internal code due to this problem. It should be noted that
5591 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5592 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5593 of these calls have also been analysed too and it is believed there are no
5594 instances in internal usage where an overflow could occur.
5596 This issue was reported by Guido Vranken.
5601 * Prevent ASN.1 BIO excessive memory allocation
5603 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5604 a short invalid encoding can cause allocation of large amounts of memory
5605 potentially consuming excessive resources or exhausting memory.
5607 Any application parsing untrusted data through d2i BIO functions is
5608 affected. The memory based functions such as d2i_X509() are *not* affected.
5609 Since the memory based functions are used by the TLS library, TLS
5610 applications are not affected.
5612 This issue was reported by Brian Carpenter.
5619 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5620 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5621 in arbitrary stack data being returned in the buffer.
5623 This issue was reported by Guido Vranken.
5628 * Modify behavior of ALPN to invoke callback after SNI/servername
5629 callback, such that updates to the SSL_CTX affect ALPN.
5633 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5638 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5639 methods are enabled and ssl2 is disabled the methods return NULL.
5643 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5645 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5646 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5647 provide any "EXPORT" or "LOW" strength ciphers.
5651 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5652 is by default disabled at build-time. Builds that are not configured with
5653 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5654 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5655 will need to explicitly call either of:
5657 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5659 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5661 as appropriate. Even if either of those is used, or the application
5662 explicitly uses the version-specific SSLv2_method() or its client and
5663 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5664 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5665 ciphers, and SSLv2 56-bit DES are no longer available.
5670 * Fix a double-free in DSA code
5672 A double free bug was discovered when OpenSSL parses malformed DSA private
5673 keys and could lead to a DoS attack or memory corruption for applications
5674 that receive DSA private keys from untrusted sources. This scenario is
5677 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5683 * Disable SRP fake user seed to address a server memory leak.
5685 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5687 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5688 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5689 was changed to ignore the "fake user" SRP seed, even if the seed
5692 Users should use SRP_VBASE_get1_by_user instead. Note that in
5693 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5694 also that even though configuring the SRP seed attempts to hide
5695 invalid usernames by continuing the handshake with fake
5696 credentials, this behaviour is not constant time and no strong
5697 guarantees are made that the handshake is indistinguishable from
5698 that of a valid user.
5703 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5705 In the BN_hex2bn function the number of hex digits is calculated using an
5706 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5707 large values of `i` this can result in `bn_expand` not allocating any
5708 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5709 field as NULL leading to a subsequent NULL ptr deref. For very large values
5710 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5711 In this case memory is allocated to the internal BIGNUM data field, but it
5712 is insufficiently sized leading to heap corruption. A similar issue exists
5713 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5714 is ever called by user applications with very large untrusted hex/dec data.
5715 This is anticipated to be a rare occurrence.
5717 All OpenSSL internal usage of these functions use data that is not expected
5718 to be untrusted, e.g. config file data or application command line
5719 arguments. If user developed applications generate config file data based
5720 on untrusted data then it is possible that this could also lead to security
5721 consequences. This is also anticipated to be rare.
5723 This issue was reported to OpenSSL by Guido Vranken.
5728 * Fix memory issues in `BIO_*printf` functions
5730 The internal `fmtstr` function used in processing a "%s" format string in
5731 the `BIO_*printf` functions could overflow while calculating the length of a
5732 string and cause an OOB read when printing very long strings.
5734 Additionally the internal `doapr_outch` function can attempt to write to an
5735 OOB memory location (at an offset from the NULL pointer) in the event of a
5736 memory allocation failure. In 1.0.2 and below this could be caused where
5737 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5738 could be in processing a very long "%s" format string. Memory leaks can
5741 The first issue may mask the second issue dependent on compiler behaviour.
5742 These problems could enable attacks where large amounts of untrusted data
5743 is passed to the `BIO_*printf` functions. If applications use these functions
5744 in this way then they could be vulnerable. OpenSSL itself uses these
5745 functions when printing out human-readable dumps of ASN.1 data. Therefore
5746 applications that print this data could be vulnerable if the data is from
5747 untrusted sources. OpenSSL command line applications could also be
5748 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5749 as command line arguments.
5751 Libssl is not considered directly vulnerable. Additionally certificates etc
5752 received via remote connections via libssl are also unlikely to be able to
5753 trigger these issues because of message size limits enforced within libssl.
5755 This issue was reported to OpenSSL Guido Vranken.
5760 * Side channel attack on modular exponentiation
5762 A side-channel attack was found which makes use of cache-bank conflicts on
5763 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5764 of RSA keys. The ability to exploit this issue is limited as it relies on
5765 an attacker who has control of code in a thread running on the same
5766 hyper-threaded core as the victim thread which is performing decryptions.
5768 This issue was reported to OpenSSL by Yuval Yarom, The University of
5769 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5770 Nadia Heninger, University of Pennsylvania with more information at
5771 <http://cachebleed.info>.
5776 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5777 if no keysize is specified with default_bits. This fixes an
5778 omission in an earlier change that changed all RSA/DSA key generation
5779 commands to use 2048 bits by default.
5783 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5785 * DH small subgroups
5787 Historically OpenSSL only ever generated DH parameters based on "safe"
5788 primes. More recently (in version 1.0.2) support was provided for
5789 generating X9.42 style parameter files such as those required for RFC 5114
5790 support. The primes used in such files may not be "safe". Where an
5791 application is using DH configured with parameters based on primes that are
5792 not "safe" then an attacker could use this fact to find a peer's private
5793 DH exponent. This attack requires that the attacker complete multiple
5794 handshakes in which the peer uses the same private DH exponent. For example
5795 this could be used to discover a TLS server's private DH exponent if it's
5796 reusing the private DH exponent or it's using a static DH ciphersuite.
5798 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5799 TLS. It is not on by default. If the option is not set then the server
5800 reuses the same private DH exponent for the life of the server process and
5801 would be vulnerable to this attack. It is believed that many popular
5802 applications do set this option and would therefore not be at risk.
5804 The fix for this issue adds an additional check where a "q" parameter is
5805 available (as is the case in X9.42 based parameters). This detects the
5806 only known attack, and is the only possible defense for static DH
5807 ciphersuites. This could have some performance impact.
5809 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5810 default and cannot be disabled. This could have some performance impact.
5812 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5817 * SSLv2 doesn't block disabled ciphers
5819 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5820 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5821 been disabled, provided that the SSLv2 protocol was not also disabled via
5824 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5825 and Sebastian Schinzel.
5830 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5832 * BN_mod_exp may produce incorrect results on x86_64
5834 There is a carry propagating bug in the x86_64 Montgomery squaring
5835 procedure. No EC algorithms are affected. Analysis suggests that attacks
5836 against RSA and DSA as a result of this defect would be very difficult to
5837 perform and are not believed likely. Attacks against DH are considered just
5838 feasible (although very difficult) because most of the work necessary to
5839 deduce information about a private key may be performed offline. The amount
5840 of resources required for such an attack would be very significant and
5841 likely only accessible to a limited number of attackers. An attacker would
5842 additionally need online access to an unpatched system using the target
5843 private key in a scenario with persistent DH parameters and a private
5844 key that is shared between multiple clients. For example this can occur by
5845 default in OpenSSL DHE based SSL/TLS ciphersuites.
5847 This issue was reported to OpenSSL by Hanno Böck.
5852 * Certificate verify crash with missing PSS parameter
5854 The signature verification routines will crash with a NULL pointer
5855 dereference if presented with an ASN.1 signature using the RSA PSS
5856 algorithm and absent mask generation function parameter. Since these
5857 routines are used to verify certificate signature algorithms this can be
5858 used to crash any certificate verification operation and exploited in a
5859 DoS attack. Any application which performs certificate verification is
5860 vulnerable including OpenSSL clients and servers which enable client
5863 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5868 * X509_ATTRIBUTE memory leak
5870 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5871 memory. This structure is used by the PKCS#7 and CMS routines so any
5872 application which reads PKCS#7 or CMS data from untrusted sources is
5873 affected. SSL/TLS is not affected.
5875 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5881 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5882 This changes the decoding behaviour for some invalid messages,
5883 though the change is mostly in the more lenient direction, and
5884 legacy behaviour is preserved as much as possible.
5888 * In DSA_generate_parameters_ex, if the provided seed is too short,
5891 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5893 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5895 * Alternate chains certificate forgery
5897 During certificate verification, OpenSSL will attempt to find an
5898 alternative certificate chain if the first attempt to build such a chain
5899 fails. An error in the implementation of this logic can mean that an
5900 attacker could cause certain checks on untrusted certificates to be
5901 bypassed, such as the CA flag, enabling them to use a valid leaf
5902 certificate to act as a CA and "issue" an invalid certificate.
5904 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5909 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5911 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5912 incompatibility in the handling of HMAC. The previous ABI has now been
5917 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5919 * Malformed ECParameters causes infinite loop
5921 When processing an ECParameters structure OpenSSL enters an infinite loop
5922 if the curve specified is over a specially malformed binary polynomial
5925 This can be used to perform denial of service against any
5926 system which processes public keys, certificate requests or
5927 certificates. This includes TLS clients and TLS servers with
5928 client authentication enabled.
5930 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5935 * Exploitable out-of-bounds read in X509_cmp_time
5937 X509_cmp_time does not properly check the length of the ASN1_TIME
5938 string and can read a few bytes out of bounds. In addition,
5939 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5942 An attacker can use this to craft malformed certificates and CRLs of
5943 various sizes and potentially cause a segmentation fault, resulting in
5944 a DoS on applications that verify certificates or CRLs. TLS clients
5945 that verify CRLs are affected. TLS clients and servers with client
5946 authentication enabled may be affected if they use custom verification
5949 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5950 independently by Hanno Böck.
5955 * PKCS7 crash with missing EnvelopedContent
5957 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5958 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5959 with missing content and trigger a NULL pointer dereference on parsing.
5961 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5962 structures from untrusted sources are affected. OpenSSL clients and
5963 servers are not affected.
5965 This issue was reported to OpenSSL by Michal Zalewski (Google).
5970 * CMS verify infinite loop with unknown hash function
5972 When verifying a signedData message the CMS code can enter an infinite loop
5973 if presented with an unknown hash function OID. This can be used to perform
5974 denial of service against any system which verifies signedData messages using
5976 This issue was reported to OpenSSL by Johannes Bauer.
5981 * Race condition handling NewSessionTicket
5983 If a NewSessionTicket is received by a multi-threaded client when attempting to
5984 reuse a previous ticket then a race condition can occur potentially leading to
5985 a double free of the ticket data.
5990 * Only support 256-bit or stronger elliptic curves with the
5991 'ecdh_auto' setting (server) or by default (client). Of supported
5992 curves, prefer P-256 (both).
5996 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5998 * ClientHello sigalgs DoS fix
6000 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6001 invalid signature algorithms extension a NULL pointer dereference will
6002 occur. This can be exploited in a DoS attack against the server.
6004 This issue was was reported to OpenSSL by David Ramos of Stanford
6008 *Stephen Henson and Matt Caswell*
6010 * Multiblock corrupted pointer fix
6012 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6013 feature only applies on 64 bit x86 architecture platforms that support AES
6014 NI instructions. A defect in the implementation of "multiblock" can cause
6015 OpenSSL's internal write buffer to become incorrectly set to NULL when
6016 using non-blocking IO. Typically, when the user application is using a
6017 socket BIO for writing, this will only result in a failed connection.
6018 However if some other BIO is used then it is likely that a segmentation
6019 fault will be triggered, thus enabling a potential DoS attack.
6021 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6026 * Segmentation fault in DTLSv1_listen fix
6028 The DTLSv1_listen function is intended to be stateless and processes the
6029 initial ClientHello from many peers. It is common for user code to loop
6030 over the call to DTLSv1_listen until a valid ClientHello is received with
6031 an associated cookie. A defect in the implementation of DTLSv1_listen means
6032 that state is preserved in the SSL object from one invocation to the next
6033 that can lead to a segmentation fault. Errors processing the initial
6034 ClientHello can trigger this scenario. An example of such an error could be
6035 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6038 This issue was reported to OpenSSL by Per Allansson.
6043 * Segmentation fault in ASN1_TYPE_cmp fix
6045 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6046 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6047 certificate signature algorithm consistency this can be used to crash any
6048 certificate verification operation and exploited in a DoS attack. Any
6049 application which performs certificate verification is vulnerable including
6050 OpenSSL clients and servers which enable client authentication.
6055 * Segmentation fault for invalid PSS parameters fix
6057 The signature verification routines will crash with a NULL pointer
6058 dereference if presented with an ASN.1 signature using the RSA PSS
6059 algorithm and invalid parameters. Since these routines are used to verify
6060 certificate signature algorithms this can be used to crash any
6061 certificate verification operation and exploited in a DoS attack. Any
6062 application which performs certificate verification is vulnerable including
6063 OpenSSL clients and servers which enable client authentication.
6065 This issue was was reported to OpenSSL by Brian Carpenter.
6070 * ASN.1 structure reuse memory corruption fix
6072 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6073 memory corruption via an invalid write. Such reuse is and has been
6074 strongly discouraged and is believed to be rare.
6076 Applications that parse structures containing CHOICE or ANY DEFINED BY
6077 components may be affected. Certificate parsing (d2i_X509 and related
6078 functions) are however not affected. OpenSSL clients and servers are
6084 * PKCS7 NULL pointer dereferences fix
6086 The PKCS#7 parsing code does not handle missing outer ContentInfo
6087 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6088 missing content and trigger a NULL pointer dereference on parsing.
6090 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6091 otherwise parse PKCS#7 structures from untrusted sources are
6092 affected. OpenSSL clients and servers are not affected.
6094 This issue was reported to OpenSSL by Michal Zalewski (Google).
6099 * DoS via reachable assert in SSLv2 servers fix
6101 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6102 servers that both support SSLv2 and enable export cipher suites by sending
6103 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6105 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6106 (OpenSSL development team).
6111 * Empty CKE with client auth and DHE fix
6113 If client auth is used then a server can seg fault in the event of a DHE
6114 ciphersuite being selected and a zero length ClientKeyExchange message
6115 being sent by the client. This could be exploited in a DoS attack.
6120 * Handshake with unseeded PRNG fix
6122 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6123 with an unseeded PRNG. The conditions are:
6124 - The client is on a platform where the PRNG has not been seeded
6125 automatically, and the user has not seeded manually
6126 - A protocol specific client method version has been used (i.e. not
6127 SSL_client_methodv23)
6128 - A ciphersuite is used that does not require additional random data from
6129 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6131 If the handshake succeeds then the client random that has been used will
6132 have been generated from a PRNG with insufficient entropy and therefore the
6133 output may be predictable.
6135 For example using the following command with an unseeded openssl will
6136 succeed on an unpatched platform:
6138 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6143 * Use After Free following d2i_ECPrivatekey error fix
6145 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6146 could cause a use after free condition. This, in turn, could cause a double
6147 free in several private key parsing functions (such as d2i_PrivateKey
6148 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6149 for applications that receive EC private keys from untrusted
6150 sources. This scenario is considered rare.
6152 This issue was discovered by the BoringSSL project and fixed in their
6158 * X509_to_X509_REQ NULL pointer deref fix
6160 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6161 the certificate key is invalid. This function is rarely used in practice.
6163 This issue was discovered by Brian Carpenter.
6168 * Removed the export ciphers from the DEFAULT ciphers
6172 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6174 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6175 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6176 So far those who have to target multiple platforms would compromise
6177 and argue that binary targeting say ARMv5 would still execute on
6178 ARMv8. "Universal" build resolves this compromise by providing
6179 near-optimal performance even on newer platforms.
6183 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6184 (other platforms pending).
6186 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6188 * Add support for the SignedCertificateTimestampList certificate and
6189 OCSP response extensions from RFC6962.
6193 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6194 for corner cases. (Certain input points at infinity could lead to
6195 bogus results, with non-infinity inputs mapped to infinity too.)
6199 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6200 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6201 common cases are optimized and there still is room for further
6202 improvements. Vector Permutation AES for Altivec is also added.
6206 * Add support for little-endian ppc64 Linux target.
6208 *Marcelo Cerri (IBM)*
6210 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6211 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6212 are optimized and there still is room for further improvements.
6213 Both 32- and 64-bit modes are supported.
6215 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6217 * Improved ARMv7 NEON support.
6221 * Support for SPARC Architecture 2011 crypto extensions, first
6222 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6223 SHA256/512, MD5, GHASH and modular exponentiation.
6225 *Andy Polyakov, David Miller*
6227 * Accelerated modular exponentiation for Intel processors, a.k.a.
6230 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6232 * Support for new and upcoming Intel processors, including AVX2,
6233 BMI and SHA ISA extensions. This includes additional "stitched"
6234 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6237 This work was sponsored by Intel Corp.
6241 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6242 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6243 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6247 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6248 this fixes a limitation in previous versions of OpenSSL.
6252 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6253 MGF1 digest and OAEP label.
6257 * Add EVP support for key wrapping algorithms, to avoid problems with
6258 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6259 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6260 algorithms and include tests cases.
6264 * Add functions to allocate and set the fields of an ECDSA_METHOD
6267 *Douglas E. Engert, Steve Henson*
6269 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6270 difference in days and seconds between two tm or ASN1_TIME structures.
6274 * Add -rev test option to s_server to just reverse order of characters
6275 received by client and send back to server. Also prints an abbreviated
6276 summary of the connection parameters.
6280 * New option -brief for s_client and s_server to print out a brief summary
6281 of connection parameters.
6285 * Add callbacks for arbitrary TLS extensions.
6287 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6289 * New option -crl_download in several openssl utilities to download CRLs
6290 from CRLDP extension in certificates.
6294 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6298 * New function X509_CRL_diff to generate a delta CRL from the difference
6299 of two full CRLs. Add support to "crl" utility.
6303 * New functions to set lookup_crls function and to retrieve
6304 X509_STORE from X509_STORE_CTX.
6308 * Print out deprecated issuer and subject unique ID fields in
6313 * Extend OCSP I/O functions so they can be used for simple general purpose
6314 HTTP as well as OCSP. New wrapper function which can be used to download
6315 CRLs using the OCSP API.
6319 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6323 * `SSL_CONF*` functions. These provide a common framework for application
6324 configuration using configuration files or command lines.
6328 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6329 message callback and prints the results. Needs compile time option
6330 "enable-ssl-trace". New options to s_client and s_server to enable
6335 * New ctrl and macro to retrieve supported points extensions.
6336 Print out extension in s_server and s_client.
6340 * New functions to retrieve certificate signature and signature
6345 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6350 * New Suite B modes for TLS code. These use and enforce the requirements
6351 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6352 only use Suite B curves. The Suite B modes can be set by using the
6353 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6357 * New chain verification flags for Suite B levels of security. Check
6358 algorithms are acceptable when flags are set in X509_verify_cert.
6362 * Make tls1_check_chain return a set of flags indicating checks passed
6363 by a certificate chain. Add additional tests to handle client
6364 certificates: checks for matching certificate type and issuer name
6369 * If an attempt is made to use a signature algorithm not in the peer
6370 preference list abort the handshake. If client has no suitable
6371 signature algorithms in response to a certificate request do not
6372 use the certificate.
6376 * If server EC tmp key is not in client preference list abort handshake.
6380 * Add support for certificate stores in CERT structure. This makes it
6381 possible to have different stores per SSL structure or one store in
6382 the parent SSL_CTX. Include distinct stores for certificate chain
6383 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6384 to build and store a certificate chain in CERT structure: returning
6385 an error if the chain cannot be built: this will allow applications
6386 to test if a chain is correctly configured.
6388 Note: if the CERT based stores are not set then the parent SSL_CTX
6389 store is used to retain compatibility with existing behaviour.
6393 * New function ssl_set_client_disabled to set a ciphersuite disabled
6394 mask based on the current session, check mask when sending client
6395 hello and checking the requested ciphersuite.
6399 * New ctrls to retrieve and set certificate types in a certificate
6400 request message. Print out received values in s_client. If certificate
6401 types is not set with custom values set sensible values based on
6402 supported signature algorithms.
6406 * Support for distinct client and server supported signature algorithms.
6410 * Add certificate callback. If set this is called whenever a certificate
6411 is required by client or server. An application can decide which
6412 certificate chain to present based on arbitrary criteria: for example
6413 supported signature algorithms. Add very simple example to s_server.
6414 This fixes many of the problems and restrictions of the existing client
6415 certificate callback: for example you can now clear an existing
6416 certificate and specify the whole chain.
6420 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6421 the certificate can be used for (if anything). Set valid_flags field
6422 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6423 to have similar checks in it.
6425 Add new "cert_flags" field to CERT structure and include a "strict mode".
6426 This enforces some TLS certificate requirements (such as only permitting
6427 certificate signature algorithms contained in the supported algorithms
6428 extension) which some implementations ignore: this option should be used
6429 with caution as it could cause interoperability issues.
6433 * Update and tidy signature algorithm extension processing. Work out
6434 shared signature algorithms based on preferences and peer algorithms
6435 and print them out in s_client and s_server. Abort handshake if no
6436 shared signature algorithms.
6440 * Add new functions to allow customised supported signature algorithms
6441 for SSL and SSL_CTX structures. Add options to s_client and s_server
6446 * New function SSL_certs_clear() to delete all references to certificates
6447 from an SSL structure. Before this once a certificate had been added
6448 it couldn't be removed.
6452 * Integrate hostname, email address and IP address checking with certificate
6453 verification. New verify options supporting checking in openssl utility.
6457 * Fixes and wildcard matching support to hostname and email checking
6458 functions. Add manual page.
6460 *Florian Weimer (Red Hat Product Security Team)*
6462 * New functions to check a hostname email or IP address against a
6463 certificate. Add options x509 utility to print results of checks against
6468 * Fix OCSP checking.
6470 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6472 * Initial experimental support for explicitly trusted non-root CAs.
6473 OpenSSL still tries to build a complete chain to a root but if an
6474 intermediate CA has a trust setting included that is used. The first
6475 setting is used: whether to trust (e.g., -addtrust option to the x509
6480 * Add -trusted_first option which attempts to find certificates in the
6481 trusted store even if an untrusted chain is also supplied.
6485 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6486 platform support for Linux and Android.
6490 * Support for linux-x32, ILP32 environment in x86_64 framework.
6494 * Experimental multi-implementation support for FIPS capable OpenSSL.
6495 When in FIPS mode the approved implementations are used as normal,
6496 when not in FIPS mode the internal unapproved versions are used instead.
6497 This means that the FIPS capable OpenSSL isn't forced to use the
6498 (often lower performance) FIPS implementations outside FIPS mode.
6502 * Transparently support X9.42 DH parameters when calling
6503 PEM_read_bio_DHparameters. This means existing applications can handle
6504 the new parameter format automatically.
6508 * Initial experimental support for X9.42 DH parameter format: mainly
6509 to support use of 'q' parameter for RFC5114 parameters.
6513 * Add DH parameters from RFC5114 including test data to dhtest.
6517 * Support for automatic EC temporary key parameter selection. If enabled
6518 the most preferred EC parameters are automatically used instead of
6519 hardcoded fixed parameters. Now a server just has to call:
6520 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6521 support ECDH and use the most appropriate parameters.
6525 * Enhance and tidy EC curve and point format TLS extension code. Use
6526 static structures instead of allocation if default values are used.
6527 New ctrls to set curves we wish to support and to retrieve shared curves.
6528 Print out shared curves in s_server. New options to s_server and s_client
6529 to set list of supported curves.
6533 * New ctrls to retrieve supported signature algorithms and
6534 supported curve values as an array of NIDs. Extend openssl utility
6535 to print out received values.
6539 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6540 between NIDs and the more common NIST names such as "P-256". Enhance
6541 ecparam utility and ECC method to recognise the NIST names for curves.
6545 * Enhance SSL/TLS certificate chain handling to support different
6546 chains for each certificate instead of one chain in the parent SSL_CTX.
6550 * Support for fixed DH ciphersuite client authentication: where both
6551 server and client use DH certificates with common parameters.
6555 * Support for fixed DH ciphersuites: those requiring DH server
6560 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6562 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6563 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6564 X509_CINF_get_signature were reverted post internal team review.
6569 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6571 * OCSP Status Request extension unbounded memory growth
6573 A malicious client can send an excessively large OCSP Status Request
6574 extension. If that client continually requests renegotiation, sending a
6575 large OCSP Status Request extension each time, then there will be unbounded
6576 memory growth on the server. This will eventually lead to a Denial Of
6577 Service attack through memory exhaustion. Servers with a default
6578 configuration are vulnerable even if they do not support OCSP. Builds using
6579 the "no-ocsp" build time option are not affected.
6581 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6586 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6589 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6595 * OOB write in MDC2_Update()
6597 An overflow can occur in MDC2_Update() either if called directly or
6598 through the EVP_DigestUpdate() function using MDC2. If an attacker
6599 is able to supply very large amounts of input data after a previous
6600 call to EVP_EncryptUpdate() with a partial block then a length check
6601 can overflow resulting in a heap corruption.
6603 The amount of data needed is comparable to SIZE_MAX which is impractical
6606 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6611 * Malformed SHA512 ticket DoS
6613 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6614 DoS attack where a malformed ticket will result in an OOB read which will
6617 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6618 a custom server callback and ticket lookup mechanism.
6620 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6625 * OOB write in BN_bn2dec()
6627 The function BN_bn2dec() does not check the return value of BN_div_word().
6628 This can cause an OOB write if an application uses this function with an
6629 overly large BIGNUM. This could be a problem if an overly large certificate
6630 or CRL is printed out from an untrusted source. TLS is not affected because
6631 record limits will reject an oversized certificate before it is parsed.
6633 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6638 * OOB read in TS_OBJ_print_bio()
6640 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6641 the total length the OID text representation would use and not the amount
6642 of data written. This will result in OOB reads when large OIDs are
6645 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6650 * Pointer arithmetic undefined behaviour
6652 Avoid some undefined pointer arithmetic
6654 A common idiom in the codebase is to check limits in the following manner:
6657 Where "p" points to some malloc'd data of SIZE bytes and
6660 "len" here could be from some externally supplied data (e.g. from a TLS
6663 The rules of C pointer arithmetic are such that "p + len" is only well
6664 defined where len <= SIZE. Therefore the above idiom is actually
6665 undefined behaviour.
6667 For example this could cause problems if some malloc implementation
6668 provides an address for "p" such that "p + len" actually overflows for
6669 values of len that are too big and therefore p + len < limit.
6671 This issue was reported to OpenSSL by Guido Vranken
6676 * Constant time flag not preserved in DSA signing
6678 Operations in the DSA signing algorithm should run in constant time in
6679 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6680 implementation means that a non-constant time codepath is followed for
6681 certain operations. This has been demonstrated through a cache-timing
6682 attack to be sufficient for an attacker to recover the private DSA key.
6684 This issue was reported by César Pereida (Aalto University), Billy Brumley
6685 (Tampere University of Technology), and Yuval Yarom (The University of
6686 Adelaide and NICTA).
6691 * DTLS buffered message DoS
6693 In a DTLS connection where handshake messages are delivered out-of-order
6694 those messages that OpenSSL is not yet ready to process will be buffered
6695 for later use. Under certain circumstances, a flaw in the logic means that
6696 those messages do not get removed from the buffer even though the handshake
6697 has been completed. An attacker could force up to approx. 15 messages to
6698 remain in the buffer when they are no longer required. These messages will
6699 be cleared when the DTLS connection is closed. The default maximum size for
6700 a message is 100k. Therefore the attacker could force an additional 1500k
6701 to be consumed per connection. By opening many simulataneous connections an
6702 attacker could cause a DoS attack through memory exhaustion.
6704 This issue was reported to OpenSSL by Quan Luo.
6709 * DTLS replay protection DoS
6711 A flaw in the DTLS replay attack protection mechanism means that records
6712 that arrive for future epochs update the replay protection "window" before
6713 the MAC for the record has been validated. This could be exploited by an
6714 attacker by sending a record for the next epoch (which does not have to
6715 decrypt or have a valid MAC), with a very large sequence number. This means
6716 that all subsequent legitimate packets are dropped causing a denial of
6717 service for a specific DTLS connection.
6719 This issue was reported to OpenSSL by the OCAP audit team.
6724 * Certificate message OOB reads
6726 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6727 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6728 theoretical DoS risk but this has not been observed in practice on common
6731 The messages affected are client certificate, client certificate request
6732 and server certificate. As a result the attack can only be performed
6733 against a client or a server which enables client authentication.
6735 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6740 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6742 * Prevent padding oracle in AES-NI CBC MAC check
6744 A MITM attacker can use a padding oracle attack to decrypt traffic
6745 when the connection uses an AES CBC cipher and the server support
6748 This issue was introduced as part of the fix for Lucky 13 padding
6749 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6750 constant time by making sure that always the same bytes are read and
6751 compared against either the MAC or padding bytes. But it no longer
6752 checked that there was enough data to have both the MAC and padding
6755 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6760 * Fix EVP_EncodeUpdate overflow
6762 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6763 Base64 encoding of binary data. If an attacker is able to supply very large
6764 amounts of input data then a length check can overflow resulting in a heap
6767 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6768 the `PEM_write_bio*` family of functions. These are mainly used within the
6769 OpenSSL command line applications, so any application which processes data
6770 from an untrusted source and outputs it as a PEM file should be considered
6771 vulnerable to this issue. User applications that call these APIs directly
6772 with large amounts of untrusted data may also be vulnerable.
6774 This issue was reported by Guido Vranken.
6779 * Fix EVP_EncryptUpdate overflow
6781 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6782 is able to supply very large amounts of input data after a previous call to
6783 EVP_EncryptUpdate() with a partial block then a length check can overflow
6784 resulting in a heap corruption. Following an analysis of all OpenSSL
6785 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6786 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6787 the first called function after an EVP_EncryptInit(), and therefore that
6788 specific call must be safe. The second form is where the length passed to
6789 EVP_EncryptUpdate() can be seen from the code to be some small value and
6790 therefore there is no possibility of an overflow. Since all instances are
6791 one of these two forms, it is believed that there can be no overflows in
6792 internal code due to this problem. It should be noted that
6793 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6794 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6795 of these calls have also been analysed too and it is believed there are no
6796 instances in internal usage where an overflow could occur.
6798 This issue was reported by Guido Vranken.
6803 * Prevent ASN.1 BIO excessive memory allocation
6805 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6806 a short invalid encoding can casuse allocation of large amounts of memory
6807 potentially consuming excessive resources or exhausting memory.
6809 Any application parsing untrusted data through d2i BIO functions is
6810 affected. The memory based functions such as d2i_X509() are *not* affected.
6811 Since the memory based functions are used by the TLS library, TLS
6812 applications are not affected.
6814 This issue was reported by Brian Carpenter.
6821 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6822 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6823 in arbitrary stack data being returned in the buffer.
6825 This issue was reported by Guido Vranken.
6830 * Modify behavior of ALPN to invoke callback after SNI/servername
6831 callback, such that updates to the SSL_CTX affect ALPN.
6835 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6840 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6841 methods are enabled and ssl2 is disabled the methods return NULL.
6845 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6847 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6848 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6849 provide any "EXPORT" or "LOW" strength ciphers.
6853 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6854 is by default disabled at build-time. Builds that are not configured with
6855 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6856 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6857 will need to explicitly call either of:
6859 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6861 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6863 as appropriate. Even if either of those is used, or the application
6864 explicitly uses the version-specific SSLv2_method() or its client and
6865 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6866 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6867 ciphers, and SSLv2 56-bit DES are no longer available.
6872 * Fix a double-free in DSA code
6874 A double free bug was discovered when OpenSSL parses malformed DSA private
6875 keys and could lead to a DoS attack or memory corruption for applications
6876 that receive DSA private keys from untrusted sources. This scenario is
6879 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6885 * Disable SRP fake user seed to address a server memory leak.
6887 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6889 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6890 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6891 was changed to ignore the "fake user" SRP seed, even if the seed
6894 Users should use SRP_VBASE_get1_by_user instead. Note that in
6895 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6896 also that even though configuring the SRP seed attempts to hide
6897 invalid usernames by continuing the handshake with fake
6898 credentials, this behaviour is not constant time and no strong
6899 guarantees are made that the handshake is indistinguishable from
6900 that of a valid user.
6905 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6907 In the BN_hex2bn function the number of hex digits is calculated using an
6908 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6909 large values of `i` this can result in `bn_expand` not allocating any
6910 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6911 field as NULL leading to a subsequent NULL ptr deref. For very large values
6912 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6913 In this case memory is allocated to the internal BIGNUM data field, but it
6914 is insufficiently sized leading to heap corruption. A similar issue exists
6915 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6916 is ever called by user applications with very large untrusted hex/dec data.
6917 This is anticipated to be a rare occurrence.
6919 All OpenSSL internal usage of these functions use data that is not expected
6920 to be untrusted, e.g. config file data or application command line
6921 arguments. If user developed applications generate config file data based
6922 on untrusted data then it is possible that this could also lead to security
6923 consequences. This is also anticipated to be rare.
6925 This issue was reported to OpenSSL by Guido Vranken.
6930 * Fix memory issues in `BIO_*printf` functions
6932 The internal `fmtstr` function used in processing a "%s" format string in
6933 the `BIO_*printf` functions could overflow while calculating the length of a
6934 string and cause an OOB read when printing very long strings.
6936 Additionally the internal `doapr_outch` function can attempt to write to an
6937 OOB memory location (at an offset from the NULL pointer) in the event of a
6938 memory allocation failure. In 1.0.2 and below this could be caused where
6939 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6940 could be in processing a very long "%s" format string. Memory leaks can
6943 The first issue may mask the second issue dependent on compiler behaviour.
6944 These problems could enable attacks where large amounts of untrusted data
6945 is passed to the `BIO_*printf` functions. If applications use these functions
6946 in this way then they could be vulnerable. OpenSSL itself uses these
6947 functions when printing out human-readable dumps of ASN.1 data. Therefore
6948 applications that print this data could be vulnerable if the data is from
6949 untrusted sources. OpenSSL command line applications could also be
6950 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6951 as command line arguments.
6953 Libssl is not considered directly vulnerable. Additionally certificates etc
6954 received via remote connections via libssl are also unlikely to be able to
6955 trigger these issues because of message size limits enforced within libssl.
6957 This issue was reported to OpenSSL Guido Vranken.
6962 * Side channel attack on modular exponentiation
6964 A side-channel attack was found which makes use of cache-bank conflicts on
6965 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6966 of RSA keys. The ability to exploit this issue is limited as it relies on
6967 an attacker who has control of code in a thread running on the same
6968 hyper-threaded core as the victim thread which is performing decryptions.
6970 This issue was reported to OpenSSL by Yuval Yarom, The University of
6971 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6972 Nadia Heninger, University of Pennsylvania with more information at
6973 <http://cachebleed.info>.
6978 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6979 if no keysize is specified with default_bits. This fixes an
6980 omission in an earlier change that changed all RSA/DSA key generation
6981 commands to use 2048 bits by default.
6985 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6987 * Protection for DH small subgroup attacks
6989 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6990 switched on by default and cannot be disabled. This could have some
6995 * SSLv2 doesn't block disabled ciphers
6997 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6998 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6999 been disabled, provided that the SSLv2 protocol was not also disabled via
7002 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7003 and Sebastian Schinzel.
7008 * Reject DH handshakes with parameters shorter than 1024 bits.
7012 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7014 * Certificate verify crash with missing PSS parameter
7016 The signature verification routines will crash with a NULL pointer
7017 dereference if presented with an ASN.1 signature using the RSA PSS
7018 algorithm and absent mask generation function parameter. Since these
7019 routines are used to verify certificate signature algorithms this can be
7020 used to crash any certificate verification operation and exploited in a
7021 DoS attack. Any application which performs certificate verification is
7022 vulnerable including OpenSSL clients and servers which enable client
7025 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7030 * X509_ATTRIBUTE memory leak
7032 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7033 memory. This structure is used by the PKCS#7 and CMS routines so any
7034 application which reads PKCS#7 or CMS data from untrusted sources is
7035 affected. SSL/TLS is not affected.
7037 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7043 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7044 This changes the decoding behaviour for some invalid messages,
7045 though the change is mostly in the more lenient direction, and
7046 legacy behaviour is preserved as much as possible.
7050 * In DSA_generate_parameters_ex, if the provided seed is too short,
7051 use a random seed, as already documented.
7053 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7055 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7057 * Alternate chains certificate forgery
7059 During certificate verfification, OpenSSL will attempt to find an
7060 alternative certificate chain if the first attempt to build such a chain
7061 fails. An error in the implementation of this logic can mean that an
7062 attacker could cause certain checks on untrusted certificates to be
7063 bypassed, such as the CA flag, enabling them to use a valid leaf
7064 certificate to act as a CA and "issue" an invalid certificate.
7066 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7072 * Race condition handling PSK identify hint
7074 If PSK identity hints are received by a multi-threaded client then
7075 the values are wrongly updated in the parent SSL_CTX structure. This can
7076 result in a race condition potentially leading to a double free of the
7082 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7084 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7085 incompatibility in the handling of HMAC. The previous ABI has now been
7088 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7090 * Malformed ECParameters causes infinite loop
7092 When processing an ECParameters structure OpenSSL enters an infinite loop
7093 if the curve specified is over a specially malformed binary polynomial
7096 This can be used to perform denial of service against any
7097 system which processes public keys, certificate requests or
7098 certificates. This includes TLS clients and TLS servers with
7099 client authentication enabled.
7101 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7106 * Exploitable out-of-bounds read in X509_cmp_time
7108 X509_cmp_time does not properly check the length of the ASN1_TIME
7109 string and can read a few bytes out of bounds. In addition,
7110 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7113 An attacker can use this to craft malformed certificates and CRLs of
7114 various sizes and potentially cause a segmentation fault, resulting in
7115 a DoS on applications that verify certificates or CRLs. TLS clients
7116 that verify CRLs are affected. TLS clients and servers with client
7117 authentication enabled may be affected if they use custom verification
7120 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7121 independently by Hanno Böck.
7126 * PKCS7 crash with missing EnvelopedContent
7128 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7129 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7130 with missing content and trigger a NULL pointer dereference on parsing.
7132 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7133 structures from untrusted sources are affected. OpenSSL clients and
7134 servers are not affected.
7136 This issue was reported to OpenSSL by Michal Zalewski (Google).
7141 * CMS verify infinite loop with unknown hash function
7143 When verifying a signedData message the CMS code can enter an infinite loop
7144 if presented with an unknown hash function OID. This can be used to perform
7145 denial of service against any system which verifies signedData messages using
7147 This issue was reported to OpenSSL by Johannes Bauer.
7152 * Race condition handling NewSessionTicket
7154 If a NewSessionTicket is received by a multi-threaded client when attempting to
7155 reuse a previous ticket then a race condition can occur potentially leading to
7156 a double free of the ticket data.
7161 * Reject DH handshakes with parameters shorter than 768 bits.
7163 *Kurt Roeckx and Emilia Kasper*
7165 * dhparam: generate 2048-bit parameters by default.
7167 *Kurt Roeckx and Emilia Kasper*
7169 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7171 * Segmentation fault in ASN1_TYPE_cmp fix
7173 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7174 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7175 certificate signature algorithm consistency this can be used to crash any
7176 certificate verification operation and exploited in a DoS attack. Any
7177 application which performs certificate verification is vulnerable including
7178 OpenSSL clients and servers which enable client authentication.
7183 * ASN.1 structure reuse memory corruption fix
7185 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7186 memory corruption via an invalid write. Such reuse is and has been
7187 strongly discouraged and is believed to be rare.
7189 Applications that parse structures containing CHOICE or ANY DEFINED BY
7190 components may be affected. Certificate parsing (d2i_X509 and related
7191 functions) are however not affected. OpenSSL clients and servers are
7197 * PKCS7 NULL pointer dereferences fix
7199 The PKCS#7 parsing code does not handle missing outer ContentInfo
7200 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7201 missing content and trigger a NULL pointer dereference on parsing.
7203 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7204 otherwise parse PKCS#7 structures from untrusted sources are
7205 affected. OpenSSL clients and servers are not affected.
7207 This issue was reported to OpenSSL by Michal Zalewski (Google).
7212 * DoS via reachable assert in SSLv2 servers fix
7214 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7215 servers that both support SSLv2 and enable export cipher suites by sending
7216 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7218 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7219 (OpenSSL development team).
7224 * Use After Free following d2i_ECPrivatekey error fix
7226 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7227 could cause a use after free condition. This, in turn, could cause a double
7228 free in several private key parsing functions (such as d2i_PrivateKey
7229 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7230 for applications that receive EC private keys from untrusted
7231 sources. This scenario is considered rare.
7233 This issue was discovered by the BoringSSL project and fixed in their
7239 * X509_to_X509_REQ NULL pointer deref fix
7241 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7242 the certificate key is invalid. This function is rarely used in practice.
7244 This issue was discovered by Brian Carpenter.
7249 * Removed the export ciphers from the DEFAULT ciphers
7253 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7255 * Build fixes for the Windows and OpenVMS platforms
7257 *Matt Caswell and Richard Levitte*
7259 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7261 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7262 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7263 dereference. This could lead to a Denial Of Service attack. Thanks to
7264 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7269 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7270 dtls1_buffer_record function under certain conditions. In particular this
7271 could occur if an attacker sent repeated DTLS records with the same
7272 sequence number but for the next epoch. The memory leak could be exploited
7273 by an attacker in a Denial of Service attack through memory exhaustion.
7274 Thanks to Chris Mueller for reporting this issue.
7279 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7280 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7281 method would be set to NULL which could later result in a NULL pointer
7282 dereference. Thanks to Frank Schmirler for reporting this issue.
7287 * Abort handshake if server key exchange message is omitted for ephemeral
7290 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7291 reporting this issue.
7296 * Remove non-export ephemeral RSA code on client and server. This code
7297 violated the TLS standard by allowing the use of temporary RSA keys in
7298 non-export ciphersuites and could be used by a server to effectively
7299 downgrade the RSA key length used to a value smaller than the server
7300 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7301 INRIA or reporting this issue.
7306 * Fixed issue where DH client certificates are accepted without verification.
7307 An OpenSSL server will accept a DH certificate for client authentication
7308 without the certificate verify message. This effectively allows a client to
7309 authenticate without the use of a private key. This only affects servers
7310 which trust a client certificate authority which issues certificates
7311 containing DH keys: these are extremely rare and hardly ever encountered.
7312 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7318 * Ensure that the session ID context of an SSL is updated when its
7319 SSL_CTX is updated via SSL_set_SSL_CTX.
7321 The session ID context is typically set from the parent SSL_CTX,
7322 and can vary with the CTX.
7326 * Fix various certificate fingerprint issues.
7328 By using non-DER or invalid encodings outside the signed portion of a
7329 certificate the fingerprint can be changed without breaking the signature.
7330 Although no details of the signed portion of the certificate can be changed
7331 this can cause problems with some applications: e.g. those using the
7332 certificate fingerprint for blacklists.
7334 1. Reject signatures with non zero unused bits.
7336 If the BIT STRING containing the signature has non zero unused bits reject
7337 the signature. All current signature algorithms require zero unused bits.
7339 2. Check certificate algorithm consistency.
7341 Check the AlgorithmIdentifier inside TBS matches the one in the
7342 certificate signature. NB: this will result in signature failure
7343 errors for some broken certificates.
7345 Thanks to Konrad Kraszewski from Google for reporting this issue.
7347 3. Check DSA/ECDSA signatures use DER.
7349 Re-encode DSA/ECDSA signatures and compare with the original received
7350 signature. Return an error if there is a mismatch.
7352 This will reject various cases including garbage after signature
7353 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7354 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7355 (negative or with leading zeroes).
7357 Further analysis was conducted and fixes were developed by Stephen Henson
7358 of the OpenSSL core team.
7364 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7365 results on some platforms, including x86_64. This bug occurs at random
7366 with a very low probability, and is not known to be exploitable in any
7367 way, though its exact impact is difficult to determine. Thanks to Pieter
7368 Wuille (Blockstream) who reported this issue and also suggested an initial
7369 fix. Further analysis was conducted by the OpenSSL development team and
7370 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7371 the OpenSSL core team.
7376 * Do not resume sessions on the server if the negotiated protocol
7377 version does not match the session's version. Resuming with a different
7378 version, while not strictly forbidden by the RFC, is of questionable
7379 sanity and breaks all known clients.
7381 *David Benjamin, Emilia Käsper*
7383 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7384 early CCS messages during renegotiation. (Note that because
7385 renegotiation is encrypted, this early CCS was not exploitable.)
7389 * Tighten client-side session ticket handling during renegotiation:
7390 ensure that the client only accepts a session ticket if the server sends
7391 the extension anew in the ServerHello. Previously, a TLS client would
7392 reuse the old extension state and thus accept a session ticket if one was
7393 announced in the initial ServerHello.
7395 Similarly, ensure that the client requires a session ticket if one
7396 was advertised in the ServerHello. Previously, a TLS client would
7397 ignore a missing NewSessionTicket message.
7401 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7405 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7406 sends a carefully crafted handshake message, to cause OpenSSL to fail
7407 to free up to 64k of memory causing a memory leak. This could be
7408 exploited in a Denial Of Service attack. This issue affects OpenSSL
7409 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7410 whether SRTP is used or configured. Implementations of OpenSSL that
7411 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7413 The fix was developed by the OpenSSL team.
7418 * Session Ticket Memory Leak.
7420 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7421 integrity of that ticket is first verified. In the event of a session
7422 ticket integrity check failing, OpenSSL will fail to free memory
7423 causing a memory leak. By sending a large number of invalid session
7424 tickets an attacker could exploit this issue in a Denial Of Service
7430 * Build option no-ssl3 is incomplete.
7432 When OpenSSL is configured with "no-ssl3" as a build option, servers
7433 could accept and complete a SSL 3.0 handshake, and clients could be
7434 configured to send them.
7437 *Akamai and the OpenSSL team*
7439 * Add support for TLS_FALLBACK_SCSV.
7440 Client applications doing fallback retries should call
7441 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7444 *Adam Langley, Bodo Moeller*
7446 * Add additional DigestInfo checks.
7448 Re-encode DigestInto in DER and check against the original when
7449 verifying RSA signature: this will reject any improperly encoded
7450 DigestInfo structures.
7452 Note: this is a precautionary measure and no attacks are currently known.
7456 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7458 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7459 SRP code can be overrun an internal buffer. Add sanity check that
7460 g, A, B < N to SRP code.
7462 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7463 Group for discovering this issue.
7468 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7469 TLS 1.0 instead of higher protocol versions when the ClientHello message
7470 is badly fragmented. This allows a man-in-the-middle attacker to force a
7471 downgrade to TLS 1.0 even if both the server and the client support a
7472 higher protocol version, by modifying the client's TLS records.
7474 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7475 researching this issue.
7480 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7481 to a denial of service attack. A malicious server can crash the client
7482 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7483 ciphersuite and sending carefully crafted handshake messages.
7485 Thanks to Felix Gröbert (Google) for discovering and researching this
7491 * By sending carefully crafted DTLS packets an attacker could cause openssl
7492 to leak memory. This can be exploited through a Denial of Service attack.
7493 Thanks to Adam Langley for discovering and researching this issue.
7498 * An attacker can force openssl to consume large amounts of memory whilst
7499 processing DTLS handshake messages. This can be exploited through a
7500 Denial of Service attack.
7501 Thanks to Adam Langley for discovering and researching this issue.
7506 * An attacker can force an error condition which causes openssl to crash
7507 whilst processing DTLS packets due to memory being freed twice. This
7508 can be exploited through a Denial of Service attack.
7509 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7515 * If a multithreaded client connects to a malicious server using a resumed
7516 session and the server sends an ec point format extension it could write
7517 up to 255 bytes to freed memory.
7519 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7525 * A malicious server can crash an OpenSSL client with a null pointer
7526 dereference (read) by specifying an SRP ciphersuite even though it was not
7527 properly negotiated with the client. This can be exploited through a
7528 Denial of Service attack.
7530 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7531 discovering and researching this issue.
7536 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7537 X509_name_oneline, X509_name_print_ex et al. to leak some information
7538 from the stack. Applications may be affected if they echo pretty printing
7539 output to the attacker.
7541 Thanks to Ivan Fratric (Google) for discovering this issue.
7544 *Emilia Käsper, and Steve Henson*
7546 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7547 for corner cases. (Certain input points at infinity could lead to
7548 bogus results, with non-infinity inputs mapped to infinity too.)
7552 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7554 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7555 handshake can force the use of weak keying material in OpenSSL
7556 SSL/TLS clients and servers.
7558 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7559 researching this issue. ([CVE-2014-0224])
7561 *KIKUCHI Masashi, Steve Henson*
7563 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7564 OpenSSL DTLS client the code can be made to recurse eventually crashing
7567 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7570 *Imre Rad, Steve Henson*
7572 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7573 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7574 client or server. This is potentially exploitable to run arbitrary
7575 code on a vulnerable client or server.
7577 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7579 *Jüri Aedla, Steve Henson*
7581 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7582 are subject to a denial of service attack.
7584 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7585 this issue. ([CVE-2014-3470])
7587 *Felix Gröbert, Ivan Fratric, Steve Henson*
7589 * Harmonize version and its documentation. -f flag is used to display
7592 *mancha <mancha1@zoho.com>*
7594 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7595 in i2d_ECPrivateKey.
7597 *mancha <mancha1@zoho.com>*
7599 * Fix some double frees. These are not thought to be exploitable.
7601 *mancha <mancha1@zoho.com>*
7603 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7605 * A missing bounds check in the handling of the TLS heartbeat extension
7606 can be used to reveal up to 64k of memory to a connected client or
7609 Thanks for Neel Mehta of Google Security for discovering this bug and to
7610 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7611 preparing the fix ([CVE-2014-0160])
7613 *Adam Langley, Bodo Moeller*
7615 * Fix for the attack described in the paper "Recovering OpenSSL
7616 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7617 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7618 <http://eprint.iacr.org/2014/140>
7620 Thanks to Yuval Yarom and Naomi Benger for discovering this
7621 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7623 *Yuval Yarom and Naomi Benger*
7625 * TLS pad extension: draft-agl-tls-padding-03
7627 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7628 TLS client Hello record length value would otherwise be > 255 and
7629 less that 512 pad with a dummy extension containing zeroes so it
7630 is at least 512 bytes long.
7632 *Adam Langley, Steve Henson*
7634 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7636 * Fix for TLS record tampering bug. A carefully crafted invalid
7637 handshake could crash OpenSSL with a NULL pointer exception.
7638 Thanks to Anton Johansson for reporting this issues.
7641 * Keep original DTLS digest and encryption contexts in retransmission
7642 structures so we can use the previous session parameters if they need
7643 to be resent. ([CVE-2013-6450])
7647 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7648 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7649 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7650 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7651 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7652 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7654 *Rob Stradling, Adam Langley*
7656 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7658 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7659 supporting platforms or when small records were transferred.
7661 *Andy Polyakov, Steve Henson*
7663 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7665 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7667 This addresses the flaw in CBC record processing discovered by
7668 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7669 at: <http://www.isg.rhul.ac.uk/tls/>
7671 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7672 Security Group at Royal Holloway, University of London
7673 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7674 Emilia Käsper for the initial patch.
7677 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7679 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7680 ciphersuites which can be exploited in a denial of service attack.
7681 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7682 and detecting this bug and to Wolfgang Ettlinger
7683 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7688 * Return an error when checking OCSP signatures when key is NULL.
7689 This fixes a DoS attack. ([CVE-2013-0166])
7693 * Make openssl verify return errors.
7695 *Chris Palmer <palmer@google.com> and Ben Laurie*
7697 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7698 the right response is stapled. Also change SSL_get_certificate()
7699 so it returns the certificate actually sent.
7700 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7702 *Rob Stradling <rob.stradling@comodo.com>*
7704 * Fix possible deadlock when decoding public keys.
7708 * Don't use TLS 1.0 record version number in initial client hello
7713 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7715 * Sanity check record length before skipping explicit IV in TLS
7716 1.2, 1.1 and DTLS to fix DoS attack.
7718 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7719 fuzzing as a service testing platform.
7724 * Initialise tkeylen properly when encrypting CMS messages.
7725 Thanks to Solar Designer of Openwall for reporting this issue.
7729 * In FIPS mode don't try to use composite ciphers as they are not
7734 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7736 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7737 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7738 mean any application compiled against OpenSSL 1.0.0 headers setting
7739 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7740 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7741 0x10000000L Any application which was previously compiled against
7742 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7743 will need to be recompiled as a result. Letting be results in
7744 inability to disable specifically TLS 1.1 and in client context,
7745 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7749 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7750 disable just protocol X, but all protocols above X *if* there are
7751 protocols *below* X still enabled. In more practical terms it means
7752 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7753 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7754 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7759 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7761 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7762 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7763 in CRYPTO_realloc_clean.
7765 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7766 issue and to Adam Langley <agl@chromium.org> for fixing it.
7769 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7771 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7775 * Workarounds for some broken servers that "hang" if a client hello
7776 record length exceeds 255 bytes.
7778 1. Do not use record version number > TLS 1.0 in initial client
7779 hello: some (but not all) hanging servers will now work.
7780 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7781 the number of ciphers sent in the client hello. This should be
7782 set to an even number, such as 50, for example by passing:
7783 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7784 Most broken servers should now work.
7785 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7786 TLS 1.2 client support entirely.
7790 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7794 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7796 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7797 STRING form instead of a DigestInfo.
7801 * The format used for MDC2 RSA signatures is inconsistent between EVP
7802 and the RSA_sign/RSA_verify functions. This was made more apparent when
7803 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7804 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7805 the correct format in RSA_verify so both forms transparently work.
7809 * Some servers which support TLS 1.0 can choke if we initially indicate
7810 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7811 encrypted premaster secret. As a workaround use the maximum permitted
7812 client version in client hello, this should keep such servers happy
7813 and still work with previous versions of OpenSSL.
7817 * Add support for TLS/DTLS heartbeats.
7819 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7821 * Add support for SCTP.
7823 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7825 * Improved PRNG seeding for VOS.
7827 *Paul Green <Paul.Green@stratus.com>*
7829 * Extensive assembler packs updates, most notably:
7831 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7832 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7833 - x86_64: bit-sliced AES implementation;
7834 - ARM: NEON support, contemporary platforms optimizations;
7835 - s390x: z196 support;
7836 - `*`: GHASH and GF(2^m) multiplication implementations;
7840 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7841 (removal of unnecessary code)
7843 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7845 * Add TLS key material exporter from RFC 5705.
7849 * Add DTLS-SRTP negotiation from RFC 5764.
7853 * Add Next Protocol Negotiation,
7854 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7855 disabled with a no-npn flag to config or Configure. Code donated
7858 *Adam Langley <agl@google.com> and Ben Laurie*
7860 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7861 NIST-P256, NIST-P521, with constant-time single point multiplication on
7862 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7863 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7864 Code made available under Apache License version 2.0.
7866 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7867 line to include this in your build of OpenSSL, and run "make depend" (or
7868 "make update"). This enables the following EC_METHODs:
7870 EC_GFp_nistp224_method()
7871 EC_GFp_nistp256_method()
7872 EC_GFp_nistp521_method()
7874 EC_GROUP_new_by_curve_name() will automatically use these (while
7875 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7878 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7880 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7881 all platforms. Move ssize_t definition from e_os.h to the public
7882 header file e_os2.h as it now appears in public header file cms.h
7886 * New -sigopt option to the ca, req and x509 utilities. Additional
7887 signature parameters can be passed using this option and in
7892 * Add RSA PSS signing function. This will generate and set the
7893 appropriate AlgorithmIdentifiers for PSS based on those in the
7894 corresponding EVP_MD_CTX structure. No application support yet.
7898 * Support for companion algorithm specific ASN1 signing routines.
7899 New function ASN1_item_sign_ctx() signs a pre-initialised
7900 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7901 the appropriate parameters.
7905 * Add new algorithm specific ASN1 verification initialisation function
7906 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7907 handling will be the same no matter what EVP_PKEY_METHOD is used.
7908 Add a PSS handler to support verification of PSS signatures: checked
7909 against a number of sample certificates.
7913 * Add signature printing for PSS. Add PSS OIDs.
7915 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7917 * Add algorithm specific signature printing. An individual ASN1 method
7918 can now print out signatures instead of the standard hex dump.
7920 More complex signatures (e.g. PSS) can print out more meaningful
7921 information. Include DSA version that prints out the signature
7926 * Password based recipient info support for CMS library: implementing
7931 * Split password based encryption into PBES2 and PBKDF2 functions. This
7932 neatly separates the code into cipher and PBE sections and is required
7933 for some algorithms that split PBES2 into separate pieces (such as
7934 password based CMS).
7938 * Session-handling fixes:
7939 - Fix handling of connections that are resuming with a session ID,
7940 but also support Session Tickets.
7941 - Fix a bug that suppressed issuing of a new ticket if the client
7942 presented a ticket with an expired session.
7943 - Try to set the ticket lifetime hint to something reasonable.
7944 - Make tickets shorter by excluding irrelevant information.
7945 - On the client side, don't ignore renewed tickets.
7947 *Adam Langley, Bodo Moeller (Google)*
7949 * Fix PSK session representation.
7953 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7955 This work was sponsored by Intel.
7959 * Add GCM support to TLS library. Some custom code is needed to split
7960 the IV between the fixed (from PRF) and explicit (from TLS record)
7961 portions. This adds all GCM ciphersuites supported by RFC5288 and
7962 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7963 add a special AESGCM string for GCM only.
7967 * Expand range of ctrls for AES GCM. Permit setting invocation
7968 field on decrypt and retrieval of invocation field only on encrypt.
7972 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7973 As required by RFC5289 these ciphersuites cannot be used if for
7974 versions of TLS earlier than 1.2.
7978 * For FIPS capable OpenSSL interpret a NULL default public key method
7979 as unset and return the appropriate default but do *not* set the default.
7980 This means we can return the appropriate method in applications that
7981 switch between FIPS and non-FIPS modes.
7985 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7986 ENGINE is used then we cannot handle that in the FIPS module so we
7987 keep original code iff non-FIPS operations are allowed.
7991 * Add -attime option to openssl utilities.
7993 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7995 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7999 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8000 FIPS EC methods unconditionally for now.
8004 * New build option no-ec2m to disable characteristic 2 code.
8008 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8009 all cases can be covered as some introduce binary incompatibilities.
8013 * Redirect RSA operations to FIPS module including keygen,
8014 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8018 * Add similar low-level API blocking to ciphers.
8022 * low-level digest APIs are not approved in FIPS mode: any attempt
8023 to use these will cause a fatal error. Applications that *really* want
8024 to use them can use the `private_*` version instead.
8028 * Redirect cipher operations to FIPS module for FIPS builds.
8032 * Redirect digest operations to FIPS module for FIPS builds.
8036 * Update build system to add "fips" flag which will link in fipscanister.o
8037 for static and shared library builds embedding a signature if needed.
8041 * Output TLS supported curves in preference order instead of numerical
8042 order. This is currently hardcoded for the highest order curves first.
8043 This should be configurable so applications can judge speed vs strength.
8047 * Add TLS v1.2 server support for client authentication.
8051 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8056 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8057 FIPS modules versions.
8061 * Add TLS v1.2 client side support for client authentication. Keep cache
8062 of handshake records longer as we don't know the hash algorithm to use
8063 until after the certificate request message is received.
8067 * Initial TLS v1.2 client support. Add a default signature algorithms
8068 extension including all the algorithms we support. Parse new signature
8069 format in client key exchange. Relax some ECC signing restrictions for
8070 TLS v1.2 as indicated in RFC5246.
8074 * Add server support for TLS v1.2 signature algorithms extension. Switch
8075 to new signature format when needed using client digest preference.
8076 All server ciphersuites should now work correctly in TLS v1.2. No client
8077 support yet and no support for client certificates.
8081 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8082 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8083 ciphersuites. At present only RSA key exchange ciphersuites work with
8084 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8085 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8086 and version checking.
8090 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8091 with this defined it will not be affected by any changes to ssl internal
8092 structures. Add several utility functions to allow openssl application
8093 to work with OPENSSL_NO_SSL_INTERN defined.
8097 * A long standing patch to add support for SRP from EdelWeb (Peter
8098 Sylvester and Christophe Renou) was integrated.
8099 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8100 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8103 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8107 * Permit abbreviated handshakes when renegotiating using the function
8108 SSL_renegotiate_abbreviated().
8110 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8112 * Add call to ENGINE_register_all_complete() to
8113 ENGINE_load_builtin_engines(), so some implementations get used
8114 automatically instead of needing explicit application support.
8118 * Add support for TLS key exporter as described in RFC5705.
8120 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8122 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8123 a few changes are required:
8125 Add SSL_OP_NO_TLSv1_1 flag.
8126 Add TLSv1_1 methods.
8127 Update version checking logic to handle version 1.1.
8128 Add explicit IV handling (ported from DTLS code).
8129 Add command line options to s_client/s_server.
8136 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8138 * X509_ATTRIBUTE memory leak
8140 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8141 memory. This structure is used by the PKCS#7 and CMS routines so any
8142 application which reads PKCS#7 or CMS data from untrusted sources is
8143 affected. SSL/TLS is not affected.
8145 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8151 * Race condition handling PSK identify hint
8153 If PSK identity hints are received by a multi-threaded client then
8154 the values are wrongly updated in the parent SSL_CTX structure. This can
8155 result in a race condition potentially leading to a double free of the
8161 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8163 * Malformed ECParameters causes infinite loop
8165 When processing an ECParameters structure OpenSSL enters an infinite loop
8166 if the curve specified is over a specially malformed binary polynomial
8169 This can be used to perform denial of service against any
8170 system which processes public keys, certificate requests or
8171 certificates. This includes TLS clients and TLS servers with
8172 client authentication enabled.
8174 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8179 * Exploitable out-of-bounds read in X509_cmp_time
8181 X509_cmp_time does not properly check the length of the ASN1_TIME
8182 string and can read a few bytes out of bounds. In addition,
8183 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8186 An attacker can use this to craft malformed certificates and CRLs of
8187 various sizes and potentially cause a segmentation fault, resulting in
8188 a DoS on applications that verify certificates or CRLs. TLS clients
8189 that verify CRLs are affected. TLS clients and servers with client
8190 authentication enabled may be affected if they use custom verification
8193 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8194 independently by Hanno Böck.
8199 * PKCS7 crash with missing EnvelopedContent
8201 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8202 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8203 with missing content and trigger a NULL pointer dereference on parsing.
8205 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8206 structures from untrusted sources are affected. OpenSSL clients and
8207 servers are not affected.
8209 This issue was reported to OpenSSL by Michal Zalewski (Google).
8214 * CMS verify infinite loop with unknown hash function
8216 When verifying a signedData message the CMS code can enter an infinite loop
8217 if presented with an unknown hash function OID. This can be used to perform
8218 denial of service against any system which verifies signedData messages using
8220 This issue was reported to OpenSSL by Johannes Bauer.
8225 * Race condition handling NewSessionTicket
8227 If a NewSessionTicket is received by a multi-threaded client when attempting to
8228 reuse a previous ticket then a race condition can occur potentially leading to
8229 a double free of the ticket data.
8234 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8236 * Segmentation fault in ASN1_TYPE_cmp fix
8238 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8239 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8240 certificate signature algorithm consistency this can be used to crash any
8241 certificate verification operation and exploited in a DoS attack. Any
8242 application which performs certificate verification is vulnerable including
8243 OpenSSL clients and servers which enable client authentication.
8248 * ASN.1 structure reuse memory corruption fix
8250 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8251 memory corruption via an invalid write. Such reuse is and has been
8252 strongly discouraged and is believed to be rare.
8254 Applications that parse structures containing CHOICE or ANY DEFINED BY
8255 components may be affected. Certificate parsing (d2i_X509 and related
8256 functions) are however not affected. OpenSSL clients and servers are
8262 * PKCS7 NULL pointer dereferences fix
8264 The PKCS#7 parsing code does not handle missing outer ContentInfo
8265 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8266 missing content and trigger a NULL pointer dereference on parsing.
8268 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8269 otherwise parse PKCS#7 structures from untrusted sources are
8270 affected. OpenSSL clients and servers are not affected.
8272 This issue was reported to OpenSSL by Michal Zalewski (Google).
8277 * DoS via reachable assert in SSLv2 servers fix
8279 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8280 servers that both support SSLv2 and enable export cipher suites by sending
8281 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8283 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8284 (OpenSSL development team).
8289 * Use After Free following d2i_ECPrivatekey error fix
8291 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8292 could cause a use after free condition. This, in turn, could cause a double
8293 free in several private key parsing functions (such as d2i_PrivateKey
8294 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8295 for applications that receive EC private keys from untrusted
8296 sources. This scenario is considered rare.
8298 This issue was discovered by the BoringSSL project and fixed in their
8304 * X509_to_X509_REQ NULL pointer deref fix
8306 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8307 the certificate key is invalid. This function is rarely used in practice.
8309 This issue was discovered by Brian Carpenter.
8314 * Removed the export ciphers from the DEFAULT ciphers
8318 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8320 * Build fixes for the Windows and OpenVMS platforms
8322 *Matt Caswell and Richard Levitte*
8324 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8326 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8327 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8328 dereference. This could lead to a Denial Of Service attack. Thanks to
8329 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8334 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8335 dtls1_buffer_record function under certain conditions. In particular this
8336 could occur if an attacker sent repeated DTLS records with the same
8337 sequence number but for the next epoch. The memory leak could be exploited
8338 by an attacker in a Denial of Service attack through memory exhaustion.
8339 Thanks to Chris Mueller for reporting this issue.
8344 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8345 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8346 method would be set to NULL which could later result in a NULL pointer
8347 dereference. Thanks to Frank Schmirler for reporting this issue.
8352 * Abort handshake if server key exchange message is omitted for ephemeral
8355 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8356 reporting this issue.
8361 * Remove non-export ephemeral RSA code on client and server. This code
8362 violated the TLS standard by allowing the use of temporary RSA keys in
8363 non-export ciphersuites and could be used by a server to effectively
8364 downgrade the RSA key length used to a value smaller than the server
8365 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8366 INRIA or reporting this issue.
8371 * Fixed issue where DH client certificates are accepted without verification.
8372 An OpenSSL server will accept a DH certificate for client authentication
8373 without the certificate verify message. This effectively allows a client to
8374 authenticate without the use of a private key. This only affects servers
8375 which trust a client certificate authority which issues certificates
8376 containing DH keys: these are extremely rare and hardly ever encountered.
8377 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8383 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8384 results on some platforms, including x86_64. This bug occurs at random
8385 with a very low probability, and is not known to be exploitable in any
8386 way, though its exact impact is difficult to determine. Thanks to Pieter
8387 Wuille (Blockstream) who reported this issue and also suggested an initial
8388 fix. Further analysis was conducted by the OpenSSL development team and
8389 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8390 the OpenSSL core team.
8395 * Fix various certificate fingerprint issues.
8397 By using non-DER or invalid encodings outside the signed portion of a
8398 certificate the fingerprint can be changed without breaking the signature.
8399 Although no details of the signed portion of the certificate can be changed
8400 this can cause problems with some applications: e.g. those using the
8401 certificate fingerprint for blacklists.
8403 1. Reject signatures with non zero unused bits.
8405 If the BIT STRING containing the signature has non zero unused bits reject
8406 the signature. All current signature algorithms require zero unused bits.
8408 2. Check certificate algorithm consistency.
8410 Check the AlgorithmIdentifier inside TBS matches the one in the
8411 certificate signature. NB: this will result in signature failure
8412 errors for some broken certificates.
8414 Thanks to Konrad Kraszewski from Google for reporting this issue.
8416 3. Check DSA/ECDSA signatures use DER.
8418 Reencode DSA/ECDSA signatures and compare with the original received
8419 signature. Return an error if there is a mismatch.
8421 This will reject various cases including garbage after signature
8422 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8423 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8424 (negative or with leading zeroes).
8426 Further analysis was conducted and fixes were developed by Stephen Henson
8427 of the OpenSSL core team.
8433 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8435 * Session Ticket Memory Leak.
8437 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8438 integrity of that ticket is first verified. In the event of a session
8439 ticket integrity check failing, OpenSSL will fail to free memory
8440 causing a memory leak. By sending a large number of invalid session
8441 tickets an attacker could exploit this issue in a Denial Of Service
8447 * Build option no-ssl3 is incomplete.
8449 When OpenSSL is configured with "no-ssl3" as a build option, servers
8450 could accept and complete a SSL 3.0 handshake, and clients could be
8451 configured to send them.
8454 *Akamai and the OpenSSL team*
8456 * Add support for TLS_FALLBACK_SCSV.
8457 Client applications doing fallback retries should call
8458 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8461 *Adam Langley, Bodo Moeller*
8463 * Add additional DigestInfo checks.
8465 Reencode DigestInto in DER and check against the original when
8466 verifying RSA signature: this will reject any improperly encoded
8467 DigestInfo structures.
8469 Note: this is a precautionary measure and no attacks are currently known.
8473 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8475 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8476 to a denial of service attack. A malicious server can crash the client
8477 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8478 ciphersuite and sending carefully crafted handshake messages.
8480 Thanks to Felix Gröbert (Google) for discovering and researching this
8486 * By sending carefully crafted DTLS packets an attacker could cause openssl
8487 to leak memory. This can be exploited through a Denial of Service attack.
8488 Thanks to Adam Langley for discovering and researching this issue.
8493 * An attacker can force openssl to consume large amounts of memory whilst
8494 processing DTLS handshake messages. This can be exploited through a
8495 Denial of Service attack.
8496 Thanks to Adam Langley for discovering and researching this issue.
8501 * An attacker can force an error condition which causes openssl to crash
8502 whilst processing DTLS packets due to memory being freed twice. This
8503 can be exploited through a Denial of Service attack.
8504 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8510 * If a multithreaded client connects to a malicious server using a resumed
8511 session and the server sends an ec point format extension it could write
8512 up to 255 bytes to freed memory.
8514 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8520 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8521 X509_name_oneline, X509_name_print_ex et al. to leak some information
8522 from the stack. Applications may be affected if they echo pretty printing
8523 output to the attacker.
8525 Thanks to Ivan Fratric (Google) for discovering this issue.
8528 *Emilia Käsper, and Steve Henson*
8530 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8531 for corner cases. (Certain input points at infinity could lead to
8532 bogus results, with non-infinity inputs mapped to infinity too.)
8536 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8538 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8539 handshake can force the use of weak keying material in OpenSSL
8540 SSL/TLS clients and servers.
8542 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8543 researching this issue. ([CVE-2014-0224])
8545 *KIKUCHI Masashi, Steve Henson*
8547 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8548 OpenSSL DTLS client the code can be made to recurse eventually crashing
8551 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8554 *Imre Rad, Steve Henson*
8556 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8557 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8558 client or server. This is potentially exploitable to run arbitrary
8559 code on a vulnerable client or server.
8561 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8563 *Jüri Aedla, Steve Henson*
8565 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8566 are subject to a denial of service attack.
8568 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8569 this issue. ([CVE-2014-3470])
8571 *Felix Gröbert, Ivan Fratric, Steve Henson*
8573 * Harmonize version and its documentation. -f flag is used to display
8576 *mancha <mancha1@zoho.com>*
8578 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8579 in i2d_ECPrivateKey.
8581 *mancha <mancha1@zoho.com>*
8583 * Fix some double frees. These are not thought to be exploitable.
8585 *mancha <mancha1@zoho.com>*
8587 * Fix for the attack described in the paper "Recovering OpenSSL
8588 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8589 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8590 <http://eprint.iacr.org/2014/140>
8592 Thanks to Yuval Yarom and Naomi Benger for discovering this
8593 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8595 *Yuval Yarom and Naomi Benger*
8597 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8599 * Keep original DTLS digest and encryption contexts in retransmission
8600 structures so we can use the previous session parameters if they need
8601 to be resent. ([CVE-2013-6450])
8605 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8606 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8607 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8608 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8609 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8610 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8612 *Rob Stradling, Adam Langley*
8614 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8616 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8618 This addresses the flaw in CBC record processing discovered by
8619 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8620 at: <http://www.isg.rhul.ac.uk/tls/>
8622 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8623 Security Group at Royal Holloway, University of London
8624 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8625 Emilia Käsper for the initial patch.
8628 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8630 * Return an error when checking OCSP signatures when key is NULL.
8631 This fixes a DoS attack. ([CVE-2013-0166])
8635 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8636 the right response is stapled. Also change SSL_get_certificate()
8637 so it returns the certificate actually sent.
8638 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8639 (This is a backport)
8641 *Rob Stradling <rob.stradling@comodo.com>*
8643 * Fix possible deadlock when decoding public keys.
8647 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8649 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8652 * Sanity check record length before skipping explicit IV in DTLS
8655 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8656 fuzzing as a service testing platform.
8661 * Initialise tkeylen properly when encrypting CMS messages.
8662 Thanks to Solar Designer of Openwall for reporting this issue.
8666 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8668 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8669 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8670 in CRYPTO_realloc_clean.
8672 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8673 issue and to Adam Langley <agl@chromium.org> for fixing it.
8676 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8678 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8680 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8681 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8682 content decryption and always return the same error. Note: this attack
8683 needs on average 2^20 messages so it only affects automated senders. The
8684 old behaviour can be re-enabled in the CMS code by setting the
8685 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8686 an MMA defence is not necessary.
8687 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8688 this issue. ([CVE-2012-0884])
8692 * Fix CVE-2011-4619: make sure we really are receiving a
8693 client hello before rejecting multiple SGC restarts. Thanks to
8694 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8698 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8700 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8701 Thanks to Antonio Martin, Enterprise Secure Access Research and
8702 Development, Cisco Systems, Inc. for discovering this bug and
8703 preparing a fix. ([CVE-2012-0050])
8707 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8709 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8710 of the Vaudenay padding oracle attack on CBC mode encryption
8711 which enables an efficient plaintext recovery attack against
8712 the OpenSSL implementation of DTLS. Their attack exploits timing
8713 differences arising during decryption processing. A research
8714 paper describing this attack can be found at:
8715 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8716 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8717 Security Group at Royal Holloway, University of London
8718 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8719 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8720 for preparing the fix. ([CVE-2011-4108])
8722 *Robin Seggelmann, Michael Tuexen*
8724 * Clear bytes used for block padding of SSL 3.0 records.
8727 *Adam Langley (Google)*
8729 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8730 Kadianakis <desnacked@gmail.com> for discovering this issue and
8731 Adam Langley for preparing the fix. ([CVE-2011-4619])
8733 *Adam Langley (Google)*
8735 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8737 *Andrey Kulikov <amdeich@gmail.com>*
8739 * Prevent malformed RFC3779 data triggering an assertion failure.
8740 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8741 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8743 *Rob Austein <sra@hactrn.net>*
8745 * Improved PRNG seeding for VOS.
8747 *Paul Green <Paul.Green@stratus.com>*
8749 * Fix ssl_ciph.c set-up race.
8751 *Adam Langley (Google)*
8753 * Fix spurious failures in ecdsatest.c.
8755 *Emilia Käsper (Google)*
8757 * Fix the BIO_f_buffer() implementation (which was mixing different
8758 interpretations of the `..._len` fields).
8760 *Adam Langley (Google)*
8762 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8763 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8764 threads won't reuse the same blinding coefficients.
8766 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8767 lock to call BN_BLINDING_invert_ex, and avoids one use of
8768 BN_BLINDING_update for each BN_BLINDING structure (previously,
8769 the last update always remained unused).
8771 *Emilia Käsper (Google)*
8773 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8775 *Bob Buckholz (Google)*
8777 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8779 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8780 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8782 *Kaspar Brand <ossl@velox.ch>*
8784 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8785 for multi-threaded use of ECDH. ([CVE-2011-3210])
8787 *Adam Langley (Google)*
8789 * Fix x509_name_ex_d2i memory leak on bad inputs.
8793 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8794 signature public key algorithm by using OID xref utilities instead.
8795 Before this you could only use some ECC ciphersuites with SHA1 only.
8799 * Add protection against ECDSA timing attacks as mentioned in the paper
8800 by Billy Bob Brumley and Nicola Tuveri, see:
8801 <http://eprint.iacr.org/2011/232.pdf>
8803 *Billy Bob Brumley and Nicola Tuveri*
8805 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8807 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8809 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8811 * Fix bug in string printing code: if *any* escaping is enabled we must
8812 escape the escape character (backslash) or the resulting string is
8817 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8819 * Disable code workaround for ancient and obsolete Netscape browsers
8820 and servers: an attacker can use it in a ciphersuite downgrade attack.
8821 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8825 * Fixed J-PAKE implementation error, originally discovered by
8826 Sebastien Martini, further info and confirmation from Stefan
8827 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8831 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8833 * Fix extension code to avoid race conditions which can result in a buffer
8834 overrun vulnerability: resumed sessions must not be modified as they can
8835 be shared by multiple threads. CVE-2010-3864
8839 * Fix WIN32 build system to correctly link an ENGINE directory into
8844 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8846 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8849 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8851 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8853 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8854 context. The operation can be customised via the ctrl mechanism in
8855 case ENGINEs want to include additional functionality.
8859 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8863 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8864 output hashes compatible with older versions of OpenSSL.
8866 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8868 * Fix compression algorithm handling: if resuming a session use the
8869 compression algorithm of the resumed session instead of determining
8870 it from client hello again. Don't allow server to change algorithm.
8874 * Add load_crls() function to commands tidying load_certs() too. Add option
8875 to verify utility to allow additional CRLs to be included.
8879 * Update OCSP request code to permit adding custom headers to the request:
8880 some responders need this.
8884 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8887 *Julia Lawall <julia@diku.dk>*
8889 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8890 needlessly dereferenced structures, used obsolete functions and
8891 didn't handle all updated verify codes correctly.
8895 * Disable MD2 in the default configuration.
8899 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8900 indicate the initial BIO being pushed or popped. This makes it possible
8901 to determine whether the BIO is the one explicitly called or as a result
8902 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8903 it handles reference counts correctly and doesn't zero out the I/O bio
8904 when it is not being explicitly popped. WARNING: applications which
8905 included workarounds for the old buggy behaviour will need to be modified
8906 or they could free up already freed BIOs.
8910 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8911 renaming to all platforms (within the 0.9.8 branch, this was
8912 done conditionally on Netware platforms to avoid a name clash).
8914 *Guenter <lists@gknw.net>*
8916 * Add ECDHE and PSK support to DTLS.
8918 *Michael Tuexen <tuexen@fh-muenster.de>*
8920 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8925 * Add "missing" function EVP_MD_flags() (without this the only way to
8926 retrieve a digest flags is by accessing the structure directly. Update
8927 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8928 or cipher is registered as in the "from" argument. Print out all
8929 registered digests in the dgst usage message instead of manually
8930 attempting to work them out.
8934 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8935 this allows the use of compression and extensions. Change default cipher
8936 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8937 by default unless an application cipher string requests it.
8941 * Alter match criteria in PKCS12_parse(). It used to try to use local
8942 key ids to find matching certificates and keys but some PKCS#12 files
8943 don't follow the (somewhat unwritten) rules and this strategy fails.
8944 Now just gather all certificates together and the first private key
8945 then look for the first certificate that matches the key.
8949 * Support use of registered digest and cipher names for dgst and cipher
8950 commands instead of having to add each one as a special case. So now
8957 openssl dgst -sha256 foo
8959 and this works for ENGINE based algorithms too.
8963 * Update Gost ENGINE to support parameter files.
8965 *Victor B. Wagner <vitus@cryptocom.ru>*
8967 * Support GeneralizedTime in ca utility.
8969 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8971 * Enhance the hash format used for certificate directory links. The new
8972 form uses the canonical encoding (meaning equivalent names will work
8973 even if they aren't identical) and uses SHA1 instead of MD5. This form
8974 is incompatible with the older format and as a result c_rehash should
8975 be used to rebuild symbolic links.
8979 * Make PKCS#8 the default write format for private keys, replacing the
8980 traditional format. This form is standardised, more secure and doesn't
8981 include an implicit MD5 dependency.
8985 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8986 committed to OpenSSL should pass this lot as a minimum.
8990 * Add session ticket override functionality for use by EAP-FAST.
8992 *Jouni Malinen <j@w1.fi>*
8994 * Modify HMAC functions to return a value. Since these can be implemented
8995 in an ENGINE errors can occur.
8999 * Type-checked OBJ_bsearch_ex.
9003 * Type-checked OBJ_bsearch. Also some constification necessitated
9004 by type-checking. Still to come: TXT_DB, bsearch(?),
9005 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9010 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9011 seconds to a tm structure directly, instead of going through OS
9012 specific date routines. This avoids any issues with OS routines such
9013 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9014 and X509_time_adj_ex() to cover the extended range. The existing
9015 X509_time_adj() is still usable and will no longer have any date issues.
9019 * Delta CRL support. New use deltas option which will attempt to locate
9020 and search any appropriate delta CRLs available.
9022 This work was sponsored by Google.
9026 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9027 code and add additional score elements. Validate alternate CRL paths
9028 as part of the CRL checking and indicate a new error "CRL path validation
9029 error" in this case. Applications wanting additional details can use
9030 the verify callback and check the new "parent" field. If this is not
9031 NULL CRL path validation is taking place. Existing applications won't
9032 see this because it requires extended CRL support which is off by
9035 This work was sponsored by Google.
9039 * Support for freshest CRL extension.
9041 This work was sponsored by Google.
9045 * Initial indirect CRL support. Currently only supported in the CRLs
9046 passed directly and not via lookup. Process certificate issuer
9047 CRL entry extension and lookup CRL entries by bother issuer name
9048 and serial number. Check and process CRL issuer entry in IDP extension.
9050 This work was sponsored by Google.
9054 * Add support for distinct certificate and CRL paths. The CRL issuer
9055 certificate is validated separately in this case. Only enabled if
9056 an extended CRL support flag is set: this flag will enable additional
9057 CRL functionality in future.
9059 This work was sponsored by Google.
9063 * Add support for policy mappings extension.
9065 This work was sponsored by Google.
9069 * Fixes to pathlength constraint, self issued certificate handling,
9070 policy processing to align with RFC3280 and PKITS tests.
9072 This work was sponsored by Google.
9076 * Support for name constraints certificate extension. DN, email, DNS
9077 and URI types are currently supported.
9079 This work was sponsored by Google.
9083 * To cater for systems that provide a pointer-based thread ID rather
9084 than numeric, deprecate the current numeric thread ID mechanism and
9085 replace it with a structure and associated callback type. This
9086 mechanism allows a numeric "hash" to be extracted from a thread ID in
9087 either case, and on platforms where pointers are larger than 'long',
9088 mixing is done to help ensure the numeric 'hash' is usable even if it
9089 can't be guaranteed unique. The default mechanism is to use "&errno"
9090 as a pointer-based thread ID to distinguish between threads.
9092 Applications that want to provide their own thread IDs should now use
9093 CRYPTO_THREADID_set_callback() to register a callback that will call
9094 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9096 Note that ERR_remove_state() is now deprecated, because it is tied
9097 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9098 to free the current thread's error state should be replaced by
9099 ERR_remove_thread_state(NULL).
9101 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9102 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9103 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9104 application was previously providing a numeric thread callback that
9105 was inappropriate for distinguishing threads, then uniqueness might
9106 have been obtained with &errno that happened immediately in the
9107 intermediate development versions of OpenSSL; this is no longer the
9108 case, the numeric thread callback will now override the automatic use
9111 *Geoff Thorpe, with help from Bodo Moeller*
9113 * Initial support for different CRL issuing certificates. This covers a
9114 simple case where the self issued certificates in the chain exist and
9115 the real CRL issuer is higher in the existing chain.
9117 This work was sponsored by Google.
9121 * Removed effectively defunct crypto/store from the build.
9125 * Revamp of STACK to provide stronger type-checking. Still to come:
9126 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9127 ASN1_STRING, CONF_VALUE.
9131 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9132 RAM on SSL connections. This option can save about 34k per idle SSL.
9136 * Revamp of LHASH to provide stronger type-checking. Still to come:
9137 STACK, TXT_DB, bsearch, qsort.
9141 * Initial support for Cryptographic Message Syntax (aka CMS) based
9142 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9143 support for data, signedData, compressedData, digestedData and
9144 encryptedData, envelopedData types included. Scripts to check against
9145 RFC4134 examples draft and interop and consistency checks of many
9146 content types and variants.
9150 * Add options to enc utility to support use of zlib compression BIO.
9154 * Extend mk1mf to support importing of options and assembly language
9155 files from Configure script, currently only included in VC-WIN32.
9156 The assembly language rules can now optionally generate the source
9157 files from the associated perl scripts.
9161 * Implement remaining functionality needed to support GOST ciphersuites.
9162 Interop testing has been performed using CryptoPro implementations.
9164 *Victor B. Wagner <vitus@cryptocom.ru>*
9166 * s390x assembler pack.
9170 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9175 * Implement Opaque PRF Input TLS extension as specified in
9176 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9177 official specification yet and no extension type assignment by
9178 IANA exists, this extension (for now) will have to be explicitly
9179 enabled when building OpenSSL by providing the extension number
9180 to use. For example, specify an option
9182 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9184 to the "config" or "Configure" script to enable the extension,
9185 assuming extension number 0x9527 (which is a completely arbitrary
9186 and unofficial assignment based on the MD5 hash of the Internet
9187 Draft). Note that by doing so, you potentially lose
9188 interoperability with other TLS implementations since these might
9189 be using the same extension number for other purposes.
9191 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9192 opaque PRF input value to use in the handshake. This will create
9193 an internal copy of the length-'len' string at 'src', and will
9194 return non-zero for success.
9196 To get more control and flexibility, provide a callback function
9199 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9200 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9204 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9207 Callback function 'cb' will be called in handshakes, and is
9208 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9209 Argument 'arg' is for application purposes (the value as given to
9210 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9211 be provided to the callback function). The callback function
9212 has to return non-zero to report success: usually 1 to use opaque
9213 PRF input just if possible, or 2 to enforce use of the opaque PRF
9214 input. In the latter case, the library will abort the handshake
9215 if opaque PRF input is not successfully negotiated.
9217 Arguments 'peerinput' and 'len' given to the callback function
9218 will always be NULL and 0 in the case of a client. A server will
9219 see the client's opaque PRF input through these variables if
9220 available (NULL and 0 otherwise). Note that if the server
9221 provides an opaque PRF input, the length must be the same as the
9222 length of the client's opaque PRF input.
9224 Note that the callback function will only be called when creating
9225 a new session (session resumption can resume whatever was
9226 previously negotiated), and will not be called in SSL 2.0
9227 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9228 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9229 for applications that need to enforce opaque PRF input.
9233 * Update ssl code to support digests other than SHA1+MD5 for handshake
9236 *Victor B. Wagner <vitus@cryptocom.ru>*
9238 * Add RFC4507 support to OpenSSL. This includes the corrections in
9239 RFC4507bis. The encrypted ticket format is an encrypted encoded
9240 SSL_SESSION structure, that way new session features are automatically
9243 If a client application caches session in an SSL_SESSION structure
9244 support is transparent because tickets are now stored in the encoded
9247 The SSL_CTX structure automatically generates keys for ticket
9248 protection in servers so again support should be possible
9249 with no application modification.
9251 If a client or server wishes to disable RFC4507 support then the option
9252 SSL_OP_NO_TICKET can be set.
9254 Add a TLS extension debugging callback to allow the contents of any client
9255 or server extensions to be examined.
9257 This work was sponsored by Google.
9261 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9262 OpenSSL should now compile cleanly on gcc 4.2
9264 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9266 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9267 support including streaming MAC support: this is required for GOST
9268 ciphersuite support.
9270 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9272 * Add option -stream to use PKCS#7 streaming in smime utility. New
9273 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9274 to output in BER and PEM format.
9278 * Experimental support for use of HMAC via EVP_PKEY interface. This
9279 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9280 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9281 ENGINE support for HMAC keys which are unextractable. New -mac and
9282 -macopt options to dgst utility.
9286 * New option -sigopt to dgst utility. Update dgst to use
9287 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9288 alternative signing parameters such as X9.31 or PSS in the dgst
9293 * Change ssl_cipher_apply_rule(), the internal function that does
9294 the work each time a ciphersuite string requests enabling
9295 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9296 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9297 the order of disabled ciphersuites such that those ciphersuites
9298 that most recently went from enabled to disabled not only stay
9299 in order with respect to each other, but also have higher priority
9300 than other disabled ciphersuites the next time ciphersuites are
9303 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9304 the same ciphersuites as with "HIGH" alone, but in a specific
9305 order where the PSK ciphersuites come first (since they are the
9306 most recently disabled ciphersuites when "HIGH" is parsed).
9308 Also, change ssl_create_cipher_list() (using this new
9309 functionality) such that between otherwise identical
9310 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9315 * Change ssl_create_cipher_list() so that it automatically
9316 arranges the ciphersuites in reasonable order before starting
9317 to process the rule string. Thus, the definition for "DEFAULT"
9318 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9319 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9320 This makes it much easier to arrive at a reasonable default order
9321 in applications for which anonymous ciphers are OK (meaning
9322 that you can't actually use DEFAULT).
9324 *Bodo Moeller; suggested by Victor Duchovni*
9326 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9327 processing) into multiple integers instead of setting
9328 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9329 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9330 (These masks as well as the individual bit definitions are hidden
9331 away into the non-exported interface ssl/ssl_locl.h, so this
9332 change to the definition of the SSL_CIPHER structure shouldn't
9333 affect applications.) This give us more bits for each of these
9334 categories, so there is no longer a need to coagulate AES128 and
9335 AES256 into a single algorithm bit, and to coagulate Camellia128
9336 and Camellia256 into a single algorithm bit, which has led to all
9339 Thus, among other things, the kludge introduced in 0.9.7m and
9340 0.9.8e for masking out AES256 independently of AES128 or masking
9341 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9343 With the change, we also introduce new ciphersuite aliases that
9344 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9349 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9350 Use the leftmost N bytes of the signature input if the input is
9351 larger than the prime q (with N being the size in bytes of q).
9355 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9356 it yet and it is largely untested.
9360 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9364 * Initial incomplete changes to avoid need for function casts in OpenSSL
9365 some compilers (gcc 4.2 and later) reject their use. Safestack is
9366 reimplemented. Update ASN1 to avoid use of legacy functions.
9370 * Win32/64 targets are linked with Winsock2.
9374 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9375 to external functions. This can be used to increase CRL handling
9376 efficiency especially when CRLs are very large by (for example) storing
9377 the CRL revoked certificates in a database.
9381 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9382 new CRLs added to a directory can be used. New command line option
9383 -verify_return_error to s_client and s_server. This causes real errors
9384 to be returned by the verify callback instead of carrying on no matter
9385 what. This reflects the way a "real world" verify callback would behave.
9389 * GOST engine, supporting several GOST algorithms and public key formats.
9390 Kindly donated by Cryptocom.
9394 * Partial support for Issuing Distribution Point CRL extension. CRLs
9395 partitioned by DP are handled but no indirect CRL or reason partitioning
9396 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9397 selected via a scoring technique which handles IDP and AKID in CRLs.
9401 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9402 will ultimately be used for all verify operations: this will remove the
9403 X509_STORE dependency on certificate verification and allow alternative
9404 lookup methods. X509_STORE based implementations of these two callbacks.
9408 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9409 Modify get_crl() to find a valid (unexpired) CRL if possible.
9413 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9414 this would be called X509_CRL_cmp() but that name is already used by
9415 a function that just compares CRL issuer names. Cache several CRL
9416 extensions in X509_CRL structure and cache CRLDP in X509.
9420 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9421 this maps equivalent X509_NAME structures into a consistent structure.
9422 Name comparison can then be performed rapidly using memcmp().
9426 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9431 * Allow digests to supply their own micalg string for S/MIME type using
9432 the ctrl EVP_MD_CTRL_MICALG.
9436 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9437 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9438 ctrl. It can then customise the structure before and/or after signing
9443 * New function OBJ_add_sigid() to allow application defined signature OIDs
9444 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9445 to free up any added signature OIDs.
9449 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9450 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9451 digest and cipher tables. New options added to openssl utility:
9452 list-message-digest-algorithms and list-cipher-algorithms.
9456 * Change the array representation of binary polynomials: the list
9457 of degrees of non-zero coefficients is now terminated with -1.
9458 Previously it was terminated with 0, which was also part of the
9459 value; thus, the array representation was not applicable to
9460 polynomials where t^0 has coefficient zero. This change makes
9461 the array representation useful in a more general context.
9465 * Various modifications and fixes to SSL/TLS cipher string
9466 handling. For ECC, the code now distinguishes between fixed ECDH
9467 with RSA certificates on the one hand and with ECDSA certificates
9468 on the other hand, since these are separate ciphersuites. The
9469 unused code for Fortezza ciphersuites has been removed.
9471 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9472 (not "ECDHE"). For consistency with the code for DH
9473 certificates, use of ECDH certificates is now considered ECDH
9474 authentication, not RSA or ECDSA authentication (the latter is
9475 merely the CA's signing algorithm and not actively used in the
9478 The temporary ciphersuite alias "ECCdraft" is no longer
9479 available, and ECC ciphersuites are no longer excluded from "ALL"
9480 and "DEFAULT". The following aliases now exist for RFC 4492
9481 ciphersuites, most of these by analogy with the DH case:
9483 kECDHr - ECDH cert, signed with RSA
9484 kECDHe - ECDH cert, signed with ECDSA
9485 kECDH - ECDH cert (signed with either RSA or ECDSA)
9486 kEECDH - ephemeral ECDH
9487 ECDH - ECDH cert or ephemeral ECDH
9493 AECDH - anonymous ECDH
9494 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9498 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9499 Use correct micalg parameters depending on digest(s) in signed message.
9503 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9504 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9508 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9509 an engine to register a method. Add ENGINE lookups for methods and
9510 functional reference processing.
9514 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9515 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9520 * New -resign option to smime utility. This adds one or more signers
9521 to an existing PKCS#7 signedData structure. Also -md option to use an
9522 alternative message digest algorithm for signing.
9526 * Tidy up PKCS#7 routines and add new functions to make it easier to
9527 create PKCS7 structures containing multiple signers. Update smime
9528 application to support multiple signers.
9532 * New -macalg option to pkcs12 utility to allow setting of an alternative
9537 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9538 Reorganize PBE internals to lookup from a static table using NIDs,
9539 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9540 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9541 PRF which will be automatically used with PBES2.
9545 * Replace the algorithm specific calls to generate keys in "req" with the
9550 * Update PKCS#7 enveloped data routines to use new API. This is now
9551 supported by any public key method supporting the encrypt operation. A
9552 ctrl is added to allow the public key algorithm to examine or modify
9553 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9558 * Add a ctrl to asn1 method to allow a public key algorithm to express
9559 a default digest type to use. In most cases this will be SHA1 but some
9560 algorithms (such as GOST) need to specify an alternative digest. The
9561 return value indicates how strong the preference is 1 means optional and
9562 2 is mandatory (that is it is the only supported type). Modify
9563 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9564 use the default md. Update openssl utilities to use the default digest
9565 type for signing if it is not explicitly indicated.
9569 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9570 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9571 signing method from the key type. This effectively removes the link
9572 between digests and public key types.
9576 * Add an OID cross reference table and utility functions. Its purpose is to
9577 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9578 rsaEncryption. This will allow some of the algorithm specific hackery
9579 needed to use the correct OID to be removed.
9583 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9584 structures for PKCS7_sign(). They are now set up by the relevant public
9589 * Add provisional EC pkey method with support for ECDSA and ECDH.
9593 * Add support for key derivation (agreement) in the API, DH method and
9598 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9599 public and private key formats. As a side effect these add additional
9600 command line functionality not previously available: DSA signatures can be
9601 generated and verified using pkeyutl and DH key support and generation in
9608 *Oliver Tappe <zooey@hirschkaefer.de>*
9610 * New make target "install_html_docs" installs HTML renditions of the
9613 *Oliver Tappe <zooey@hirschkaefer.de>*
9615 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9616 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9617 support key and parameter generation and add initial key generation
9618 functionality for RSA.
9622 * Add functions for main EVP_PKEY_method operations. The undocumented
9623 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9624 `EVP_PKEY_{encrypt,decrypt}_old`.
9628 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9629 key API, doesn't do much yet.
9633 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9634 public key algorithms. New option to openssl utility:
9635 "list-public-key-algorithms" to print out info.
9639 * Implement the Supported Elliptic Curves Extension for
9640 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9644 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9645 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9649 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9650 utilities such as rsa, dsa, dsaparam etc except they process any key
9655 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9656 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9657 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9662 * Initial support for pluggable public key ASN1.
9663 De-spaghettify the public key ASN1 handling. Move public and private
9664 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9665 algorithm specific handling to a single module within the relevant
9666 algorithm directory. Add functions to allow (near) opaque processing
9667 of public and private key structures.
9671 * Implement the Supported Point Formats Extension for
9672 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9676 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9677 for the psk identity [hint] and the psk callback functions to the
9678 SSL_SESSION, SSL and SSL_CTX structure.
9681 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9685 SSL_CTX_use_psk_identity_hint
9686 SSL_get_psk_identity_hint
9687 SSL_get_psk_identity
9688 SSL_use_psk_identity_hint
9690 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9692 * Add RFC 3161 compliant time stamp request creation, response generation
9693 and response verification functionality.
9695 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9697 * Add initial support for TLS extensions, specifically for the server_name
9698 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9699 have new members for a host name. The SSL data structure has an
9700 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9701 stored in that context to allow for session resumption, even after the
9702 SSL has been switched to a new SSL_CTX in reaction to a client's
9703 server_name extension.
9705 New functions (subject to change):
9707 SSL_get_servername()
9708 SSL_get_servername_type()
9711 New CTRL codes and macros (subject to change):
9713 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9714 - SSL_CTX_set_tlsext_servername_callback()
9715 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9716 - SSL_CTX_set_tlsext_servername_arg()
9717 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9719 openssl s_client has a new '-servername ...' option.
9721 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9722 '-key2 ...', '-servername_fatal' (subject to change). This allows
9723 testing the HostName extension for a specific single host name ('-cert'
9724 and '-key' remain fallbacks for handshakes without HostName
9725 negotiation). If the unrecognized_name alert has to be sent, this by
9726 default is a warning; it becomes fatal with the '-servername_fatal'
9729 *Peter Sylvester, Remy Allais, Christophe Renou*
9731 * Whirlpool hash implementation is added.
9735 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9736 bn(64,32). Because of instruction set limitations it doesn't have
9737 any negative impact on performance. This was done mostly in order
9738 to make it possible to share assembler modules, such as bn_mul_mont
9739 implementations, between 32- and 64-bit builds without hassle.
9743 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9744 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9749 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9750 dedicated Montgomery multiplication procedure, is introduced.
9751 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9752 "64-bit" performance on certain 32-bit targets.
9756 * New option SSL_OP_NO_COMP to disable use of compression selectively
9757 in SSL structures. New SSL ctrl to set maximum send fragment size.
9758 Save memory by setting the I/O buffer sizes dynamically instead of
9759 using the maximum available value.
9763 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9764 in addition to the text details.
9768 * Very, very preliminary EXPERIMENTAL support for printing of general
9769 ASN1 structures. This currently produces rather ugly output and doesn't
9770 handle several customised structures at all.
9774 * Integrated support for PVK file format and some related formats such
9775 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9776 these in the 'rsa' and 'dsa' utilities.
9780 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9784 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9785 place for the (very old) "NETSCAPE" format certificates which are now
9786 handled using new ASN1 code equivalents.
9790 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9791 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9792 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9796 * Modify CRL distribution points extension code to print out previously
9797 unsupported fields. Enhance extension setting code to allow setting of
9802 * Add print and set support for Issuing Distribution Point CRL extension.
9806 * Change 'Configure' script to enable Camellia by default.
9813 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9815 * When rejecting SSL/TLS records due to an incorrect version number, never
9816 update s->server with a new major version number. As of
9817 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9818 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9819 the previous behavior could result in a read attempt at NULL when
9820 receiving specific incorrect SSL/TLS records once record payload
9821 protection is active. ([CVE-2010-0740])
9823 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9825 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9826 could be crashed if the relevant tables were not present (e.g. chrooted).
9828 *Tomas Hoger <thoger@redhat.com>*
9830 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9832 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9834 *Martin Olsson, Neel Mehta*
9836 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9837 accommodate for stack sorting, always a write lock!).
9841 * On some versions of WIN32 Heap32Next is very slow. This can cause
9842 excessive delays in the RAND_poll(): over a minute. As a workaround
9843 include a time check in the inner Heap32Next loop too.
9847 * The code that handled flushing of data in SSL/TLS originally used the
9848 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9849 the problem outlined in PR#1949. The fix suggested there however can
9850 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9851 of Apache). So instead simplify the code to flush unconditionally.
9852 This should be fine since flushing with no data to flush is a no op.
9856 * Handle TLS versions 2.0 and later properly and correctly use the
9857 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9858 off ancient servers have a habit of sticking around for a while...
9862 * Modify compression code so it frees up structures without using the
9863 ex_data callbacks. This works around a problem where some applications
9864 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9865 restarting) then use compression (e.g. SSL with compression) later.
9866 This results in significant per-connection memory leaks and
9867 has caused some security issues including CVE-2008-1678 and
9872 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9873 change when encrypting or decrypting.
9877 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9878 connect and renegotiate with servers which do not support RI.
9879 Until RI is more widely deployed this option is enabled by default.
9883 * Add "missing" ssl ctrls to clear options and mode.
9887 * If client attempts to renegotiate and doesn't support RI respond with
9888 a no_renegotiation alert as required by RFC5746. Some renegotiating
9889 TLS clients will continue a connection gracefully when they receive
9890 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9891 waiting for a server hello which it will never receive. Now we treat a
9892 received no_renegotiation alert as a fatal error. This is because
9893 applications requesting a renegotiation might well expect it to succeed
9894 and would have no code in place to handle the server denying it so the
9895 only safe thing to do is to terminate the connection.
9899 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9900 peer supports secure renegotiation and 0 otherwise. Print out peer
9901 renegotiation support in s_client/s_server.
9905 * Replace the highly broken and deprecated SPKAC certification method with
9906 the updated NID creation version. This should correctly handle UTF8.
9910 * Implement RFC5746. Re-enable renegotiation but require the extension
9911 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9912 turns out to be a bad idea. It has been replaced by
9913 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9914 SSL_CTX_set_options(). This is really not recommended unless you
9915 know what you are doing.
9917 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9919 * Fixes to stateless session resumption handling. Use initial_ctx when
9920 issuing and attempting to decrypt tickets in case it has changed during
9921 servername handling. Use a non-zero length session ID when attempting
9922 stateless session resumption: this makes it possible to determine if
9923 a resumption has occurred immediately after receiving server hello
9924 (several places in OpenSSL subtly assume this) instead of later in
9929 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9930 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9931 fixes for a few places where the return code is not checked
9934 *Julia Lawall <julia@diku.dk>*
9936 * Add --strict-warnings option to Configure script to include devteam
9937 warnings in other configurations.
9941 * Add support for --libdir option and LIBDIR variable in makefiles. This
9942 makes it possible to install openssl libraries in locations which
9943 have names other than "lib", for example "/usr/lib64" which some
9946 *Steve Henson, based on patch from Jeremy Utley*
9948 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9949 X690 8.9.12 and can produce some misleading textual output of OIDs.
9951 *Steve Henson, reported by Dan Kaminsky*
9953 * Delete MD2 from algorithm tables. This follows the recommendation in
9954 several standards that it is not used in new applications due to
9955 several cryptographic weaknesses. For binary compatibility reasons
9956 the MD2 API is still compiled in by default.
9960 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9965 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9966 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9969 *Guenter <lists@gknw.net>*
9971 * Fix the server certificate chain building code to use X509_verify_cert(),
9972 it used to have an ad-hoc builder which was unable to cope with anything
9973 other than a simple chain.
9975 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9977 * Don't check self signed certificate signatures in X509_verify_cert()
9978 by default (a flag can override this): it just wastes time without
9979 adding any security. As a useful side effect self signed root CAs
9980 with non-FIPS digests are now usable in FIPS mode.
9984 * In dtls1_process_out_of_seq_message() the check if the current message
9985 is already buffered was missing. For every new message was memory
9986 allocated, allowing an attacker to perform an denial of service attack
9987 with sending out of seq handshake messages until there is no memory
9988 left. Additionally every future message was buffered, even if the
9989 sequence number made no sense and would be part of another handshake.
9990 So only messages with sequence numbers less than 10 in advance will be
9991 buffered. ([CVE-2009-1378])
9993 *Robin Seggelmann, discovered by Daniel Mentz*
9995 * Records are buffered if they arrive with a future epoch to be
9996 processed after finishing the corresponding handshake. There is
9997 currently no limitation to this buffer allowing an attacker to perform
9998 a DOS attack with sending records with future epochs until there is no
9999 memory left. This patch adds the pqueue_size() function to determine
10000 the size of a buffer and limits the record buffer to 100 entries.
10003 *Robin Seggelmann, discovered by Daniel Mentz*
10005 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10006 parent structure is freed. ([CVE-2009-1379])
10010 * Handle non-blocking I/O properly in SSL_shutdown() call.
10012 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10014 * Add `2.5.4.*` OIDs
10016 *Ilya O. <vrghost@gmail.com>*
10018 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10020 * Disable renegotiation completely - this fixes a severe security
10021 problem ([CVE-2009-3555]) at the cost of breaking all
10022 renegotiation. Renegotiation can be re-enabled by setting
10023 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10024 run-time. This is really not recommended unless you know what
10029 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10031 * Don't set val to NULL when freeing up structures, it is freed up by
10032 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10033 zeroing past the valid field. ([CVE-2009-0789])
10035 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10037 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10038 checked correctly. This would allow some invalid signed attributes to
10039 appear to verify correctly. ([CVE-2009-0591])
10041 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10043 * Reject UniversalString and BMPString types with invalid lengths. This
10044 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10045 a legal length. ([CVE-2009-0590])
10049 * Set S/MIME signing as the default purpose rather than setting it
10050 unconditionally. This allows applications to override it at the store
10055 * Permit restricted recursion of ASN1 strings. This is needed in practice
10056 to handle some structures.
10060 * Improve efficiency of mem_gets: don't search whole buffer each time
10063 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10065 * New -hex option for openssl rand.
10069 * Print out UTF8String and NumericString when parsing ASN1.
10073 * Support NumericString type for name components.
10077 * Allow CC in the environment to override the automatically chosen
10078 compiler. Note that nothing is done to ensure flags work with the
10083 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10085 * Properly check EVP_VerifyFinal() and similar return values
10088 *Ben Laurie, Bodo Moeller, Google Security Team*
10090 * Enable TLS extensions by default.
10094 * Allow the CHIL engine to be loaded, whether the application is
10095 multithreaded or not. (This does not release the developer from the
10096 obligation to set up the dynamic locking callbacks.)
10098 *Sander Temme <sander@temme.net>*
10100 * Use correct exit code if there is an error in dgst command.
10102 *Steve Henson; problem pointed out by Roland Dirlewanger*
10104 * Tweak Configure so that you need to say "experimental-jpake" to enable
10105 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10109 * Add experimental JPAKE support, including demo authentication in
10110 s_client and s_server.
10114 * Set the comparison function in v3_addr_canonize().
10116 *Rob Austein <sra@hactrn.net>*
10118 * Add support for XMPP STARTTLS in s_client.
10120 *Philip Paeps <philip@freebsd.org>*
10122 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10123 to ensure that even with this option, only ciphersuites in the
10124 server's preference list will be accepted. (Note that the option
10125 applies only when resuming a session, so the earlier behavior was
10126 just about the algorithm choice for symmetric cryptography.)
10130 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10132 * Fix NULL pointer dereference if a DTLS server received
10133 ChangeCipherSpec as first record ([CVE-2009-1386]).
10137 * Fix a state transition in s3_srvr.c and d1_srvr.c
10138 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10140 *Nagendra Modadugu*
10142 * The fix in 0.9.8c that supposedly got rid of unsafe
10143 double-checked locking was incomplete for RSA blinding,
10144 addressing just one layer of what turns out to have been
10145 doubly unsafe triple-checked locking.
10147 So now fix this for real by retiring the MONT_HELPER macro
10148 in crypto/rsa/rsa_eay.c.
10150 *Bodo Moeller; problem pointed out by Marius Schilder*
10152 * Various precautionary measures:
10154 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10156 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10157 (NB: This would require knowledge of the secret session ticket key
10158 to exploit, in which case you'd be SOL either way.)
10160 - Change bn_nist.c so that it will properly handle input BIGNUMs
10161 outside the expected range.
10163 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10166 *Neel Mehta, Bodo Moeller*
10168 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10169 the load fails. Useful for distros.
10171 *Ben Laurie and the FreeBSD team*
10173 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10177 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10181 * Expand ENGINE to support engine supplied SSL client certificate functions.
10183 This work was sponsored by Logica.
10187 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10188 keystores. Support for SSL/TLS client authentication too.
10189 Not compiled unless enable-capieng specified to Configure.
10191 This work was sponsored by Logica.
10195 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10196 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10197 attribute creation routines such as certificate requests and PKCS#12
10202 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10204 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10205 handshake which could lead to a client crash as found using the
10206 Codenomicon TLS test suite ([CVE-2008-1672])
10208 *Steve Henson, Mark Cox*
10210 * Fix double free in TLS server name extensions which could lead to
10211 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10215 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10217 Clear the error queue to ensure that error entries left from
10218 older function calls do not interfere with the correct operation.
10220 *Lutz Jaenicke, Erik de Castro Lopo*
10222 * Remove root CA certificates of commercial CAs:
10224 The OpenSSL project does not recommend any specific CA and does not
10225 have any policy with respect to including or excluding any CA.
10226 Therefore it does not make any sense to ship an arbitrary selection
10227 of root CA certificates with the OpenSSL software.
10231 * RSA OAEP patches to fix two separate invalid memory reads.
10232 The first one involves inputs when 'lzero' is greater than
10233 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10234 before the beginning of from). The second one involves inputs where
10235 the 'db' section contains nothing but zeroes (there is a one-byte
10236 invalid read after the end of 'db').
10238 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10240 * Partial backport from 0.9.9-dev:
10242 Introduce bn_mul_mont (dedicated Montgomery multiplication
10243 procedure) as a candidate for BIGNUM assembler implementation.
10244 While 0.9.9-dev uses assembler for various architectures, only
10245 x86_64 is available by default here in the 0.9.8 branch, and
10246 32-bit x86 is available through a compile-time setting.
10248 To try the 32-bit x86 assembler implementation, use Configure
10249 option "enable-montasm" (which exists only for this backport).
10251 As "enable-montasm" for 32-bit x86 disclaims code stability
10252 anyway, in this constellation we activate additional code
10253 backported from 0.9.9-dev for further performance improvements,
10254 namely BN_from_montgomery_word. (To enable this otherwise,
10255 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10257 *Andy Polyakov (backport partially by Bodo Moeller)*
10259 * Add TLS session ticket callback. This allows an application to set
10260 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10261 values. This is useful for key rollover for example where several key
10262 sets may exist with different names.
10266 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10267 This was broken until now in 0.9.8 releases, such that the only way
10268 a registered ENGINE could be used (assuming it initialises
10269 successfully on the host) was to explicitly set it as the default
10270 for the relevant algorithms. This is in contradiction with 0.9.7
10271 behaviour and the documentation. With this fix, when an ENGINE is
10272 registered into a given algorithm's table of implementations, the
10273 'uptodate' flag is reset so that auto-discovery will be used next
10274 time a new context for that algorithm attempts to select an
10277 *Ian Lister (tweaked by Geoff Thorpe)*
10279 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10280 implementation in the following ways:
10282 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10285 Lack of BER streaming support means one pass streaming processing is
10286 only supported if data is detached: setting the streaming flag is
10287 ignored for embedded content.
10289 CMS support is disabled by default and must be explicitly enabled
10290 with the enable-cms configuration option.
10294 * Update the GMP engine glue to do direct copies between BIGNUM and
10295 mpz_t when openssl and GMP use the same limb size. Otherwise the
10296 existing "conversion via a text string export" trick is still used.
10298 *Paul Sheer <paulsheer@gmail.com>*
10300 * Zlib compression BIO. This is a filter BIO which compressed and
10301 uncompresses any data passed through it.
10305 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10306 RFC3394 compatible AES key wrapping.
10310 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10311 sets string data without copying. X509_ALGOR_set0() and
10312 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10313 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10314 from an X509_ATTRIBUTE structure optionally checking it occurs only
10315 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10320 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10321 to get the expected BN_FLG_CONSTTIME behavior.
10323 *Bodo Moeller (Google)*
10327 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10328 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10329 - added some more tests to do_tests.pl
10330 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10331 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10332 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10333 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10334 - various changes to netware.pl to enable gcc-cross builds on Win32
10336 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10337 - various changes to fix missing prototype warnings
10338 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10339 - added AES, WHIRLPOOL and CPUID assembler code to build files
10340 - added missing AES assembler make rules to mk1mf.pl
10341 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10343 *Guenter Knauf <eflash@gmx.net>*
10345 * Implement certificate status request TLS extension defined in RFC3546.
10346 A client can set the appropriate parameters and receive the encoded
10347 OCSP response via a callback. A server can query the supplied parameters
10348 and set the encoded OCSP response in the callback. Add simplified examples
10349 to s_client and s_server.
10353 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10355 * Fix various bugs:
10356 + Binary incompatibility of ssl_ctx_st structure
10357 + DTLS interoperation with non-compliant servers
10358 + Don't call get_session_cb() without proposed session
10359 + Fix ia64 assembler code
10361 *Andy Polyakov, Steve Henson*
10363 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10365 * DTLS Handshake overhaul. There were longstanding issues with
10366 OpenSSL DTLS implementation, which were making it impossible for
10367 RFC 4347 compliant client to communicate with OpenSSL server.
10368 Unfortunately just fixing these incompatibilities would "cut off"
10369 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10370 server keeps tolerating non RFC compliant syntax. The opposite is
10371 not true, 0.9.8f client can not communicate with earlier server.
10372 This update even addresses CVE-2007-4995.
10376 * Changes to avoid need for function casts in OpenSSL: some compilers
10377 (gcc 4.2 and later) reject their use.
10378 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10381 * Add RFC4507 support to OpenSSL. This includes the corrections in
10382 RFC4507bis. The encrypted ticket format is an encrypted encoded
10383 SSL_SESSION structure, that way new session features are automatically
10386 If a client application caches session in an SSL_SESSION structure
10387 support is transparent because tickets are now stored in the encoded
10390 The SSL_CTX structure automatically generates keys for ticket
10391 protection in servers so again support should be possible
10392 with no application modification.
10394 If a client or server wishes to disable RFC4507 support then the option
10395 SSL_OP_NO_TICKET can be set.
10397 Add a TLS extension debugging callback to allow the contents of any client
10398 or server extensions to be examined.
10400 This work was sponsored by Google.
10404 * Add initial support for TLS extensions, specifically for the server_name
10405 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10406 have new members for a host name. The SSL data structure has an
10407 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10408 stored in that context to allow for session resumption, even after the
10409 SSL has been switched to a new SSL_CTX in reaction to a client's
10410 server_name extension.
10412 New functions (subject to change):
10414 SSL_get_servername()
10415 SSL_get_servername_type()
10418 New CTRL codes and macros (subject to change):
10420 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10421 - SSL_CTX_set_tlsext_servername_callback()
10422 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10423 - SSL_CTX_set_tlsext_servername_arg()
10424 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10426 openssl s_client has a new '-servername ...' option.
10428 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10429 '-key2 ...', '-servername_fatal' (subject to change). This allows
10430 testing the HostName extension for a specific single host name ('-cert'
10431 and '-key' remain fallbacks for handshakes without HostName
10432 negotiation). If the unrecognized_name alert has to be sent, this by
10433 default is a warning; it becomes fatal with the '-servername_fatal'
10436 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10438 * Add AES and SSE2 assembly language support to VC++ build.
10442 * Mitigate attack on final subtraction in Montgomery reduction.
10446 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10447 (which previously caused an internal error).
10451 * Squeeze another 10% out of IGE mode when in != out.
10455 * AES IGE mode speedup.
10457 *Dean Gaudet (Google)*
10459 * Add the Korean symmetric 128-bit cipher SEED (see
10460 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10461 add SEED ciphersuites from RFC 4162:
10463 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10464 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10465 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10466 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10468 To minimize changes between patchlevels in the OpenSSL 0.9.8
10469 series, SEED remains excluded from compilation unless OpenSSL
10470 is configured with 'enable-seed'.
10472 *KISA, Bodo Moeller*
10474 * Mitigate branch prediction attacks, which can be practical if a
10475 single processor is shared, allowing a spy process to extract
10476 information. For detailed background information, see
10477 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10478 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10479 and Necessary Software Countermeasures"). The core of the change
10480 are new versions BN_div_no_branch() and
10481 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10482 respectively, which are slower, but avoid the security-relevant
10483 conditional branches. These are automatically called by BN_div()
10484 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10485 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10486 remove a conditional branch.
10488 BN_FLG_CONSTTIME is the new name for the previous
10489 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10490 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10491 in the exponent causes BN_mod_exp_mont() to use the alternative
10492 implementation in BN_mod_exp_mont_consttime().) The old name
10493 remains as a deprecated alias.
10495 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10496 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10497 constant-time implementations for more than just exponentiation.
10498 Here too the old name is kept as a deprecated alias.
10500 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10501 the BN_BLINDING structure gets an independent copy of the
10502 modulus. This means that the previous `BIGNUM *m` argument to
10503 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10504 essentially becomes `const BIGNUM *m`, although we can't actually
10505 change this in the header file before 0.9.9. It allows
10506 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10507 enable BN_FLG_CONSTTIME.
10509 *Matthew D Wood (Intel Corp)*
10511 * In the SSL/TLS server implementation, be strict about session ID
10512 context matching (which matters if an application uses a single
10513 external cache for different purposes). Previously,
10514 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10515 set. This did ensure strict client verification, but meant that,
10516 with applications using a single external cache for quite
10517 different requirements, clients could circumvent ciphersuite
10518 restrictions for a given session ID context by starting a session
10519 in a different context.
10523 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10524 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10525 authentication-only ciphersuites.
10529 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10530 not complete and could lead to a possible single byte overflow
10531 ([CVE-2007-5135]) [Ben Laurie]
10533 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10535 * Since AES128 and AES256 (and similarly Camellia128 and
10536 Camellia256) share a single mask bit in the logic of
10537 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10538 kludge to work properly if AES128 is available and AES256 isn't
10539 (or if Camellia128 is available and Camellia256 isn't).
10543 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10544 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10545 When a point or a seed is encoded in a BIT STRING, we need to
10546 prevent the removal of trailing zero bits to get the proper DER
10547 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10548 of a NamedBitList, for which trailing 0 bits need to be removed.)
10552 * Have SSL/TLS server implementation tolerate "mismatched" record
10553 protocol version while receiving ClientHello even if the
10554 ClientHello is fragmented. (The server can't insist on the
10555 particular protocol version it has chosen before the ServerHello
10556 message has informed the client about his choice.)
10560 * Add RFC 3779 support.
10562 *Rob Austein for ARIN, Ben Laurie*
10564 * Load error codes if they are not already present instead of using a
10565 static variable. This allows them to be cleanly unloaded and reloaded.
10566 Improve header file function name parsing.
10570 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10571 or CAPABILITY handshake as required by RFCs.
10573 *Goetz Babin-Ebell*
10575 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10577 * Introduce limits to prevent malicious keys being able to
10578 cause a denial of service. ([CVE-2006-2940])
10580 *Steve Henson, Bodo Moeller*
10582 * Fix ASN.1 parsing of certain invalid structures that can result
10583 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10585 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10586 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10588 * Fix SSL client code which could crash if connecting to a
10589 malicious SSLv2 server. ([CVE-2006-4343])
10591 *Tavis Ormandy and Will Drewry, Google Security Team*
10593 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10594 match only those. Before that, "AES256-SHA" would be interpreted
10595 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10596 the same strength classification in 0.9.7h) as we currently only
10597 have a single AES bit in the ciphersuite description bitmap.
10598 That change, however, also applied to ciphersuite strings such as
10599 "RC4-MD5" that intentionally matched multiple ciphersuites --
10600 namely, SSL 2.0 ciphersuites in addition to the more common ones
10601 from SSL 3.0/TLS 1.0.
10603 So we change the selection algorithm again: Naming an explicit
10604 ciphersuite selects this one ciphersuite, and any other similar
10605 ciphersuite (same bitmap) from *other* protocol versions.
10606 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10607 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10609 Since SSL 2.0 does not have any ciphersuites for which the
10610 128/256 bit distinction would be relevant, this works for now.
10611 The proper fix will be to use different bits for AES128 and
10612 AES256, which would have avoided the problems from the beginning;
10613 however, bits are scarce, so we can only do this in a new release
10614 (not just a patchlevel) when we can change the SSL_CIPHER
10615 definition to split the single 'unsigned long mask' bitmap into
10616 multiple values to extend the available space.
10620 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10622 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10623 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10625 * Add AES IGE and biIGE modes.
10629 * Change the Unix randomness entropy gathering to use poll() when
10630 possible instead of select(), since the latter has some
10631 undesirable limitations.
10633 *Darryl Miles via Richard Levitte and Bodo Moeller*
10635 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10636 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10637 cannot be implicitly activated as part of, e.g., the "AES" alias.
10638 However, please upgrade to OpenSSL 0.9.9[-dev] for
10639 non-experimental use of the ECC ciphersuites to get TLS extension
10640 support, which is required for curve and point format negotiation
10641 to avoid potential handshake problems.
10645 * Disable rogue ciphersuites:
10647 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10648 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10649 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10651 The latter two were purportedly from
10652 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10655 Also deactivate the remaining ciphersuites from
10656 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10657 unofficial, and the ID has long expired.
10661 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10662 dual-core machines) and other potential thread-safety issues.
10666 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10667 versions), which is now available for royalty-free use
10668 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10669 Also, add Camellia TLS ciphersuites from RFC 4132.
10671 To minimize changes between patchlevels in the OpenSSL 0.9.8
10672 series, Camellia remains excluded from compilation unless OpenSSL
10673 is configured with 'enable-camellia'.
10677 * Disable the padding bug check when compression is in use. The padding
10678 bug check assumes the first packet is of even length, this is not
10679 necessarily true if compression is enabled and can result in false
10680 positives causing handshake failure. The actual bug test is ancient
10681 code so it is hoped that implementations will either have fixed it by
10682 now or any which still have the bug do not support compression.
10686 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10688 * When applying a cipher rule check to see if string match is an explicit
10689 cipher suite and only match that one cipher suite if it is.
10693 * Link in manifests for VC++ if needed.
10695 *Austin Ziegler <halostatue@gmail.com>*
10697 * Update support for ECC-based TLS ciphersuites according to
10698 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10699 TLS extensions, which are supported starting with the 0.9.9
10700 branch, not in the OpenSSL 0.9.8 branch).
10704 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10705 opaque EVP_CIPHER_CTX handling.
10709 * Fixes and enhancements to zlib compression code. We now only use
10710 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10711 to conform with the standards mentioned here:
10712 <http://www.zlib.net/DLL_FAQ.txt>
10713 Static zlib linking now works on Windows and the new --with-zlib-include
10714 --with-zlib-lib options to Configure can be used to supply the location
10715 of the headers and library. Gracefully handle case where zlib library
10720 * Several fixes and enhancements to the OID generation code. The old code
10721 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10722 handle numbers larger than ULONG_MAX, truncated printing and had a
10723 non standard OBJ_obj2txt() behaviour.
10727 * Add support for building of engines under engine/ as shared libraries
10728 under VC++ build system.
10732 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10733 Hopefully, we will not see any false combination of paths any more.
10737 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10739 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10740 (part of SSL_OP_ALL). This option used to disable the
10741 countermeasure against man-in-the-middle protocol-version
10742 rollback in the SSL 2.0 server implementation, which is a bad
10743 idea. ([CVE-2005-2969])
10745 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10746 for Information Security, National Institute of Advanced Industrial
10747 Science and Technology [AIST], Japan)*
10749 * Add two function to clear and return the verify parameter flags.
10753 * Keep cipherlists sorted in the source instead of sorting them at
10754 runtime, thus removing the need for a lock.
10758 * Avoid some small subgroup attacks in Diffie-Hellman.
10760 *Nick Mathewson and Ben Laurie*
10762 * Add functions for well-known primes.
10766 * Extended Windows CE support.
10768 *Satoshi Nakamura and Andy Polyakov*
10770 * Initialize SSL_METHOD structures at compile time instead of during
10771 runtime, thus removing the need for a lock.
10775 * Make PKCS7_decrypt() work even if no certificate is supplied by
10776 attempting to decrypt each encrypted key in turn. Add support to
10781 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10783 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10786 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10790 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10791 key into the same file any more.
10795 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10799 * Add -utf8 command line and config file option to 'ca'.
10801 *Stefan <stf@udoma.org*
10803 * Removed the macro des_crypt(), as it seems to conflict with some
10804 libraries. Use DES_crypt().
10808 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10809 involves renaming the source and generated shared-libs for
10810 both. The engines will accept the corrected or legacy ids
10811 ('ncipher' and '4758_cca' respectively) when binding. NB,
10812 this only applies when building 'shared'.
10814 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10816 * Add attribute functions to EVP_PKEY structure. Modify
10817 PKCS12_create() to recognize a CSP name attribute and
10818 use it. Make -CSP option work again in pkcs12 utility.
10822 * Add new functionality to the bn blinding code:
10823 - automatic re-creation of the BN_BLINDING parameters after
10824 a fixed number of uses (currently 32)
10825 - add new function for parameter creation
10826 - introduce flags to control the update behaviour of the
10827 BN_BLINDING parameters
10828 - hide BN_BLINDING structure
10829 Add a second BN_BLINDING slot to the RSA structure to improve
10830 performance when a single RSA object is shared among several
10835 * Add support for DTLS.
10837 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10839 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10840 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10844 * Remove buggy and incomplete DH cert support from
10845 ssl/ssl_rsa.c and ssl/s3_both.c
10849 * Use SHA-1 instead of MD5 as the default digest algorithm for
10850 the `apps/openssl` commands.
10854 * Compile clean with "-Wall -Wmissing-prototypes
10855 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10856 DEBUG_SAFESTACK must also be set.
10860 * Change ./Configure so that certain algorithms can be disabled by default.
10861 The new counterpiece to "no-xxx" is "enable-xxx".
10863 The patented RC5 and MDC2 algorithms will now be disabled unless
10864 "enable-rc5" and "enable-mdc2", respectively, are specified.
10866 (IDEA remains enabled despite being patented. This is because IDEA
10867 is frequently required for interoperability, and there is no license
10868 fee for non-commercial use. As before, "no-idea" can be used to
10869 avoid this algorithm.)
10873 * Add processing of proxy certificates (see RFC 3820). This work was
10874 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10875 EGEE (Enabling Grids for E-science in Europe).
10879 * RC4 performance overhaul on modern architectures/implementations, such
10880 as Intel P4, IA-64 and AMD64.
10884 * New utility extract-section.pl. This can be used specify an alternative
10885 section number in a pod file instead of having to treat each file as
10886 a separate case in Makefile. This can be done by adding two lines to the
10889 =for comment openssl_section:XXX
10891 The blank line is mandatory.
10895 * New arguments -certform, -keyform and -pass for s_client and s_server
10896 to allow alternative format key and certificate files and passphrase
10901 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10902 update associated structures and add various utility functions.
10904 Add new policy related verify parameters, include policy checking in
10905 standard verify code. Enhance 'smime' application with extra parameters
10906 to support policy checking and print out.
10910 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10911 Nehemiah processors. These extensions support AES encryption in hardware
10912 as well as RNG (though RNG support is currently disabled).
10914 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10916 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10920 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10922 *Andy Polyakov and a number of other people*
10924 * Improved PowerPC platform support. Most notably BIGNUM assembler
10925 implementation contributed by IBM.
10927 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10929 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10930 exponent rather than 'unsigned long'. There is a corresponding change to
10931 the new 'rsa_keygen' element of the RSA_METHOD structure.
10933 *Jelte Jansen, Geoff Thorpe*
10935 * Functionality for creating the initial serial number file is now
10936 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10938 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10939 number file to 1, which is bound to cause problems. To avoid
10940 the problems while respecting compatibility between different 0.9.7
10941 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10942 CA.pl for serial number initialization. With the new release 0.9.8,
10943 we can fix the problem directly in the 'ca' utility.)
10947 * Reduced header interdependencies by declaring more opaque objects in
10948 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10949 give fewer recursive includes, which could break lazy source code - so
10950 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10951 developers should define this symbol when building and using openssl to
10952 ensure they track the recommended behaviour, interfaces, [etc], but
10953 backwards-compatible behaviour prevails when this isn't defined.
10957 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10961 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10962 This will generate a random key of the appropriate length based on the
10963 cipher context. The EVP_CIPHER can provide its own random key generation
10964 routine to support keys of a specific form. This is used in the des and
10965 3des routines to generate a key of the correct parity. Update S/MIME
10966 code to use new functions and hence generate correct parity DES keys.
10967 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10968 valid (weak or incorrect parity).
10972 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10973 as looking them up. This is useful when the verified structure may contain
10974 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10975 present unless the new PKCS7_NO_CRL flag is asserted.
10979 * Extend ASN1 oid configuration module. It now additionally accepts the
10982 shortName = some long name, 1.2.3.4
10986 * Reimplemented the BN_CTX implementation. There is now no more static
10987 limitation on the number of variables it can handle nor the depth of the
10988 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10989 information can now expand as required, and rather than having a single
10990 static array of bignums, BN_CTX now uses a linked-list of such arrays
10991 allowing it to expand on demand whilst maintaining the usefulness of
10992 BN_CTX's "bundling".
10996 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10997 to allow all RSA operations to function using a single BN_CTX.
11001 * Preliminary support for certificate policy evaluation and checking. This
11002 is initially intended to pass the tests outlined in "Conformance Testing
11003 of Relying Party Client Certificate Path Processing Logic" v1.07.
11007 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11008 remained unused and not that useful. A variety of other little bignum
11009 tweaks and fixes have also been made continuing on from the audit (see
11014 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11015 associated ASN1, EVP and SSL functions and old ASN1 macros.
11019 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11020 and this should never fail. So the return value from the use of
11021 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11022 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11026 * BN_CTX_get() should return zero-valued bignums, providing the same
11027 initialised value as BN_new().
11029 *Geoff Thorpe, suggested by Ulf Möller*
11031 * Support for inhibitAnyPolicy certificate extension.
11035 * An audit of the BIGNUM code is underway, for which debugging code is
11036 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11037 is considered valid when processing BIGNUMs, and causes execution to
11038 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11039 further steps are taken to deliberately pollute unused data in BIGNUM
11040 structures to try and expose faulty code further on. For now, openssl will
11041 (in its default mode of operation) continue to tolerate the inconsistent
11042 forms that it has tolerated in the past, but authors and packagers should
11043 consider trying openssl and their own applications when compiled with
11044 these debugging symbols defined. It will help highlight potential bugs in
11045 their own code, and will improve the test coverage for OpenSSL itself. At
11046 some point, these tighter rules will become openssl's default to improve
11047 maintainability, though the assert()s and other overheads will remain only
11048 in debugging configurations. See bn.h for more details.
11050 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11052 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11053 that can only be obtained through BN_CTX_new() (which implicitly
11054 initialises it). The presence of this function only made it possible
11055 to overwrite an existing structure (and cause memory leaks).
11059 * Because of the callback-based approach for implementing LHASH as a
11060 template type, lh_insert() adds opaque objects to hash-tables and
11061 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11062 to clean up those corresponding objects before destroying the hash table
11063 (and losing the object pointers). So some over-zealous constifications in
11064 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11065 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11066 prototyped to have "const" restrictions on the object pointers they are
11067 given (and so aren't required to cast them away any more).
11071 * The tmdiff.h API was so ugly and minimal that our own timing utility
11072 (speed) prefers to use its own implementation. The two implementations
11073 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11074 its object type properly exposed (MS_TM) instead of casting to/from
11075 `char *`. This may still change yet if someone realises MS_TM and
11077 aren't necessarily the greatest nomenclatures - but this is what was used
11078 internally to the implementation so I've used that for now.
11082 * Ensure that deprecated functions do not get compiled when
11083 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11084 the self-tests were still using deprecated key-generation functions so
11085 these have been updated also.
11089 * Reorganise PKCS#7 code to separate the digest location functionality
11090 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11091 New function PKCS7_set_digest() to set the digest type for PKCS#7
11092 digestedData type. Add additional code to correctly generate the
11093 digestedData type and add support for this type in PKCS7 initialization
11098 * New function PKCS7_set0_type_other() this initializes a PKCS7
11099 structure of type "other".
11103 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11104 sure the loop does correctly stop and breaking ("division by zero")
11105 modulus operations are not performed. The (pre-generated) prime
11106 table crypto/bn/bn_prime.h was already correct, but it could not be
11107 re-generated on some platforms because of the "division by zero"
11108 situation in the script.
11110 *Ralf S. Engelschall*
11112 * Update support for ECC-based TLS ciphersuites according to
11113 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11114 SHA-1 now is only used for "small" curves (where the
11115 representation of a field element takes up to 24 bytes); for
11116 larger curves, the field element resulting from ECDH is directly
11117 used as premaster secret.
11119 *Douglas Stebila (Sun Microsystems Laboratories)*
11121 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11122 curve secp160r1 to the tests.
11124 *Douglas Stebila (Sun Microsystems Laboratories)*
11126 * Add the possibility to load symbols globally with DSO.
11128 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11130 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11131 control of the error stack.
11135 * Add support for STORE in ENGINE.
11139 * Add the STORE type. The intention is to provide a common interface
11140 to certificate and key stores, be they simple file-based stores, or
11141 HSM-type store, or LDAP stores, or...
11142 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11146 * Add a generic structure called OPENSSL_ITEM. This can be used to
11147 pass a list of arguments to any function as well as provide a way
11148 for a function to pass data back to the caller.
11152 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11153 works like BUF_strdup() but can be used to duplicate a portion of
11154 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11159 * Add the function sk_find_ex() which works like sk_find(), but will
11160 return an index to an element even if an exact match couldn't be
11161 found. The index is guaranteed to point at the element where the
11162 searched-for key would be inserted to preserve sorting order.
11166 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11167 takes an extra flags argument for optional functionality. Currently,
11168 the following flags are defined:
11170 OBJ_BSEARCH_VALUE_ON_NOMATCH
11171 This one gets OBJ_bsearch_ex() to return a pointer to the first
11172 element where the comparing function returns a negative or zero
11175 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11176 This one gets OBJ_bsearch_ex() to return a pointer to the first
11177 element where the comparing function returns zero. This is useful
11178 if there are more than one element where the comparing function
11183 * Make it possible to create self-signed certificates with 'openssl ca'
11184 in such a way that the self-signed certificate becomes part of the
11185 CA database and uses the same mechanisms for serial number generation
11186 as all other certificate signing. The new flag '-selfsign' enables
11187 this functionality. Adapt CA.sh and CA.pl.in.
11191 * Add functionality to check the public key of a certificate request
11192 against a given private. This is useful to check that a certificate
11193 request can be signed by that key (self-signing).
11197 * Make it possible to have multiple active certificates with the same
11198 subject in the CA index file. This is done only if the keyword
11199 'unique_subject' is set to 'no' in the main CA section (default
11200 if 'CA_default') of the configuration file. The value is saved
11201 with the database itself in a separate index attribute file,
11202 named like the index file with '.attr' appended to the name.
11206 * Generate multi-valued AVAs using '+' notation in config files for
11211 * Support for nameConstraints certificate extension.
11215 * Support for policyConstraints certificate extension.
11219 * Support for policyMappings certificate extension.
11223 * Make sure the default DSA_METHOD implementation only uses its
11224 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11225 and change its own handlers to be NULL so as to remove unnecessary
11226 indirection. This lets alternative implementations fallback to the
11227 default implementation more easily.
11231 * Support for directoryName in GeneralName related extensions
11236 * Make it possible to link applications using Makefile.shared.
11237 Make that possible even when linking against static libraries!
11241 * Support for single pass processing for S/MIME signing. This now
11242 means that S/MIME signing can be done from a pipe, in addition
11243 cleartext signing (multipart/signed type) is effectively streaming
11244 and the signed data does not need to be all held in memory.
11246 This is done with a new flag PKCS7_STREAM. When this flag is set
11247 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11248 is done after the data is output (and digests calculated) in
11249 SMIME_write_PKCS7().
11253 * Add full support for -rpath/-R, both in shared libraries and
11254 applications, at least on the platforms where it's known how
11259 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11260 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11261 will now compute a table of multiples of the generator that
11262 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11263 faster (notably in the case of a single point multiplication,
11264 scalar * generator).
11266 *Nils Larsch, Bodo Moeller*
11268 * IPv6 support for certificate extensions. The various extensions
11269 which use the IP:a.b.c.d can now take IPv6 addresses using the
11270 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11275 * Added an ENGINE that implements RSA by performing private key
11276 exponentiations with the GMP library. The conversions to and from
11277 GMP's mpz_t format aren't optimised nor are any montgomery forms
11278 cached, and on x86 it appears OpenSSL's own performance has caught up.
11279 However there are likely to be other architectures where GMP could
11280 provide a boost. This ENGINE is not built in by default, but it can be
11281 specified at Configure time and should be accompanied by the necessary
11282 linker additions, eg;
11283 ./config -DOPENSSL_USE_GMP -lgmp
11287 * "openssl engine" will not display ENGINE/DSO load failure errors when
11288 testing availability of engines with "-t" - the old behaviour is
11289 produced by increasing the feature's verbosity with "-tt".
11293 * ECDSA routines: under certain error conditions uninitialized BN objects
11294 could be freed. Solution: make sure initialization is performed early
11295 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11300 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11301 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11302 software implementations. For DSA and DH, parameter generation can
11303 also be overridden by providing the appropriate method callbacks.
11307 * Change the "progress" mechanism used in key-generation and
11308 primality testing to functions that take a new BN_GENCB pointer in
11309 place of callback/argument pairs. The new API functions have `_ex`
11310 postfixes and the older functions are reimplemented as wrappers for
11311 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11312 declarations of the old functions to help (graceful) attempts to
11313 migrate to the new functions. Also, the new key-generation API
11314 functions operate on a caller-supplied key-structure and return
11315 success/failure rather than returning a key or NULL - this is to
11316 help make "keygen" another member function of RSA_METHOD etc.
11318 Example for using the new callback interface:
11320 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11321 void *my_arg = ...;
11324 BN_GENCB_set(&my_cb, my_callback, my_arg);
11326 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11327 /* For the meaning of a, b in calls to my_callback(), see the
11328 * documentation of the function that calls the callback.
11329 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11330 * my_callback should return 1 if it wants BN_is_prime_ex()
11331 * to continue, or 0 to stop.
11336 * Change the ZLIB compression method to be stateful, and make it
11337 available to TLS with the number defined in
11338 draft-ietf-tls-compression-04.txt.
11342 * Add the ASN.1 structures and functions for CertificatePair, which
11343 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11345 CertificatePair ::= SEQUENCE {
11346 forward [0] Certificate OPTIONAL,
11347 reverse [1] Certificate OPTIONAL,
11348 -- at least one of the pair shall be present -- }
11350 Also implement the PEM functions to read and write certificate
11351 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11353 This needed to be defined, mostly for the sake of the LDAP
11354 attribute crossCertificatePair, but may prove useful elsewhere as
11359 * Make it possible to inhibit symlinking of shared libraries in
11360 Makefile.shared, for Cygwin's sake.
11364 * Extend the BIGNUM API by creating a function
11365 void BN_set_negative(BIGNUM *a, int neg);
11366 and a macro that behave like
11367 int BN_is_negative(const BIGNUM *a);
11369 to avoid the need to access 'a->neg' directly in applications.
11373 * Implement fast modular reduction for pseudo-Mersenne primes
11374 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11375 EC_GROUP_new_curve_GFp() will now automatically use this
11378 *Nils Larsch <nla@trustcenter.de>*
11380 * Add new lock type (CRYPTO_LOCK_BN).
11384 * Change the ENGINE framework to automatically load engines
11385 dynamically from specific directories unless they could be
11386 found to already be built in or loaded. Move all the
11387 current engines except for the cryptodev one to a new
11388 directory engines/.
11389 The engines in engines/ are built as shared libraries if
11390 the "shared" options was given to ./Configure or ./config.
11391 Otherwise, they are inserted in libcrypto.a.
11392 /usr/local/ssl/engines is the default directory for dynamic
11393 engines, but that can be overridden at configure time through
11394 the usual use of --prefix and/or --openssldir, and at run
11395 time with the environment variable OPENSSL_ENGINES.
11397 *Geoff Thorpe and Richard Levitte*
11399 * Add Makefile.shared, a helper makefile to build shared
11400 libraries. Adapt Makefile.org.
11404 * Add version info to Win32 DLLs.
11406 *Peter 'Luna' Runestig" <peter@runestig.com>*
11408 * Add new 'medium level' PKCS#12 API. Certificates and keys
11409 can be added using this API to created arbitrary PKCS#12
11410 files while avoiding the low-level API.
11412 New options to PKCS12_create(), key or cert can be NULL and
11413 will then be omitted from the output file. The encryption
11414 algorithm NIDs can be set to -1 for no encryption, the mac
11415 iteration count can be set to 0 to omit the mac.
11417 Enhance pkcs12 utility by making the -nokeys and -nocerts
11418 options work when creating a PKCS#12 file. New option -nomac
11419 to omit the mac, NONE can be set for an encryption algorithm.
11420 New code is modified to use the enhanced PKCS12_create()
11421 instead of the low-level API.
11425 * Extend ASN1 encoder to support indefinite length constructed
11426 encoding. This can output sequences tags and octet strings in
11427 this form. Modify pk7_asn1.c to support indefinite length
11428 encoding. This is experimental and needs additional code to
11429 be useful, such as an ASN1 bio and some enhanced streaming
11432 Extend template encode functionality so that tagging is passed
11433 down to the template encoder.
11437 * Let 'openssl req' fail if an argument to '-newkey' is not
11438 recognized instead of using RSA as a default.
11442 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11443 As these are not official, they are not included in "ALL";
11444 the "ECCdraft" ciphersuite group alias can be used to select them.
11446 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11448 * Add ECDH engine support.
11450 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11452 * Add ECDH in new directory crypto/ecdh/.
11454 *Douglas Stebila (Sun Microsystems Laboratories)*
11456 * Let BN_rand_range() abort with an error after 100 iterations
11457 without success (which indicates a broken PRNG).
11461 * Change BN_mod_sqrt() so that it verifies that the input value
11462 is really the square of the return value. (Previously,
11463 BN_mod_sqrt would show GIGO behaviour.)
11467 * Add named elliptic curves over binary fields from X9.62, SECG,
11468 and WAP/WTLS; add OIDs that were still missing.
11470 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11472 * Extend the EC library for elliptic curves over binary fields
11473 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11476 EC_GF2m_simple_method
11480 EC_GROUP_new_curve_GF2m
11481 EC_GROUP_set_curve_GF2m
11482 EC_GROUP_get_curve_GF2m
11483 EC_POINT_set_affine_coordinates_GF2m
11484 EC_POINT_get_affine_coordinates_GF2m
11485 EC_POINT_set_compressed_coordinates_GF2m
11487 Point compression for binary fields is disabled by default for
11488 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11491 As binary polynomials are represented as BIGNUMs, various members
11492 of the EC_GROUP and EC_POINT data structures can be shared
11493 between the implementations for prime fields and binary fields;
11494 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11495 are essentially identical to their `..._GFp` counterparts.
11496 (For simplicity, the `..._GFp` prefix has been dropped from
11497 various internal method names.)
11499 An internal 'field_div' method (similar to 'field_mul' and
11500 'field_sqr') has been added; this is used only for binary fields.
11502 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11504 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11505 through methods ('mul', 'precompute_mult').
11507 The generic implementations (now internally called 'ec_wNAF_mul'
11508 and 'ec_wNAF_precomputed_mult') remain the default if these
11509 methods are undefined.
11511 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11513 * New function EC_GROUP_get_degree, which is defined through
11514 EC_METHOD. For curves over prime fields, this returns the bit
11515 length of the modulus.
11517 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11519 * New functions EC_GROUP_dup, EC_POINT_dup.
11520 (These simply call ..._new and ..._copy).
11522 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11524 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11525 Polynomials are represented as BIGNUMs (where the sign bit is not
11526 used) in the following functions [macros]:
11529 BN_GF2m_sub [= BN_GF2m_add]
11530 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11531 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11532 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11534 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11535 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11536 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11537 BN_GF2m_cmp [= BN_ucmp]
11539 (Note that only the 'mod' functions are actually for fields GF(2^m).
11540 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11542 For some functions, an the irreducible polynomial defining a
11543 field can be given as an 'unsigned int[]' with strictly
11544 decreasing elements giving the indices of those bits that are set;
11545 i.e., p[] represents the polynomial
11546 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11548 p[0] > p[1] > ... > p[k] = 0.
11549 This applies to the following functions:
11552 BN_GF2m_mod_mul_arr
11553 BN_GF2m_mod_sqr_arr
11554 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11555 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11556 BN_GF2m_mod_exp_arr
11557 BN_GF2m_mod_sqrt_arr
11558 BN_GF2m_mod_solve_quad_arr
11562 Conversion can be performed by the following functions:
11567 bntest.c has additional tests for binary polynomial arithmetic.
11569 Two implementations for BN_GF2m_mod_div() are available.
11570 The default algorithm simply uses BN_GF2m_mod_inv() and
11571 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11572 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11573 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11575 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11577 * Add new error code 'ERR_R_DISABLED' that can be used when some
11578 functionality is disabled at compile-time.
11580 *Douglas Stebila <douglas.stebila@sun.com>*
11582 * Change default behaviour of 'openssl asn1parse' so that more
11583 information is visible when viewing, e.g., a certificate:
11585 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11586 mode the content of non-printable OCTET STRINGs is output in a
11587 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11588 avoid the appearance of a printable string.
11590 *Nils Larsch <nla@trustcenter.de>*
11592 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11594 EC_GROUP_set_asn1_flag()
11595 EC_GROUP_get_asn1_flag()
11596 EC_GROUP_set_point_conversion_form()
11597 EC_GROUP_get_point_conversion_form()
11598 These control ASN1 encoding details:
11599 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11600 has been set to OPENSSL_EC_NAMED_CURVE.
11601 - Points are encoded in uncompressed form by default; options for
11602 asn1_for are as for point2oct, namely
11603 POINT_CONVERSION_COMPRESSED
11604 POINT_CONVERSION_UNCOMPRESSED
11605 POINT_CONVERSION_HYBRID
11607 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11609 EC_GROUP_set_seed()
11610 EC_GROUP_get0_seed()
11611 EC_GROUP_get_seed_len()
11612 This is used only for ASN1 purposes (so far).
11614 *Nils Larsch <nla@trustcenter.de>*
11616 * Add 'field_type' member to EC_METHOD, which holds the NID
11617 of the appropriate field type OID. The new function
11618 EC_METHOD_get_field_type() returns this value.
11620 *Nils Larsch <nla@trustcenter.de>*
11623 EC_POINT_point2bn()
11624 EC_POINT_bn2point()
11625 EC_POINT_point2hex()
11626 EC_POINT_hex2point()
11627 providing useful interfaces to EC_POINT_point2oct() and
11628 EC_POINT_oct2point().
11630 *Nils Larsch <nla@trustcenter.de>*
11632 * Change internals of the EC library so that the functions
11633 EC_GROUP_set_generator()
11634 EC_GROUP_get_generator()
11635 EC_GROUP_get_order()
11636 EC_GROUP_get_cofactor()
11637 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11638 to methods, which would lead to unnecessary code duplication when
11639 adding different types of curves.
11641 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11643 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11644 arithmetic, and such that modified wNAFs are generated
11645 (which avoid length expansion in many cases).
11649 * Add a function EC_GROUP_check_discriminant() (defined via
11650 EC_METHOD) that verifies that the curve discriminant is non-zero.
11652 Add a function EC_GROUP_check() that makes some sanity tests
11653 on a EC_GROUP, its generator and order. This includes
11654 EC_GROUP_check_discriminant().
11656 *Nils Larsch <nla@trustcenter.de>*
11658 * Add ECDSA in new directory crypto/ecdsa/.
11660 Add applications 'openssl ecparam' and 'openssl ecdsa'
11661 (these are based on 'openssl dsaparam' and 'openssl dsa').
11663 ECDSA support is also included in various other files across the
11664 library. Most notably,
11665 - 'openssl req' now has a '-newkey ecdsa:file' option;
11666 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11667 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11668 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11669 them suitable for ECDSA where domain parameters must be
11670 extracted before the specific public key;
11671 - ECDSA engine support has been added.
11673 *Nils Larsch <nla@trustcenter.de>*
11675 * Include some named elliptic curves, and add OIDs from X9.62,
11676 SECG, and WAP/WTLS. Each curve can be obtained from the new
11678 EC_GROUP_new_by_curve_name(),
11679 and the list of available named curves can be obtained with
11680 EC_get_builtin_curves().
11681 Also add a 'curve_name' member to EC_GROUP objects, which can be
11683 EC_GROUP_set_curve_name()
11684 EC_GROUP_get_curve_name()
11686 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11688 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11689 was actually never needed) and in BN_mul(). The removal in BN_mul()
11690 required a small change in bn_mul_part_recursive() and the addition
11691 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11692 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11693 bn_sub_words() and bn_add_words() except they take arrays with
11698 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11700 * Cleanse PEM buffers before freeing them since they may contain
11703 *Benjamin Bennett <ben@psc.edu>*
11705 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11706 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11707 authentication-only ciphersuites.
11711 * Since AES128 and AES256 share a single mask bit in the logic of
11712 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11713 kludge to work properly if AES128 is available and AES256 isn't.
11717 * Expand security boundary to match 1.1.1 module.
11721 * Remove redundant features: hash file source, editing of test vectors
11722 modify fipsld to use external fips_premain.c signature.
11726 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11727 run algorithm test programs.
11731 * Make algorithm test programs more tolerant of whitespace.
11735 * Have SSL/TLS server implementation tolerate "mismatched" record
11736 protocol version while receiving ClientHello even if the
11737 ClientHello is fragmented. (The server can't insist on the
11738 particular protocol version it has chosen before the ServerHello
11739 message has informed the client about his choice.)
11743 * Load error codes if they are not already present instead of using a
11744 static variable. This allows them to be cleanly unloaded and reloaded.
11748 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11750 * Introduce limits to prevent malicious keys being able to
11751 cause a denial of service. ([CVE-2006-2940])
11753 *Steve Henson, Bodo Moeller*
11755 * Fix ASN.1 parsing of certain invalid structures that can result
11756 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11758 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11759 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11761 * Fix SSL client code which could crash if connecting to a
11762 malicious SSLv2 server. ([CVE-2006-4343])
11764 *Tavis Ormandy and Will Drewry, Google Security Team*
11766 * Change ciphersuite string processing so that an explicit
11767 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11768 will no longer include "AES128-SHA"), and any other similar
11769 ciphersuite (same bitmap) from *other* protocol versions (so that
11770 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11771 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11772 changes from 0.9.8b and 0.9.8d.
11776 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11778 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11779 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11781 * Change the Unix randomness entropy gathering to use poll() when
11782 possible instead of select(), since the latter has some
11783 undesirable limitations.
11785 *Darryl Miles via Richard Levitte and Bodo Moeller*
11787 * Disable rogue ciphersuites:
11789 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11790 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11791 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11793 The latter two were purportedly from
11794 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11797 Also deactivate the remaining ciphersuites from
11798 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11799 unofficial, and the ID has long expired.
11803 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11804 dual-core machines) and other potential thread-safety issues.
11808 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11810 * Adapt fipsld and the build system to link against the validated FIPS
11811 module in FIPS mode.
11815 * Fixes for VC++ 2005 build under Windows.
11819 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11820 from a Windows bash shell such as MSYS. It is autodetected from the
11821 "config" script when run from a VC++ environment. Modify standard VC++
11822 build to use fipscanister.o from the GNU make build.
11826 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11828 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11829 The value now differs depending on if you build for FIPS or not.
11830 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11831 safely run with a non-FIPSed libcrypto, as it may crash because of
11832 the difference induced by this change.
11836 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11838 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11839 (part of SSL_OP_ALL). This option used to disable the
11840 countermeasure against man-in-the-middle protocol-version
11841 rollback in the SSL 2.0 server implementation, which is a bad
11842 idea. ([CVE-2005-2969])
11844 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11845 for Information Security, National Institute of Advanced Industrial
11846 Science and Technology [AIST, Japan)]*
11848 * Minimal support for X9.31 signatures and PSS padding modes. This is
11849 mainly for FIPS compliance and not fully integrated at this stage.
11853 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11854 the exponentiation using a fixed-length exponent. (Otherwise,
11855 the information leaked through timing could expose the secret key
11856 after many signatures; cf. Bleichenbacher's attack on DSA with
11861 * Make a new fixed-window mod_exp implementation the default for
11862 RSA, DSA, and DH private-key operations so that the sequence of
11863 squares and multiplies and the memory access pattern are
11864 independent of the particular secret key. This will mitigate
11865 cache-timing and potential related attacks.
11867 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11868 and this is automatically used by BN_mod_exp_mont() if the new flag
11869 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11870 will use this BN flag for private exponents unless the flag
11871 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11872 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11874 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11876 * Change the client implementation for SSLv23_method() and
11877 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11878 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11879 (Previously, the SSL 2.0 backwards compatible Client Hello
11880 message format would be used even with SSL_OP_NO_SSLv2.)
11884 * Add support for smime-type MIME parameter in S/MIME messages which some
11889 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11890 a threadsafe manner. Modify rsa code to use new function and add calls
11891 to dsa and dh code (which had race conditions before).
11895 * Include the fixed error library code in the C error file definitions
11896 instead of fixing them up at runtime. This keeps the error code
11897 structures constant.
11901 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11903 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11906 * Fixes for newer kerberos headers. NB: the casts are needed because
11907 the 'length' field is signed on one version and unsigned on another
11908 with no (?) obvious way to tell the difference, without these VC++
11909 complains. Also the "definition" of FAR (blank) is no longer included
11910 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11911 some needed definitions.
11915 * Undo Cygwin change.
11919 * Added support for proxy certificates according to RFC 3820.
11920 Because they may be a security thread to unaware applications,
11921 they must be explicitly allowed in run-time. See
11922 docs/HOWTO/proxy_certificates.txt for further information.
11926 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11928 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11929 server and client random values. Previously
11930 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11931 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11933 This change has negligible security impact because:
11935 1. Server and client random values still have 24 bytes of pseudo random
11938 2. Server and client random values are sent in the clear in the initial
11941 3. The master secret is derived using the premaster secret (48 bytes in
11942 size for static RSA ciphersuites) as well as client server and random
11945 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11948 *Stephen Henson, reported by UK NISCC*
11950 * Use Windows randomness collection on Cygwin.
11954 * Fix hang in EGD/PRNGD query when communication socket is closed
11955 prematurely by EGD/PRNGD.
11957 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11959 * Prompt for pass phrases when appropriate for PKCS12 input format.
11963 * Back-port of selected performance improvements from development
11964 branch, as well as improved support for PowerPC platforms.
11968 * Add lots of checks for memory allocation failure, error codes to indicate
11969 failure and freeing up memory if a failure occurs.
11971 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11973 * Add new -passin argument to dgst.
11977 * Perform some character comparisons of different types in X509_NAME_cmp:
11978 this is needed for some certificates that re-encode DNs into UTF8Strings
11979 (in violation of RFC3280) and can't or won't issue name rollover
11984 * Make an explicit check during certificate validation to see that
11985 the CA setting in each certificate on the chain is correct. As a
11986 side effect always do the following basic checks on extensions,
11987 not just when there's an associated purpose to the check:
11989 - if there is an unhandled critical extension (unless the user
11990 has chosen to ignore this fault)
11991 - if the path length has been exceeded (if one is set at all)
11992 - that certain extensions fit the associated purpose (if one has
11997 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11999 * Avoid a race condition when CRLs are checked in a multi threaded
12000 environment. This would happen due to the reordering of the revoked
12001 entries during signature checking and serial number lookup. Now the
12002 encoding is cached and the serial number sort performed under a lock.
12003 Add new STACK function sk_is_sorted().
12007 * Add Delta CRL to the extension code.
12011 * Various fixes to s3_pkt.c so alerts are sent properly.
12013 *David Holmes <d.holmes@f5.com>*
12015 * Reduce the chances of duplicate issuer name and serial numbers (in
12016 violation of RFC3280) using the OpenSSL certificate creation utilities.
12017 This is done by creating a random 64 bit value for the initial serial
12018 number when a serial number file is created or when a self signed
12019 certificate is created using 'openssl req -x509'. The initial serial
12020 number file is created using 'openssl x509 -next_serial' in CA.pl
12021 rather than being initialized to 1.
12025 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12027 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12028 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12030 *Joe Orton, Steve Henson*
12032 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12035 *Joe Orton, Steve Henson*
12037 * Make it possible to have multiple active certificates with the same
12038 subject in the CA index file. This is done only if the keyword
12039 'unique_subject' is set to 'no' in the main CA section (default
12040 if 'CA_default') of the configuration file. The value is saved
12041 with the database itself in a separate index attribute file,
12042 named like the index file with '.attr' appended to the name.
12046 * X509 verify fixes. Disable broken certificate workarounds when
12047 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12048 keyUsage extension present. Don't accept CRLs with unhandled critical
12049 extensions: since verify currently doesn't process CRL extensions this
12050 rejects a CRL with *any* critical extensions. Add new verify error codes
12055 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12056 A clarification of RFC2560 will require the use of OCTET STRINGs and
12057 some implementations cannot handle the current raw format. Since OpenSSL
12058 copies and compares OCSP nonces as opaque blobs without any attempt at
12059 parsing them this should not create any compatibility issues.
12063 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12064 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12065 this HMAC (and other) operations are several times slower than OpenSSL
12070 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12072 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12074 * Use the correct content when signing type "other".
12078 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12080 * Fix various bugs revealed by running the NISCC test suite:
12082 Stop out of bounds reads in the ASN1 code when presented with
12083 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12085 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12087 If verify callback ignores invalid public key errors don't try to check
12088 certificate signature with the NULL public key.
12092 * New -ignore_err option in ocsp application to stop the server
12093 exiting on the first error in a request.
12097 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12098 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12103 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12104 extra data after the compression methods not only for TLS 1.0
12105 but also for SSL 3.0 (as required by the specification).
12107 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12109 * Change X509_certificate_type() to mark the key as exported/exportable
12110 when it's 512 *bits* long, not 512 bytes.
12114 * Change AES_cbc_encrypt() so it outputs exact multiple of
12115 blocks during encryption.
12119 * Various fixes to base64 BIO and non blocking I/O. On write
12120 flushes were not handled properly if the BIO retried. On read
12121 data was not being buffered properly and had various logic bugs.
12122 This also affects blocking I/O when the data being decoded is a
12127 * Various S/MIME bugfixes and compatibility changes:
12128 output correct application/pkcs7 MIME type if
12129 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12130 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12131 of files as .eml work). Correctly handle very long lines in MIME
12136 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12138 * Countermeasure against the Klima-Pokorny-Rosa extension of
12139 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12140 a protocol version number mismatch like a decryption error
12141 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12145 * Turn on RSA blinding by default in the default implementation
12146 to avoid a timing attack. Applications that don't want it can call
12147 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12148 They would be ill-advised to do so in most cases.
12150 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12152 * Change RSA blinding code so that it works when the PRNG is not
12153 seeded (in this case, the secret RSA exponent is abused as
12154 an unpredictable seed -- if it is not unpredictable, there
12155 is no point in blinding anyway). Make RSA blinding thread-safe
12156 by remembering the creator's thread ID in rsa->blinding and
12157 having all other threads use local one-time blinding factors
12158 (this requires more computation than sharing rsa->blinding, but
12159 avoids excessive locking; and if an RSA object is not shared
12160 between threads, blinding will still be very fast).
12164 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12165 ENGINE as defaults for all supported algorithms irrespective of
12166 the 'flags' parameter. 'flags' is now honoured, so applications
12167 should make sure they are passing it correctly.
12171 * Target "mingw" now allows native Windows code to be generated in
12172 the Cygwin environment as well as with the MinGW compiler.
12176 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12178 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12179 via timing by performing a MAC computation even if incorrect
12180 block cipher padding has been found. This is a countermeasure
12181 against active attacks where the attacker has to distinguish
12182 between bad padding and a MAC verification error. ([CVE-2003-0078])
12184 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12185 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12186 Martin Vuagnoux (EPFL, Ilion)*
12188 * Make the no-err option work as intended. The intention with no-err
12189 is not to have the whole error stack handling routines removed from
12190 libcrypto, it's only intended to remove all the function name and
12191 reason texts, thereby removing some of the footprint that may not
12192 be interesting if those errors aren't displayed anyway.
12194 NOTE: it's still possible for any application or module to have its
12195 own set of error texts inserted. The routines are there, just not
12196 used by default when no-err is given.
12200 * Add support for FreeBSD on IA64.
12202 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12204 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12205 Kerberos function mit_des_cbc_cksum(). Before this change,
12206 the value returned by DES_cbc_cksum() was like the one from
12207 mit_des_cbc_cksum(), except the bytes were swapped.
12209 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12211 * Allow an application to disable the automatic SSL chain building.
12212 Before this a rather primitive chain build was always performed in
12213 ssl3_output_cert_chain(): an application had no way to send the
12214 correct chain if the automatic operation produced an incorrect result.
12216 Now the chain builder is disabled if either:
12218 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12220 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12222 The reasoning behind this is that an application would not want the
12223 auto chain building to take place if extra chain certificates are
12224 present and it might also want a means of sending no additional
12225 certificates (for example the chain has two certificates and the
12230 * Add the possibility to build without the ENGINE framework.
12232 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12234 * Under Win32 gmtime() can return NULL: check return value in
12235 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12239 * DSA routines: under certain error conditions uninitialized BN objects
12240 could be freed. Solution: make sure initialization is performed early
12241 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12242 Nils Larsch <nla@trustcenter.de> via PR#459)
12246 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12247 checked on reconnect on the client side, therefore session resumption
12248 could still fail with a "ssl session id is different" error. This
12249 behaviour is masked when SSL_OP_ALL is used due to
12250 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12251 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12252 followup to PR #377.
12256 * IA-32 assembler support enhancements: unified ELF targets, support
12257 for SCO/Caldera platforms, fix for Cygwin shared build.
12261 * Add support for FreeBSD on sparc64. As a consequence, support for
12262 FreeBSD on non-x86 processors is separate from x86 processors on
12263 the config script, much like the NetBSD support.
12265 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12267 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12269 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12272 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12273 code (06) was taken as the first octet of the session ID and the last
12274 octet was ignored consequently. As a result SSLv2 client side session
12275 caching could not have worked due to the session ID mismatch between
12277 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12282 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12283 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12288 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12289 seems that in spite of existing for more than a year, many application
12290 author have done nothing to provide the necessary callbacks, which
12291 means that this particular engine will not work properly anywhere.
12292 This is a very unfortunate situation which forces us, in the name
12293 of usability, to give the hw_ncipher.c a static lock, which is part
12295 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12296 appear in 0.9.8 or later. We EXPECT application authors to have
12297 dealt properly with this when 0.9.8 is released (unless we actually
12298 make such changes in the libcrypto locking code that changes will
12299 have to be made anyway).
12303 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12304 octets have been read, EOF or an error occurs. Without this change
12305 some truncated ASN1 structures will not produce an error.
12309 * Disable Heimdal support, since it hasn't been fully implemented.
12310 Still give the possibility to force the use of Heimdal, but with
12311 warnings and a request that patches get sent to openssl-dev.
12315 * Add the VC-CE target, introduce the WINCE sysname, and add
12316 INSTALL.WCE and appropriate conditionals to make it build.
12318 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12320 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12321 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12322 edit numbers of the version.
12324 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12326 * Introduce safe string copy and catenation functions
12327 (BUF_strlcpy() and BUF_strlcat()).
12329 *Ben Laurie (CHATS) and Richard Levitte*
12331 * Avoid using fixed-size buffers for one-line DNs.
12333 *Ben Laurie (CHATS)*
12335 * Add BUF_MEM_grow_clean() to avoid information leakage when
12336 resizing buffers containing secrets, and use where appropriate.
12338 *Ben Laurie (CHATS)*
12340 * Avoid using fixed size buffers for configuration file location.
12342 *Ben Laurie (CHATS)*
12344 * Avoid filename truncation for various CA files.
12346 *Ben Laurie (CHATS)*
12348 * Use sizeof in preference to magic numbers.
12350 *Ben Laurie (CHATS)*
12352 * Avoid filename truncation in cert requests.
12354 *Ben Laurie (CHATS)*
12356 * Add assertions to check for (supposedly impossible) buffer
12359 *Ben Laurie (CHATS)*
12361 * Don't cache truncated DNS entries in the local cache (this could
12362 potentially lead to a spoofing attack).
12364 *Ben Laurie (CHATS)*
12366 * Fix various buffers to be large enough for hex/decimal
12367 representations in a platform independent manner.
12369 *Ben Laurie (CHATS)*
12371 * Add CRYPTO_realloc_clean() to avoid information leakage when
12372 resizing buffers containing secrets, and use where appropriate.
12374 *Ben Laurie (CHATS)*
12376 * Add BIO_indent() to avoid much slightly worrying code to do
12379 *Ben Laurie (CHATS)*
12381 * Convert sprintf()/BIO_puts() to BIO_printf().
12383 *Ben Laurie (CHATS)*
12385 * buffer_gets() could terminate with the buffer only half
12388 *Ben Laurie (CHATS)*
12390 * Add assertions to prevent user-supplied crypto functions from
12391 overflowing internal buffers by having large block sizes, etc.
12393 *Ben Laurie (CHATS)*
12395 * New OPENSSL_assert() macro (similar to assert(), but enabled
12398 *Ben Laurie (CHATS)*
12400 * Eliminate unused copy of key in RC4.
12402 *Ben Laurie (CHATS)*
12404 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12406 *Ben Laurie (CHATS)*
12408 * Fix off-by-one error in EGD path.
12410 *Ben Laurie (CHATS)*
12412 * If RANDFILE path is too long, ignore instead of truncating.
12414 *Ben Laurie (CHATS)*
12416 * Eliminate unused and incorrectly sized X.509 structure
12419 *Ben Laurie (CHATS)*
12421 * Eliminate unused and dangerous function knumber().
12423 *Ben Laurie (CHATS)*
12425 * Eliminate unused and dangerous structure, KSSL_ERR.
12427 *Ben Laurie (CHATS)*
12429 * Protect against overlong session ID context length in an encoded
12430 session object. Since these are local, this does not appear to be
12433 *Ben Laurie (CHATS)*
12435 * Change from security patch (see 0.9.6e below) that did not affect
12436 the 0.9.6 release series:
12438 Remote buffer overflow in SSL3 protocol - an attacker could
12439 supply an oversized master key in Kerberos-enabled versions.
12442 *Ben Laurie (CHATS)*
12444 * Change the SSL kerb5 codes to match RFC 2712.
12448 * Make -nameopt work fully for req and add -reqopt switch.
12450 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12452 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12454 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12456 * Make sure tests can be performed even if the corresponding algorithms
12457 have been removed entirely. This was also the last step to make
12458 OpenSSL compilable with DJGPP under all reasonable conditions.
12460 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12462 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12463 to allow version independent disabling of normally unselected ciphers,
12464 which may be activated as a side-effect of selecting a single cipher.
12466 (E.g., cipher list string "RSA" enables ciphersuites that are left
12467 out of "ALL" because they do not provide symmetric encryption.
12468 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12470 *Lutz Jaenicke, Bodo Moeller*
12472 * Add appropriate support for separate platform-dependent build
12473 directories. The recommended way to make a platform-dependent
12474 build directory is the following (tested on Linux), maybe with
12477 # Place yourself outside of the OpenSSL source tree. In
12478 # this example, the environment variable OPENSSL_SOURCE
12479 # is assumed to contain the absolute OpenSSL source directory.
12480 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12481 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12482 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12483 mkdir -p `dirname $F`
12484 ln -s $OPENSSL_SOURCE/$F $F
12487 To be absolutely sure not to disturb the source tree, a "make clean"
12488 is a good thing. If it isn't successful, don't worry about it,
12489 it probably means the source directory is very clean.
12493 * Make sure any ENGINE control commands make local copies of string
12494 pointers passed to them whenever necessary. Otherwise it is possible
12495 the caller may have overwritten (or deallocated) the original string
12496 data when a later ENGINE operation tries to use the stored values.
12498 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12500 * Improve diagnostics in file reading and command-line digests.
12502 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12504 * Add AES modes CFB and OFB to the object database. Correct an
12505 error in AES-CFB decryption.
12509 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12510 allows existing EVP_CIPHER_CTX structures to be reused after
12511 calling `EVP_*Final()`. This behaviour is used by encryption
12512 BIOs and some applications. This has the side effect that
12513 applications must explicitly clean up cipher contexts with
12514 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12518 * Check the values of dna and dnb in bn_mul_recursive before calling
12519 bn_mul_comba (a non zero value means the a or b arrays do not contain
12520 n2 elements) and fallback to bn_mul_normal if either is not zero.
12524 * Fix escaping of non-ASCII characters when using the -subj option
12525 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12529 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12530 form for "surname", serialNumber has no short form.
12531 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12532 therefore remove "mail" short name for "internet 7".
12533 The OID for unique identifiers in X509 certificates is
12534 x500UniqueIdentifier, not uniqueIdentifier.
12535 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12539 * Add an "init" command to the ENGINE config module and auto initialize
12540 ENGINEs. Without any "init" command the ENGINE will be initialized
12541 after all ctrl commands have been executed on it. If init=1 the
12542 ENGINE is initialized at that point (ctrls before that point are run
12543 on the uninitialized ENGINE and after on the initialized one). If
12544 init=0 then the ENGINE will not be initialized at all.
12548 * Fix the 'app_verify_callback' interface so that the user-defined
12549 argument is actually passed to the callback: In the
12550 SSL_CTX_set_cert_verify_callback() prototype, the callback
12551 declaration has been changed from
12554 int (*cb)(X509_STORE_CTX *,void *);
12555 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12556 i=s->ctx->app_verify_callback(&ctx)
12557 has been changed into
12558 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12560 To update applications using SSL_CTX_set_cert_verify_callback(),
12561 a dummy argument can be added to their callback functions.
12563 *D. K. Smetters <smetters@parc.xerox.com>*
12565 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12567 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12569 * Add and OPENSSL_LOAD_CONF define which will cause
12570 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12571 This allows older applications to transparently support certain
12572 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12573 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12574 load the config file and OPENSSL_add_all_algorithms_conf() which will
12575 always load it have also been added.
12579 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12580 Adjust NIDs and EVP layer.
12582 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12584 * Config modules support in openssl utility.
12586 Most commands now load modules from the config file,
12587 though in a few (such as version) this isn't done
12588 because it couldn't be used for anything.
12590 In the case of ca and req the config file used is
12591 the same as the utility itself: that is the -config
12592 command line option can be used to specify an
12597 * Move default behaviour from OPENSSL_config(). If appname is NULL
12598 use "openssl_conf" if filename is NULL use default openssl config file.
12602 * Add an argument to OPENSSL_config() to allow the use of an alternative
12603 config section name. Add a new flag to tolerate a missing config file
12604 and move code to CONF_modules_load_file().
12608 * Support for crypto accelerator cards from Accelerated Encryption
12609 Processing, www.aep.ie. (Use engine 'aep')
12610 The support was copied from 0.9.6c [engine] and adapted/corrected
12611 to work with the new engine framework.
12613 *AEP Inc. and Richard Levitte*
12615 * Support for SureWare crypto accelerator cards from Baltimore
12616 Technologies. (Use engine 'sureware')
12617 The support was copied from 0.9.6c [engine] and adapted
12618 to work with the new engine framework.
12622 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12623 make the newer ENGINE framework commands for the CHIL engine work.
12625 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12627 * Make it possible to produce shared libraries on ReliantUNIX.
12629 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12631 * Add the configuration target debug-linux-ppro.
12632 Make 'openssl rsa' use the general key loading routines
12633 implemented in `apps.c`, and make those routines able to
12634 handle the key format FORMAT_NETSCAPE and the variant
12637 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12639 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12641 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12643 * Add -keyform to rsautl, and document -engine.
12645 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12647 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12648 BIO_R_NO_SUCH_FILE error code rather than the generic
12649 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12653 * Add new functions
12654 ERR_peek_last_error
12655 ERR_peek_last_error_line
12656 ERR_peek_last_error_line_data.
12657 These are similar to
12659 ERR_peek_error_line
12660 ERR_peek_error_line_data,
12661 but report on the latest error recorded rather than the first one
12662 still in the error queue.
12664 *Ben Laurie, Bodo Moeller*
12666 * default_algorithms option in ENGINE config module. This allows things
12668 default_algorithms = ALL
12669 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12673 * Preliminary ENGINE config module.
12677 * New experimental application configuration code.
12681 * Change the AES code to follow the same name structure as all other
12682 symmetric ciphers, and behave the same way. Move everything to
12683 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12685 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12687 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12689 *Ben Laurie and Theo de Raadt*
12691 * Add option to output public keys in req command.
12693 *Massimiliano Pala madwolf@openca.org*
12695 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12696 (up to about 10% better than before for P-192 and P-224).
12700 * New functions/macros
12702 SSL_CTX_set_msg_callback(ctx, cb)
12703 SSL_CTX_set_msg_callback_arg(ctx, arg)
12704 SSL_set_msg_callback(ssl, cb)
12705 SSL_set_msg_callback_arg(ssl, arg)
12707 to request calling a callback function
12709 void cb(int write_p, int version, int content_type,
12710 const void *buf, size_t len, SSL *ssl, void *arg)
12712 whenever a protocol message has been completely received
12713 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12714 protocol version according to which the SSL library interprets
12715 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12716 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12717 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12718 specification (change_cipher_spec(20), alert(21), handshake(22)).
12719 'buf' and 'len' point to the actual message, 'ssl' to the
12720 SSL object, and 'arg' is the application-defined value set by
12721 SSL[_CTX]_set_msg_callback_arg().
12723 'openssl s_client' and 'openssl s_server' have new '-msg' options
12724 to enable a callback that displays all protocol messages.
12728 * Change the shared library support so shared libraries are built as
12729 soon as the corresponding static library is finished, and thereby get
12730 openssl and the test programs linked against the shared library.
12731 This still only happens when the keyword "shard" has been given to
12732 the configuration scripts.
12734 NOTE: shared library support is still an experimental thing, and
12735 backward binary compatibility is still not guaranteed.
12737 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12739 * Add support for Subject Information Access extension.
12741 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12743 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12744 additional bytes when new memory had to be allocated, not just
12745 when reusing an existing buffer.
12749 * New command line and configuration option 'utf8' for the req command.
12750 This allows field values to be specified as UTF8 strings.
12754 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12755 runs for the former and machine-readable output for the latter.
12759 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12760 of the e-mail address in the DN (i.e., it will go into a certificate
12761 extension only). The new configuration file option 'email_in_dn = no'
12762 has the same effect.
12764 *Massimiliano Pala madwolf@openca.org*
12766 * Change all functions with names starting with `des_` to be starting
12767 with `DES_` instead. Add wrappers that are compatible with libdes,
12768 but are named `_ossl_old_des_*`. Finally, add macros that map the
12769 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12770 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12771 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12774 Since we provide two compatibility mappings, the user needs to
12775 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12776 compatibility is desired. The default (i.e., when that macro
12777 isn't defined) is OpenSSL 0.9.6c compatibility.
12779 There are also macros that enable and disable the support of old
12780 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12781 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12782 are defined, the default will apply: to support the old des routines.
12784 In either case, one must include openssl/des.h to get the correct
12785 definitions. Do not try to just include openssl/des_old.h, that
12788 NOTE: This is a major break of an old API into a new one. Software
12789 authors are encouraged to switch to the `DES_` style functions. Some
12790 time in the future, des_old.h and the libdes compatibility functions
12791 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12792 default), and then completely removed.
12796 * Test for certificates which contain unsupported critical extensions.
12797 If such a certificate is found during a verify operation it is
12798 rejected by default: this behaviour can be overridden by either
12799 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12800 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12801 X509_supported_extension() has also been added which returns 1 if a
12802 particular extension is supported.
12806 * Modify the behaviour of EVP cipher functions in similar way to digests
12807 to retain compatibility with existing code.
12811 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12812 compatibility with existing code. In particular the 'ctx' parameter does
12813 not have to be to be initialized before the call to EVP_DigestInit() and
12814 it is tidied up after a call to EVP_DigestFinal(). New function
12815 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12816 EVP_MD_CTX_copy() changed to not require the destination to be
12817 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12818 requires the destination to be valid.
12820 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12821 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12825 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12826 so that complete 'Handshake' protocol structures are kept in memory
12827 instead of overwriting 'msg_type' and 'length' with 'body' data.
12831 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12833 *Massimo Santin via Richard Levitte*
12835 * Major restructuring to the underlying ENGINE code. This includes
12836 reduction of linker bloat, separation of pure "ENGINE" manipulation
12837 (initialisation, etc) from functionality dealing with implementations
12838 of specific crypto interfaces. This change also introduces integrated
12839 support for symmetric ciphers and digest implementations - so ENGINEs
12840 can now accelerate these by providing EVP_CIPHER and EVP_MD
12841 implementations of their own. This is detailed in
12842 [crypto/engine/README.md](crypto/engine/README.md)
12843 as it couldn't be adequately described here. However, there are a few
12844 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12845 were changed in the original introduction of ENGINE code have now
12846 reverted back - the hooking from this code to ENGINE is now a good
12847 deal more passive and at run-time, operations deal directly with
12848 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12849 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12850 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12851 they were not being used by the framework as there is no concept of a
12852 BIGNUM_METHOD and they could not be generalised to the new
12853 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12854 ENGINE_cpy() has been removed as it cannot be consistently defined in
12859 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12863 * Change mkdef.pl to sort symbols that get the same entry number,
12864 and make sure the automatically generated functions `ERR_load_*`
12865 become part of libeay.num as well.
12869 * New function SSL_renegotiate_pending(). This returns true once
12870 renegotiation has been requested (either SSL_renegotiate() call
12871 or HelloRequest/ClientHello received from the peer) and becomes
12872 false once a handshake has been completed.
12873 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12874 sends a HelloRequest, but does not ensure that a handshake takes
12875 place. SSL_renegotiate_pending() is useful for checking if the
12876 client has followed the request.)
12880 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12881 By default, clients may request session resumption even during
12882 renegotiation (if session ID contexts permit); with this option,
12883 session resumption is possible only in the first handshake.
12885 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12886 more bits available for options that should not be part of
12887 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12891 * Add some demos for certificate and certificate request creation.
12895 * Make maximum certificate chain size accepted from the peer application
12896 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12897 "Douglas E. Engert" <deengert@anl.gov>.
12901 * Add support for shared libraries for Unixware-7
12902 (Boyd Lynn Gerber <gerberb@zenez.com>).
12906 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12907 be done prior to destruction. Use this to unload error strings from
12908 ENGINEs that load their own error strings. NB: This adds two new API
12909 functions to "get" and "set" this destroy handler in an ENGINE.
12913 * Alter all existing ENGINE implementations (except "openssl" and
12914 "openbsd") to dynamically instantiate their own error strings. This
12915 makes them more flexible to be built both as statically-linked ENGINEs
12916 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12917 Also, add stub code to each that makes building them as self-contained
12918 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12922 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12923 implementations into applications that are completely implemented in
12924 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12925 commands that can be used to configure what shared-library to load and
12926 to control aspects of the way it is handled. Also, made an update to
12927 the [README-Engine.md](README-Engine.md) file
12928 that brings its information up-to-date and
12929 provides some information and instructions on the "dynamic" ENGINE
12930 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12934 * Make it possible to unload ranges of ERR strings with a new
12935 "ERR_unload_strings" function.
12939 * Add a copy() function to EVP_MD.
12943 * Make EVP_MD routines take a context pointer instead of just the
12944 md_data void pointer.
12948 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12949 that the digest can only process a single chunk of data
12950 (typically because it is provided by a piece of
12951 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12952 is only going to provide a single chunk of data, and hence the
12953 framework needn't accumulate the data for oneshot drivers.
12957 * As with "ERR", make it possible to replace the underlying "ex_data"
12958 functions. This change also alters the storage and management of global
12959 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12960 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12961 index counters. The API functions that use this state have been changed
12962 to take a "class_index" rather than pointers to the class's local STACK
12963 and counter, and there is now an API function to dynamically create new
12964 classes. This centralisation allows us to (a) plug a lot of the
12965 thread-safety problems that existed, and (b) makes it possible to clean
12966 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12967 such data would previously have always leaked in application code and
12968 workarounds were in place to make the memory debugging turn a blind eye
12969 to it. Application code that doesn't use this new function will still
12970 leak as before, but their memory debugging output will announce it now
12971 rather than letting it slide.
12973 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12974 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12975 has a return value to indicate success or failure.
12979 * Make it possible to replace the underlying "ERR" functions such that the
12980 global state (2 LHASH tables and 2 locks) is only used by the "default"
12981 implementation. This change also adds two functions to "get" and "set"
12982 the implementation prior to it being automatically set the first time
12983 any other ERR function takes place. Ie. an application can call "get",
12984 pass the return value to a module it has just loaded, and that module
12985 can call its own "set" function using that value. This means the
12986 module's "ERR" operations will use (and modify) the error state in the
12987 application and not in its own statically linked copy of OpenSSL code.
12991 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12992 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12993 the operation, and provides a more encapsulated way for external code
12994 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12995 to use these functions rather than manually incrementing the counts.
12997 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13001 * Add EVP test program.
13005 * Add symmetric cipher support to ENGINE. Expect the API to change!
13009 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13010 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13011 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13012 These allow a CRL to be built without having to access X509_CRL fields
13013 directly. Modify 'ca' application to use new functions.
13017 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13018 bug workarounds. Rollback attack detection is a security feature.
13019 The problem will only arise on OpenSSL servers when TLSv1 is not
13020 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13021 Software authors not wanting to support TLSv1 will have special reasons
13022 for their choice and can explicitly enable this option.
13024 *Bodo Moeller, Lutz Jaenicke*
13026 * Rationalise EVP so it can be extended: don't include a union of
13027 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13028 (similar to those existing for EVP_CIPHER_CTX).
13033 EVP_MD_CTX_init(&md); /* new function call */
13034 EVP_DigestInit(&md, EVP_sha1());
13035 EVP_DigestUpdate(&md, in, len);
13036 EVP_DigestFinal(&md, out, NULL);
13037 EVP_MD_CTX_cleanup(&md); /* new function call */
13041 * Make DES key schedule conform to the usual scheme, as well as
13042 correcting its structure. This means that calls to DES functions
13043 now have to pass a pointer to a des_key_schedule instead of a
13044 plain des_key_schedule (which was actually always a pointer
13047 des_key_schedule ks;
13049 des_set_key_checked(..., &ks);
13050 des_ncbc_encrypt(..., &ks, ...);
13052 (Note that a later change renames 'des_...' into 'DES_...'.)
13056 * Initial reduction of linker bloat: the use of some functions, such as
13057 PEM causes large amounts of unused functions to be linked in due to
13058 poor organisation. For example pem_all.c contains every PEM function
13059 which has a knock on effect of linking in large amounts of (unused)
13060 ASN1 code. Grouping together similar functions and splitting unrelated
13061 functions prevents this.
13065 * Cleanup of EVP macros.
13069 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13070 correct `_ecb suffix`.
13074 * Add initial OCSP responder support to ocsp application. The
13075 revocation information is handled using the text based index
13076 use by the ca application. The responder can either handle
13077 requests generated internally, supplied in files (for example
13078 via a CGI script) or using an internal minimal server.
13082 * Add configuration choices to get zlib compression for TLS.
13086 * Changes to Kerberos SSL for RFC 2712 compliance:
13087 1. Implemented real KerberosWrapper, instead of just using
13088 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13089 2. Implemented optional authenticator field of KerberosWrapper.
13091 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13092 and authenticator structs; see crypto/krb5/.
13094 Generalized Kerberos calls to support multiple Kerberos libraries.
13095 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13096 via Richard Levitte*
13098 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13099 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13100 values for each of the key sizes rather than having just
13101 parameters (and 'speed' generating keys each time).
13105 * Speed up EVP routines.
13108 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13109 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13110 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13111 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13113 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13114 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13115 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13118 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13120 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13124 * Added the OS2-EMX target.
13126 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13128 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13129 New functions to support `NCONF` routines in extension code.
13130 New function `CONF_set_nconf()`
13131 to allow functions which take an `NCONF` to also handle the old `LHASH`
13132 structure: this means that the old `CONF` compatible routines can be
13133 retained (in particular w.rt. extensions) without having to duplicate the
13134 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13138 * Enhance the general user interface with mechanisms for inner control
13139 and with possibilities to have yes/no kind of prompts.
13143 * Change all calls to low-level digest routines in the library and
13144 applications to use EVP. Add missing calls to HMAC_cleanup() and
13145 don't assume HMAC_CTX can be copied using memcpy().
13147 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13149 * Add the possibility to control engines through control names but with
13150 arbitrary arguments instead of just a string.
13151 Change the key loaders to take a UI_METHOD instead of a callback
13152 function pointer. NOTE: this breaks binary compatibility with earlier
13153 versions of OpenSSL [engine].
13154 Adapt the nCipher code for these new conditions and add a card insertion
13159 * Enhance the general user interface with mechanisms to better support
13160 dialog box interfaces, application-defined prompts, the possibility
13161 to use defaults (for example default passwords from somewhere else)
13162 and interrupts/cancellations.
13166 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13167 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13171 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13172 tidy up some unnecessarily weird code in 'sk_new()').
13174 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13176 * Change the key loading routines for ENGINEs to use the same kind
13177 callback (pem_password_cb) as all other routines that need this
13182 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13183 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13184 than this minimum value is recommended.
13188 * New random seeder for OpenVMS, using the system process statistics
13189 that are easily reachable.
13193 * Windows apparently can't transparently handle global
13194 variables defined in DLLs. Initialisations such as:
13196 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13198 won't compile. This is used by the any applications that need to
13199 declare their own ASN1 modules. This was fixed by adding the option
13200 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13201 needed for static libraries under Win32.
13205 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13206 setting of purpose and trust fields. New X509_STORE trust and
13207 purpose functions and tidy up setting in other SSL functions.
13211 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13212 structure. These are inherited by X509_STORE_CTX when it is
13213 initialised. This allows various defaults to be set in the
13214 X509_STORE structure (such as flags for CRL checking and custom
13215 purpose or trust settings) for functions which only use X509_STORE_CTX
13216 internally such as S/MIME.
13218 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13219 trust settings if they are not set in X509_STORE. This allows X509_STORE
13220 purposes and trust (in S/MIME for example) to override any set by default.
13222 Add command line options for CRL checking to smime, s_client and s_server
13227 * Initial CRL based revocation checking. If the CRL checking flag(s)
13228 are set then the CRL is looked up in the X509_STORE structure and
13229 its validity and signature checked, then if the certificate is found
13230 in the CRL the verify fails with a revoked error.
13232 Various new CRL related callbacks added to X509_STORE_CTX structure.
13234 Command line options added to 'verify' application to support this.
13236 This needs some additional work, such as being able to handle multiple
13237 CRLs with different times, extension based lookup (rather than just
13238 by subject name) and ultimately more complete V2 CRL extension
13243 * Add a general user interface API (crypto/ui/). This is designed
13244 to replace things like des_read_password and friends (backward
13245 compatibility functions using this new API are provided).
13246 The purpose is to remove prompting functions from the DES code
13247 section as well as provide for prompting through dialog boxes in
13248 a window system and the like.
13252 * Add "ex_data" support to ENGINE so implementations can add state at a
13253 per-structure level rather than having to store it globally.
13257 * Make it possible for ENGINE structures to be copied when retrieved by
13258 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13259 This causes the "original" ENGINE structure to act like a template,
13260 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13261 operational state can be localised to each ENGINE structure, despite the
13262 fact they all share the same "methods". New ENGINE structures returned in
13263 this case have no functional references and the return value is the single
13264 structural reference. This matches the single structural reference returned
13265 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13270 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13271 needs to match any other type at all we need to manually clear the
13276 * Changes to the "openssl engine" utility to include;
13277 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13278 about an ENGINE's available control commands.
13279 - executing control commands from command line arguments using the
13280 '-pre' and '-post' switches. '-post' is only used if '-t' is
13281 specified and the ENGINE is successfully initialised. The syntax for
13282 the individual commands are colon-separated, for example;
13283 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13287 * New dynamic control command support for ENGINEs. ENGINEs can now
13288 declare their own commands (numbers), names (strings), descriptions,
13289 and input types for run-time discovery by calling applications. A
13290 subset of these commands are implicitly classed as "executable"
13291 depending on their input type, and only these can be invoked through
13292 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13293 can be based on user input, config files, etc). The distinction is
13294 that "executable" commands cannot return anything other than a boolean
13295 result and can only support numeric or string input, whereas some
13296 discoverable commands may only be for direct use through
13297 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13298 pointers, or other custom uses. The "executable" commands are to
13299 support parameterisations of ENGINE behaviour that can be
13300 unambiguously defined by ENGINEs and used consistently across any
13301 OpenSSL-based application. Commands have been added to all the
13302 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13303 control over shared-library paths without source code alterations.
13307 * Changed all ENGINE implementations to dynamically allocate their
13308 ENGINEs rather than declaring them statically. Apart from this being
13309 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13310 this also allows the implementations to compile without using the
13311 internal engine_int.h header.
13315 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13316 'const' value. Any code that should be able to modify a RAND_METHOD
13317 should already have non-const pointers to it (ie. they should only
13318 modify their own ones).
13322 * Made a variety of little tweaks to the ENGINE code.
13323 - "atalla" and "ubsec" string definitions were moved from header files
13324 to C code. "nuron" string definitions were placed in variables
13325 rather than hard-coded - allowing parameterisation of these values
13326 later on via ctrl() commands.
13327 - Removed unused "#if 0"'d code.
13328 - Fixed engine list iteration code so it uses ENGINE_free() to release
13329 structural references.
13330 - Constified the RAND_METHOD element of ENGINE structures.
13331 - Constified various get/set functions as appropriate and added
13332 missing functions (including a catch-all ENGINE_cpy that duplicates
13333 all ENGINE values onto a new ENGINE except reference counts/state).
13334 - Removed NULL parameter checks in get/set functions. Setting a method
13335 or function to NULL is a way of cancelling out a previously set
13336 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13337 and doesn't justify the extra error symbols and code.
13338 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13339 flags from engine_int.h to engine.h.
13340 - Changed prototypes for ENGINE handler functions (init(), finish(),
13341 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13345 * Implement binary inversion algorithm for BN_mod_inverse in addition
13346 to the algorithm using long division. The binary algorithm can be
13347 used only if the modulus is odd. On 32-bit systems, it is faster
13348 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13349 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13350 up to 450 bits. In 64-bit environments, the binary algorithm
13351 appears to be advantageous for much longer moduli; here we use it
13352 for moduli up to 2048 bits.
13356 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13357 could not support the combine flag in choice fields.
13361 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13362 extensions from a certificate request to the certificate.
13366 * Allow multiple 'certopt' and 'nameopt' options to be separated
13367 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13368 file: this allows the display of the certificate about to be
13369 signed to be customised, to allow certain fields to be included
13370 or excluded and extension details. The old system didn't display
13371 multicharacter strings properly, omitted fields not in the policy
13372 and couldn't display additional details such as extensions.
13376 * Function EC_POINTs_mul for multiple scalar multiplication
13377 of an arbitrary number of elliptic curve points
13378 \sum scalars[i]*points[i],
13379 optionally including the generator defined for the EC_GROUP:
13380 scalar*generator + \sum scalars[i]*points[i].
13382 EC_POINT_mul is a simple wrapper function for the typical case
13383 that the point list has just one item (besides the optional
13388 * First EC_METHODs for curves over GF(p):
13390 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13391 operations and provides various method functions that can also
13392 operate with faster implementations of modular arithmetic.
13394 EC_GFp_mont_method() reuses most functions that are part of
13395 EC_GFp_simple_method, but uses Montgomery arithmetic.
13397 *Bodo Moeller; point addition and point doubling
13398 implementation directly derived from source code provided by
13399 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13401 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13402 crypto/ec/ec_lib.c):
13404 Curves are EC_GROUP objects (with an optional group generator)
13405 based on EC_METHODs that are built into the library.
13407 Points are EC_POINT objects based on EC_GROUP objects.
13409 Most of the framework would be able to handle curves over arbitrary
13410 finite fields, but as there are no obvious types for fields other
13411 than GF(p), some functions are limited to that for now.
13415 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13416 that the file contains a complete HTTP response.
13420 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13421 change the def and num file printf format specifier from "%-40sXXX"
13422 to "%-39s XXX". The latter will always guarantee a space after the
13423 field while the former will cause them to run together if the field
13424 is 40 of more characters long.
13428 * Constify the cipher and digest 'method' functions and structures
13429 and modify related functions to take constant EVP_MD and EVP_CIPHER
13434 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13435 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13439 * Modify `EVP_Digest*()` routines so they now return values. Although the
13440 internal software routines can never fail additional hardware versions
13445 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13447 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13448 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13451 ERR_R_NESTED_ASN1_ERROR
13453 ERR_R_MISSING_ASN1_EOS
13454 were 4 .. 9, conflicting with
13455 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13457 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13458 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13460 Add new error code 'ERR_R_INTERNAL_ERROR'.
13464 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13469 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13470 sets the subject name for a new request or supersedes the
13471 subject name in a given request. Formats that can be parsed are
13472 'CN=Some Name, OU=myOU, C=IT'
13474 'CN=Some Name/OU=myOU/C=IT'.
13476 Add options '-batch' and '-verbose' to 'openssl req'.
13478 *Massimiliano Pala <madwolf@hackmasters.net>*
13480 * Introduce the possibility to access global variables through
13481 functions on platform were that's the best way to handle exporting
13482 global variables in shared libraries. To enable this functionality,
13483 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13484 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13485 is normally done by Configure or something similar).
13487 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13488 in the source file (foo.c) like this:
13490 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13491 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13493 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13494 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13496 OPENSSL_DECLARE_GLOBAL(int,foo);
13497 #define foo OPENSSL_GLOBAL_REF(foo)
13498 OPENSSL_DECLARE_GLOBAL(double,bar);
13499 #define bar OPENSSL_GLOBAL_REF(bar)
13501 The #defines are very important, and therefore so is including the
13502 header file everywhere where the defined globals are used.
13504 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13505 of ASN.1 items, but that structure is a bit different.
13507 The largest change is in util/mkdef.pl which has been enhanced with
13508 better and easier to understand logic to choose which symbols should
13509 go into the Windows .def files as well as a number of fixes and code
13510 cleanup (among others, algorithm keywords are now sorted
13511 lexicographically to avoid constant rewrites).
13515 * In BN_div() keep a copy of the sign of 'num' before writing the
13516 result to 'rm' because if rm==num the value will be overwritten
13517 and produce the wrong result if 'num' is negative: this caused
13518 problems with BN_mod() and BN_nnmod().
13522 * Function OCSP_request_verify(). This checks the signature on an
13523 OCSP request and verifies the signer certificate. The signer
13524 certificate is just checked for a generic purpose and OCSP request
13529 * Add OCSP_check_validity() function to check the validity of OCSP
13530 responses. OCSP responses are prepared in real time and may only
13531 be a few seconds old. Simply checking that the current time lies
13532 between thisUpdate and nextUpdate max reject otherwise valid responses
13533 caused by either OCSP responder or client clock inaccuracy. Instead
13534 we allow thisUpdate and nextUpdate to fall within a certain period of
13535 the current time. The age of the response can also optionally be
13536 checked. Two new options -validity_period and -status_age added to
13541 * If signature or public key algorithm is unrecognized print out its
13542 OID rather that just UNKNOWN.
13546 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13547 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13548 ID to be generated from the issuer certificate alone which can then be
13549 passed to OCSP_id_issuer_cmp().
13553 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13554 ASN1 modules to export functions returning ASN1_ITEM pointers
13555 instead of the ASN1_ITEM structures themselves. This adds several
13556 new macros which allow the underlying ASN1 function/structure to
13557 be accessed transparently. As a result code should not use ASN1_ITEM
13558 references directly (such as &X509_it) but instead use the relevant
13559 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13560 use of the new ASN1 code on platforms where exporting structures
13561 is problematical (for example in shared libraries) but exporting
13562 functions returning pointers to structures is not.
13566 * Add support for overriding the generation of SSL/TLS session IDs.
13567 These callbacks can be registered either in an SSL_CTX or per SSL.
13568 The purpose of this is to allow applications to control, if they wish,
13569 the arbitrary values chosen for use as session IDs, particularly as it
13570 can be useful for session caching in multiple-server environments. A
13571 command-line switch for testing this (and any client code that wishes
13572 to use such a feature) has been added to "s_server".
13574 *Geoff Thorpe, Lutz Jaenicke*
13576 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13577 of the form `#if defined(...) || defined(...) || ...` and
13578 `#if !defined(...) && !defined(...) && ...`. This also avoids
13579 the growing number of special cases it was previously handling.
13583 * Make all configuration macros available for application by making
13584 sure they are available in opensslconf.h, by giving them names starting
13585 with `OPENSSL_` to avoid conflicts with other packages and by making
13586 sure e_os2.h will cover all platform-specific cases together with
13588 Additionally, it is now possible to define configuration/platform-
13589 specific names (called "system identities"). In the C code, these
13590 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13591 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13592 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13597 * New option -set_serial to 'req' and 'x509' this allows the serial
13598 number to use to be specified on the command line. Previously self
13599 signed certificates were hard coded with serial number 0 and the
13600 CA options of 'x509' had to use a serial number in a file which was
13605 * New options to 'ca' utility to support V2 CRL entry extensions.
13606 Currently CRL reason, invalidity date and hold instruction are
13607 supported. Add new CRL extensions to V3 code and some new objects.
13611 * New function EVP_CIPHER_CTX_set_padding() this is used to
13612 disable standard block padding (aka PKCS#5 padding) in the EVP
13613 API, which was previously mandatory. This means that the data is
13614 not padded in any way and so the total length much be a multiple
13615 of the block size, otherwise an error occurs.
13619 * Initial (incomplete) OCSP SSL support.
13623 * New function OCSP_parse_url(). This splits up a URL into its host,
13624 port and path components: primarily to parse OCSP URLs. New -url
13625 option to ocsp utility.
13629 * New nonce behavior. The return value of OCSP_check_nonce() now
13630 reflects the various checks performed. Applications can decide
13631 whether to tolerate certain situations such as an absent nonce
13632 in a response when one was present in a request: the ocsp application
13633 just prints out a warning. New function OCSP_add1_basic_nonce()
13634 this is to allow responders to include a nonce in a response even if
13635 the request is nonce-less.
13639 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13640 skipped when using openssl x509 multiple times on a single input file,
13641 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13645 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13646 set string type: to handle setting ASN1_TIME structures. Fix ca
13647 utility to correctly initialize revocation date of CRLs.
13651 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13652 the clients preferred ciphersuites and rather use its own preferences.
13653 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13654 Internet Explorer by ensuring unchanged hash method during stepup.
13655 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13659 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13660 to aes and add a new 'exist' option to print out symbols that don't
13665 * Additional options to ocsp utility to allow flags to be set and
13666 additional certificates supplied.
13670 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13671 OCSP client a number of certificate to only verify the response
13676 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13677 handle the new API. Currently only ECB, CBC modes supported. Add new
13680 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13681 Encryption Standard (AES) Ciphersuites for Transport Layer
13682 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13683 not enabled by default and were not part of the "ALL" ciphersuite
13684 alias because they were not yet official; they could be
13685 explicitly requested by specifying the "AESdraft" ciphersuite
13686 group alias. In the final release of OpenSSL 0.9.7, the group
13687 alias is called "AES" and is part of "ALL".)
13689 *Ben Laurie, Steve Henson, Bodo Moeller*
13691 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13692 request to response.
13696 * Functions for OCSP responders. OCSP_request_onereq_count(),
13697 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13698 extract information from a certificate request. OCSP_response_create()
13699 creates a response and optionally adds a basic response structure.
13700 OCSP_basic_add1_status() adds a complete single response to a basic
13701 response and returns the OCSP_SINGLERESP structure just added (to allow
13702 extensions to be included for example). OCSP_basic_add1_cert() adds a
13703 certificate to a basic response and OCSP_basic_sign() signs a basic
13704 response with various flags. New helper functions ASN1_TIME_check()
13705 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13706 (converts ASN1_TIME to GeneralizedTime).
13710 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13711 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13712 structure from a certificate. X509_pubkey_digest() digests the public_key
13713 contents: this is used in various key identifiers.
13717 * Make sk_sort() tolerate a NULL argument.
13719 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13721 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13722 passed by the function are trusted implicitly. If any of them signed the
13723 response then it is assumed to be valid and is not verified.
13727 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13728 to data. This was previously part of the PKCS7 ASN1 code. This
13729 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13730 *Steve Henson, reported by Kenneth R. Robinette
13731 <support@securenetterm.com>*
13733 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13734 routines: without these tracing memory leaks is very painful.
13735 Fix leaks in PKCS12 and PKCS7 routines.
13739 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13740 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13741 effectively meant GeneralizedTime would never be used. Now it
13742 is initialised to -1 but X509_time_adj() now has to check the value
13743 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13744 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13745 *Steve Henson, reported by Kenneth R. Robinette
13746 <support@securenetterm.com>*
13748 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13749 result in a zero length in the ASN1_INTEGER structure which was
13750 not consistent with the structure when d2i_ASN1_INTEGER() was used
13751 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13752 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13753 where it did not print out a minus for negative ASN1_INTEGER.
13757 * Add summary printout to ocsp utility. The various functions which
13758 convert status values to strings have been renamed to:
13759 OCSP_response_status_str(), OCSP_cert_status_str() and
13760 OCSP_crl_reason_str() and are no longer static. New options
13761 to verify nonce values and to disable verification. OCSP response
13762 printout format cleaned up.
13766 * Add additional OCSP certificate checks. These are those specified
13767 in RFC2560. This consists of two separate checks: the CA of the
13768 certificate being checked must either be the OCSP signer certificate
13769 or the issuer of the OCSP signer certificate. In the latter case the
13770 OCSP signer certificate must contain the OCSP signing extended key
13771 usage. This check is performed by attempting to match the OCSP
13772 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13773 in the OCSP_CERTID structures of the response.
13777 * Initial OCSP certificate verification added to OCSP_basic_verify()
13778 and related routines. This uses the standard OpenSSL certificate
13779 verify routines to perform initial checks (just CA validity) and
13780 to obtain the certificate chain. Then additional checks will be
13781 performed on the chain. Currently the root CA is checked to see
13782 if it is explicitly trusted for OCSP signing. This is used to set
13783 a root CA as a global signing root: that is any certificate that
13784 chains to that CA is an acceptable OCSP signing certificate.
13788 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13789 extensions from a separate configuration file.
13790 As when reading extensions from the main configuration file,
13791 the '-extensions ...' option may be used for specifying the
13794 *Massimiliano Pala <madwolf@comune.modena.it>*
13796 * New OCSP utility. Allows OCSP requests to be generated or
13797 read. The request can be sent to a responder and the output
13798 parsed, outputted or printed in text form. Not complete yet:
13799 still needs to check the OCSP response validity.
13803 * New subcommands for 'openssl ca':
13804 `openssl ca -status <serial>` prints the status of the cert with
13805 the given serial number (according to the index file).
13806 `openssl ca -updatedb` updates the expiry status of certificates
13809 *Massimiliano Pala <madwolf@comune.modena.it>*
13811 * New '-newreq-nodes' command option to CA.pl. This is like
13812 '-newreq', but calls 'openssl req' with the '-nodes' option
13813 so that the resulting key is not encrypted.
13815 *Damien Miller <djm@mindrot.org>*
13817 * New configuration for the GNU Hurd.
13819 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13821 * Initial code to implement OCSP basic response verify. This
13822 is currently incomplete. Currently just finds the signer's
13823 certificate and verifies the signature on the response.
13827 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13828 value of OPENSSLDIR. This is available via the new '-d' option
13829 to 'openssl version', and is also included in 'openssl version -a'.
13833 * Allowing defining memory allocation callbacks that will be given
13834 file name and line number information in additional arguments
13835 (a `const char*` and an int). The basic functionality remains, as
13836 well as the original possibility to just replace malloc(),
13837 realloc() and free() by functions that do not know about these
13838 additional arguments. To register and find out the current
13839 settings for extended allocation functions, the following
13840 functions are provided:
13842 CRYPTO_set_mem_ex_functions
13843 CRYPTO_set_locked_mem_ex_functions
13844 CRYPTO_get_mem_ex_functions
13845 CRYPTO_get_locked_mem_ex_functions
13847 These work the same way as CRYPTO_set_mem_functions and friends.
13848 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13849 extended allocation function is enabled.
13850 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13851 a conventional allocation function is enabled.
13853 *Richard Levitte, Bodo Moeller*
13855 * Finish off removing the remaining LHASH function pointer casts.
13856 There should no longer be any prototype-casting required when using
13857 the LHASH abstraction, and any casts that remain are "bugs". See
13858 the callback types and macros at the head of lhash.h for details
13859 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13863 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13864 If /dev/[u]random devices are not available or do not return enough
13865 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13867 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13868 /etc/entropy will be queried once each in this sequence, querying stops
13869 when enough entropy was collected without querying more sockets.
13873 * Change the Unix RAND_poll() variant to be able to poll several
13874 random devices, as specified by DEVRANDOM, until a sufficient amount
13875 of data has been collected. We spend at most 10 ms on each file
13876 (select timeout) and read in non-blocking mode. DEVRANDOM now
13877 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13878 (previously it was just the string "/dev/urandom"), so on typical
13879 platforms the 10 ms delay will never occur.
13880 Also separate out the Unix variant to its own file, rand_unix.c.
13881 For VMS, there's a currently-empty rand_vms.c.
13885 * Move OCSP client related routines to ocsp_cl.c. These
13886 provide utility functions which an application needing
13887 to issue a request to an OCSP responder and analyse the
13888 response will typically need: as opposed to those which an
13889 OCSP responder itself would need which will be added later.
13891 OCSP_request_sign() signs an OCSP request with an API similar
13892 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13893 response. OCSP_response_get1_basic() extracts basic response
13894 from response. OCSP_resp_find_status(): finds and extracts status
13895 information from an OCSP_CERTID structure (which will be created
13896 when the request structure is built). These are built from lower
13897 level functions which work on OCSP_SINGLERESP structures but
13898 won't normally be used unless the application wishes to examine
13899 extensions in the OCSP response for example.
13901 Replace nonce routines with a pair of functions.
13902 OCSP_request_add1_nonce() adds a nonce value and optionally
13903 generates a random value. OCSP_check_nonce() checks the
13904 validity of the nonce in an OCSP response.
13908 * Change function OCSP_request_add() to OCSP_request_add0_id().
13909 This doesn't copy the supplied OCSP_CERTID and avoids the
13910 need to free up the newly created id. Change return type
13911 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13912 This can then be used to add extensions to the request.
13913 Deleted OCSP_request_new(), since most of its functionality
13914 is now in OCSP_REQUEST_new() (and the case insensitive name
13915 clash) apart from the ability to set the request name which
13916 will be added elsewhere.
13920 * Update OCSP API. Remove obsolete extensions argument from
13921 various functions. Extensions are now handled using the new
13922 OCSP extension code. New simple OCSP HTTP function which
13923 can be used to send requests and parse the response.
13927 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13928 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13929 uses the special reorder version of SET OF to sort the attributes
13930 and reorder them to match the encoded order. This resolves a long
13931 standing problem: a verify on a PKCS7 structure just after signing
13932 it used to fail because the attribute order did not match the
13933 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13934 it uses the received order. This is necessary to tolerate some broken
13935 software that does not order SET OF. This is handled by encoding
13936 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13937 to produce the required SET OF.
13941 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13942 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13943 files to get correct declarations of the ASN.1 item variables.
13947 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13948 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13949 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13950 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13951 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13952 ASN1_ITEM and no wrapper functions.
13956 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13957 replace the old function pointer based I/O routines. Change most of
13958 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13962 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13963 lines, recognize more "algorithms" that can be deselected, and make
13964 it complain about algorithm deselection that isn't recognised.
13968 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13969 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13970 to use new functions. Add NO_ASN1_OLD which can be set to remove
13971 some old style ASN1 functions: this can be used to determine if old
13972 code will still work when these eventually go away.
13976 * New extension functions for OCSP structures, these follow the
13977 same conventions as certificates and CRLs.
13981 * New function X509V3_add1_i2d(). This automatically encodes and
13982 adds an extension. Its behaviour can be customised with various
13983 flags to append, replace or delete. Various wrappers added for
13984 certificates and CRLs.
13988 * Fix to avoid calling the underlying ASN1 print routine when
13989 an extension cannot be parsed. Correct a typo in the
13990 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13994 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13995 entries for variables.
13999 * Add functionality to `apps/openssl.c` for detecting locking
14000 problems: As the program is single-threaded, all we have
14001 to do is register a locking callback using an array for
14002 storing which locks are currently held by the program.
14006 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14007 SSL_get_ex_data_X509_STORE_idx(), which is used in
14008 ssl_verify_cert_chain() and thus can be called at any time
14009 during TLS/SSL handshakes so that thread-safety is essential.
14010 Unfortunately, the ex_data design is not at all suited
14011 for multi-threaded use, so it probably should be abolished.
14015 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14017 *Broadcom, tweaked and integrated by Geoff Thorpe*
14019 * Move common extension printing code to new function
14020 X509V3_print_extensions(). Reorganise OCSP print routines and
14021 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14025 * New function X509_signature_print() to remove duplication in some
14030 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14031 set (this was treated exactly the same as SET OF previously). This
14032 is used to reorder the STACK representing the structure to match the
14033 encoding. This will be used to get round a problem where a PKCS7
14034 structure which was signed could not be verified because the STACK
14035 order did not reflect the encoded order.
14039 * Reimplement the OCSP ASN1 module using the new code.
14043 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14044 for its ASN1 operations. The old style function pointers still exist
14045 for now but they will eventually go away.
14049 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14050 completely replaces the old ASN1 functionality with a table driven
14051 encoder and decoder which interprets an ASN1_ITEM structure describing
14052 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14053 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14054 has also been converted to the new form.
14058 * Change BN_mod_exp_recp so that negative moduli are tolerated
14059 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14060 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14061 for negative moduli.
14065 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14066 of not touching the result's sign bit.
14070 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14075 * Changed the LHASH code to use prototypes for callbacks, and created
14076 macros to declare and implement thin (optionally static) functions
14077 that provide type-safety and avoid function pointer casting for the
14078 type-specific callbacks.
14082 * Added Kerberos Cipher Suites to be used with TLS, as written in
14084 *Veers Staats <staatsvr@asc.hpc.mil>,
14085 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14087 * Reformat the FAQ so the different questions and answers can be divided
14088 in sections depending on the subject.
14092 * Have the zlib compression code load ZLIB.DLL dynamically under
14097 * New function BN_mod_sqrt for computing square roots modulo a prime
14098 (using the probabilistic Tonelli-Shanks algorithm unless
14099 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14100 be handled deterministically).
14102 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14104 * Make BN_mod_inverse faster by explicitly handling small quotients
14105 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14106 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14110 * New function BN_kronecker.
14114 * Fix BN_gcd so that it works on negative inputs; the result is
14115 positive unless both parameters are zero.
14116 Previously something reasonably close to an infinite loop was
14117 possible because numbers could be growing instead of shrinking
14118 in the implementation of Euclid's algorithm.
14122 * Fix BN_is_word() and BN_is_one() macros to take into account the
14123 sign of the number in question.
14125 Fix BN_is_word(a,w) to work correctly for w == 0.
14127 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14128 because its test if the absolute value of 'a' equals 'w'.
14129 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14130 it exists mostly for use in the implementations of BN_is_zero(),
14131 BN_is_one(), and BN_is_word().
14135 * New function BN_swap.
14139 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14140 the exponentiation functions are more likely to produce reasonable
14141 results on negative inputs.
14145 * Change BN_mod_mul so that the result is always non-negative.
14146 Previously, it could be negative if one of the factors was negative;
14147 I don't think anyone really wanted that behaviour.
14151 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14152 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14153 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14154 and add new functions:
14163 BN_mod_lshift1_quick
14165 BN_mod_lshift_quick
14167 These functions always generate non-negative results.
14169 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14170 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14172 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14173 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14174 be reduced modulo `m`.
14176 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14179 The following entry accidentally appeared in the CHANGES file
14180 distributed with OpenSSL 0.9.7. The modifications described in
14181 it do *not* apply to OpenSSL 0.9.7.
14183 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14184 was actually never needed) and in BN_mul(). The removal in BN_mul()
14185 required a small change in bn_mul_part_recursive() and the addition
14186 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14187 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14188 bn_sub_words() and bn_add_words() except they take arrays with
14194 * In 'openssl passwd', verify passwords read from the terminal
14195 unless the '-salt' option is used (which usually means that
14196 verification would just waste user's time since the resulting
14197 hash is going to be compared with some given password hash)
14198 or the new '-noverify' option is used.
14200 This is an incompatible change, but it does not affect
14201 non-interactive use of 'openssl passwd' (passwords on the command
14202 line, '-stdin' option, '-in ...' option) and thus should not
14203 cause any problems.
14207 * Remove all references to RSAref, since there's no more need for it.
14211 * Make DSO load along a path given through an environment variable
14212 (SHLIB_PATH) with shl_load().
14216 * Constify the ENGINE code as a result of BIGNUM constification.
14217 Also constify the RSA code and most things related to it. In a
14218 few places, most notable in the depth of the ASN.1 code, ugly
14219 casts back to non-const were required (to be solved at a later
14224 * Make it so the openssl application has all engines loaded by default.
14228 * Constify the BIGNUM routines a little more.
14232 * Add the following functions:
14234 ENGINE_load_cswift()
14236 ENGINE_load_atalla()
14237 ENGINE_load_nuron()
14238 ENGINE_load_builtin_engines()
14240 That way, an application can itself choose if external engines that
14241 are built-in in OpenSSL shall ever be used or not. The benefit is
14242 that applications won't have to be linked with libdl or other dso
14243 libraries unless it's really needed.
14245 Changed 'openssl engine' to load all engines on demand.
14246 Changed the engine header files to avoid the duplication of some
14247 declarations (they differed!).
14251 * 'openssl engine' can now list capabilities.
14255 * Better error reporting in 'openssl engine'.
14259 * Never call load_dh_param(NULL) in s_server.
14263 * Add engine application. It can currently list engines by name and
14264 identity, and test if they are actually available.
14268 * Improve RPM specification file by forcing symbolic linking and making
14269 sure the installed documentation is also owned by root.root.
14271 *Damien Miller <djm@mindrot.org>*
14273 * Give the OpenSSL applications more possibilities to make use of
14274 keys (public as well as private) handled by engines.
14278 * Add OCSP code that comes from CertCo.
14282 * Add VMS support for the Rijndael code.
14286 * Added untested support for Nuron crypto accelerator.
14290 * Add support for external cryptographic devices. This code was
14291 previously distributed separately as the "engine" branch.
14293 *Geoff Thorpe, Richard Levitte*
14295 * Rework the filename-translation in the DSO code. It is now possible to
14296 have far greater control over how a "name" is turned into a filename
14297 depending on the operating environment and any oddities about the
14298 different shared library filenames on each system.
14302 * Support threads on FreeBSD-elf in Configure.
14306 * Fix for SHA1 assembly problem with MASM: it produces
14307 warnings about corrupt line number information when assembling
14308 with debugging information. This is caused by the overlapping
14311 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14314 NCONF_get_number() has no error checking at all. As a replacement,
14315 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14316 promoted strongly. The old NCONF_get_number is kept around for
14317 binary backward compatibility.
14318 Make it possible for methods to load from something other than a BIO,
14319 by providing a function pointer that is given a name instead of a BIO.
14320 For example, this could be used to load configuration data from an
14325 * Fix for non blocking accept BIOs. Added new I/O special reason
14326 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14327 with non blocking I/O was not possible because no retry code was
14328 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14333 * Added the beginnings of Rijndael support.
14337 * Fix for bug in DirectoryString mask setting. Add support for
14338 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14339 to allow certificate printing to more controllable, additional
14340 'certopt' option to 'x509' to allow new printing options to be
14345 * Clean old EAY MD5 hack from e_os.h.
14349 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14351 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14352 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14354 *Joe Orton, Steve Henson*
14356 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14358 * Fix additional bug revealed by the NISCC test suite:
14360 Stop bug triggering large recursion when presented with
14361 certain ASN.1 tags ([CVE-2003-0851])
14365 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14367 * Fix various bugs revealed by running the NISCC test suite:
14369 Stop out of bounds reads in the ASN1 code when presented with
14370 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14372 If verify callback ignores invalid public key errors don't try to check
14373 certificate signature with the NULL public key.
14377 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14378 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14383 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14384 extra data after the compression methods not only for TLS 1.0
14385 but also for SSL 3.0 (as required by the specification).
14387 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14389 * Change X509_certificate_type() to mark the key as exported/exportable
14390 when it's 512 *bits* long, not 512 bytes.
14394 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14396 * Countermeasure against the Klima-Pokorny-Rosa extension of
14397 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14398 a protocol version number mismatch like a decryption error
14399 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14403 * Turn on RSA blinding by default in the default implementation
14404 to avoid a timing attack. Applications that don't want it can call
14405 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14406 They would be ill-advised to do so in most cases.
14408 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14410 * Change RSA blinding code so that it works when the PRNG is not
14411 seeded (in this case, the secret RSA exponent is abused as
14412 an unpredictable seed -- if it is not unpredictable, there
14413 is no point in blinding anyway). Make RSA blinding thread-safe
14414 by remembering the creator's thread ID in rsa->blinding and
14415 having all other threads use local one-time blinding factors
14416 (this requires more computation than sharing rsa->blinding, but
14417 avoids excessive locking; and if an RSA object is not shared
14418 between threads, blinding will still be very fast).
14422 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14424 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14425 via timing by performing a MAC computation even if incorrect
14426 block cipher padding has been found. This is a countermeasure
14427 against active attacks where the attacker has to distinguish
14428 between bad padding and a MAC verification error. ([CVE-2003-0078])
14430 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14431 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14432 Martin Vuagnoux (EPFL, Ilion)*
14434 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14436 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14437 memory from its contents. This is done with a counter that will
14438 place alternating values in each byte. This can be used to solve
14439 two issues: 1) the removal of calls to memset() by highly optimizing
14440 compilers, and 2) cleansing with other values than 0, since those can
14441 be read through on certain media, for example a swap space on disk.
14445 * Bugfix: client side session caching did not work with external caching,
14446 because the session->cipher setting was not restored when reloading
14447 from the external cache. This problem was masked, when
14448 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14449 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14453 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14454 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14456 *Zeev Lieber <zeev-l@yahoo.com>*
14458 * Undo an undocumented change introduced in 0.9.6e which caused
14459 repeated calls to OpenSSL_add_all_ciphers() and
14460 OpenSSL_add_all_digests() to be ignored, even after calling
14465 * Change the default configuration reader to deal with last line not
14466 being properly terminated.
14470 * Change X509_NAME_cmp() so it applies the special rules on handling
14471 DN values that are of type PrintableString, as well as RDNs of type
14472 emailAddress where the value has the type ia5String.
14474 *stefank@valicert.com via Richard Levitte*
14476 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14477 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14478 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14479 the bitwise-OR of the two for use by the majority of applications
14480 wanting this behaviour, and update the docs. The documented
14481 behaviour and actual behaviour were inconsistent and had been
14482 changing anyway, so this is more a bug-fix than a behavioural
14485 *Geoff Thorpe, diagnosed by Nadav Har'El*
14487 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14488 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14492 * Fix initialization code race conditions in
14493 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14494 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14495 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14496 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14497 ssl2_get_cipher_by_char(),
14498 ssl3_get_cipher_by_char().
14500 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14502 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14503 the cached sessions are flushed, as the remove_cb() might use ex_data
14504 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14505 (see [openssl.org #212]).
14507 *Geoff Thorpe, Lutz Jaenicke*
14509 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14510 length, instead of the encoding length to d2i_ASN1_OBJECT.
14514 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14516 * [In 0.9.6g-engine release:]
14517 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14519 *Lynn Gazis <lgazis@rainbow.com>*
14521 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14523 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14524 and get fix the header length calculation.
14525 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14526 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14528 * Use proper error handling instead of 'assertions' in buffer
14529 overflow checks added in 0.9.6e. This prevents DoS (the
14530 assertions could call abort()).
14532 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14534 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14536 * Add various sanity checks to asn1_get_length() to reject
14537 the ASN1 length bytes if they exceed sizeof(long), will appear
14538 negative or the content length exceeds the length of the
14541 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14543 * Fix cipher selection routines: ciphers without encryption had no flags
14544 for the cipher strength set and where therefore not handled correctly
14545 by the selection routines (PR #130).
14549 * Fix EVP_dsa_sha macro.
14554 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14555 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14556 that was added in OpenSSL 0.9.6d.
14558 As the countermeasure turned out to be incompatible with some
14559 broken SSL implementations, the new option is part of SSL_OP_ALL.
14560 SSL_OP_ALL is usually employed when compatibility with weird SSL
14561 implementations is desired (e.g. '-bugs' option to 's_client' and
14562 's_server'), so the new option is automatically set in many
14567 * Changes in security patch:
14569 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14570 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14571 Air Force Materiel Command, USAF, under agreement number
14574 * Add various sanity checks to asn1_get_length() to reject
14575 the ASN1 length bytes if they exceed sizeof(long), will appear
14576 negative or the content length exceeds the length of the
14577 supplied buffer. ([CVE-2002-0659])
14579 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14581 * Assertions for various potential buffer overflows, not known to
14582 happen in practice.
14584 *Ben Laurie (CHATS)*
14586 * Various temporary buffers to hold ASCII versions of integers were
14587 too small for 64 bit platforms. ([CVE-2002-0655])
14588 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14590 * Remote buffer overflow in SSL3 protocol - an attacker could
14591 supply an oversized session ID to a client. ([CVE-2002-0656])
14593 *Ben Laurie (CHATS)*
14595 * Remote buffer overflow in SSL2 protocol - an attacker could
14596 supply an oversized client master key. ([CVE-2002-0656])
14598 *Ben Laurie (CHATS)*
14600 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14602 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14603 encoded as NULL) with id-dsa-with-sha1.
14605 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14607 * Check various `X509_...()` return values in `apps/req.c`.
14609 *Nils Larsch <nla@trustcenter.de>*
14611 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14612 an end-of-file condition would erroneously be flagged, when the CRLF
14613 was just at the end of a processed block. The bug was discovered when
14614 processing data through a buffering memory BIO handing the data to a
14615 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14616 <ptsekov@syntrex.com> and Nedelcho Stanev.
14620 * Implement a countermeasure against a vulnerability recently found
14621 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14622 before application data chunks to avoid the use of known IVs
14623 with data potentially chosen by the attacker.
14627 * Fix length checks in ssl3_get_client_hello().
14631 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14632 to prevent ssl3_read_internal() from incorrectly assuming that
14633 ssl3_read_bytes() found application data while handshake
14634 processing was enabled when in fact s->s3->in_read_app_data was
14635 merely automatically cleared during the initial handshake.
14637 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14639 * Fix object definitions for Private and Enterprise: they were not
14640 recognized in their shortname (=lowercase) representation. Extend
14641 obj_dat.pl to issue an error when using undefined keywords instead
14642 of silently ignoring the problem (Svenning Sorensen
14643 <sss@sss.dnsalias.net>).
14647 * Fix DH_generate_parameters() so that it works for 'non-standard'
14648 generators, i.e. generators other than 2 and 5. (Previously, the
14649 code did not properly initialise the 'add' and 'rem' values to
14650 BN_generate_prime().)
14652 In the new general case, we do not insist that 'generator' is
14653 actually a primitive root: This requirement is rather pointless;
14654 a generator of the order-q subgroup is just as good, if not
14659 * Map new X509 verification errors to alerts. Discovered and submitted by
14660 Tom Wu <tom@arcot.com>.
14664 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14665 returning non-zero before the data has been completely received
14666 when using non-blocking I/O.
14668 *Bodo Moeller; problem pointed out by John Hughes*
14670 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14672 *Ben Laurie, Lutz Jaenicke*
14674 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14675 Yoram Zahavi <YoramZ@gilian.com>).
14679 * Add information about CygWin 1.3 and on, and preserve proper
14680 configuration for the versions before that.
14682 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14684 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14685 check whether we deal with a copy of a session and do not delete from
14686 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14687 <izhar@checkpoint.com>.
14691 * Do not store session data into the internal session cache, if it
14692 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14693 flag is set). Proposed by Aslam <aslam@funk.com>.
14697 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14702 * [In 0.9.6d-engine release:]
14703 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14705 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14707 * Add the configuration target linux-s390x.
14709 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14711 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14712 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14713 variable as an indication that a ClientHello message has been
14714 received. As the flag value will be lost between multiple
14715 invocations of ssl3_accept when using non-blocking I/O, the
14716 function may not be aware that a handshake has actually taken
14717 place, thus preventing a new session from being added to the
14720 To avoid this problem, we now set s->new_session to 2 instead of
14721 using a local variable.
14723 *Lutz Jaenicke, Bodo Moeller*
14725 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14726 if the SSL_R_LENGTH_MISMATCH error is detected.
14728 *Geoff Thorpe, Bodo Moeller*
14730 * New 'shared_ldflag' column in Configure platform table.
14734 * Fix EVP_CIPHER_mode macro.
14736 *"Dan S. Camper" <dan@bti.net>*
14738 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14739 type, we must throw them away by setting rr->length to 0.
14741 *D P Chang <dpc@qualys.com>*
14743 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14745 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14746 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14747 worked incorrectly for those cases where range = `10..._2` and
14748 `3*range` is two bits longer than range.)
14752 * Only add signing time to PKCS7 structures if it is not already
14757 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14758 OBJ_ld_ce should be OBJ_id_ce.
14759 Also some ip-pda OIDs in crypto/objects/objects.txt were
14760 incorrect (cf. RFC 3039).
14762 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14764 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14765 returns early because it has nothing to do.
14767 *Andy Schneider <andy.schneider@bjss.co.uk>*
14769 * [In 0.9.6c-engine release:]
14770 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14772 *Andy Schneider <andy.schneider@bjss.co.uk>*
14774 * [In 0.9.6c-engine release:]
14775 Add support for Cryptographic Appliance's keyserver technology.
14776 (Use engine 'keyclient')
14778 *Cryptographic Appliances and Geoff Thorpe*
14780 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14781 is called via tools/c89.sh because arguments have to be
14782 rearranged (all '-L' options must appear before the first object
14785 *Richard Shapiro <rshapiro@abinitio.com>*
14787 * [In 0.9.6c-engine release:]
14788 Add support for Broadcom crypto accelerator cards, backported
14791 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14793 * [In 0.9.6c-engine release:]
14794 Add support for SureWare crypto accelerator cards from
14795 Baltimore Technologies. (Use engine 'sureware')
14797 *Baltimore Technologies and Mark Cox*
14799 * [In 0.9.6c-engine release:]
14800 Add support for crypto accelerator cards from Accelerated
14801 Encryption Processing, www.aep.ie. (Use engine 'aep')
14803 *AEP Inc. and Mark Cox*
14805 * Add a configuration entry for gcc on UnixWare.
14807 *Gary Benson <gbenson@redhat.com>*
14809 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14810 messages are stored in a single piece (fixed-length part and
14811 variable-length part combined) and fix various bugs found on the way.
14815 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14816 instead. BIO_gethostbyname() does not know what timeouts are
14817 appropriate, so entries would stay in cache even when they have
14819 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14821 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14822 faced with a pathologically small ClientHello fragment that does
14823 not contain client_version: Instead of aborting with an error,
14824 simply choose the highest available protocol version (i.e.,
14825 TLS 1.0 unless it is disabled). In practice, ClientHello
14826 messages are never sent like this, but this change gives us
14827 strictly correct behaviour at least for TLS.
14831 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14832 never resets s->method to s->ctx->method when called from within
14833 one of the SSL handshake functions.
14835 *Bodo Moeller; problem pointed out by Niko Baric*
14837 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14838 (sent using the client's version number) if client_version is
14839 smaller than the protocol version in use. Also change
14840 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14841 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14842 the client will at least see that alert.
14846 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14851 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14852 client receives HelloRequest while in a handshake.
14854 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14856 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14857 should end in 'break', not 'goto end' which circumvents various
14858 cleanups done in state SSL_ST_OK. But session related stuff
14859 must be disabled for SSL_ST_OK in the case that we just sent a
14862 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14863 before just sending a HelloRequest.
14865 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14867 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14868 reveal whether illegal block cipher padding was found or a MAC
14869 verification error occurred. (Neither SSLerr() codes nor alerts
14870 are directly visible to potential attackers, but the information
14871 may leak via logfiles.)
14873 Similar changes are not required for the SSL 2.0 implementation
14874 because the number of padding bytes is sent in clear for SSL 2.0,
14875 and the extra bytes are just ignored. However ssl/s2_pkt.c
14876 failed to verify that the purported number of padding bytes is in
14881 * Add OpenUNIX-8 support including shared libraries
14882 (Boyd Lynn Gerber <gerberb@zenez.com>).
14886 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14887 'wristwatch attack' using huge encoding parameters (cf.
14888 James H. Manger's CRYPTO 2001 paper). Note that the
14889 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14890 encoding parameters and hence was not vulnerable.
14894 * BN_sqr() bug fix.
14896 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14898 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14899 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14900 followed by modular reduction.
14902 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14904 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14905 equivalent based on BN_pseudo_rand() instead of BN_rand().
14909 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14910 This function was broken, as the check for a new client hello message
14911 to handle SGC did not allow these large messages.
14912 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14916 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14920 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14921 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14925 * Rework the configuration and shared library support for Tru64 Unix.
14926 The configuration part makes use of modern compiler features and
14927 still retains old compiler behavior for those that run older versions
14928 of the OS. The shared library support part includes a variant that
14929 uses the RPATH feature, and is available through the special
14930 configuration target "alpha-cc-rpath", which will never be selected
14933 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14935 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14936 with the same message size as in ssl3_get_certificate_request().
14937 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14938 messages might inadvertently be reject as too long.
14940 *Petr Lampa <lampa@fee.vutbr.cz>*
14942 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14946 * Modified SSL library such that the verify_callback that has been set
14947 specifically for an SSL object with SSL_set_verify() is actually being
14948 used. Before the change, a verify_callback set with this function was
14949 ignored and the verify_callback() set in the SSL_CTX at the time of
14950 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14951 to allow the necessary settings.
14955 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14956 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14957 done automatically (in contradiction to the requirements of the C
14958 standard). This made problems when used from OpenSSH.
14962 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14963 dh->length and always used
14965 BN_rand_range(priv_key, dh->p).
14967 BN_rand_range() is not necessary for Diffie-Hellman, and this
14968 specific range makes Diffie-Hellman unnecessarily inefficient if
14969 dh->length (recommended exponent length) is much smaller than the
14970 length of dh->p. We could use BN_rand_range() if the order of
14971 the subgroup was stored in the DH structure, but we only have
14976 BN_rand(priv_key, l, ...)
14978 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14985 RSA_eay_public_encrypt
14986 RSA_eay_private_decrypt
14987 RSA_eay_private_encrypt (signing)
14988 RSA_eay_public_decrypt (signature verification)
14990 (default implementations for RSA_public_encrypt,
14991 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14992 always reject numbers >= n.
14996 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14997 to synchronize access to 'locking_thread'. This is necessary on
14998 systems where access to 'locking_thread' (an 'unsigned long'
14999 variable) is not atomic.
15003 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15004 *before* setting the 'crypto_lock_rand' flag. The previous code had
15005 a race condition if 0 is a valid thread ID.
15007 *Travis Vitek <vitek@roguewave.com>*
15009 * Add support for shared libraries under Irix.
15011 *Albert Chin-A-Young <china@thewrittenword.com>*
15013 * Add configuration option to build on Linux on both big-endian and
15014 little-endian MIPS.
15016 *Ralf Baechle <ralf@uni-koblenz.de>*
15018 * Add the possibility to create shared libraries on HP-UX.
15022 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15024 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15025 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15026 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15027 PRNG state recovery was possible based on the output of
15028 one PRNG request appropriately sized to gain knowledge on
15029 'md' followed by enough consecutive 1-byte PRNG requests
15030 to traverse all of 'state'.
15032 1. When updating 'md_local' (the current thread's copy of 'md')
15033 during PRNG output generation, hash all of the previous
15034 'md_local' value, not just the half used for PRNG output.
15036 2. Make the number of bytes from 'state' included into the hash
15037 independent from the number of PRNG bytes requested.
15039 The first measure alone would be sufficient to avoid
15040 Markku-Juhani's attack. (Actually it had never occurred
15041 to me that the half of 'md_local' used for chaining was the
15042 half from which PRNG output bytes were taken -- I had always
15043 assumed that the secret half would be used.) The second
15044 measure makes sure that additional data from 'state' is never
15045 mixed into 'md_local' in small portions; this heuristically
15046 further strengthens the PRNG.
15050 * Fix crypto/bn/asm/mips3.s.
15054 * When only the key is given to "enc", the IV is undefined. Print out
15055 an error message in this case.
15059 * Handle special case when X509_NAME is empty in X509 printing routines.
15063 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15064 positive and less than q.
15068 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15069 used: it isn't thread safe and the add_lock_callback should handle
15072 *Paul Rose <Paul.Rose@bridge.com>*
15074 * Verify that incoming data obeys the block size in
15075 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15081 *Ulf Möller, Bodo Möller*
15083 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15084 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15085 when fixing the server behaviour for backwards-compatible 'client
15086 hello' messages. (Note that the attack is impractical against
15087 SSL 3.0 and TLS 1.0 anyway because length and version checking
15088 means that the probability of guessing a valid ciphertext is
15089 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15092 Before 0.9.5, the countermeasure (hide the error by generating a
15093 random 'decryption result') did not work properly because
15094 ERR_clear_error() was missing, meaning that SSL_get_error() would
15095 detect the supposedly ignored error.
15097 Both problems are now fixed.
15101 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15102 (previously it was 1024).
15106 * Fix for compatibility mode trust settings: ignore trust settings
15107 unless some valid trust or reject settings are present.
15111 * Fix for blowfish EVP: its a variable length cipher.
15115 * Fix various bugs related to DSA S/MIME verification. Handle missing
15116 parameters in DSA public key structures and return an error in the
15117 DSA routines if parameters are absent.
15121 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15122 in the current directory if neither $RANDFILE nor $HOME was set.
15123 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15124 caused some confusion to Windows users who haven't defined $HOME.
15125 Thus RAND_file_name() is changed again: e_os.h can define a
15126 DEFAULT_HOME, which will be used if $HOME is not set.
15127 For Windows, we use "C:"; on other platforms, we still require
15128 environment variables.
15130 * Move 'if (!initialized) RAND_poll()' into regions protected by
15131 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15132 having multiple threads call RAND_poll() concurrently.
15136 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15137 combination of a flag and a thread ID variable.
15138 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15139 flag), *other* threads can enter ssleay_add_bytes without obeying
15140 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15141 that they do not hold after the first thread unsets add_do_not_lock).
15145 * Change bctest again: '-x' expressions are not available in all
15146 versions of 'test'.
15150 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15152 * Fix a couple of memory leaks in PKCS7_dataDecode()
15154 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15156 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15157 the default extension for executables, if any. Also, make the perl
15158 scripts that use symlink() to test if it really exists and use "cp"
15159 if it doesn't. All this made OpenSSL compilable and installable in
15164 * Fix for asn1_GetSequence() for indefinite length constructed data.
15165 If SEQUENCE is length is indefinite just set c->slen to the total
15166 amount of data available.
15168 *Steve Henson, reported by shige@FreeBSD.org*
15170 *This change does not apply to 0.9.7.*
15172 * Change bctest to avoid here-documents inside command substitution
15173 (workaround for FreeBSD /bin/sh bug).
15174 For compatibility with Ultrix, avoid shell functions (introduced
15175 in the bctest version that searches along $PATH).
15179 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15180 with des_encrypt() defined on some operating systems, like Solaris
15185 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15186 On the Importance of Eliminating Errors in Cryptographic
15187 Computations, J. Cryptology 14 (2001) 2, 101-119,
15188 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15192 * MIPS assembler BIGNUM division bug fix.
15196 * Disabled incorrect Alpha assembler code.
15200 * Fix PKCS#7 decode routines so they correctly update the length
15201 after reading an EOC for the EXPLICIT tag.
15205 *This change does not apply to 0.9.7.*
15207 * Fix bug in PKCS#12 key generation routines. This was triggered
15208 if a 3DES key was generated with a 0 initial byte. Include
15209 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15210 (but broken) behaviour.
15214 * Enhance bctest to search for a working bc along $PATH and print
15217 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15219 * Fix memory leaks in err.c: free err_data string if necessary;
15220 don't write to the wrong index in ERR_set_error_data.
15224 * Implement ssl23_peek (analogous to ssl23_read), which previously
15229 * Replace rdtsc with `_emit` statements for VC++ version 5.
15231 *Jeremy Cooper <jeremy@baymoo.org>*
15233 * Make it possible to reuse SSLv2 sessions.
15237 * In copy_email() check for >= 0 as a return value for
15238 X509_NAME_get_index_by_NID() since 0 is a valid index.
15240 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15242 * Avoid coredump with unsupported or invalid public keys by checking if
15243 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15244 PKCS7_verify() fails with non detached data.
15248 * Don't use getenv in library functions when run as setuid/setgid.
15249 New function OPENSSL_issetugid().
15253 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15254 due to incorrect handling of multi-threading:
15256 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15258 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15260 3. Count how many times MemCheck_off() has been called so that
15261 nested use can be treated correctly. This also avoids
15262 inband-signalling in the previous code (which relied on the
15263 assumption that thread ID 0 is impossible).
15267 * Add "-rand" option also to s_client and s_server.
15271 * Fix CPU detection on Irix 6.x.
15272 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15273 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15275 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15280 *This change does not apply to 0.9.7.*
15282 * Use the cached encoding of an X509_NAME structure rather than
15283 copying it. This is apparently the reason for the libsafe "errors"
15284 but the code is actually correct.
15288 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15289 Bleichenbacher's DSA attack.
15290 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15291 to be set and top=0 forces the highest bit to be set; top=-1 is new
15292 and leaves the highest bit random.
15294 *Ulf Moeller, Bodo Moeller*
15296 * In the `NCONF_...`-based implementations for `CONF_...` queries
15297 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15298 a temporary CONF structure with the data component set to NULL
15299 (which gives segmentation faults in lh_retrieve).
15300 Instead, use NULL for the CONF pointer in CONF_get_string and
15301 CONF_get_number (which may use environment variables) and directly
15302 return NULL from CONF_get_section.
15306 * Fix potential buffer overrun for EBCDIC.
15310 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15311 keyUsage if basicConstraints absent for a CA.
15315 * Make SMIME_write_PKCS7() write mail header values with a format that
15316 is more generally accepted (no spaces before the semicolon), since
15317 some programs can't parse those values properly otherwise. Also make
15318 sure BIO's that break lines after each write do not create invalid
15323 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15324 macros previously used would not encode an empty SEQUENCE OF
15325 and break the signature.
15329 *This change does not apply to 0.9.7.*
15331 * Zero the premaster secret after deriving the master secret in
15336 * Add some EVP_add_digest_alias registrations (as found in
15337 OpenSSL_add_all_digests()) to SSL_library_init()
15338 aka OpenSSL_add_ssl_algorithms(). This provides improved
15339 compatibility with peers using X.509 certificates
15340 with unconventional AlgorithmIdentifier OIDs.
15344 * Fix for Irix with NO_ASM.
15346 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15348 * ./config script fixes.
15350 *Ulf Moeller, Richard Levitte*
15352 * Fix 'openssl passwd -1'.
15356 * Change PKCS12_key_gen_asc() so it can cope with non null
15357 terminated strings whose length is passed in the passlen
15358 parameter, for example from PEM callbacks. This was done
15359 by adding an extra length parameter to asc2uni().
15361 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15363 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15364 call failed, free the DSA structure.
15368 * Fix to uni2asc() to cope with zero length Unicode strings.
15369 These are present in some PKCS#12 files.
15373 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15374 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15375 when writing a 32767 byte record.
15377 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15379 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15380 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15382 (RSA objects have a reference count access to which is protected
15383 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15384 so they are meant to be shared between threads.)
15385 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15386 "Reddie, Steven" <Steven.Reddie@ca.com>*
15388 * Fix a deadlock in CRYPTO_mem_leaks().
15392 * Use better test patterns in bntest.
15396 * rand_win.c fix for Borland C.
15400 * BN_rshift bugfix for n == 0.
15404 * Add a 'bctest' script that checks for some known 'bc' bugs
15405 so that 'make test' does not abort just because 'bc' is broken.
15409 * Store verify_result within SSL_SESSION also for client side to
15410 avoid potential security hole. (Re-used sessions on the client side
15411 always resulted in verify_result==X509_V_OK, not using the original
15412 result of the server certificate verification.)
15416 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15417 SSL3_RT_APPLICATION_DATA, return 0.
15418 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15423 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15424 releases, have been re-implemented by renaming the previous
15425 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15426 and ssl3_read_internal, respectively, and adding 'peek' parameters
15427 to them. The new ssl[23]_{read,peek} functions are calls to
15428 ssl[23]_read_internal with the 'peek' flag set appropriately.
15429 A 'peek' parameter has also been added to ssl3_read_bytes, which
15430 does the actual work for ssl3_read_internal.
15434 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15435 the method-specific "init()" handler. Also clean up ex_data after
15436 calling the method-specific "finish()" handler. Previously, this was
15437 happening the other way round.
15441 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15442 The previous value, 12, was not always sufficient for BN_mod_exp().
15446 * Make sure that shared libraries get the internal name engine with
15447 the full version number and not just 0. This should mark the
15448 shared libraries as not backward compatible. Of course, this should
15449 be changed again when we can guarantee backward binary compatibility.
15453 * Fix typo in get_cert_by_subject() in by_dir.c
15455 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15457 * Rework the system to generate shared libraries:
15459 - Make note of the expected extension for the shared libraries and
15460 if there is a need for symbolic links from for example libcrypto.so.0
15461 to libcrypto.so.0.9.7. There is extended info in Configure for
15464 - Make as few rebuilds of the shared libraries as possible.
15466 - Still avoid linking the OpenSSL programs with the shared libraries.
15468 - When installing, install the shared libraries separately from the
15473 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15475 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15476 and not in SSL_clear because the latter is also used by the
15477 accept/connect functions; previously, the settings made by
15478 SSL_set_read_ahead would be lost during the handshake.
15480 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15482 * Correct util/mkdef.pl to be selective about disabled algorithms.
15483 Previously, it would create entries for disabled algorithms no
15488 * Added several new manual pages for SSL_* function.
15492 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15494 * In ssl23_get_client_hello, generate an error message when faced
15495 with an initial SSL 3.0/TLS record that is too small to contain the
15496 first two bytes of the ClientHello message, i.e. client_version.
15497 (Note that this is a pathologic case that probably has never happened
15498 in real life.) The previous approach was to use the version number
15499 from the record header as a substitute; but our protocol choice
15500 should not depend on that one because it is not authenticated
15501 by the Finished messages.
15505 * More robust randomness gathering functions for Windows.
15507 *Jeffrey Altman <jaltman@columbia.edu>*
15509 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15510 not set then we don't setup the error code for issuer check errors
15511 to avoid possibly overwriting other errors which the callback does
15512 handle. If an application does set the flag then we assume it knows
15513 what it is doing and can handle the new informational codes
15518 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15519 a general "ANY" type, as such it should be able to decode anything
15520 including tagged types. However it didn't check the class so it would
15521 wrongly interpret tagged types in the same way as their universal
15522 counterpart and unknown types were just rejected. Changed so that the
15523 tagged and unknown types are handled in the same way as a SEQUENCE:
15524 that is the encoding is stored intact. There is also a new type
15525 "V_ASN1_OTHER" which is used when the class is not universal, in this
15526 case we have no idea what the actual type is so we just lump them all
15531 * On VMS, stdout may very well lead to a file that is written to
15532 in a record-oriented fashion. That means that every write() will
15533 write a separate record, which will be read separately by the
15534 programs trying to read from it. This can be very confusing.
15536 The solution is to put a BIO filter in the way that will buffer
15537 text until a linefeed is reached, and then write everything a
15538 line at a time, so every record written will be an actual line,
15539 not chunks of lines and not (usually doesn't happen, but I've
15540 seen it once) several lines in one record. BIO_f_linebuffer() is
15543 Currently, it's a VMS-only method, because that's where it has
15544 been tested well enough.
15548 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15549 it can return incorrect results.
15550 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15551 but it was in 0.9.6-beta[12].)
15555 * Disable the check for content being present when verifying detached
15556 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15557 include zero length content when signing messages.
15561 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15562 BIO_ctrl (for BIO pairs).
15566 * Add DSO method for VMS.
15570 * Bug fix: Montgomery multiplication could produce results with the
15575 * Add RPM specification openssl.spec and modify it to build three
15576 packages. The default package contains applications, application
15577 documentation and run-time libraries. The devel package contains
15578 include files, static libraries and function documentation. The
15579 doc package contains the contents of the doc directory. The original
15580 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15584 * Add a large number of documentation files for many SSL routines.
15586 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15588 * Add a configuration entry for Sony News 4.
15590 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15592 * Don't set the two most significant bits to one when generating a
15593 random number < q in the DSA library.
15597 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15598 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15599 the underlying transport is blocking) if a handshake took place.
15600 (The default behaviour is needed by applications such as s_client
15601 and s_server that use select() to determine when to use SSL_read;
15602 but for applications that know in advance when to expect data, it
15603 just makes things more complicated.)
15607 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15612 * Add a few more EBCDIC conditionals that make `req` and `x509`
15613 work better on such systems.
15615 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15617 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15618 Update PKCS12_parse() so it copies the friendlyName and the
15619 keyid to the certificates aux info.
15623 * Fix bug in PKCS7_verify() which caused an infinite loop
15624 if there was more than one signature.
15626 *Sven Uszpelkat <su@celocom.de>*
15628 * Major change in util/mkdef.pl to include extra information
15629 about each symbol, as well as presenting variables as well
15630 as functions. This change means that there's n more need
15631 to rebuild the .num files when some algorithms are excluded.
15635 * Allow the verify time to be set by an application,
15636 rather than always using the current time.
15640 * Phase 2 verify code reorganisation. The certificate
15641 verify code now looks up an issuer certificate by a
15642 number of criteria: subject name, authority key id
15643 and key usage. It also verifies self signed certificates
15644 by the same criteria. The main comparison function is
15645 X509_check_issued() which performs these checks.
15647 Lot of changes were necessary in order to support this
15648 without completely rewriting the lookup code.
15650 Authority and subject key identifier are now cached.
15652 The LHASH 'certs' is X509_STORE has now been replaced
15653 by a STACK_OF(X509_OBJECT). This is mainly because an
15654 LHASH can't store or retrieve multiple objects with
15655 the same hash value.
15657 As a result various functions (which were all internal
15658 use only) have changed to handle the new X509_STORE
15659 structure. This will break anything that messed round
15660 with X509_STORE internally.
15662 The functions X509_STORE_add_cert() now checks for an
15663 exact match, rather than just subject name.
15665 The X509_STORE API doesn't directly support the retrieval
15666 of multiple certificates matching a given criteria, however
15667 this can be worked round by performing a lookup first
15668 (which will fill the cache with candidate certificates)
15669 and then examining the cache for matches. This is probably
15670 the best we can do without throwing out X509_LOOKUP
15671 entirely (maybe later...).
15673 The X509_VERIFY_CTX structure has been enhanced considerably.
15675 All certificate lookup operations now go via a get_issuer()
15676 callback. Although this currently uses an X509_STORE it
15677 can be replaced by custom lookups. This is a simple way
15678 to bypass the X509_STORE hackery necessary to make this
15679 work and makes it possible to use more efficient techniques
15680 in future. A very simple version which uses a simple
15681 STACK for its trusted certificate store is also provided
15682 using X509_STORE_CTX_trusted_stack().
15684 The verify_cb() and verify() callbacks now have equivalents
15685 in the X509_STORE_CTX structure.
15687 X509_STORE_CTX also has a 'flags' field which can be used
15688 to customise the verify behaviour.
15692 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15693 excludes S/MIME capabilities.
15697 * When a certificate request is read in keep a copy of the
15698 original encoding of the signed data and use it when outputting
15699 again. Signatures then use the original encoding rather than
15700 a decoded, encoded version which may cause problems if the
15701 request is improperly encoded.
15705 * For consistency with other BIO_puts implementations, call
15706 buffer_write(b, ...) directly in buffer_puts instead of calling
15709 In BIO_puts, increment b->num_write as in BIO_write.
15711 *Peter.Sylvester@EdelWeb.fr*
15713 * Fix BN_mul_word for the case where the word is 0. (We have to use
15714 BN_zero, we may not return a BIGNUM with an array consisting of
15715 words set to zero.)
15719 * Avoid calling abort() from within the library when problems are
15720 detected, except if preprocessor symbols have been defined
15721 (such as REF_CHECK, BN_DEBUG etc.).
15725 * New openssl application 'rsautl'. This utility can be
15726 used for low-level RSA operations. DER public key
15727 BIO/fp routines also added.
15731 * New Configure entry and patches for compiling on QNX 4.
15733 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15735 * A demo state-machine implementation was sponsored by
15736 Nuron (<http://www.nuron.com/>) and is now available in
15737 demos/state_machine.
15741 * New options added to the 'dgst' utility for signature
15742 generation and verification.
15746 * Unrecognized PKCS#7 content types are now handled via a
15747 catch all ASN1_TYPE structure. This allows unsupported
15748 types to be stored as a "blob" and an application can
15749 encode and decode it manually.
15753 * Fix various signed/unsigned issues to make a_strex.c
15754 compile under VC++.
15756 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15758 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15759 length if passed a buffer. ASN1_INTEGER_to_BN failed
15760 if passed a NULL BN and its argument was negative.
15762 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15764 * Modification to PKCS#7 encoding routines to output definite
15765 length encoding. Since currently the whole structures are in
15766 memory there's not real point in using indefinite length
15767 constructed encoding. However if OpenSSL is compiled with
15768 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15772 * Added BIO_vprintf() and BIO_vsnprintf().
15776 * Added more prefixes to parse for in the strings written
15777 through a logging bio, to cover all the levels that are available
15778 through syslog. The prefixes are now:
15780 PANIC, EMERG, EMR => LOG_EMERG
15781 ALERT, ALR => LOG_ALERT
15782 CRIT, CRI => LOG_CRIT
15783 ERROR, ERR => LOG_ERR
15784 WARNING, WARN, WAR => LOG_WARNING
15785 NOTICE, NOTE, NOT => LOG_NOTICE
15786 INFO, INF => LOG_INFO
15787 DEBUG, DBG => LOG_DEBUG
15789 and as before, if none of those prefixes are present at the
15790 beginning of the string, LOG_ERR is chosen.
15792 On Win32, the `LOG_*` levels are mapped according to this:
15794 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15795 LOG_WARNING => EVENTLOG_WARNING_TYPE
15796 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15800 * Made it possible to reconfigure with just the configuration
15801 argument "reconf" or "reconfigure". The command line arguments
15802 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15803 and are retrieved from there when reconfiguring.
15809 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15811 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15815 * The obj_dat.pl script was messing up the sorting of object
15816 names. The reason was that it compared the quoted version
15817 of strings as a result "OCSP" > "OCSP Signing" because
15818 " > SPACE. Changed script to store unquoted versions of
15819 names and add quotes on output. It was also omitting some
15820 names from the lookup table if they were given a default
15821 value (that is if SN is missing it is given the same
15822 value as LN and vice versa), these are now added on the
15823 grounds that if an object has a name we should be able to
15824 look it up. Finally added warning output when duplicate
15825 short or long names are found.
15829 * Changes needed for Tandem NSK.
15831 *Scott Uroff <scott@xypro.com>*
15833 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15834 RSA_padding_check_SSLv23(), special padding was never detected
15835 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15836 version rollback attacks was not effective.
15838 In s23_clnt.c, don't use special rollback-attack detection padding
15839 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15840 client; similarly, in s23_srvr.c, don't do the rollback check if
15841 SSL 2.0 is the only protocol enabled in the server.
15845 * Make it possible to get hexdumps of unprintable data with 'openssl
15846 asn1parse'. By implication, the functions ASN1_parse_dump() and
15847 BIO_dump_indent() are added.
15851 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15852 these print out strings and name structures based on various
15853 flags including RFC2253 support and proper handling of
15854 multibyte characters. Added options to the 'x509' utility
15855 to allow the various flags to be set.
15859 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15860 Also change the functions X509_cmp_current_time() and
15861 X509_gmtime_adj() work with an ASN1_TIME structure,
15862 this will enable certificates using GeneralizedTime in validity
15863 dates to be checked.
15867 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15868 negative public key encodings) on by default,
15869 NO_NEG_PUBKEY_BUG can be set to disable it.
15873 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15874 content octets. An i2c_ASN1_OBJECT is unnecessary because
15875 the encoding can be trivially obtained from the structure.
15879 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15880 not read locks (`CRYPTO_r_[un]lock`).
15884 * A first attempt at creating official support for shared
15885 libraries through configuration. I've kept it so the
15886 default is static libraries only, and the OpenSSL programs
15887 are always statically linked for now, but there are
15888 preparations for dynamic linking in place.
15889 This has been tested on Linux and Tru64.
15893 * Randomness polling function for Win9x, as described in:
15894 Peter Gutmann, Software Generation of Practically Strong
15899 * Fix so PRNG is seeded in req if using an already existing
15904 * New options to smime application. -inform and -outform
15905 allow alternative formats for the S/MIME message including
15906 PEM and DER. The -content option allows the content to be
15907 specified separately. This should allow things like Netscape
15908 form signing output easier to verify.
15912 * Fix the ASN1 encoding of tags using the 'long form'.
15916 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15917 STRING types. These convert content octets to and from the
15918 underlying type. The actual tag and length octets are
15919 already assumed to have been read in and checked. These
15920 are needed because all other string types have virtually
15921 identical handling apart from the tag. By having versions
15922 of the ASN1 functions that just operate on content octets
15923 IMPLICIT tagging can be handled properly. It also allows
15924 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15925 and ASN1_INTEGER are identical apart from the tag.
15929 * Change the handling of OID objects as follows:
15931 - New object identifiers are inserted in objects.txt, following
15932 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15933 - objects.pl is used to process obj_mac.num and create a new
15935 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15938 This is currently kind of a hack, and the perl code in objects.pl
15939 isn't very elegant, but it works as I intended. The simplest way
15940 to check that it worked correctly is to look in obj_dat.h and
15941 check the array nid_objs and make sure the objects haven't moved
15942 around (this is important!). Additions are OK, as well as
15943 consistent name changes.
15947 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15951 * Addition of the command line parameter '-rand file' to 'openssl req'.
15952 The given file adds to whatever has already been seeded into the
15953 random pool through the RANDFILE configuration file option or
15954 environment variable, or the default random state file.
15958 * mkstack.pl now sorts each macro group into lexical order.
15959 Previously the output order depended on the order the files
15960 appeared in the directory, resulting in needless rewriting
15965 * Patches to make OpenSSL compile under Win32 again. Mostly
15966 work arounds for the VC++ problem that it treats func() as
15967 func(void). Also stripped out the parts of mkdef.pl that
15968 added extra typesafe functions: these no longer exist.
15972 * Reorganisation of the stack code. The macros are now all
15973 collected in safestack.h . Each macro is defined in terms of
15974 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15975 DEBUG_SAFESTACK is now handled in terms of function casts,
15976 this has the advantage of retaining type safety without the
15977 use of additional functions. If DEBUG_SAFESTACK is not defined
15978 then the non typesafe macros are used instead. Also modified the
15979 mkstack.pl script to handle the new form. Needs testing to see
15980 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15981 the default if no major problems. Similar behaviour for ASN1_SET_OF
15982 and PKCS12_STACK_OF.
15986 * When some versions of IIS use the 'NET' form of private key the
15987 key derivation algorithm is different. Normally MD5(password) is
15988 used as a 128 bit RC4 key. In the modified case
15989 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15990 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15991 as the old Netscape_RSA functions except they have an additional
15992 'sgckey' parameter which uses the modified algorithm. Also added
15993 an -sgckey command line option to the rsa utility. Thanks to
15994 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15995 algorithm to openssl-dev.
15999 * The evp_local.h macros were using 'c.##kname' which resulted in
16000 invalid expansion on some systems (SCO 5.0.5 for example).
16001 Corrected to 'c.kname'.
16003 *Phillip Porch <root@theporch.com>*
16005 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16006 a STACK of email addresses from a certificate or request, these look
16007 in the subject name and the subject alternative name extensions and
16008 omit any duplicate addresses.
16012 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16013 This makes DSA verification about 2 % faster.
16017 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16018 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16019 plus overhead for 1024 bit moduli).
16020 This makes exponentiations about 0.5 % faster for 1024 bit
16021 exponents (as measured by "openssl speed rsa2048").
16025 * Rename memory handling macros to avoid conflicts with other
16027 Malloc => OPENSSL_malloc
16028 Malloc_locked => OPENSSL_malloc_locked
16029 Realloc => OPENSSL_realloc
16030 Free => OPENSSL_free
16034 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16035 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16039 * CygWin32 support.
16041 *John Jarvie <jjarvie@newsguy.com>*
16043 * The type-safe stack code has been rejigged. It is now only compiled
16044 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16045 by default all type-specific stack functions are "#define"d back to
16046 standard stack functions. This results in more streamlined output
16047 but retains the type-safety checking possibilities of the original
16052 * The STACK code has been cleaned up, and certain type declarations
16053 that didn't make a lot of sense have been brought in line. This has
16054 also involved a cleanup of sorts in safestack.h to more correctly
16055 map type-safe stack functions onto their plain stack counterparts.
16056 This work has also resulted in a variety of "const"ifications of
16057 lots of the code, especially `_cmp` operations which should normally
16058 be prototyped with "const" parameters anyway.
16062 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16063 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16064 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16065 where all of 'md' is used each time the PRNG is used, but 'state'
16066 is used only indexed by a cyclic counter. As entropy may not be
16067 well distributed from the beginning, 'md' is important as a
16068 chaining variable. However, the output function chains only half
16069 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16070 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16071 in all of 'state' being rewritten, with the new values depending
16072 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16076 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16077 the handshake is continued after ssl_verify_cert_chain();
16078 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16079 can lead to 'unexplainable' connection aborts later.
16081 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16083 * Major EVP API cipher revision.
16084 Add hooks for extra EVP features. This allows various cipher
16085 parameters to be set in the EVP interface. Support added for variable
16086 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16087 setting of RC2 and RC5 parameters.
16089 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16092 Remove lots of duplicated code from the EVP library. For example *every*
16093 cipher init() function handles the 'iv' in the same way according to the
16094 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16095 for CFB and OFB modes they zero ctx->num.
16097 New functionality allows removal of S/MIME code RC2 hack.
16099 Most of the routines have the same form and so can be declared in terms
16102 By shifting this to the top level EVP_CipherInit() it can be removed from
16103 all individual ciphers. If the cipher wants to handle IVs or keys
16104 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16107 Change lots of functions like EVP_EncryptUpdate() to now return a
16108 value: although software versions of the algorithms cannot fail
16109 any installed hardware versions can.
16113 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16114 this option is set, tolerate broken clients that send the negotiated
16115 protocol version number instead of the requested protocol version
16120 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16121 i.e. non-zero for export ciphersuites, zero otherwise.
16122 Previous versions had this flag inverted, inconsistent with
16123 rsa_tmp_cb (..._TMP_RSA_CB).
16125 *Bodo Moeller; problem reported by Amit Chopra*
16127 * Add missing DSA library text string. Work around for some IIS
16128 key files with invalid SEQUENCE encoding.
16132 * Add a document (doc/standards.txt) that list all kinds of standards
16133 and so on that are implemented in OpenSSL.
16137 * Enhance c_rehash script. Old version would mishandle certificates
16138 with the same subject name hash and wouldn't handle CRLs at all.
16139 Added -fingerprint option to crl utility, to support new c_rehash
16144 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16148 * Fix for SSL server purpose checking. Server checking was
16149 rejecting certificates which had extended key usage present
16150 but no ssl client purpose.
16152 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16154 * Make PKCS#12 code work with no password. The PKCS#12 spec
16155 is a little unclear about how a blank password is handled.
16156 Since the password in encoded as a BMPString with terminating
16157 double NULL a zero length password would end up as just the
16158 double NULL. However no password at all is different and is
16159 handled differently in the PKCS#12 key generation code. NS
16160 treats a blank password as zero length. MSIE treats it as no
16161 password on export: but it will try both on import. We now do
16162 the same: PKCS12_parse() tries zero length and no password if
16163 the password is set to "" or NULL (NULL is now a valid password:
16164 it wasn't before) as does the pkcs12 application.
16168 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16169 perror when PEM_read_bio_X509_REQ fails, the error message must
16170 be obtained from the error queue.
16174 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16175 it in ERR_remove_state if appropriate, and change ERR_get_state
16176 accordingly to avoid race conditions (this is necessary because
16177 thread_hash is no longer constant once set).
16181 * Bugfix for linux-elf makefile.one.
16185 * RSA_get_default_method() will now cause a default
16186 RSA_METHOD to be chosen if one doesn't exist already.
16187 Previously this was only set during a call to RSA_new()
16188 or RSA_new_method(NULL) meaning it was possible for
16189 RSA_get_default_method() to return NULL.
16193 * Added native name translation to the existing DSO code
16194 that will convert (if the flag to do so is set) filenames
16195 that are sufficiently small and have no path information
16196 into a canonical native form. Eg. "blah" converted to
16197 "libblah.so" or "blah.dll" etc.
16201 * New function ERR_error_string_n(e, buf, len) which is like
16202 ERR_error_string(e, buf), but writes at most 'len' bytes
16203 including the 0 terminator. For ERR_error_string_n, 'buf'
16206 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16208 * CONF library reworked to become more general. A new CONF
16209 configuration file reader "class" is implemented as well as a
16210 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16211 old `CONF_*` functions are still there, but are reimplemented to
16212 work in terms of the new functions. Also, a set of functions
16213 to handle the internal storage of the configuration data is
16214 provided to make it easier to write new configuration file
16215 reader "classes" (I can definitely see something reading a
16216 configuration file in XML format, for example), called `_CONF_*`,
16217 or "the configuration storage API"...
16219 The new configuration file reading functions are:
16221 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16222 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16224 NCONF_default, NCONF_WIN32
16226 NCONF_dump_fp, NCONF_dump_bio
16228 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16229 NCONF_new creates a new CONF object. This works in the same way
16230 as other interfaces in OpenSSL, like the BIO interface.
16231 `NCONF_dump_*` dump the internal storage of the configuration file,
16232 which is useful for debugging. All other functions take the same
16233 arguments as the old `CONF_*` functions with the exception of the
16234 first that must be a `CONF *` instead of a `LHASH *`.
16236 To make it easier to use the new classes with the old `CONF_*` functions,
16237 the function CONF_set_default_method is provided.
16241 * Add '-tls1' option to 'openssl ciphers', which was already
16242 mentioned in the documentation but had not been implemented.
16243 (This option is not yet really useful because even the additional
16244 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16248 * Initial DSO code added into libcrypto for letting OpenSSL (and
16249 OpenSSL-based applications) load shared libraries and bind to
16250 them in a portable way.
16252 *Geoff Thorpe, with contributions from Richard Levitte*
16254 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16256 * Make sure _lrotl and _lrotr are only used with MSVC.
16258 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16259 (the default implementation of RAND_status).
16261 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16262 to '-clrext' (= clear extensions), as intended and documented.
16263 *Bodo Moeller; inconsistency pointed out by Michael Attili
16264 <attili@amaxo.com>*
16266 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16267 was larger than the MD block size.
16269 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16271 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16272 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16273 using the passed key: if the passed key was a private key the result
16274 of X509_print(), for example, would be to print out all the private key
16279 * des_quad_cksum() byte order bug fix.
16280 *Ulf Möller, using the problem description in krb4-0.9.7, where
16281 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16283 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16286 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16288 * For easily testing in shell scripts whether some command
16289 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16290 returns with exit code 0 iff no command of the given name is available.
16291 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16292 the output goes to stdout and nothing is printed to stderr.
16293 Additional arguments are always ignored.
16295 Since for each cipher there is a command of the same name,
16296 the 'no-cipher' compilation switches can be tested this way.
16298 ('openssl no-XXX' is not able to detect pseudo-commands such
16299 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16303 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16307 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16308 is set; it will be thrown away anyway because each handshake creates
16310 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16311 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16312 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16313 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16317 * New s_client option -ign_eof: EOF at stdin is ignored, and
16318 'Q' and 'R' lose their special meanings (quit/renegotiate).
16319 This is part of what -quiet does; unlike -quiet, -ign_eof
16320 does not suppress any output.
16324 * Add compatibility options to the purpose and trust code. The
16325 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16326 accepts a certificate or CA, this was the previous behaviour,
16327 with all the associated security issues.
16329 X509_TRUST_COMPAT is the old trust behaviour: only and
16330 automatically trust self signed roots in certificate store. A
16331 new trust setting X509_TRUST_DEFAULT is used to specify that
16332 a purpose has no associated trust setting and it should instead
16333 use the value in the default purpose.
16337 * Fix the PKCS#8 DSA private key code so it decodes keys again
16338 and fix a memory leak.
16342 * In util/mkerr.pl (which implements 'make errors'), preserve
16343 reason strings from the previous version of the .c file, as
16344 the default to have only downcase letters (and digits) in
16345 automatically generated reasons codes is not always appropriate.
16349 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16350 using strerror. Previously, ERR_reason_error_string() returned
16351 library names as reason strings for SYSerr; but SYSerr is a special
16352 case where small numbers are errno values, not library numbers.
16356 * Add '-dsaparam' option to 'openssl dhparam' application. This
16357 converts DSA parameters into DH parameters. (When creating parameters,
16358 DSA_generate_parameters is used.)
16362 * Include 'length' (recommended exponent length) in C code generated
16363 by 'openssl dhparam -C'.
16367 * The second argument to set_label in perlasm was already being used
16368 so couldn't be used as a "file scope" flag. Moved to third argument
16373 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16374 instead of RAND_bytes for encryption IVs and salts.
16378 * Include RAND_status() into RAND_METHOD instead of implementing
16379 it only for md_rand.c Otherwise replacing the PRNG by calling
16380 RAND_set_rand_method would be impossible.
16384 * Don't let DSA_generate_key() enter an infinite loop if the random
16385 number generation fails.
16389 * New 'rand' application for creating pseudo-random output.
16393 * Added configuration support for Linux/IA64
16395 *Rolf Haberrecker <rolf@suse.de>*
16397 * Assembler module support for Mingw32.
16401 * Shared library support for HPUX (in shlib/).
16403 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16405 * Shared library support for Solaris gcc.
16407 *Lutz Behnke <behnke@trustcenter.de>*
16409 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16411 * PKCS7_encrypt() was adding text MIME headers twice because they
16412 were added manually and by SMIME_crlf_copy().
16416 * In bntest.c don't call BN_rand with zero bits argument.
16418 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16420 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16421 case was implemented. This caused BN_div_recp() to fail occasionally.
16425 * Add an optional second argument to the set_label() in the perl
16426 assembly language builder. If this argument exists and is set
16427 to 1 it signals that the assembler should use a symbol whose
16428 scope is the entire file, not just the current function. This
16429 is needed with MASM which uses the format label:: for this scope.
16431 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16433 * Change the ASN1 types so they are typedefs by default. Before
16434 almost all types were #define'd to ASN1_STRING which was causing
16435 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16440 * Change names of new functions to the new get1/get0 naming
16441 convention: After 'get1', the caller owns a reference count
16442 and has to call `..._free`; 'get0' returns a pointer to some
16443 data structure without incrementing reference counters.
16444 (Some of the existing 'get' functions increment a reference
16445 counter, some don't.)
16446 Similarly, 'set1' and 'add1' functions increase reference
16447 counters or duplicate objects.
16451 * Allow for the possibility of temp RSA key generation failure:
16452 the code used to assume it always worked and crashed on failure.
16456 * Fix potential buffer overrun problem in BIO_printf().
16457 *Ulf Möller, using public domain code by Patrick Powell; problem
16458 pointed out by David Sacerdote <das33@cornell.edu>*
16460 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16461 RAND_egd() and RAND_status(). In the command line application,
16462 the EGD socket can be specified like a seed file using RANDFILE
16467 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16468 Some CAs (e.g. Verisign) distribute certificates in this form.
16472 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16473 list to exclude them. This means that no special compilation option
16474 is needed to use anonymous DH: it just needs to be included in the
16479 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16480 EVP_MD_type. The old functionality is available in a new macro called
16481 EVP_MD_md(). Change code that uses it and update docs.
16485 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16486 where the `void *` argument is replaced by a function pointer argument.
16487 Previously `void *` was abused to point to functions, which works on
16488 many platforms, but is not correct. As these functions are usually
16489 called by macros defined in OpenSSL header files, most source code
16490 should work without changes.
16494 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16495 sections with information on -D... compiler switches used for
16496 compiling the library so that applications can see them. To enable
16497 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16498 must be defined. E.g.,
16499 #define OPENSSL_ALGORITHM_DEFINES
16500 #include <openssl/opensslconf.h>
16501 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16503 *Richard Levitte, Ulf and Bodo Möller*
16505 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16510 * Change the 'other' type in certificate aux info to a STACK_OF
16511 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16512 the required ASN1 format: arbitrary types determined by an OID.
16516 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16517 argument to 'req'. This is not because the function is newer or
16518 better than others it just uses the work 'NEW' in the certificate
16519 request header lines. Some software needs this.
16523 * Reorganise password command line arguments: now passwords can be
16524 obtained from various sources. Delete the PEM_cb function and make
16525 it the default behaviour: i.e. if the callback is NULL and the
16526 usrdata argument is not NULL interpret it as a null terminated pass
16527 phrase. If usrdata and the callback are NULL then the pass phrase
16528 is prompted for as usual.
16532 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16533 the support is automatically enabled. The resulting binaries will
16534 autodetect the card and use it if present.
16536 *Ben Laurie and Compaq Inc.*
16538 * Work around for Netscape hang bug. This sends certificate request
16539 and server done in one record. Since this is perfectly legal in the
16540 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16541 the bugs/SSLv3 entry for more info.
16545 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16549 * Add -rand argument to smime and pkcs12 applications and read/write
16554 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16558 * Add command line password options to the remaining applications.
16562 * Bug fix for BN_div_recp() for numerators with an even number of
16567 * More tests in bntest.c, and changed test_bn output.
16571 * ./config recognizes MacOS X now.
16575 * Bug fix for BN_div() when the first words of num and divisor are
16576 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16580 * Add support for various broken PKCS#8 formats, and command line
16581 options to produce them.
16585 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16586 get temporary BIGNUMs from a BN_CTX.
16590 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16595 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16596 include a #define from the old name to the new. The original intent
16597 was that statically linked binaries could for example just call
16598 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16599 link with digests. This never worked because SSLeay_add_all_digests()
16600 and SSLeay_add_all_ciphers() were in the same source file so calling
16601 one would link with the other. They are now in separate source files.
16605 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16609 * Use a less unusual form of the Miller-Rabin primality test (it used
16610 a binary algorithm for exponentiation integrated into the Miller-Rabin
16611 loop, our standard modexp algorithms are faster).
16615 * Support for the EBCDIC character set completed.
16617 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16619 * Source code cleanups: use const where appropriate, eliminate casts,
16620 use `void *` instead of `char *` in lhash.
16624 * Bugfix: ssl3_send_server_key_exchange was not restartable
16625 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16626 this the server could overwrite ephemeral keys that the client
16631 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16632 using 50 iterations of the Rabin-Miller test.
16634 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16635 iterations of the Rabin-Miller test as required by the appendix
16636 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16637 As BN_is_prime_fasttest includes trial division, DSA parameter
16638 generation becomes much faster.
16640 This implies a change for the callback functions in DSA_is_prime
16641 and DSA_generate_parameters: The callback function is called once
16642 for each positive witness in the Rabin-Miller test, not just
16643 occasionally in the inner loop; and the parameters to the
16644 callback function now provide an iteration count for the outer
16645 loop rather than for the current invocation of the inner loop.
16646 DSA_generate_parameters additionally can call the callback
16647 function with an 'iteration count' of -1, meaning that a
16648 candidate has passed the trial division test (when q is generated
16649 from an application-provided seed, trial division is skipped).
16653 * New function BN_is_prime_fasttest that optionally does trial
16654 division before starting the Rabin-Miller test and has
16655 an additional BN_CTX * argument (whereas BN_is_prime always
16656 has to allocate at least one BN_CTX).
16657 'callback(1, -1, cb_arg)' is called when a number has passed the
16658 trial division stage.
16662 * Fix for bug in CRL encoding. The validity dates weren't being handled
16667 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16671 * New function BN_pseudo_rand().
16675 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16676 bignum version of BN_from_montgomery() with the working code from
16677 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16682 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16683 made it impossible to use the same SSL_SESSION data structure in
16684 SSL2 clients in multiple threads.
16688 * The return value of RAND_load_file() no longer counts bytes obtained
16689 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16690 to seed the PRNG (previously an explicit byte count was required).
16692 *Ulf Möller, Bodo Möller*
16694 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16695 used `char *` instead of `void *` and had casts all over the place.
16699 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16703 * Retain source code compatibility for BN_prime_checks macro:
16704 BN_is_prime(..., BN_prime_checks, ...) now uses
16705 BN_prime_checks_for_size to determine the appropriate number of
16706 Rabin-Miller iterations.
16710 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16711 DH_CHECK_P_NOT_SAFE_PRIME.
16712 (Check if this is true? OpenPGP calls them "strong".)
16716 * Merge the functionality of "dh" and "gendh" programs into a new program
16717 "dhparam". The old programs are retained for now but will handle DH keys
16718 (instead of parameters) in future.
16722 * Make the ciphers, s_server and s_client programs check the return values
16723 when a new cipher list is set.
16727 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16728 ciphers. Before when the 56bit ciphers were enabled the sorting was
16731 The syntax for the cipher sorting has been extended to support sorting by
16732 cipher-strength (using the strength_bits hard coded in the tables).
16733 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16735 Fix a bug in the cipher-command parser: when supplying a cipher command
16736 string with an "undefined" symbol (neither command nor alphanumeric
16737 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16738 an error is flagged.
16740 Due to the strength-sorting extension, the code of the
16741 ssl_create_cipher_list() function was completely rearranged. I hope that
16742 the readability was also increased :-)
16744 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16746 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16747 for the first serial number and places 2 in the serial number file. This
16748 avoids problems when the root CA is created with serial number zero and
16749 the first user certificate has the same issuer name and serial number
16754 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16755 the new code. Add documentation for this stuff.
16759 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16760 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16761 structures and behave in an analogous way to the X509v3 functions:
16762 they shouldn't be called directly but wrapper functions should be used
16765 So we also now have some wrapper functions that call the X509at functions
16766 when passed certificate requests. (TO DO: similar things can be done with
16767 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16768 things. Some of these need some d2i or i2d and print functionality
16769 because they handle more complex structures.)
16773 * Add missing #ifndefs that caused missing symbols when building libssl
16774 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16775 NO_RSA in `ssl/s2*.c`.
16777 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16779 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16780 has a return value which indicates the quality of the random data
16781 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16782 error queue. New function RAND_pseudo_bytes() generates output that is
16783 guaranteed to be unique but not unpredictable. RAND_add is like
16784 RAND_seed, but takes an extra argument for an entropy estimate
16785 (RAND_seed always assumes full entropy).
16789 * Do more iterations of Rabin-Miller probable prime test (specifically,
16790 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16791 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16792 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16793 false-positive rate of at most 2^-80 for random input.
16797 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16801 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16802 in the 0.9.5 release), this returns the chain
16803 from an X509_CTX structure with a dup of the stack and all
16804 the X509 reference counts upped: so the stack will exist
16805 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16808 Also make SSL_SESSION_print() print out the verify return
16813 * Add manpage for the pkcs12 command. Also change the default
16814 behaviour so MAC iteration counts are used unless the new
16815 -nomaciter option is used. This improves file security and
16816 only older versions of MSIE (4.0 for example) need it.
16820 * Honor the no-xxx Configure options when creating .DEF files.
16824 * Add PKCS#10 attributes to field table: challengePassword,
16825 unstructuredName and unstructuredAddress. These are taken from
16826 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16827 international characters are used.
16829 More changes to X509_ATTRIBUTE code: allow the setting of types
16830 based on strings. Remove the 'loc' parameter when adding
16831 attributes because these will be a SET OF encoding which is sorted
16836 * Initial changes to the 'req' utility to allow request generation
16837 automation. This will allow an application to just generate a template
16838 file containing all the field values and have req construct the
16841 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16842 used all over the place including certificate requests and PKCS#7
16843 structures. They are currently handled manually where necessary with
16844 some primitive wrappers for PKCS#7. The new functions behave in a
16845 manner analogous to the X509 extension functions: they allow
16846 attributes to be looked up by NID and added.
16848 Later something similar to the X509V3 code would be desirable to
16849 automatically handle the encoding, decoding and printing of the
16850 more complex types. The string types like challengePassword can
16851 be handled by the string table functions.
16853 Also modified the multi byte string table handling. Now there is
16854 a 'global mask' which masks out certain types. The table itself
16855 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16856 is useful when for example there is only one permissible type
16857 (as in countryName) and using the mask might result in no valid
16862 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16863 SSL_get_peer_finished to allow applications to obtain the latest
16864 Finished messages sent to the peer or expected from the peer,
16865 respectively. (SSL_get_peer_finished is usually the Finished message
16866 actually received from the peer, otherwise the protocol will be aborted.)
16868 As the Finished message are message digests of the complete handshake
16869 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16870 be used for external authentication procedures when the authentication
16871 provided by SSL/TLS is not desired or is not enough.
16875 * Enhanced support for Alpha Linux is added. Now ./config checks if
16876 the host supports BWX extension and if Compaq C is present on the
16877 $PATH. Just exploiting of the BWX extension results in 20-30%
16878 performance kick for some algorithms, e.g. DES and RC4 to mention
16879 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16884 * Add support for MS "fast SGC". This is arguably a violation of the
16885 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16886 weak crypto and after checking the certificate is SGC a second one
16887 with strong crypto. MS SGC stops the first handshake after receiving
16888 the server certificate message and sends a second client hello. Since
16889 a server will typically do all the time consuming operations before
16890 expecting any further messages from the client (server key exchange
16891 is the most expensive) there is little difference between the two.
16893 To get OpenSSL to support MS SGC we have to permit a second client
16894 hello message after we have sent server done. In addition we have to
16895 reset the MAC if we do get this second client hello.
16899 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16900 if a DER encoded private key is RSA or DSA traditional format. Changed
16901 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16902 format DER encoded private key. Newer code should use PKCS#8 format which
16903 has the key type encoded in the ASN1 structure. Added DER private key
16904 support to pkcs8 application.
16908 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16909 ciphersuites has been selected (as required by the SSL 3/TLS 1
16910 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16911 is set, we interpret this as a request to violate the specification
16912 (the worst that can happen is a handshake failure, and 'correct'
16913 behaviour would result in a handshake failure anyway).
16917 * In SSL_CTX_add_session, take into account that there might be multiple
16918 SSL_SESSION structures with the same session ID (e.g. when two threads
16919 concurrently obtain them from an external cache).
16920 The internal cache can handle only one SSL_SESSION with a given ID,
16921 so if there's a conflict, we now throw out the old one to achieve
16926 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16927 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16928 some routines that use cipher OIDs: some ciphers do not have OIDs
16929 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16934 * Simplify the trust setting structure and code. Now we just have
16935 two sequences of OIDs for trusted and rejected settings. These will
16936 typically have values the same as the extended key usage extension
16937 and any application specific purposes.
16939 The trust checking code now has a default behaviour: it will just
16940 check for an object with the same NID as the passed id. Functions can
16941 be provided to override either the default behaviour or the behaviour
16942 for a given id. SSL client, server and email already have functions
16943 in place for compatibility: they check the NID and also return "trusted"
16944 if the certificate is self signed.
16948 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16949 traditional format into an EVP_PKEY structure.
16953 * Add a password callback function PEM_cb() which either prompts for
16954 a password if usr_data is NULL or otherwise assumes it is a null
16955 terminated password. Allow passwords to be passed on command line
16956 environment or config files in a few more utilities.
16960 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16961 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16962 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16963 Update documentation.
16967 * Support for ASN1 "NULL" type. This could be handled before by using
16968 ASN1_TYPE but there wasn't any function that would try to read a NULL
16969 and produce an error if it couldn't. For compatibility we also have
16970 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16971 don't allocate anything because they don't need to.
16975 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16978 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16980 * Rebuild of the memory allocation routines used by OpenSSL code and
16981 possibly others as well. The purpose is to make an interface that
16982 provide hooks so anyone can build a separate set of allocation and
16983 deallocation routines to be used by OpenSSL, for example memory
16984 pool implementations, or something else, which was previously hard
16985 since Malloc(), Realloc() and Free() were defined as macros having
16986 the values malloc, realloc and free, respectively (except for Win32
16987 compilations). The same is provided for memory debugging code.
16988 OpenSSL already comes with functionality to find memory leaks, but
16989 this gives people a chance to debug other memory problems.
16991 With these changes, a new set of functions and macros have appeared:
16993 CRYPTO_set_mem_debug_functions() [F]
16994 CRYPTO_get_mem_debug_functions() [F]
16995 CRYPTO_dbg_set_options() [F]
16996 CRYPTO_dbg_get_options() [F]
16997 CRYPTO_malloc_debug_init() [M]
16999 The memory debug functions are NULL by default, unless the library
17000 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17001 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17002 gives the standard debugging functions that come with OpenSSL) or
17003 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17004 provided by the library user) must be used. When the standard
17005 debugging functions are used, CRYPTO_dbg_set_options can be used to
17006 request additional information:
17007 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17008 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17010 Also, things like CRYPTO_set_mem_functions will always give the
17011 expected result (the new set of functions is used for allocation
17012 and deallocation) at all times, regardless of platform and compiler
17015 To finish it up, some functions that were never use in any other
17016 way than through macros have a new API and new semantic:
17018 CRYPTO_dbg_malloc()
17019 CRYPTO_dbg_realloc()
17022 All macros of value have retained their old syntax.
17024 *Richard Levitte and Bodo Moeller*
17026 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17027 ordering of SMIMECapabilities wasn't in "strength order" and there
17028 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17033 * Some ASN1 types with illegal zero length encoding (INTEGER,
17034 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17036 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17038 * Merge in my S/MIME library for OpenSSL. This provides a simple
17039 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17040 functionality to handle multipart/signed properly) and a utility
17041 called 'smime' to call all this stuff. This is based on code I
17042 originally wrote for Celo who have kindly allowed it to be
17043 included in OpenSSL.
17047 * Add variants des_set_key_checked and des_set_key_unchecked of
17048 des_set_key (aka des_key_sched). Global variable des_check_key
17049 decides which of these is called by des_set_key; this way
17050 des_check_key behaves as it always did, but applications and
17051 the library itself, which was buggy for des_check_key == 1,
17052 have a cleaner way to pick the version they need.
17056 * New function PKCS12_newpass() which changes the password of a
17061 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17062 dynamic mix. In both cases the ids can be used as an index into the
17063 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17064 functions so they accept a list of the field values and the
17065 application doesn't need to directly manipulate the X509_TRUST
17070 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17075 * Modify the way the V3 extension code looks up extensions. This now
17076 works in a similar way to the object code: we have some "standard"
17077 extensions in a static table which is searched with OBJ_bsearch()
17078 and the application can add dynamic ones if needed. The file
17079 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17080 updated whenever a new extension is added to the core code and kept
17081 in ext_nid order. There is a simple program 'tabtest.c' which checks
17082 this. New extensions are not added too often so this file can readily
17083 be maintained manually.
17085 There are two big advantages in doing things this way. The extensions
17086 can be looked up immediately and no longer need to be "added" using
17087 X509V3_add_standard_extensions(): this function now does nothing.
17088 Side note: I get *lots* of email saying the extension code doesn't
17089 work because people forget to call this function.
17090 Also no dynamic allocation is done unless new extensions are added:
17091 so if we don't add custom extensions there is no need to call
17092 X509V3_EXT_cleanup().
17096 * Modify enc utility's salting as follows: make salting the default. Add a
17097 magic header, so unsalted files fail gracefully instead of just decrypting
17098 to garbage. This is because not salting is a big security hole, so people
17099 should be discouraged from doing it.
17103 * Fixes and enhancements to the 'x509' utility. It allowed a message
17104 digest to be passed on the command line but it only used this
17105 parameter when signing a certificate. Modified so all relevant
17106 operations are affected by the digest parameter including the
17107 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17108 DSA key was used because it didn't fix the digest.
17112 * Initial certificate chain verify code. Currently tests the untrusted
17113 certificates for consistency with the verify purpose (which is set
17114 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17116 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17117 this is because it will reject chains with invalid extensions whereas
17118 every previous version of OpenSSL and SSLeay made no checks at all.
17120 Trust code: checks the root CA for the relevant trust settings. Trust
17121 settings have an initial value consistent with the verify purpose: e.g.
17122 if the verify purpose is for SSL client use it expects the CA to be
17123 trusted for SSL client use. However the default value can be changed to
17124 permit custom trust settings: one example of this would be to only trust
17125 certificates from a specific "secure" set of CAs.
17127 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17128 which should be used for version portability: especially since the
17129 verify structure is likely to change more often now.
17131 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17132 to set them. If not set then assume SSL clients will verify SSL servers
17135 Two new options to the verify program: -untrusted allows a set of
17136 untrusted certificates to be passed in and -purpose which sets the
17137 intended purpose of the certificate. If a purpose is set then the
17138 new chain verify code is used to check extension consistency.
17142 * Support for the authority information access extension.
17146 * Modify RSA and DSA PEM read routines to transparently handle
17147 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17148 public keys in a format compatible with certificate
17149 SubjectPublicKeyInfo structures. Unfortunately there were already
17150 functions called *_PublicKey_* which used various odd formats so
17151 these are retained for compatibility: however the DSA variants were
17152 never in a public release so they have been deleted. Changed dsa/rsa
17153 utilities to handle the new format: note no releases ever handled public
17154 keys so we should be OK.
17156 The primary motivation for this change is to avoid the same fiasco
17157 that dogs private keys: there are several incompatible private key
17158 formats some of which are standard and some OpenSSL specific and
17159 require various evil hacks to allow partial transparent handling and
17160 even then it doesn't work with DER formats. Given the option anything
17161 other than PKCS#8 should be dumped: but the other formats have to
17162 stay in the name of compatibility.
17164 With public keys and the benefit of hindsight one standard format
17165 is used which works with EVP_PKEY, RSA or DSA structures: though
17166 it clearly returns an error if you try to read the wrong kind of key.
17168 Added a -pubkey option to the 'x509' utility to output the public key.
17169 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17170 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17171 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17172 that do the same as the `EVP_PKEY_assign_*()` except they up the
17173 reference count of the added key (they don't "swallow" the
17178 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17179 CRLs would fail if the file contained no certificates or no CRLs:
17180 added a new function to read in both types and return the number
17181 read: this means that if none are read it will be an error. The
17182 DER versions of the certificate and CRL reader would always fail
17183 because it isn't possible to mix certificates and CRLs in DER format
17184 without choking one or the other routine. Changed this to just read
17185 a certificate: this is the best we can do. Also modified the code
17186 in `apps/verify.c` to take notice of return codes: it was previously
17187 attempting to read in certificates from NULL pointers and ignoring
17188 any errors: this is one reason why the cert and CRL reader seemed
17189 to work. It doesn't check return codes from the default certificate
17190 routines: these may well fail if the certificates aren't installed.
17194 * Code to support otherName option in GeneralName.
17198 * First update to verify code. Change the verify utility
17199 so it warns if it is passed a self signed certificate:
17200 for consistency with the normal behaviour. X509_verify
17201 has been modified to it will now verify a self signed
17202 certificate if *exactly* the same certificate appears
17203 in the store: it was previously impossible to trust a
17204 single self signed certificate. This means that:
17205 openssl verify ss.pem
17206 now gives a warning about a self signed certificate but
17207 openssl verify -CAfile ss.pem ss.pem
17212 * For servers, store verify_result in SSL_SESSION data structure
17213 (and add it to external session representation).
17214 This is needed when client certificate verifications fails,
17215 but an application-provided verification callback (set by
17216 SSL_CTX_set_cert_verify_callback) allows accepting the session
17217 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17218 but returns 1): When the session is reused, we have to set
17219 ssl->verify_result to the appropriate error code to avoid
17222 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17224 * Fix a bug in the new PKCS#7 code: it didn't consider the
17225 case in PKCS7_dataInit() where the signed PKCS7 structure
17226 didn't contain any existing data because it was being created.
17228 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17230 * Add a salt to the key derivation routines in enc.c. This
17231 forms the first 8 bytes of the encrypted file. Also add a
17232 -S option to allow a salt to be input on the command line.
17236 * New function X509_cmp(). Oddly enough there wasn't a function
17237 to compare two certificates. We do this by working out the SHA1
17238 hash and comparing that. X509_cmp() will be needed by the trust
17243 * SSL_get1_session() is like SSL_get_session(), but increments
17244 the reference count in the SSL_SESSION returned.
17246 *Geoff Thorpe <geoff@eu.c2.net>*
17248 * Fix for 'req': it was adding a null to request attributes.
17249 Also change the X509_LOOKUP and X509_INFO code to handle
17250 certificate auxiliary information.
17254 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17259 * Add the possibility to add extra information to the memory leak
17260 detecting output, to form tracebacks, showing from where each
17261 allocation was originated: CRYPTO_push_info("constant string") adds
17262 the string plus current file name and line number to a per-thread
17263 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17264 is like calling CYRPTO_pop_info() until the stack is empty.
17265 Also updated memory leak detection code to be multi-thread-safe.
17269 * Add options -text and -noout to pkcs7 utility and delete the
17270 encryption options which never did anything. Update docs.
17274 * Add options to some of the utilities to allow the pass phrase
17275 to be included on either the command line (not recommended on
17276 OSes like Unix) or read from the environment. Update the
17277 manpages and fix a few bugs.
17281 * Add a few manpages for some of the openssl commands.
17285 * Fix the -revoke option in ca. It was freeing up memory twice,
17286 leaking and not finding already revoked certificates.
17290 * Extensive changes to support certificate auxiliary information.
17291 This involves the use of X509_CERT_AUX structure and X509_AUX
17292 functions. An X509_AUX function such as PEM_read_X509_AUX()
17293 can still read in a certificate file in the usual way but it
17294 will also read in any additional "auxiliary information". By
17295 doing things this way a fair degree of compatibility can be
17296 retained: existing certificates can have this information added
17297 using the new 'x509' options.
17299 Current auxiliary information includes an "alias" and some trust
17300 settings. The trust settings will ultimately be used in enhanced
17301 certificate chain verification routines: currently a certificate
17302 can only be trusted if it is self signed and then it is trusted
17307 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17308 The problem was that one of the replacement routines had not been working
17309 since SSLeay releases. For now the offending routine has been replaced
17310 with non-optimised assembler. Even so, this now gives around 95%
17311 performance improvement for 1024 bit RSA signs.
17315 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17316 handling. Most clients have the effective key size in bits equal to
17317 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17318 A few however don't do this and instead use the size of the decrypted key
17319 to determine the RC2 key length and the AlgorithmIdentifier to determine
17320 the effective key length. In this case the effective key length can still
17321 be 40 bits but the key length can be 168 bits for example. This is fixed
17322 by manually forcing an RC2 key into the EVP_PKEY structure because the
17323 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17324 the key length and effective key length are equal.
17328 * Add a bunch of functions that should simplify the creation of
17329 X509_NAME structures. Now you should be able to do:
17330 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17331 and have it automatically work out the correct field type and fill in
17332 the structures. The more adventurous can try:
17333 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17334 and it will (hopefully) work out the correct multibyte encoding.
17338 * Change the 'req' utility to use the new field handling and multibyte
17339 copy routines. Before the DN field creation was handled in an ad hoc
17340 way in req, ca, and x509 which was rather broken and didn't support
17341 BMPStrings or UTF8Strings. Since some software doesn't implement
17342 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17343 using the dirstring_type option. See the new comment in the default
17344 openssl.cnf for more info.
17348 * Make crypto/rand/md_rand.c more robust:
17349 - Assure unique random numbers after fork().
17350 - Make sure that concurrent threads access the global counter and
17351 md serializably so that we never lose entropy in them
17352 or use exactly the same state in multiple threads.
17353 Access to the large state is not always serializable because
17354 the additional locking could be a performance killer, and
17355 md should be large enough anyway.
17359 * New file `apps/app_rand.c` with commonly needed functionality
17360 for handling the random seed file.
17362 Use the random seed file in some applications that previously did not:
17364 dsaparam -genkey (which also ignored its '-rand' option),
17367 x509 (when signing).
17368 Except on systems with /dev/urandom, it is crucial to have a random
17369 seed file at least for key creation, DSA signing, and for DH exchanges;
17370 for RSA signatures we could do without one.
17372 gendh and gendsa (unlike genrsa) used to read only the first byte
17373 of each file listed in the '-rand' option. The function as previously
17374 found in genrsa is now in app_rand.c and is used by all programs
17375 that support '-rand'.
17379 * In RAND_write_file, use mode 0600 for creating files;
17380 don't just chmod when it may be too late.
17384 * Report an error from X509_STORE_load_locations
17385 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17389 * New function ASN1_mbstring_copy() this copies a string in either
17390 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17391 into an ASN1_STRING type. A mask of permissible types is passed
17392 and it chooses the "minimal" type to use or an error if not type
17397 * Add function equivalents to the various macros in asn1.h. The old
17398 macros are retained with an `M_` prefix. Code inside the library can
17399 use the `M_` macros. External code (including the openssl utility)
17400 should *NOT* in order to be "shared library friendly".
17404 * Add various functions that can check a certificate's extensions
17405 to see if it usable for various purposes such as SSL client,
17406 server or S/MIME and CAs of these types. This is currently
17407 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17408 verification. Also added a -purpose flag to x509 utility to
17409 print out all the purposes.
17413 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17418 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17419 for, obtain and decode and extension and obtain its critical flag.
17420 This allows all the necessary extension code to be handled in a
17421 single function call.
17425 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17426 platforms. See crypto/rc4/rc4_enc.c for further details.
17430 * New -noout option to asn1parse. This causes no output to be produced
17431 its main use is when combined with -strparse and -out to extract data
17432 from a file (which may not be in ASN.1 format).
17436 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17437 when producing the local key id.
17439 *Richard Levitte <levitte@stacken.kth.se>*
17441 * New option -dhparam in s_server. This allows a DH parameter file to be
17442 stated explicitly. If it is not stated then it tries the first server
17443 certificate file. The previous behaviour hard coded the filename
17448 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17449 a public key to be input or output. For example:
17450 openssl rsa -in key.pem -pubout -out pubkey.pem
17451 Also added necessary DSA public key functions to handle this.
17455 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17456 in the message. This was handled by allowing
17457 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17459 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17461 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17462 to the end of the strings whereas this didn't. This would cause problems
17463 if strings read with d2i_ASN1_bytes() were later modified.
17465 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17467 * Fix for base64 decode bug. When a base64 bio reads only one line of
17468 data and it contains EOF it will end up returning an error. This is
17469 caused by input 46 bytes long. The cause is due to the way base64
17470 BIOs find the start of base64 encoded data. They do this by trying a
17471 trial decode on each line until they find one that works. When they
17472 do a flag is set and it starts again knowing it can pass all the
17473 data directly through the decoder. Unfortunately it doesn't reset
17474 the context it uses. This means that if EOF is reached an attempt
17475 is made to pass two EOFs through the context and this causes the
17476 resulting error. This can also cause other problems as well. As is
17477 usual with these problems it takes *ages* to find and the fix is
17478 trivial: move one line.
17480 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17482 * Ugly workaround to get s_client and s_server working under Windows. The
17483 old code wouldn't work because it needed to select() on sockets and the
17484 tty (for keypresses and to see if data could be written). Win32 only
17485 supports select() on sockets so we select() with a 1s timeout on the
17486 sockets and then see if any characters are waiting to be read, if none
17487 are present then we retry, we also assume we can always write data to
17488 the tty. This isn't nice because the code then blocks until we've
17489 received a complete line of data and it is effectively polling the
17490 keyboard at 1s intervals: however it's quite a bit better than not
17491 working at all :-) A dedicated Windows application might handle this
17492 with an event loop for example.
17496 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17497 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17498 will be called when RSA_sign() and RSA_verify() are used. This is useful
17499 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17500 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17501 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17502 This necessitated the support of an extra signature type NID_md5_sha1
17503 for SSL signatures and modifications to the SSL library to use it instead
17504 of calling RSA_public_decrypt() and RSA_private_encrypt().
17508 * Add new -verify -CAfile and -CApath options to the crl program, these
17509 will lookup a CRL issuers certificate and verify the signature in a
17510 similar way to the verify program. Tidy up the crl program so it
17511 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17512 less strict. It will now permit CRL extensions even if it is not
17513 a V2 CRL: this will allow it to tolerate some broken CRLs.
17517 * Initialize all non-automatic variables each time one of the openssl
17518 sub-programs is started (this is necessary as they may be started
17519 multiple times from the "OpenSSL>" prompt).
17521 *Lennart Bang, Bodo Moeller*
17523 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17524 removing all other RSA functionality (this is what NO_RSA does). This
17525 is so (for example) those in the US can disable those operations covered
17526 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17531 * Non-copying interface to BIO pairs.
17532 (still largely untested)
17536 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17537 ASCII string. This was handled independently in various places before.
17541 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17542 UTF8 strings a character at a time.
17546 * Use client_version from client hello to select the protocol
17547 (s23_srvr.c) and for RSA client key exchange verification
17548 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17552 * Add various utility functions to handle SPKACs, these were previously
17553 handled by poking round in the structure internals. Added new function
17554 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17555 print, verify and generate SPKACs. Based on an original idea from
17556 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17560 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17564 * Allow the config file extension section to be overwritten on the
17565 command line. Based on an original idea from Massimiliano Pala
17566 <madwolf@comune.modena.it>. The new option is called -extensions
17567 and can be applied to ca, req and x509. Also -reqexts to override
17568 the request extensions in req and -crlexts to override the crl extensions
17573 * Add new feature to the SPKAC handling in ca. Now you can include
17574 the same field multiple times by preceding it by "XXXX." for example:
17577 this is the same syntax as used in the req config file.
17581 * Allow certificate extensions to be added to certificate requests. These
17582 are specified in a 'req_extensions' option of the req section of the
17583 config file. They can be printed out with the -text option to req but
17584 are otherwise ignored at present.
17588 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17589 data read consists of only the final block it would not decrypted because
17590 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17591 A misplaced 'break' also meant the decrypted final block might not be
17592 copied until the next read.
17596 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17597 a few extra parameters to the DH structure: these will be useful if
17598 for example we want the value of 'q' or implement X9.42 DH.
17602 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17603 provides hooks that allow the default DSA functions or functions on a
17604 "per key" basis to be replaced. This allows hardware acceleration and
17605 hardware key storage to be handled without major modification to the
17606 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17607 associated functions.
17611 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17612 as "read only": it can't be written to and the buffer it points to will
17613 not be freed. Reading from a read only BIO is much more efficient than
17614 a normal memory BIO. This was added because there are several times when
17615 an area of memory needs to be read from a BIO. The previous method was
17616 to create a memory BIO and write the data to it, this results in two
17617 copies of the data and an O(n^2) reading algorithm. There is a new
17618 function BIO_new_mem_buf() which creates a read only memory BIO from
17619 an area of memory. Also modified the PKCS#7 routines to use read only
17624 * Bugfix: ssl23_get_client_hello did not work properly when called in
17625 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17626 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17627 but a retry condition occurred while trying to read the rest.
17631 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17632 NID_pkcs7_encrypted by default: this was wrong since this should almost
17633 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17634 the encrypted data type: this is a more sensible place to put it and it
17635 allows the PKCS#12 code to be tidied up that duplicated this
17640 * Changed obj_dat.pl script so it takes its input and output files on
17641 the command line. This should avoid shell escape redirection problems
17646 * Initial support for certificate extension requests, these are included
17647 in things like Xenroll certificate requests. Included functions to allow
17648 extensions to be obtained and added.
17652 * -crlf option to s_client and s_server for sending newlines as
17653 CRLF (as required by many protocols).
17657 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17659 * Install libRSAglue.a when OpenSSL is built with RSAref.
17661 *Ralf S. Engelschall*
17663 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17665 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17667 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17672 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17673 DH parameters/keys (q is lost during that conversion, but the resulting
17674 DH parameters contain its length).
17676 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17677 much faster than DH_generate_parameters (which creates parameters
17678 where `p = 2*q + 1`), and also the smaller q makes DH computations
17679 much more efficient (160-bit exponentiation instead of 1024-bit
17680 exponentiation); so this provides a convenient way to support DHE
17681 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17682 utter importance to use
17683 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17685 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17686 when such DH parameters are used, because otherwise small subgroup
17687 attacks may become possible!
17691 * Avoid memory leak in i2d_DHparams.
17695 * Allow the -k option to be used more than once in the enc program:
17696 this allows the same encrypted message to be read by multiple recipients.
17700 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17701 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17702 it will always use the numerical form of the OID, even if it has a short
17707 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17708 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17709 otherwise bn_mod_exp was called. In the case of hardware keys for example
17710 no private key components need be present and it might store extra data
17711 in the RSA structure, which cannot be accessed from bn_mod_exp.
17712 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17713 private key operations.
17717 * Added support for SPARC Linux.
17721 * pem_password_cb function type incompatibly changed from
17722 typedef int pem_password_cb(char *buf, int size, int rwflag);
17724 ....(char *buf, int size, int rwflag, void *userdata);
17725 so that applications can pass data to their callbacks:
17726 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17727 additional void * argument, which is just handed through whenever
17728 the password callback is called.
17730 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17732 New function SSL_CTX_set_default_passwd_cb_userdata.
17734 Compatibility note: As many C implementations push function arguments
17735 onto the stack in reverse order, the new library version is likely to
17736 interoperate with programs that have been compiled with the old
17737 pem_password_cb definition (PEM_whatever takes some data that
17738 happens to be on the stack as its last argument, and the callback
17739 just ignores this garbage); but there is no guarantee whatsoever that
17742 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17743 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17744 problems not only on Windows, but also on some Unix platforms.
17745 To avoid problematic command lines, these definitions are now in an
17746 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17747 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17751 * MIPS III/IV assembler module is reimplemented.
17755 * More DES library cleanups: remove references to srand/rand and
17756 delete an unused file.
17760 * Add support for the free Netwide assembler (NASM) under Win32,
17761 since not many people have MASM (ml) and it can be hard to obtain.
17762 This is currently experimental but it seems to work OK and pass all
17763 the tests. Check out INSTALL.W32 for info.
17767 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17768 without temporary keys kept an extra copy of the server key,
17769 and connections with temporary keys did not free everything in case
17774 * New function RSA_check_key and new openssl rsa option -check
17775 for verifying the consistency of RSA keys.
17777 *Ulf Moeller, Bodo Moeller*
17779 * Various changes to make Win32 compile work:
17780 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17781 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17782 comparison" warnings.
17783 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17787 * Add a debugging option to PKCS#5 v2 key generation function: when
17788 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17789 derived keys are printed to stderr.
17793 * Copy the flags in ASN1_STRING_dup().
17795 *Roman E. Pavlov <pre@mo.msk.ru>*
17797 * The x509 application mishandled signing requests containing DSA
17798 keys when the signing key was also DSA and the parameters didn't match.
17800 It was supposed to omit the parameters when they matched the signing key:
17801 the verifying software was then supposed to automatically use the CA's
17802 parameters if they were absent from the end user certificate.
17804 Omitting parameters is no longer recommended. The test was also
17805 the wrong way round! This was probably due to unusual behaviour in
17806 EVP_cmp_parameters() which returns 1 if the parameters match.
17807 This meant that parameters were omitted when they *didn't* match and
17808 the certificate was useless. Certificates signed with 'ca' didn't have
17811 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17813 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17814 The interface is as follows:
17815 Applications can use
17816 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17817 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17818 "off" is now the default.
17819 The library internally uses
17820 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17821 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17822 to disable memory-checking temporarily.
17824 Some inconsistent states that previously were possible (and were
17825 even the default) are now avoided.
17827 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17828 with each memory chunk allocated; this is occasionally more helpful
17829 than just having a counter.
17831 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17833 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17838 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17839 which largely parallels "options", but is for changing API behaviour,
17840 whereas "options" are about protocol behaviour.
17841 Initial "mode" flags are:
17843 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17844 a single record has been written.
17845 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17846 retries use the same buffer location.
17847 (But all of the contents must be
17852 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17855 * Fix problems with no-hmac etc.
17857 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17859 * New functions RSA_get_default_method(), RSA_set_method() and
17860 RSA_get_method(). These allows replacement of RSA_METHODs without having
17861 to mess around with the internals of an RSA structure.
17865 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17866 Also really enable memory leak checks in openssl.c and in some
17869 *Chad C. Mulligan, Bodo Moeller*
17871 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17872 up the length of negative integers. This has now been simplified to just
17873 store the length when it is first determined and use it later, rather
17874 than trying to keep track of where data is copied and updating it to
17876 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17878 * Add a new function PKCS7_signatureVerify. This allows the verification
17879 of a PKCS#7 signature but with the signing certificate passed to the
17880 function itself. This contrasts with PKCS7_dataVerify which assumes the
17881 certificate is present in the PKCS#7 structure. This isn't always the
17882 case: certificates can be omitted from a PKCS#7 structure and be
17883 distributed by "out of band" means (such as a certificate database).
17887 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17888 function prototypes in pem.h, also change util/mkdef.pl to add the
17889 necessary function names.
17893 * mk1mf.pl (used by Windows builds) did not properly read the
17894 options set by Configure in the top level Makefile, and Configure
17895 was not even able to write more than one option correctly.
17896 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17900 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17901 file to be loaded from a BIO or FILE pointer. The BIO version will
17902 for example allow memory BIOs to contain config info.
17906 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17907 Whoever hopes to achieve shared-library compatibility across versions
17908 must use this, not the compile-time macro.
17909 (Exercise 0.9.4: Which is the minimum library version required by
17911 Note: All this applies only to multi-threaded programs, others don't
17916 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17917 through a BIO pair triggered the default case, i.e.
17918 SSLerr(...,SSL_R_UNKNOWN_STATE).
17922 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17923 can use the SSL library even if none of the specific BIOs is
17928 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17929 for the encoded length.
17931 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17933 * Add initial documentation of the X509V3 functions.
17937 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17938 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17939 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17940 secure PKCS#8 private key format with a high iteration count.
17944 * Fix determination of Perl interpreter: A perl or perl5
17945 *directory* in $PATH was also accepted as the interpreter.
17947 *Ralf S. Engelschall*
17949 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17950 wrong with it but it was very old and did things like calling
17951 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17952 unusual formatting.
17956 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17957 to use the new extension code.
17961 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17962 with macros. This should make it easier to change their form, add extra
17963 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17968 * Add to configuration table a new entry that can specify an alternative
17969 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17970 according to Mark Crispin <MRC@Panda.COM>.
17974 * DES CBC did not update the IV. Weird.
17978 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17979 Changing the behaviour of the former might break existing programs --
17980 where IV updating is needed, des_ncbc_encrypt can be used.
17983 * When bntest is run from "make test" it drives bc to check its
17984 calculations, as well as internally checking them. If an internal check
17985 fails, it needs to cause bc to give a non-zero result or make test carries
17986 on without noticing the failure. Fixed.
17990 * DES library cleanups.
17994 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17995 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17996 ciphers. NOTE: although the key derivation function has been verified
17997 against some published test vectors it has not been extensively tested
17998 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18003 * Instead of "mkdir -p", which is not fully portable, use new
18004 Perl script "util/mkdir-p.pl".
18008 * Rewrite the way password based encryption (PBE) is handled. It used to
18009 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18010 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18011 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18012 the 'parameter' field of the AlgorithmIdentifier is passed to the
18013 underlying key generation function so it must do its own ASN1 parsing.
18014 This has also changed the EVP_PBE_CipherInit() function which now has a
18015 'parameter' argument instead of literal salt and iteration count values
18016 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18020 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18021 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18022 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18023 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18024 value was just used as a "magic string" and not used directly its
18025 value doesn't matter.
18029 * Introduce some semblance of const correctness to BN. Shame C doesn't
18034 * "linux-sparc64" configuration (ultrapenguin).
18036 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18037 "linux-sparc" configuration.
18039 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18041 * config now generates no-xxx options for missing ciphers.
18045 * Support the EBCDIC character set (work in progress).
18046 File ebcdic.c not yet included because it has a different license.
18048 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18050 * Support BS2000/OSD-POSIX.
18052 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18054 * Make callbacks for key generation use `void *` instead of `char *`.
18058 * Make S/MIME samples compile (not yet tested).
18062 * Additional typesafe stacks.
18066 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18070 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18072 * New configuration variant "sco5-gcc".
18074 * Updated some demos.
18076 *Sean O Riordain, Wade Scholine*
18078 * Add missing BIO_free at exit of pkcs12 application.
18082 * Fix memory leak in conf.c.
18086 * Updates for Win32 to assembler version of MD5.
18090 * Set #! path to perl in `apps/der_chop` to where we found it
18091 instead of using a fixed path.
18095 * SHA library changes for irix64-mips4-cc.
18099 * Improvements for VMS support.
18103 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18105 * Bignum library bug fix. IRIX 6 passes "make test" now!
18106 This also avoids the problems with SC4.2 and unpatched SC5.
18108 *Andy Polyakov <appro@fy.chalmers.se>*
18110 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18111 These are required because of the typesafe stack would otherwise break
18112 existing code. If old code used a structure member which used to be STACK
18113 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18114 sk_num or sk_value it would produce an error because the num, data members
18115 are not present in STACK_OF. Now it just produces a warning. sk_set
18116 replaces the old method of assigning a value to sk_value
18117 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18118 that does this will no longer work (and should use sk_set instead) but
18119 this could be regarded as a "questionable" behaviour anyway.
18123 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18124 correctly handle encrypted S/MIME data.
18128 * Change type of various DES function arguments from des_cblock
18129 (which means, in function argument declarations, pointer to char)
18130 to des_cblock * (meaning pointer to array with 8 char elements),
18131 which allows the compiler to do more typechecking; it was like
18132 that back in SSLeay, but with lots of ugly casts.
18134 Introduce new type const_des_cblock.
18138 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18139 problems: find RecipientInfo structure that matches recipient certificate
18140 and initialise the ASN1 structures properly based on passed cipher.
18144 * Belatedly make the BN tests actually check the results.
18148 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18149 to and from BNs: it was completely broken. New compilation option
18150 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18151 key elements as negative integers.
18155 * Reorganize and speed up MD5.
18157 *Andy Polyakov <appro@fy.chalmers.se>*
18161 *Richard Levitte <richard@levitte.org>*
18163 * New option -out to asn1parse to allow the parsed structure to be
18164 output to a file. This is most useful when combined with the -strparse
18165 option to examine the output of things like OCTET STRINGS.
18169 * Make SSL library a little more fool-proof by not requiring any longer
18170 that `SSL_set_{accept,connect}_state` be called before
18171 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18172 in many applications because usually everything *appeared* to work as
18173 intended anyway -- now it really works as intended).
18177 * Move openssl.cnf out of lib/.
18181 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18182 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18183 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18185 *Ralf S. Engelschall*
18187 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18188 handle PKCS#7 enveloped data properly.
18190 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18192 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18193 copying pointers. The cert_st handling is changed by this in
18194 various ways (and thus what used to be known as ctx->default_cert
18195 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18196 any longer when s->cert does not give us what we need).
18197 ssl_cert_instantiate becomes obsolete by this change.
18198 As soon as we've got the new code right (possibly it already is?),
18199 we have solved a couple of bugs of the earlier code where s->cert
18200 was used as if it could not have been shared with other SSL structures.
18202 Note that using the SSL API in certain dirty ways now will result
18203 in different behaviour than observed with earlier library versions:
18204 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18205 does not influence s as it used to.
18207 In order to clean up things more thoroughly, inside SSL_SESSION
18208 we don't use CERT any longer, but a new structure SESS_CERT
18209 that holds per-session data (if available); currently, this is
18210 the peer's certificate chain and, for clients, the server's certificate
18211 and temporary key. CERT holds only those values that can have
18212 meaningful defaults in an SSL_CTX.
18216 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18217 from the internal representation. Various PKCS#7 fixes: remove some
18218 evil casts and set the enc_dig_alg field properly based on the signing
18223 * Allow PKCS#12 password to be set from the command line or the
18224 environment. Let 'ca' get its config file name from the environment
18225 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18230 * Allow certificate policies extension to use an IA5STRING for the
18231 organization field. This is contrary to the PKIX definition but
18232 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18237 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18238 without disallowing inline assembler and the like for non-pedantic builds.
18242 * Support Borland C++ builder.
18244 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18250 * SHA-1 cleanups and performance enhancements.
18252 *Andy Polyakov <appro@fy.chalmers.se>*
18254 * Sparc v8plus assembler for the bignum library.
18256 *Andy Polyakov <appro@fy.chalmers.se>*
18258 * Accept any -xxx and +xxx compiler options in Configure.
18262 * Update HPUX configuration.
18266 * Add missing `sk_<type>_unshift()` function to safestack.h
18268 *Ralf S. Engelschall*
18270 * New function SSL_CTX_use_certificate_chain_file that sets the
18271 "extra_cert"s in addition to the certificate. (This makes sense
18272 only for "PEM" format files, as chains as a whole are not
18277 * Support verify_depth from the SSL API.
18278 x509_vfy.c had what can be considered an off-by-one-error:
18279 Its depth (which was not part of the external interface)
18280 was actually counting the number of certificates in a chain;
18281 now it really counts the depth.
18285 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18286 instead of X509err, which often resulted in confusing error
18287 messages since the error codes are not globally unique
18288 (e.g. an alleged error in ssl3_accept when a certificate
18289 didn't match the private key).
18291 * New function SSL_CTX_set_session_id_context that allows to set a default
18292 value (so that you don't need SSL_set_session_id_context for each
18293 connection using the SSL_CTX).
18297 * OAEP decoding bug fix.
18301 * Support INSTALL_PREFIX for package builders, as proposed by
18306 * New Configure options "threads" and "no-threads". For systems
18307 where the proper compiler options are known (currently Solaris
18308 and Linux), "threads" is the default.
18312 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18316 * Install various scripts to $(OPENSSLDIR)/misc, not to
18317 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18318 such as /usr/local/bin.
18322 * "make linux-shared" to build shared libraries.
18324 *Niels Poppe <niels@netbox.org>*
18326 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18330 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18331 extension adding in x509 utility.
18335 * Remove NOPROTO sections and error code comments.
18339 * Partial rewrite of the DEF file generator to now parse the ANSI
18344 * New Configure options --prefix=DIR and --openssldir=DIR.
18348 * Complete rewrite of the error code script(s). It is all now handled
18349 by one script at the top level which handles error code gathering,
18350 header rewriting and C source file generation. It should be much better
18351 than the old method: it now uses a modified version of Ulf's parser to
18352 read the ANSI prototypes in all header files (thus the old K&R definitions
18353 aren't needed for error creation any more) and do a better job of
18354 translating function codes into names. The old 'ASN1 error code embedded
18355 in a comment' is no longer necessary and it doesn't use .err files which
18356 have now been deleted. Also the error code call doesn't have to appear all
18357 on one line (which resulted in some large lines...).
18361 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18365 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18366 0 (which usually indicates a closed connection), but continue reading.
18370 * Fix some race conditions.
18374 * Add support for CRL distribution points extension. Add Certificate
18375 Policies and CRL distribution points documentation.
18379 * Move the autogenerated header file parts to crypto/opensslconf.h.
18383 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18384 8 of keying material. Merlin has also confirmed interop with this fix
18385 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18387 *Merlin Hughes <merlin@baltimore.ie>*
18389 * Fix lots of warnings.
18391 *Richard Levitte <levitte@stacken.kth.se>*
18393 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18394 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18396 *Richard Levitte <levitte@stacken.kth.se>*
18398 * Fix problems with sizeof(long) == 8.
18400 *Andy Polyakov <appro@fy.chalmers.se>*
18402 * Change functions to ANSI C.
18406 * Fix typos in error codes.
18408 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18410 * Remove defunct assembler files from Configure.
18414 * SPARC v8 assembler BIGNUM implementation.
18416 *Andy Polyakov <appro@fy.chalmers.se>*
18418 * Support for Certificate Policies extension: both print and set.
18419 Various additions to support the r2i method this uses.
18423 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18424 return a const string when you are expecting an allocated buffer.
18428 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18429 types DirectoryString and DisplayText.
18433 * Add code to allow r2i extensions to access the configuration database,
18434 add an LHASH database driver and add several ctx helper functions.
18438 * Fix an evil bug in bn_expand2() which caused various BN functions to
18439 fail when they extended the size of a BIGNUM.
18443 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18444 support typesafe stack.
18448 * Fix typo in SSL_[gs]et_options().
18450 *Nils Frostberg <nils@medcom.se>*
18452 * Delete various functions and files that belonged to the (now obsolete)
18453 old X509V3 handling code.
18457 * New Configure option "rsaref".
18461 * Don't auto-generate pem.h.
18465 * Introduce type-safe ASN.1 SETs.
18469 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18471 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18473 * Introduce type-safe STACKs. This will almost certainly break lots of code
18474 that links with OpenSSL (well at least cause lots of warnings), but fear
18475 not: the conversion is trivial, and it eliminates loads of evil casts. A
18476 few STACKed things have been converted already. Feel free to convert more.
18477 In the fullness of time, I'll do away with the STACK type altogether.
18481 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18482 specified in `<certfile>` by updating the entry in the index.txt file.
18483 This way one no longer has to edit the index.txt file manually for
18484 revoking a certificate. The -revoke option does the gory details now.
18486 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18488 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18489 `-text` option at all and this way the `-noout -text` combination was
18490 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18492 *Ralf S. Engelschall*
18494 * Make sure a corresponding plain text error message exists for the
18495 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18496 verify callback function determined that a certificate was revoked.
18498 *Ralf S. Engelschall*
18500 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18501 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18502 all available ciphers including rc5, which was forgotten until now.
18503 In order to let the testing shell script know which algorithms
18504 are available, a new (up to now undocumented) command
18505 `openssl list-cipher-commands` is used.
18509 * Bugfix: s_client occasionally would sleep in select() when
18510 it should have checked SSL_pending() first.
18514 * New functions DSA_do_sign and DSA_do_verify to provide access to
18515 the raw DSA values prior to ASN.1 encoding.
18519 * Tweaks to Configure
18521 *Niels Poppe <niels@netbox.org>*
18523 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18528 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18532 * New config option to avoid instructions that are illegal on the 80386.
18533 The default code is faster, but requires at least a 486.
18537 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18538 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18539 same as SSL2_VERSION anyway.
18543 * New "-showcerts" option for s_client.
18547 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18548 application. Various cleanups and fixes.
18552 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18553 modify error routines to work internally. Add error codes and PBE init
18554 to library startup routines.
18558 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18559 packing functions to asn1 and evp. Changed function names and error
18560 codes along the way.
18564 * PKCS12 integration: and so it begins... First of several patches to
18565 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18566 objects to objects.h
18570 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18571 and display support for Thawte strong extranet extension.
18575 * Add LinuxPPC support.
18577 *Jeff Dubrule <igor@pobox.org>*
18579 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18580 bn_div_words in alpha.s.
18582 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18584 * Make sure the RSA OAEP test is skipped under -DRSAref because
18585 OAEP isn't supported when OpenSSL is built with RSAref.
18587 *Ulf Moeller <ulf@fitug.de>*
18589 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18590 so they no longer are missing under -DNOPROTO.
18592 *Soren S. Jorvang <soren@t.dk>*
18594 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18596 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18597 doesn't work when the session is reused. Coming soon!
18601 * Fix a security hole, that allows sessions to be reused in the wrong
18602 context thus bypassing client cert protection! All software that uses
18603 client certs and session caches in multiple contexts NEEDS PATCHING to
18604 allow session reuse! A fuller solution is in the works.
18606 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18608 * Some more source tree cleanups (removed obsolete files
18609 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18610 permission on "config" script to be executable) and a fix for the INSTALL
18613 *Ulf Moeller <ulf@fitug.de>*
18615 * Remove some legacy and erroneous uses of malloc, free instead of
18618 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18620 * Make rsa_oaep_test return non-zero on error.
18622 *Ulf Moeller <ulf@fitug.de>*
18624 * Add support for native Solaris shared libraries. Configure
18625 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18626 if someone would make that last step automatic.
18628 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18630 * ctx_size was not built with the right compiler during "make links". Fixed.
18634 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18635 except NULL ciphers". This means the default cipher list will no longer
18636 enable NULL ciphers. They need to be specifically enabled e.g. with
18637 the string "DEFAULT:eNULL".
18641 * Fix to RSA private encryption routines: if p < q then it would
18642 occasionally produce an invalid result. This will only happen with
18643 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18647 * Be less restrictive and allow also `perl util/perlpath.pl
18648 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18649 because this way one can also use an interpreter named `perl5` (which is
18650 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18651 installed as `perl`).
18653 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18655 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18657 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18659 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18660 advapi32.lib to Win32 build and change the pem test comparison
18661 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18662 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18663 and crypto/des/ede_cbcm_enc.c.
18667 * DES quad checksum was broken on big-endian architectures. Fixed.
18671 * Comment out two functions in bio.h that aren't implemented. Fix up the
18672 Win32 test batch file so it (might) work again. The Win32 test batch file
18673 is horrible: I feel ill....
18677 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18678 in e_os.h. Audit of header files to check ANSI and non ANSI
18679 sections: 10 functions were absent from non ANSI section and not exported
18680 from Windows DLLs. Fixed up libeay.num for new functions.
18684 * Make `openssl version` output lines consistent.
18686 *Ralf S. Engelschall*
18688 * Fix Win32 symbol export lists for BIO functions: Added
18689 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18690 to ms/libeay{16,32}.def.
18692 *Ralf S. Engelschall*
18694 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18695 fine under Unix and passes some trivial tests I've now added. But the
18696 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18697 added to make sure no one expects that this stuff really works in the
18698 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18699 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18702 *Ralf S. Engelschall*
18704 * Fix the generation of two part addresses in perl.
18706 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18708 * Add config entry for Linux on MIPS.
18710 *John Tobey <jtobey@channel1.com>*
18712 * Make links whenever Configure is run, unless we are on Windoze.
18716 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18717 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18722 * Add a useful kludge to allow package maintainers to specify compiler and
18723 other platforms details on the command line without having to patch the
18724 Configure script every time: One now can use
18725 `perl Configure <id>:<details>`,
18726 i.e. platform ids are allowed to have details appended
18727 to them (separated by colons). This is treated as there would be a static
18728 pre-configured entry in Configure's %table under key `<id>` with value
18729 `<details>` and `perl Configure <id>` is called. So, when you want to
18730 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18731 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18732 now, which overrides the FreeBSD-elf entry on-the-fly.
18734 *Ralf S. Engelschall*
18736 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18740 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18741 on the `perl Configure ...` command line. This way one can compile
18742 OpenSSL libraries with Position Independent Code (PIC) which is needed
18743 for linking it into DSOs.
18745 *Ralf S. Engelschall*
18747 * Remarkably, export ciphers were totally broken and no-one had noticed!
18752 * Cleaned up the LICENSE document: The official contact for any license
18753 questions now is the OpenSSL core team under openssl-core@openssl.org.
18754 And add a paragraph about the dual-license situation to make sure people
18755 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18756 to the OpenSSL toolkit.
18758 *Ralf S. Engelschall*
18760 * General source tree makefile cleanups: Made `making xxx in yyy...`
18761 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18762 Additionally cleaned up the `make links` target: Remove unnecessary
18763 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18764 to speed processing and no longer clutter the display with confusing
18765 stuff. Instead only the actually done links are displayed.
18767 *Ralf S. Engelschall*
18769 * Permit null encryption ciphersuites, used for authentication only. It used
18770 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18771 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18776 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18777 signed attributes when verifying signatures (this would break them),
18778 the detached data encoding was wrong and public keys obtained using
18779 X509_get_pubkey() weren't freed.
18783 * Add text documentation for the BUFFER functions. Also added a work around
18784 to a Win95 console bug. This was triggered by the password read stuff: the
18785 last character typed gets carried over to the next fread(). If you were
18786 generating a new cert request using 'req' for example then the last
18787 character of the passphrase would be CR which would then enter the first
18792 * Added the new 'Includes OpenSSL Cryptography Software' button as
18793 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18794 button and can be used by applications based on OpenSSL to show the
18795 relationship to the OpenSSL project.
18797 *Ralf S. Engelschall*
18799 * Remove confusing variables in function signatures in files
18800 ssl/ssl_lib.c and ssl/ssl.h.
18802 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18804 * Don't install bss_file.c under PREFIX/include/
18806 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18808 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18809 functions that return function pointers and has support for NT specific
18810 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18811 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18812 unsigned to signed types: this was killing the Win32 compile.
18816 * Add new certificate file to stack functions,
18817 SSL_add_dir_cert_subjects_to_stack() and
18818 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18819 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18820 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18821 This means that Apache-SSL and similar packages don't have to mess around
18822 to add as many CAs as they want to the preferred list.
18826 * Experiment with doxygen documentation. Currently only partially applied to
18828 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18829 openssl.doxy as the configuration file.
18833 * Get rid of remaining C++-style comments which strict C compilers hate.
18835 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18837 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18838 compiled in by default: it has problems with large keys.
18842 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18843 DH private keys and/or callback functions which directly correspond to
18844 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18845 is needed for applications which have to configure certificates on a
18846 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18848 For the RSA certificate situation is makes no difference, but
18849 for the DSA certificate situation this fixes the "no shared cipher"
18850 problem where the OpenSSL cipher selection procedure failed because the
18851 temporary keys were not overtaken from the context and the API provided
18852 no way to reconfigure them.
18853 The new functions now let applications reconfigure the stuff and they
18854 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18855 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18856 non-public-API function ssl_cert_instantiate() is used as a helper
18857 function and also to reduce code redundancy inside ssl_rsa.c.
18859 *Ralf S. Engelschall*
18861 * Move s_server -dcert and -dkey options out of the undocumented feature
18862 area because they are useful for the DSA situation and should be
18863 recognized by the users.
18865 *Ralf S. Engelschall*
18867 * Fix the cipher decision scheme for export ciphers: the export bits are
18868 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18869 SSL_EXP_MASK. So, the original variable has to be used instead of the
18870 already masked variable.
18872 *Richard Levitte <levitte@stacken.kth.se>*
18874 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18876 *Richard Levitte <levitte@stacken.kth.se>*
18878 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18879 from `int` to `unsigned int` because it is a length and initialized by
18880 EVP_DigestFinal() which expects an `unsigned int *`.
18882 *Richard Levitte <levitte@stacken.kth.se>*
18884 * Don't hard-code path to Perl interpreter on shebang line of Configure
18885 script. Instead use the usual Shell->Perl transition trick.
18887 *Ralf S. Engelschall*
18889 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18890 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18891 -noout -modulus` as it's already the case for `openssl rsa -noout
18892 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18893 currently the public key is printed (a decision which was already done by
18894 `openssl dsa -modulus` in the past) which serves a similar purpose.
18895 Additionally the NO_RSA no longer completely removes the whole -modulus
18896 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18899 *Ralf S. Engelschall*
18901 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18902 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18904 *Arne Ansper <arne@ats.cyber.ee>*
18906 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18907 to be added. Now both 'req' and 'ca' can use new objects defined in the
18912 * Add cool BIO that does syslog (or event log on NT).
18914 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18916 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18917 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18918 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18919 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18923 * Add preliminary config info for new extension code.
18927 * Make RSA_NO_PADDING really use no padding.
18929 *Ulf Moeller <ulf@fitug.de>*
18931 * Generate errors when private/public key check is done.
18935 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18936 for some CRL extensions and new objects added.
18940 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18941 key usage extension and fuller support for authority key id.
18945 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18946 padding method for RSA, which is recommended for new applications in PKCS
18947 #1 v2.0 (RFC 2437, October 1998).
18948 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18949 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18950 against Bleichbacher's attack on RSA.
18951 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18954 * Updates to the new SSL compression code
18956 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18958 * Fix so that the version number in the master secret, when passed
18959 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18960 (because the server will not accept higher), that the version number
18961 is 0x03,0x01, not 0x03,0x00
18963 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18965 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18966 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18967 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18971 * Support for RAW extensions where an arbitrary extension can be
18972 created by including its DER encoding. See `apps/openssl.cnf` for
18977 * Make sure latest Perl versions don't interpret some generated C array
18978 code as Perl array code in the crypto/err/err_genc.pl script.
18980 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18982 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18983 not many people have the assembler. Various Win32 compilation fixes and
18984 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18985 build instructions.
18989 * Modify configure script 'Configure' to automatically create crypto/date.h
18990 file under Win32 and also build pem.h from pem.org. New script
18991 util/mkfiles.pl to create the MINFO file on environments that can't do a
18992 'make files': perl util/mkfiles.pl >MINFO should work.
18996 * Major rework of DES function declarations, in the pursuit of correctness
18997 and purity. As a result, many evil casts evaporated, and some weirdness,
18998 too. You may find this causes warnings in your code. Zapping your evil
18999 casts will probably fix them. Mostly.
19003 * Fix for a typo in asn1.h. Bug fix to object creation script
19004 obj_dat.pl. It considered a zero in an object definition to mean
19005 "end of object": none of the objects in objects.h have any zeros
19006 so it wasn't spotted.
19008 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19010 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19011 Masking (CBCM). In the absence of test vectors, the best I have been able
19012 to do is check that the decrypt undoes the encrypt, so far. Send me test
19013 vectors if you have them.
19017 * Correct calculation of key length for export ciphers (too much space was
19018 allocated for null ciphers). This has not been tested!
19022 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19023 message is now correct (it understands "crypto" and "ssl" on its
19024 command line). There is also now an "update" option. This will update
19025 the util/ssleay.num and util/libeay.num files with any new functions.
19027 perl util/mkdef.pl crypto ssl update
19028 it will update them.
19032 * Overhauled the Perl interface:
19033 - ported BN stuff to OpenSSL's different BN library
19034 - made the perl/ source tree CVS-aware
19035 - renamed the package from SSLeay to OpenSSL (the files still contain
19036 their history because I've copied them in the repository)
19037 - removed obsolete files (the test scripts will be replaced
19038 by better Test::Harness variants in the future)
19040 *Ralf S. Engelschall*
19042 * First cut for a very conservative source tree cleanup:
19043 1. merge various obsolete readme texts into doc/ssleay.txt
19044 where we collect the old documents and readme texts.
19045 2. remove the first part of files where I'm already sure that we no
19046 longer need them because of three reasons: either they are just temporary
19047 files which were left by Eric or they are preserved original files where
19048 I've verified that the diff is also available in the CVS via "cvs diff
19049 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19050 the crypto/md/ stuff).
19052 *Ralf S. Engelschall*
19054 * More extension code. Incomplete support for subject and issuer alt
19055 name, issuer and authority key id. Change the i2v function parameters
19056 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19057 what that's for :-) Fix to ASN1 macro which messed up
19058 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19062 * Preliminary support for ENUMERATED type. This is largely copied from the
19067 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19069 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19071 * Make sure `make rehash` target really finds the `openssl` program.
19073 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19075 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19076 like to hear about it if this slows down other processors.
19080 * Add CygWin32 platform information to Configure script.
19082 *Alan Batie <batie@aahz.jf.intel.com>*
19084 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19086 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19088 * New program nseq to manipulate netscape certificate sequences
19092 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19097 * Fixes to BN code. Previously the default was to define BN_RECURSION
19098 but the BN code had some problems that would cause failures when
19099 doing certificate verification and some other functions.
19101 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19103 * Add ASN1 and PEM code to support netscape certificate sequences.
19107 * Add ASN1 and PEM code to support netscape certificate sequences.
19111 * Add several PKIX and private extended key usage OIDs.
19115 * Modify the 'ca' program to handle the new extension code. Modify
19116 openssl.cnf for new extension format, add comments.
19120 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19121 and add a sample to openssl.cnf so req -x509 now adds appropriate
19126 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19127 error code, add initial support to X509_print() and x509 application.
19131 * Takes a deep breath and start adding X509 V3 extension support code. Add
19132 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19133 stuff is currently isolated and isn't even compiled yet.
19137 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19138 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19139 Removed the versions check from X509 routines when loading extensions:
19140 this allows certain broken certificates that don't set the version
19141 properly to be processed.
19145 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19146 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19147 can still be regenerated with "make depend".
19151 * Spelling mistake in C version of CAST-128.
19153 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19155 * Changes to the error generation code. The perl script err-code.pl
19156 now reads in the old error codes and retains the old numbers, only
19157 adding new ones if necessary. It also only changes the .err files if new
19158 codes are added. The makefiles have been modified to only insert errors
19159 when needed (to avoid needlessly modifying header files). This is done
19160 by only inserting errors if the .err file is newer than the auto generated
19161 C file. To rebuild all the error codes from scratch (the old behaviour)
19162 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19163 or delete all the .err files.
19167 * CAST-128 was incorrectly implemented for short keys. The C version has
19168 been fixed, but is untested. The assembler versions are also fixed, but
19169 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19170 to regenerate it if needed.
19171 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19172 Hagino <itojun@kame.net>*
19174 * File was opened incorrectly in randfile.c.
19176 *Ulf Möller <ulf@fitug.de>*
19178 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19179 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19180 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19181 al: it's just almost always a UTCTime. Note this patch adds new error
19182 codes so do a "make errors" if there are problems.
19186 * Correct Linux 1 recognition in config.
19188 *Ulf Möller <ulf@fitug.de>*
19190 * Remove pointless MD5 hash when using DSA keys in ca.
19192 *Anonymous <nobody@replay.com>*
19194 * Generate an error if given an empty string as a cert directory. Also
19195 generate an error if handed NULL (previously returned 0 to indicate an
19196 error, but didn't set one).
19198 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19200 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19204 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19205 parameters. This was causing a warning which killed off the Win32 compile.
19209 * Remove C++ style comments from crypto/bn/bn_local.h.
19211 *Neil Costigan <neil.costigan@celocom.com>*
19213 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19214 based on a text string, looking up short and long names and finally
19215 "dot" format. The "dot" format stuff didn't work. Added new function
19216 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19217 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19218 OID is not part of the table.
19222 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19223 X509_LOOKUP_by_alias().
19227 * Sort openssl functions by name.
19231 * Get the `gendsa` command working and add it to the `list` command. Remove
19232 encryption from sample DSA keys (in case anyone is interested the password
19237 * Make *all* `*_free` functions accept a NULL pointer.
19239 *Frans Heymans <fheymans@isaserver.be>*
19241 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19244 *Anonymous <nobody@replay.com>*
19246 * s_server should send the CAfile as acceptable CAs, not its own cert.
19248 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19250 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19252 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19254 * Temp key "for export" tests were wrong in s3_srvr.c.
19256 *Anonymous <nobody@replay.com>*
19258 * Add prototype for temp key callback functions
19259 SSL_CTX_set_tmp_{rsa,dh}_callback().
19263 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19264 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19268 * X509_name_add_entry() freed the wrong thing after an error.
19270 *Arne Ansper <arne@ats.cyber.ee>*
19272 * rsa_eay.c would attempt to free a NULL context.
19274 *Arne Ansper <arne@ats.cyber.ee>*
19276 * BIO_s_socket() had a broken should_retry() on Windoze.
19278 *Arne Ansper <arne@ats.cyber.ee>*
19280 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19282 *Arne Ansper <arne@ats.cyber.ee>*
19284 * Make sure the already existing X509_STORE->depth variable is initialized
19285 in X509_STORE_new(), but document the fact that this variable is still
19286 unused in the certificate verification process.
19288 *Ralf S. Engelschall*
19290 * Fix the various library and `apps/` files to free up pkeys obtained from
19291 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19295 * Fix reference counting in X509_PUBKEY_get(). This makes
19296 demos/maurice/example2.c work, amongst others, probably.
19298 *Steve Henson and Ben Laurie*
19300 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19301 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19302 are no longer created. This way we have a single and consistent command
19303 line interface `openssl <command>`, similar to `cvs <command>`.
19305 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19307 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19308 BIT STRING wrapper always have zero unused bits.
19312 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19316 * Make the top-level INSTALL documentation easier to understand.
19320 * Makefiles updated to exit if an error occurs in a sub-directory
19321 make (including if user presses ^C) [Paul Sutton]
19323 * Make Montgomery context stuff explicit in RSA data structure.
19327 * Fix build order of pem and err to allow for generated pem.h.
19331 * Fix renumbering bug in X509_NAME_delete_entry().
19335 * Enhanced the err-ins.pl script so it makes the error library number
19336 global and can add a library name. This is needed for external ASN1 and
19337 other error libraries.
19341 * Fixed sk_insert which never worked properly.
19345 * Fix ASN1 macros so they can handle indefinite length constructed
19346 EXPLICIT tags. Some non standard certificates use these: they can now
19351 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19352 into a single doc/ssleay.txt bundle. This way the information is still
19353 preserved but no longer messes up this directory. Now it's new room for
19354 the new set of documentation files.
19356 *Ralf S. Engelschall*
19358 * SETs were incorrectly DER encoded. This was a major pain, because they
19359 shared code with SEQUENCEs, which aren't coded the same. This means that
19360 almost everything to do with SETs or SEQUENCEs has either changed name or
19361 number of arguments.
19363 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19365 * Fix test data to work with the above.
19369 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19370 was already fixed by Eric for 0.9.1 it seems.
19372 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19374 * Autodetect FreeBSD3.
19378 * Fix various bugs in Configure. This affects the following platforms:
19382 unixware-2.0-pentium
19387 * Eliminate generated files from CVS. Reorder tests to regenerate files
19388 before they are needed.
19392 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19396 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19398 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19399 changed SSLeay to OpenSSL in version strings.
19401 *Ralf S. Engelschall*
19403 * Some fixups to the top-level documents.
19407 * Fixed the nasty bug where rsaref.h was not found under compile-time
19408 because the symlink to include/ was missing.
19410 *Ralf S. Engelschall*
19412 * Incorporated the popular no-RSA/DSA-only patches
19413 which allow to compile an RSA-free SSLeay.
19415 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19417 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19418 when "ssleay" is still not found.
19420 *Ralf S. Engelschall*
19422 * Added more platforms to Configure: Cray T3E, HPUX 11,
19424 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19426 * Updated the README file.
19428 *Ralf S. Engelschall*
19430 * Added various .cvsignore files in the CVS repository subdirs
19431 to make a "cvs update" really silent.
19433 *Ralf S. Engelschall*
19435 * Recompiled the error-definition header files and added
19436 missing symbols to the Win32 linker tables.
19438 *Ralf S. Engelschall*
19440 * Cleaned up the top-level documents;
19441 o new files: CHANGES and LICENSE
19442 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19443 o merged COPYRIGHT into LICENSE
19444 o removed obsolete TODO file
19445 o renamed MICROSOFT to INSTALL.W32
19447 *Ralf S. Engelschall*
19449 * Removed dummy files from the 0.9.1b source tree:
19450 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19451 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19452 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19453 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19454 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19456 *Ralf S. Engelschall*
19458 * Added various platform portability fixes.
19462 * The Genesis of the OpenSSL rpject:
19463 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19464 Young and Tim J. Hudson created while they were working for C2Net until
19467 *The OpenSSL Project*
19469 ### Changes between 0.9.0b and 0.9.1b [not released]
19471 * Updated a few CA certificates under certs/
19475 * Changed some BIGNUM api stuff.
19479 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19480 DGUX x86, Linux Alpha, etc.
19484 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19485 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19490 * Add -strparse option to asn1pars program which parses nested
19493 *Dr Stephen Henson <shenson@bigfoot.com>*
19495 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19499 * DSA fix for "ca" program.
19503 * Added "-genkey" option to "dsaparam" program.
19507 * Added RIPE MD160 (rmd160) message digest.
19511 * Added -a (all) option to "ssleay version" command.
19515 * Added PLATFORM define which is the id given to Configure.
19519 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19523 * Extended the ASN.1 parser routines.
19527 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19531 * Added a BN_CTX to the BN library.
19535 * Fixed the weak key values in DES library
19539 * Changed API in EVP library for cipher aliases.
19543 * Added support for RC2/64bit cipher.
19547 * Converted the lhash library to the crypto/mem.c functions.
19551 * Added more recognized ASN.1 object ids.
19555 * Added more RSA padding checks for SSL/TLS.
19559 * Added BIO proxy/filter functionality.
19563 * Added extra_certs to SSL_CTX which can be used
19564 send extra CA certificates to the client in the CA cert chain sending
19565 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19569 * Now Fortezza is denied in the authentication phase because
19570 this is key exchange mechanism is not supported by SSLeay at all.
19574 * Additional PKCS1 checks.
19578 * Support the string "TLSv1" for all TLS v1 ciphers.
19582 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19583 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19587 * Fixed a few memory leaks.
19591 * Fixed various code and comment typos.
19595 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19596 bytes sent in the client random.
19598 *Edward Bishop <ebishop@spyglass.com>*
19602 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19603 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19604 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19605 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19606 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19607 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19608 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19609 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19610 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19611 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19612 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19613 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19614 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19615 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19616 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19617 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19618 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19619 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19620 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19621 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19622 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19623 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19624 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19625 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19626 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19627 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19628 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19629 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19630 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19631 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19632 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19633 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19634 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19635 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19636 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19637 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19638 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19639 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19640 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19641 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19642 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19643 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19644 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19645 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19646 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19647 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19648 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19649 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19650 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19651 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19652 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19653 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19654 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19655 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19656 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19657 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19658 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19659 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19660 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19661 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19662 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19663 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19664 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19665 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19666 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19667 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19668 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19669 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19670 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19671 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19672 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19673 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19674 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19675 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19676 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19677 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19678 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19679 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19680 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19681 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19682 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19683 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19684 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19685 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19686 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19687 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19688 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19689 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19690 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19691 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19692 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19693 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19694 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19695 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19696 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19697 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19698 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19699 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19700 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19701 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19702 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19703 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19704 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19705 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19706 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19707 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19708 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19709 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19710 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19711 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19712 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19713 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19714 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19715 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19716 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19717 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19718 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19719 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19720 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19721 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19722 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19723 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19724 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19725 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19726 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19727 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19728 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19729 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19730 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19731 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19732 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19733 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19734 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19735 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19736 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19737 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19738 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19739 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19740 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19741 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19742 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19743 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19744 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19745 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19746 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19747 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19748 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19749 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19750 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19751 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19752 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19753 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19754 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19755 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19756 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19757 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19758 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19759 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19760 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19761 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19762 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19763 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19764 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19765 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19766 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19767 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19768 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19769 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19770 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19771 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19772 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19773 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19774 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19775 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19776 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655