4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.1.1 and 3.1.2 [xx XXX xxxx]
27 * Do not ignore empty associated data entries with AES-SIV.
29 The AES-SIV algorithm allows for authentication of multiple associated
30 data entries along with the encryption. To authenticate empty data the
31 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
32 with NULL pointer as the output buffer and 0 as the input buffer length.
33 The AES-SIV implementation in OpenSSL just returns success for such call
34 instead of performing the associated data authentication operation.
35 The empty data thus will not be authenticated. ([CVE-2023-2975])
37 Thanks to Juerg Wullschleger (Google) for discovering the issue.
39 The fix changes the authentication tag value and the ciphertext for
40 applications that use empty associated data entries with AES-SIV.
41 To decrypt data encrypted with previous versions of OpenSSL the application
42 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
47 * When building with the `enable-fips` option and using the resulting
48 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
49 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
50 not operate with truncated digests (FIPS 140-3 IG G.R).
54 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
56 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
57 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
59 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
60 numeric text form. For gigantic sub-identifiers, this would take a very
61 long time, the time complexity being O(n^2) where n is the size of that
62 sub-identifier. ([CVE-2023-2650])
64 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
65 IDENTIFIER to canonical numeric text form if the size of that OBJECT
66 IDENTIFIER is 586 bytes or less, and fail otherwise.
68 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
69 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
70 most 128 sub-identifiers, and that the maximum value that each sub-
71 identifier may have is 2^32-1 (4294967295 decimal).
73 For each byte of every sub-identifier, only the 7 lower bits are part of
74 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
75 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
80 * Multiple algorithm implementation fixes for ARM BE platforms.
84 * Added a -pedantic option to fipsinstall that adjusts the various
85 settings to ensure strict FIPS compliance rather than backwards
90 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
91 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
92 trigger a crash of an application using AES-XTS decryption if the memory
93 just after the buffer being decrypted is not mapped.
94 Thanks to Anton Romanov (Amazon) for discovering the issue.
99 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
100 The previous fix for this timing side channel turned out to cause
101 a severe 2-3x performance regression in the typical use case
102 compared to 3.0.7. The new fix uses existing constant time
103 code paths, and restores the previous performance level while
104 fully eliminating all existing timing side channels.
105 The fix was developed by Bernd Edlinger with testing support
110 * Add FIPS provider configuration option to disallow the use of
111 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
112 The option '-no_drbg_truncated_digests' can optionally be
113 supplied to 'openssl fipsinstall'.
117 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
118 that it does not enable policy checking. Thanks to David Benjamin for
119 discovering this issue.
124 * Fixed an issue where invalid certificate policies in leaf certificates are
125 silently ignored by OpenSSL and other certificate policy checks are skipped
126 for that certificate. A malicious CA could use this to deliberately assert
127 invalid certificate policies in order to circumvent policy checking on the
128 certificate altogether.
133 * Limited the number of nodes created in a policy tree to mitigate
134 against CVE-2023-0464. The default limit is set to 1000 nodes, which
135 should be sufficient for most installations. If required, the limit
136 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
137 time define to a desired maximum number of nodes or zero to allow
143 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
145 * Add FIPS provider configuration option to enforce the
146 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
147 The option '-ems-check' can optionally be supplied to
148 'openssl fipsinstall'.
152 * The FIPS provider includes a few non-approved algorithms for
153 backward compatibility purposes and the "fips=yes" property query
154 must be used for all algorithm fetches to ensure FIPS compliance.
156 The algorithms that are included but not approved are Triple DES ECB,
157 Triple DES CBC and EdDSA.
161 * Added support for KMAC in KBKDF.
165 * RNDR and RNDRRS support in provider functions to provide
166 random number generation for Arm CPUs (aarch64).
170 * s_client and s_server apps now explicitly say when the TLS version
171 does not include the renegotiation mechanism. This avoids confusion
172 between that scenario versus when the TLS version includes secure
173 renegotiation but the peer lacks support for it.
177 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
179 *Tomasz Kantecki, Andrey Matyukov*
181 * The various OBJ_* functions have been made thread safe.
185 * Parallel dual-prime 1536/2048-bit modular exponentiation for
186 AVX512_IFMA capable processors.
188 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
190 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
191 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
192 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
193 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
194 `OPENSSL_NO_DEPRECATED_3_1`.
196 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
197 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
198 definitions for these functions regardless of whether
199 `OPENSSL_NO_DEPRECATED_3_1` is defined.
201 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
202 functions regardless of whether they are using them. It is recommended that
203 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
207 * When generating safe-prime DH parameters set the recommended private key
208 length equivalent to minimum key lengths as in RFC 7919.
212 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
213 maximum size that is smaller or equal to the digest length to comply with
214 FIPS 186-4 section 5. This is implemented by a new option
215 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
216 `rsa_pss_saltlen` parameter, which is now the default. Signature
217 verification is not affected by this change and continues to work as before.
224 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
225 listed here are only a brief description.
226 The migration guide contains more detailed information related to new features,
227 breaking changes, and mappings for the large list of deprecated functions.
229 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
231 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
233 * Fixed NULL dereference during PKCS7 data verification.
235 A NULL pointer can be dereferenced when signatures are being
236 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
237 algorithm used for the signature is known to the OpenSSL library but
238 the implementation of the hash algorithm is not available the digest
239 initialization will fail. There is a missing check for the return
240 value from the initialization function which later leads to invalid
241 usage of the digest API most likely leading to a crash.
244 PKCS7 data is processed by the SMIME library calls and also by the
245 time stamp (TS) library calls. The TLS implementation in OpenSSL does
246 not call these functions however third party applications would be
247 affected if they call these functions to verify signatures on untrusted
252 * Fixed X.400 address type confusion in X.509 GeneralName.
254 There is a type confusion vulnerability relating to X.400 address processing
255 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
256 but the public structure definition for GENERAL_NAME incorrectly specified
257 the type of the x400Address field as ASN1_TYPE. This field is subsequently
258 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
261 When CRL checking is enabled (i.e. the application sets the
262 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
263 pass arbitrary pointers to a memcmp call, enabling them to read memory
264 contents or enact a denial of service.
269 * Fixed NULL dereference validating DSA public key.
271 An invalid pointer dereference on read can be triggered when an
272 application tries to check a malformed DSA public key by the
273 EVP_PKEY_public_check() function. This will most likely lead
274 to an application crash. This function can be called on public
275 keys supplied from untrusted sources which could allow an attacker
276 to cause a denial of service attack.
278 The TLS implementation in OpenSSL does not call this function
279 but applications might call the function if there are additional
280 security requirements imposed by standards such as FIPS 140-3.
283 *Shane Lontis, Tomáš Mráz*
285 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
287 An invalid pointer dereference on read can be triggered when an
288 application tries to load malformed PKCS7 data with the
289 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
291 The result of the dereference is an application crash which could
292 lead to a denial of service attack. The TLS implementation in OpenSSL
293 does not call this function however third party applications might
294 call these functions on untrusted data.
299 * Fixed Use-after-free following BIO_new_NDEF.
301 The public API function BIO_new_NDEF is a helper function used for
302 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
303 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
304 be called directly by end user applications.
306 The function receives a BIO from the caller, prepends a new BIO_f_asn1
307 filter BIO onto the front of it to form a BIO chain, and then returns
308 the new head of the BIO chain to the caller. Under certain conditions,
309 for example if a CMS recipient public key is invalid, the new filter BIO
310 is freed and the function returns a NULL result indicating a failure.
311 However, in this case, the BIO chain is not properly cleaned up and the
312 BIO passed by the caller still retains internal pointers to the previously
313 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
314 then a use-after-free will occur. This will most likely result in a crash.
317 *Viktor Dukhovni, Matt Caswell*
319 * Fixed Double free after calling PEM_read_bio_ex.
321 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
322 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
323 data. If the function succeeds then the "name_out", "header" and "data"
324 arguments are populated with pointers to buffers containing the relevant
325 decoded data. The caller is responsible for freeing those buffers. It is
326 possible to construct a PEM file that results in 0 bytes of payload data.
327 In this case PEM_read_bio_ex() will return a failure code but will populate
328 the header argument with a pointer to a buffer that has already been freed.
329 If the caller also frees this buffer then a double free will occur. This
330 will most likely lead to a crash.
332 The functions PEM_read_bio() and PEM_read() are simple wrappers around
333 PEM_read_bio_ex() and therefore these functions are also directly affected.
335 These functions are also called indirectly by a number of other OpenSSL
336 functions including PEM_X509_INFO_read_bio_ex() and
337 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
338 internal uses of these functions are not vulnerable because the caller does
339 not free the header argument if PEM_read_bio_ex() returns a failure code.
342 *Kurt Roeckx, Matt Caswell*
344 * Fixed Timing Oracle in RSA Decryption.
346 A timing based side channel exists in the OpenSSL RSA Decryption
347 implementation which could be sufficient to recover a plaintext across
348 a network in a Bleichenbacher style attack. To achieve a successful
349 decryption an attacker would have to be able to send a very large number
350 of trial messages for decryption. The vulnerability affects all RSA padding
351 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
354 *Dmitry Belyavsky, Hubert Kario*
356 * Fixed X.509 Name Constraints Read Buffer Overflow.
358 A read buffer overrun can be triggered in X.509 certificate verification,
359 specifically in name constraint checking. The read buffer overrun might
360 result in a crash which could lead to a denial of service attack.
361 In a TLS client, this can be triggered by connecting to a malicious
362 server. In a TLS server, this can be triggered if the server requests
363 client authentication and a malicious client connects.
368 * Fixed X.509 Policy Constraints Double Locking security issue.
370 If an X.509 certificate contains a malformed policy constraint and
371 policy processing is enabled, then a write lock will be taken twice
372 recursively. On some operating systems (most widely: Windows) this
373 results in a denial of service when the affected process hangs. Policy
374 processing being enabled on a publicly facing server is not considered
375 to be a common setup.
380 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
381 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
382 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
383 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
384 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
385 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
386 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
387 for legacy EC and SM2 keys is also changed similarly to honor the
388 equivalent conversion format flag as specified in the underlying
389 `EC_KEY` object being exported to a provider, when this function is
390 called through `EVP_PKEY_export()`.
394 * Fixed a type confusion vulnerability relating to X.400 address processing
395 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
396 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
397 vulnerability may allow an attacker who can provide a certificate chain and
398 CRL (neither of which need have a valid signature) to pass arbitrary pointers
399 to a `memcmp` call, creating a possible read primitive, subject to some
400 constraints. Refer to the advisory for more information. Thanks to David
401 Benjamin for discovering this issue. ([CVE-2023-0286])
403 This issue has been fixed by changing the public header file definition of
404 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
405 possible for any existing application to successfully use the existing
406 definition; however, if any application references the `x400Address` field
407 (e.g. in dead code), note that the type of this field has changed. There is
412 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
414 * Fixed two buffer overflows in punycode decoding functions.
416 A buffer overrun can be triggered in X.509 certificate verification,
417 specifically in name constraint checking. Note that this occurs after
418 certificate chain signature verification and requires either a CA to
419 have signed the malicious certificate or for the application to continue
420 certificate verification despite failure to construct a path to a trusted
423 In a TLS client, this can be triggered by connecting to a malicious
424 server. In a TLS server, this can be triggered if the server requests
425 client authentication and a malicious client connects.
427 An attacker can craft a malicious email address to overflow
428 an arbitrary number of bytes containing the `.` character (decimal 46)
429 on the stack. This buffer overflow could result in a crash (causing a
433 An attacker can craft a malicious email address to overflow four
434 attacker-controlled bytes on the stack. This buffer overflow could
435 result in a crash (causing a denial of service) or potentially remote code
436 execution depending on stack layout for any given platform/compiler.
441 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
442 parameters in OpenSSL code.
443 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
444 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
445 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
446 Using these invalid names may cause algorithms to use slower methods
447 that ignore the CRT parameters.
451 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
456 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
457 data to be signed before signing the certificate.
461 * Added RIPEMD160 to the default provider.
465 * Ensured that the key share group sent or accepted for the key exchange
466 is allowed for the protocol version.
470 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
472 * OpenSSL supports creating a custom cipher via the legacy
473 EVP_CIPHER_meth_new() function and associated function calls. This function
474 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
475 to use the new provider mechanism in order to implement custom ciphers.
477 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
478 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
479 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
480 and decryption initialisation functions). Instead of using the custom cipher
481 directly it incorrectly tries to fetch an equivalent cipher from the
482 available providers. An equivalent cipher is found based on the NID passed to
483 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
484 given cipher. However it is possible for an application to incorrectly pass
485 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
486 is used in this way the OpenSSL encryption/decryption initialisation function
487 will match the NULL cipher as being equivalent and will fetch this from the
488 available providers. This will succeed if the default provider has been
489 loaded (or if a third party provider has been loaded that offers this
490 cipher). Using the NULL cipher means that the plaintext is emitted as the
493 Applications are only affected by this issue if they call
494 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
495 encryption/decryption initialisation function. Applications that only use
496 SSL/TLS are not impacted by this issue.
501 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
506 * Fixed the linux-mips64 Configure target which was missing the
507 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
512 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
517 * Correctly handle a retransmitted ClientHello in DTLS
521 * Fixed detection of ktls support in cross-compile environment on Linux
525 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
530 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
531 report correct results in some cases
535 * Fix UWP builds by defining VirtualLock
539 * For known safe primes use the minimum key length according to RFC 7919.
540 Longer private key sizes unnecessarily raise the cycles needed to compute the
541 shared secret without any increase of the real security. This fixes a
542 regression from 1.1.1 where these shorter keys were generated for the known
547 * Added the loongarch64 target
551 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
552 only passed to the FIPS provider and not to the default or legacy provider.
556 * Fixed reported performance degradation on aarch64. Restored the
557 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
558 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
559 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
560 The new algorithm is still used for 32 bit targets.
564 * Added a missing header for memcmp that caused compilation failure on some
569 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
571 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
572 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
573 This issue makes the RSA implementation with 2048 bit private keys
574 incorrect on such machines and memory corruption will happen during
575 the computation. As a consequence of the memory corruption an attacker
576 may be able to trigger a remote code execution on the machine performing
579 SSL/TLS servers or other servers using 2048 bit RSA private keys running
580 on machines supporting AVX512IFMA instructions of the X86_64 architecture
581 are affected by this issue.
586 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
587 implementation would not encrypt the entirety of the data under some
588 circumstances. This could reveal sixteen bytes of data that was
589 preexisting in the memory that wasn't written. In the special case of
590 "in place" encryption, sixteen bytes of the plaintext would be revealed.
592 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
593 they are both unaffected.
596 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
598 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
600 * In addition to the c_rehash shell command injection identified in
601 CVE-2022-1292, further bugs where the c_rehash script does not
602 properly sanitise shell metacharacters to prevent command injection have been
605 When the CVE-2022-1292 was fixed it was not discovered that there
606 are other places in the script where the file names of certificates
607 being hashed were possibly passed to a command executed through the shell.
609 This script is distributed by some operating systems in a manner where
610 it is automatically executed. On such operating systems, an attacker
611 could execute arbitrary commands with the privileges of the script.
613 Use of the c_rehash script is considered obsolete and should be replaced
614 by the OpenSSL rehash command line tool.
617 *Daniel Fiala, Tomáš Mráz*
619 * Case insensitive string comparison no longer uses locales. It has instead
620 been directly implemented.
624 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
626 * Case insensitive string comparison is reimplemented via new locale-agnostic
627 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
628 comparison. The previous implementation had problems when the Turkish locale
633 * Fixed a bug in the c_rehash script which was not properly sanitising shell
634 metacharacters to prevent command injection. This script is distributed by
635 some operating systems in a manner where it is automatically executed. On
636 such operating systems, an attacker could execute arbitrary commands with the
637 privileges of the script.
639 Use of the c_rehash script is considered obsolete and should be replaced
640 by the OpenSSL rehash command line tool.
645 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
646 certificate on an OCSP response. The bug caused the function in the case
647 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
648 response (meaning a successful verification) even in the case where the
649 response signing certificate fails to verify.
651 It is anticipated that most users of `OCSP_basic_verify` will not use the
652 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
653 a negative value (indicating a fatal error) in the case of a certificate
654 verification failure. The normal expected return value in this case would be
657 This issue also impacts the command line OpenSSL "ocsp" application. When
658 verifying an ocsp response with the "-no_cert_checks" option the command line
659 application will report that the verification is successful even though it
660 has in fact failed. In this case the incorrect successful response will also
661 be accompanied by error messages showing the failure and contradicting the
662 apparently successful result.
667 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
668 AAD data as the MAC key. This made the MAC key trivially predictable.
670 An attacker could exploit this issue by performing a man-in-the-middle attack
671 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
672 that the modified data would still pass the MAC integrity check.
674 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
675 endpoint will always be rejected by the recipient and the connection will
676 fail at that point. Many application protocols require data to be sent from
677 the client to the server first. Therefore, in such a case, only an OpenSSL
678 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
680 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
681 sent in both directions. In this case both clients and servers could be
682 affected, regardless of the application protocol.
684 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
685 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
686 the handshake when using this ciphersuite.
688 The confidentiality of data is not impacted by this issue, i.e. an attacker
689 cannot decrypt data that has been encrypted using this ciphersuite - they can
692 In order for this attack to work both endpoints must legitimately negotiate
693 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
694 OpenSSL 3.0, and is not available within the default provider or the default
695 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
696 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
697 following must have occurred:
699 1) OpenSSL must have been compiled with the (non-default) compile time option
700 enable-weak-ssl-ciphers
702 2) OpenSSL must have had the legacy provider explicitly loaded (either
703 through application code or via configuration)
705 3) The ciphersuite must have been explicitly added to the ciphersuite list
707 4) The libssl security level must have been set to 0 (default is 1)
709 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
711 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
712 others that both endpoints have in common
717 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
718 occupied by the removed hash table entries.
720 This function is used when decoding certificates or keys. If a long lived
721 process periodically decodes certificates or keys its memory usage will
722 expand without bounds and the process might be terminated by the operating
723 system causing a denial of service. Also traversing the empty hash table
724 entries will take increasingly more time.
726 Typically such long lived processes might be TLS clients or TLS servers
727 configured to accept client certificate authentication.
730 *Hugo Landau, Aliaksei Levin*
732 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
733 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
734 statistics are no longer supported. For compatibility, these statistics are
735 still listed in the output but are now always reported as zero.
739 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
741 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
742 for non-prime moduli.
744 Internally this function is used when parsing certificates that contain
745 elliptic curve public keys in compressed form or explicit elliptic curve
746 parameters with a base point encoded in compressed form.
748 It is possible to trigger the infinite loop by crafting a certificate that
749 has invalid explicit curve parameters.
751 Since certificate parsing happens prior to verification of the certificate
752 signature, any process that parses an externally supplied certificate may thus
753 be subject to a denial of service attack. The infinite loop can also be
754 reached when parsing crafted private keys as they can contain explicit
755 elliptic curve parameters.
757 Thus vulnerable situations include:
759 - TLS clients consuming server certificates
760 - TLS servers consuming client certificates
761 - Hosting providers taking certificates or private keys from customers
762 - Certificate authorities parsing certification requests from subscribers
763 - Anything else which parses ASN.1 elliptic curve parameters
765 Also any other applications that use the BN_mod_sqrt() where the attacker
766 can control the parameter values are vulnerable to this DoS issue.
771 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
772 to the list of ciphersuites providing Perfect Forward Secrecy as
773 required by SECLEVEL >= 3.
775 *Dmitry Belyavskiy, Nicola Tuveri*
777 * Made the AES constant time code for no-asm configurations
778 optional due to the resulting 95% performance degradation.
779 The AES constant time code can be enabled, for no assembly
780 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
784 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
789 * The negative return value handling of the certificate verification callback
790 was reverted. The replacement is to set the verification retry state with
791 the SSL_set_retry_verify() function.
795 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
797 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
798 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
799 verify a certificate supplied by a server. That function may return a
800 negative return value to indicate an internal error (for example out of
801 memory). Such a negative return value is mishandled by OpenSSL and will cause
802 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
803 success and a subsequent call to SSL_get_error() to return the value
804 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
805 returned by OpenSSL if the application has previously called
806 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
807 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
808 totally unexpected and applications may not behave correctly as a result. The
809 exact behaviour will depend on the application but it could result in
810 crashes, infinite loops or other similar incorrect responses.
812 This issue is made more serious in combination with a separate bug in OpenSSL
813 3.0 that will cause X509_verify_cert() to indicate an internal error when
814 processing a certificate chain. This will occur where a certificate does not
815 include the Subject Alternative Name extension but where a Certificate
816 Authority has enforced name constraints. This issue can occur even with valid
822 * Corrected a few file name and file reference bugs in the build,
823 installation and setup scripts, which lead to installation verification
824 failures. Slightly enhanced the installation verification script.
828 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
833 * Fixed PVK encoder to properly query for the passphrase.
837 * Multiple fixes in the OSSL_HTTP API functions.
841 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
842 OSSL_PARAM_INTEGER data type and return error on negative numbers
843 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
844 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
848 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
852 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
856 * Multiple threading fixes.
860 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
864 * Allow fetching an operation from the provider that owns an unexportable key
865 as a fallback if that is still allowed by the property query.
869 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
871 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
876 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
877 S390X capability vector to zero. This simplifies testing of different code
878 paths on S390X architecture.
882 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
883 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
884 SP 800-38D". The communication will fail at this point.
888 * The EC_GROUP_clear_free() function is deprecated as there is nothing
889 confidential in EC_GROUP data.
893 * The byte order mark (BOM) character is ignored if encountered at the
894 beginning of a PEM-formatted file.
898 * Added CMS support for the Russian GOST algorithms.
902 * Due to move of the implementation of cryptographic operations
903 to the providers, validation of various operation parameters can
904 be postponed until the actual operation is executed where previously
905 it happened immediately when an operation parameter was set.
907 For example when setting an unsupported curve with
908 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
909 fail but later keygen operations with the EVP_PKEY_CTX will fail.
911 *OpenSSL team members and many third party contributors*
913 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
914 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
915 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
916 instead to retrieve these algorithms from a provider.
920 * On build targets where the multilib postfix is set in the build
921 configuration the libdir directory was changing based on whether
922 the lib directory with the multilib postfix exists on the system
923 or not. This unpredictable behavior was removed and eventual
924 multilib postfix is now always added to the default libdir. Use
925 `--libdir=lib` to override the libdir if adding the postfix is
930 * The triple DES key wrap functionality now conforms to RFC 3217 but is
931 no longer interoperable with OpenSSL 1.1.1.
935 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
936 function codes, this function can only cause problems for calling
941 * Add a configurable flag to output date formats as ISO 8601. Does not
942 change the default date format.
946 * Version of MSVC earlier than 1300 could get link warnings, which could
947 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
948 Support for this flag has been removed.
952 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
953 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
954 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
955 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
956 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
960 * The signatures of the functions to get and set options on SSL and
961 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
962 Some source code changes may be required.
966 * The public definitions of conf_method_st and conf_st have been
967 deprecated. They will be made opaque in a future release.
969 *Rich Salz and Tomáš Mráz*
971 * Client-initiated renegotiation is disabled by default. To allow it, use
972 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
973 flag, or the "ClientRenegotiation" config parameter as appropriate.
977 * Add "abspath" and "includedir" pragma's to config files, to prevent,
978 or modify relative pathname inclusion.
982 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
983 validated. Please consult the README-FIPS and
984 README-PROVIDERS files, as well as the migration guide.
986 *OpenSSL team members and many third party contributors*
988 * For the key types DH and DHX the allowed settable parameters are now different.
992 * The openssl commands that read keys, certificates, and CRLs now
993 automatically detect the PEM or DER format of the input files.
995 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
997 * Added enhanced PKCS#12 APIs which accept a library context.
1001 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1005 * Added support for Kernel TLS (KTLS).
1007 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1009 * Support for RFC 5746 secure renegotiation is now required by default for
1010 SSL or TLS connections to succeed.
1014 * The signature of the `copy` functional parameter of the
1015 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1016 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1017 the signature of the `pub_decode` functional parameter of the
1018 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1019 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1023 * The error return values from some control calls (ctrl) have changed.
1027 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1031 * Many functions in the EVP_ namespace that are getters of values from
1032 implementations or contexts were renamed to include get or get0 in their
1033 names. Old names are provided as macro aliases for compatibility and
1038 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1039 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1040 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1045 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1048 * The output from the command line applications may have minor
1053 * The output from numerous "printing" may have minor changes.
1057 * Windows thread synchronization uses read/write primitives (SRWLock) when
1058 supported by the OS, otherwise CriticalSection continues to be used.
1062 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1063 work on read only BIO source/sinks that do not support these functions.
1064 This allows piping or redirection of a file BIO using stdin to be buffered
1065 into memory. This is used internally in OSSL_DECODER_from_bio().
1069 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1070 this function would return one of the values OSSL_STORE_INFO_NAME,
1071 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1072 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1073 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1074 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1075 using this function should be amended to handle the changed return value.
1079 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1080 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1081 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1082 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1083 contains more than one certificate identifier: This means that all
1084 certificates referenced there MUST be part of the validation chain.
1088 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1089 RC5, DESX and DES have been moved to the legacy provider.
1093 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1094 RIPEMD-160 have been moved to the legacy provider.
1098 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1103 * The deprecated functions EVP_PKEY_get0_RSA(),
1104 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1105 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1106 well as the similarly named "get1" functions behave differently in
1111 * A number of functions handling low-level keys or engines were deprecated
1112 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1113 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1114 EVP_PKEY_get0_siphash().
1118 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1119 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1120 will need to load the legacy crypto provider. This includes these PBE
1121 algorithms which use this KDF:
1122 - NID_pbeWithMD2AndDES_CBC
1123 - NID_pbeWithMD5AndDES_CBC
1124 - NID_pbeWithSHA1AndRC2_CBC
1125 - NID_pbeWithMD2AndRC2_CBC
1126 - NID_pbeWithMD5AndRC2_CBC
1127 - NID_pbeWithSHA1AndDES_CBC
1131 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1132 BIO_debug_callback() functions.
1136 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1137 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1141 * The RAND_METHOD APIs have been deprecated.
1145 * The SRP APIs have been deprecated.
1149 * Add a compile time option to prevent the caching of provider fetched
1150 algorithms. This is enabled by including the no-cached-fetch option
1151 at configuration time.
1155 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1156 count of PKCS12_DEFAULT_ITER.
1158 *Tomáš Mráz and Sahana Prasad*
1160 * The openssl speed command does not use low-level API calls anymore.
1164 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1167 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1169 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1173 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1174 providers may supply their own group implementations (using either the "key
1175 exchange" or the "key encapsulation" methods) which will automatically be
1176 detected and used by libssl.
1178 *Matt Caswell, Nicola Tuveri*
1180 * The undocumented function X509_certificate_type() has been deprecated;
1184 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1188 * Removed RSA padding mode for SSLv23 (which was only used for
1189 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1190 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1195 * Deprecated the obsolete X9.31 RSA key generation related functions.
1197 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1198 is not allowed to return a value > 1, this is no more taken as failure.
1200 *Viktor Dukhovni and David von Oheimb*
1202 * Deprecated the obsolete X9.31 RSA key generation related functions
1203 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1204 BN_X931_generate_prime_ex().
1208 * The default key generation method for the regular 2-prime RSA keys was
1209 changed to the FIPS 186-4 B.3.6 method.
1213 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1217 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1221 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1222 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1224 *Rich Salz, Richard Levitte, and David von Oheimb*
1226 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1230 * Deprecated `OCSP_parse_url()`.
1234 * Validation of SM2 keys has been separated from the validation of regular EC
1239 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1240 switches: a validation failure triggers an early exit, returning a failure
1241 exit status to the parent process.
1245 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1246 to ignore unknown ciphers.
1250 * The `-cipher-commands` and `-digest-commands` options
1251 of the command line utility `list` have been deprecated.
1252 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1256 * Added convenience functions for generating asymmetric key pairs:
1257 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1258 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1262 * All of the low-level EC_KEY functions have been deprecated.
1264 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1266 * Deprecated all the libcrypto and libssl error string loading
1271 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1272 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1277 * The `-crypt` option to the `passwd` command line tool has been removed.
1281 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1286 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1290 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1291 EVP_PKEY_get1_tls_encodedpoint().
1295 * The security callback, which can be customised by application code, supports
1296 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1297 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1301 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1302 interface. Their functionality remains unchanged.
1306 * Added new option for 'openssl list', '-providers', which will display the
1307 list of loaded providers, their names, version and status. It optionally
1308 displays their gettable parameters.
1312 * Removed EVP_PKEY_set_alias_type().
1316 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1317 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1321 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1322 parameters are still checked for type safety at compile time via helper
1327 * Remove the RAND_DRBG API
1329 *Paul Dale and Matthias St. Pierre*
1331 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1332 as well as actual hostnames.
1336 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1337 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1338 conversely, silently ignore DTLS protocol version bounds when configuring
1339 TLS-based contexts. The commands can be repeated to set bounds of both
1340 types. The same applies with the corresponding "min_protocol" and
1341 "max_protocol" command-line switches, in case some application uses both TLS
1344 SSL_CTX instances that are created for a fixed protocol version (e.g.
1345 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1346 attempts to apply bounds to these protocol versions would result in an
1347 error. Now only the "version-flexible" SSL_CTX instances are subject to
1348 limits in configuration files in command-line options.
1352 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1357 * Reworked the recorded ERR codes to make better space for system errors.
1358 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1359 given code is a system error (true) or an OpenSSL error (false).
1363 * Reworked the test perl framework to better allow parallel testing.
1365 *Nicola Tuveri and David von Oheimb*
1367 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1368 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1372 * 'Configure' has been changed to figure out the configuration target if
1373 none is given on the command line. Consequently, the 'config' script is
1374 now only a mere wrapper. All documentation is changed to only mention
1377 *Rich Salz and Richard Levitte*
1379 * Added a library context `OSSL_LIB_CTX` that applications as well as
1380 other libraries can use to form a separate context within which
1381 libcrypto operations are performed.
1385 * Added various `_ex` functions to the OpenSSL API that support using
1386 a non-default `OSSL_LIB_CTX`.
1390 * Handshake now fails if Extended Master Secret extension is dropped
1395 * Dropped interactive mode from the `openssl` program.
1399 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1401 *David von Oheimb and Shane Lontis*
1403 * Deprecated `EC_METHOD_get_field_type()`.
1407 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1408 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1409 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1413 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1417 * Add CAdES-BES signature verification support, mostly derived
1418 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1420 *Filipe Raimundo da Silva*
1422 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1426 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1427 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1431 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1435 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1436 EC_KEY_precompute_mult().
1440 * Deprecated EC_POINTs_mul().
1444 * Removed FIPS_mode() and FIPS_mode_set().
1448 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1452 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1453 EC_POINT_get_Jprojective_coordinates_GFp().
1457 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1458 arrays to be more easily constructed via a series of utility functions.
1459 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1460 the various push functions and finally convert to a passable OSSL_PARAM
1461 array using OSSL_PARAM_BLD_to_param().
1465 * The security strength of SHA1 and MD5 based signatures in TLS has been
1470 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1471 contain a provider side internal key.
1475 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1479 * Project text documents not yet having a proper file name extension
1480 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1481 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1485 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1486 have been converted to Markdown with the goal to produce documents
1487 which not only look pretty when viewed online in the browser, but
1488 remain well readable inside a plain text editor.
1490 To achieve this goal, a 'minimalistic' Markdown style has been applied
1491 which avoids formatting elements that interfere too much with the
1492 reading flow in the text file. For example, it
1494 * avoids [ATX headings][] and uses [setext headings][] instead
1495 (which works for `<h1>` and `<h2>` headings only).
1496 * avoids [inline links][] and uses [reference links][] instead.
1497 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1499 [ATX headings]: https://github.github.com/gfm/#atx-headings
1500 [setext headings]: https://github.github.com/gfm/#setext-headings
1501 [inline links]: https://github.github.com/gfm/#inline-link
1502 [reference links]: https://github.github.com/gfm/#reference-link
1503 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1504 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1506 *Matthias St. Pierre*
1508 * The test suite is changed to preserve results of each test recipe.
1509 A new directory test-runs/ with subdirectories named like the
1510 test recipes are created in the build tree for this purpose.
1514 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1515 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1516 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1518 *David von Oheimb, Martin Peylo*
1520 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1521 It supports arbitrary request and response content types, GET redirection,
1522 TLS, connections via HTTP(S) proxies, connections and exchange via
1523 user-defined BIOs (allowing implicit connections), persistent connections,
1524 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1525 The legacy OCSP-focused (and only partly documented) API
1526 is retained for backward compatibility, while most of it is deprecated.
1530 * Added `util/check-format.pl`, a tool for checking adherence to the
1531 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1532 The checks performed are incomplete and yield some false positives.
1533 Still the tool should be useful for detecting most typical glitches.
1537 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1538 If domain name resolution yields multiple IP addresses all of them are tried
1539 after `connect()` failures.
1543 * All of the low-level RSA functions have been deprecated.
1547 * X509 certificates signed using SHA1 are no longer allowed at security
1552 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1553 modified to use PKEY APIs. These commands are now in maintenance mode
1554 and no new features will be added to them.
1558 * The command line utility rsautl has been deprecated.
1562 * The command line utilities genrsa and rsa have been modified to use PKEY
1563 APIs. They now write PKCS#8 keys by default. These commands are now in
1564 maintenance mode and no new features will be added to them.
1568 * All of the low-level DH functions have been deprecated.
1570 *Paul Dale and Matt Caswell*
1572 * All of the low-level DSA functions have been deprecated.
1576 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1577 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1581 * Deprecated low-level ECDH and ECDSA functions.
1585 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1589 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1590 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1591 a new formulation to include all the things it can be used for,
1592 as well as words of caution.
1596 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1600 * All of the low-level HMAC functions have been deprecated.
1602 *Paul Dale and David von Oheimb*
1604 * Over two thousand fixes were made to the documentation, including:
1605 - Common options (such as -rand/-writerand, TLS version control, etc)
1606 were refactored and point to newly-enhanced descriptions in openssl.pod.
1607 - Added style conformance for all options (with help from Richard Levitte),
1608 documented all reported missing options, added a CI build to check
1609 that all options are documented and that no unimplemented options
1611 - Documented some internals, such as all use of environment variables.
1612 - Addressed all internal broken L<> references.
1616 * All of the low-level CMAC functions have been deprecated.
1620 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1621 functions have been deprecated.
1623 *Paul Dale and David von Oheimb*
1625 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1626 set of functions. The documentation mentioned negative values for some
1627 errors, but this was never the case, so the mention of negative values
1630 Code that followed the documentation and thereby check with something
1631 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1635 * All of the low-level cipher functions have been deprecated.
1637 *Matt Caswell and Paul Dale*
1639 * Removed include/openssl/opensslconf.h.in and replaced it with
1640 include/openssl/configuration.h.in, which differs in not including
1641 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1642 was added to include both.
1644 This allows internal hacks where one might need to modify the set
1645 of configured macros, for example this if deprecated symbols are
1646 still supposed to be available internally:
1648 #include <openssl/configuration.h>
1650 #undef OPENSSL_NO_DEPRECATED
1651 #define OPENSSL_SUPPRESS_DEPRECATED
1653 #include <openssl/macros.h>
1655 This should not be used by applications that use the exported
1656 symbols, as that will lead to linking errors.
1660 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1661 used in exponentiation with 512-bit moduli. No EC algorithms are
1662 affected. Analysis suggests that attacks against 2-prime RSA1024,
1663 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1664 difficult to perform and are not believed likely. Attacks against DH512
1665 are considered just feasible. However, for an attack the target would
1666 have to re-use the DH512 private key, which is not recommended anyway.
1667 Also applications directly using the low-level API BN_mod_exp may be
1668 affected if they use BN_FLG_CONSTTIME.
1673 * Most memory-debug features have been deprecated, and the functionality
1674 replaced with no-ops.
1678 * Added documentation for the STACK API.
1682 * Introduced a new method type and API, OSSL_ENCODER, to represent
1683 generic encoders. These do the same sort of job that PEM writers
1684 and d2i functions do, but with support for methods supplied by
1685 providers, and the possibility for providers to support other
1690 * Introduced a new method type and API, OSSL_DECODER, to represent
1691 generic decoders. These do the same sort of job that PEM readers
1692 and i2d functions do, but with support for methods supplied by
1693 providers, and the possibility for providers to support other
1698 * Added a .pragma directive to the syntax of configuration files, to
1699 allow varying behavior in a supported and predictable manner.
1700 Currently added pragma:
1704 This allows dollar signs to be a keyword character unless it's
1705 followed by a opening brace or parenthesis. This is useful for
1706 platforms where dollar signs are commonly used in names, such as
1707 volume names and system directory names on VMS.
1711 * Added functionality to create an EVP_PKEY from user data.
1715 * Change the interpretation of the '--api' configuration option to
1716 mean that this is a desired API compatibility level with no
1717 further meaning. The previous interpretation, that this would
1718 also mean to remove all deprecated symbols up to and including
1719 the given version, no requires that 'no-deprecated' is also used
1720 in the configuration.
1722 When building applications, the desired API compatibility level
1723 can be set with the OPENSSL_API_COMPAT macro like before. For
1724 API compatibility version below 3.0, the old style numerical
1725 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1726 For version 3.0 and on, the value is expected to be the decimal
1727 value calculated from the major and minor version like this:
1729 MAJOR * 10000 + MINOR * 100
1733 -DOPENSSL_API_COMPAT=30000 For 3.0
1734 -DOPENSSL_API_COMPAT=30200 For 3.2
1736 To hide declarations that are deprecated up to and including the
1737 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1738 given when building the application as well.
1742 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1743 access to certificate and CRL stores via URIs and OSSL_STORE
1746 This adds the following functions:
1748 - X509_LOOKUP_store()
1749 - X509_STORE_load_file()
1750 - X509_STORE_load_path()
1751 - X509_STORE_load_store()
1752 - SSL_add_store_cert_subjects_to_stack()
1753 - SSL_CTX_set_default_verify_store()
1754 - SSL_CTX_load_verify_file()
1755 - SSL_CTX_load_verify_dir()
1756 - SSL_CTX_load_verify_store()
1760 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1761 The presence of this system service is determined at run-time.
1765 * Added functionality to create an EVP_PKEY context based on data
1766 for methods from providers. This takes an algorithm name and a
1767 property query string and simply stores them, with the intent
1768 that any operation that uses this context will use those strings
1769 to fetch the needed methods implicitly, thereby making the port
1770 of application written for pre-3.0 OpenSSL easier.
1774 * The undocumented function NCONF_WIN32() has been deprecated; for
1775 conversion details see the HISTORY section of doc/man5/config.pod
1779 * Introduced the new functions EVP_DigestSignInit_ex() and
1780 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1781 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1782 pages for further details.
1786 * Over two thousand fixes were made to the documentation, including:
1787 adding missing command flags, better style conformance, documentation
1790 *Rich Salz, Richard Levitte*
1792 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1793 X25519, X448, Ed25519 and Ed448.
1797 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1802 * Deprecated the public definition of `ERR_STATE` as well as the function
1803 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1808 * Added ERR functionality to give callers access to the stored function
1809 names that have replaced the older function code based functions.
1811 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1812 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1813 ERR_peek_error_all() and ERR_peek_last_error_all().
1815 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1816 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1817 ERR_func_error_string().
1821 * Extended testing to be verbose for failing tests only. The make variables
1822 VERBOSE_FAILURE or VF can be used to enable this:
1824 $ make VF=1 test # Unix
1825 $ mms /macro=(VF=1) test ! OpenVMS
1826 $ nmake VF=1 test # Windows
1830 * Added the `-copy_extensions` option to the `x509` command for use with
1831 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1832 all extensions in the request are copied to the certificate or vice versa.
1834 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1836 * Added the `-copy_extensions` option to the `req` command for use with
1837 `-x509`. When given with the `copy` or `copyall` argument,
1838 all extensions in the certification request are copied to the certificate.
1842 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1843 they generate are by default RFC 5280 compliant in the following sense:
1844 There is a subjectKeyIdentifier extension with a hash value of the public key
1845 and for not self-signed certs there is an authorityKeyIdentifier extension
1846 with a keyIdentifier field or issuer information identifying the signing key.
1847 This is done unless some configuration overrides the new default behavior,
1848 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1852 * Added several checks to `X509_verify_cert()` according to requirements in
1853 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1854 (which may be done by using the CLI option `-x509_strict`):
1855 * The basicConstraints of CA certificates must be marked critical.
1856 * CA certificates must explicitly include the keyUsage extension.
1857 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1858 * The issuer name of any certificate must not be empty.
1859 * The subject name of CA certs, certs with keyUsage crlSign,
1860 and certs without subjectAlternativeName must not be empty.
1861 * If a subjectAlternativeName extension is given it must not be empty.
1862 * The signatureAlgorithm field and the cert signature must be consistent.
1863 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1864 must not be marked critical.
1865 * The authorityKeyIdentifier must be given for X.509v3 certs
1866 unless they are self-signed.
1867 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1871 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1872 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1876 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1877 used even when parsing explicit parameters, when loading a encoded key
1878 or calling `EC_GROUP_new_from_ecpkparameters()`/
1879 `EC_GROUP_new_from_ecparameters()`.
1880 This prevents bypass of security hardening and performance gains,
1881 especially for curves with specialized EC_METHODs.
1882 By default, if a key encoded with explicit parameters is loaded and later
1883 encoded, the output is still encoded with explicit parameters, even if
1884 internally a "named" EC_GROUP is used for computation.
1888 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1889 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1890 NULL. After this change, only the cofactor parameter can be NULL. It also
1891 does some minimal sanity checks on the passed order.
1896 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1897 An attack is simple, if the first CMS_recipientInfo is valid but the
1898 second CMS_recipientInfo is chosen ciphertext. If the second
1899 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1900 encryption key will be replaced by garbage, and the message cannot be
1901 decoded, but if the RSA decryption fails, the correct encryption key is
1902 used and the recipient will not notice the attack.
1903 As a work around for this potential attack the length of the decrypted
1904 key must be equal to the cipher default key length, in case the
1905 certificate is not given and all recipientInfo are tried out.
1906 The old behaviour can be re-enabled in the CMS code by setting the
1907 CMS_DEBUG_DECRYPT flag.
1911 * Early start up entropy quality from the DEVRANDOM seed source has been
1912 improved for older Linux systems. The RAND subsystem will wait for
1913 /dev/random to be producing output before seeding from /dev/urandom.
1914 The seeded state is stored for future library initialisations using
1915 a system global shared memory segment. The shared memory identifier
1916 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1917 the desired value. The default identifier is 114.
1921 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1922 when primes for RSA keys are computed.
1923 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1924 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1925 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1926 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1927 This avoids possible fingerprinting of newly generated RSA modules.
1931 * Correct the extended master secret constant on EBCDIC systems. Without this
1932 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1933 negotiate EMS will fail. Unfortunately this also means that TLS connections
1934 between EBCDIC systems with this fix, and EBCDIC systems without this
1935 fix will fail if they negotiate EMS.
1939 * Changed the library initialisation so that the config file is now loaded
1940 by default. This was already the case for libssl. It now occurs for both
1941 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1942 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1946 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1947 where the former acts as a replacement for `ERR_put_error()`, and the
1948 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1949 `ERR_raise_data()` adds more flexibility by taking a format string and
1950 an arbitrary number of arguments following it, to be processed with
1955 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1956 to check if a named provider is loaded and available. When called, it
1957 will also activate all fallback providers if such are still present.
1961 * Enforce a minimum DH modulus size of 512 bits.
1965 * Changed DH parameters to generate the order q subgroup instead of 2q.
1966 Previously generated DH parameters are still accepted by DH_check
1967 but DH_generate_key works around that by clearing bit 0 of the
1968 private key for those. This avoids leaking bit 0 of the private key.
1972 * Significantly reduce secure memory usage by the randomness pools.
1976 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1981 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1982 algorithms. An implementation of a key exchange algorithm can be obtained
1983 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1984 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1985 the older EVP_PKEY_derive_init() function. See the man pages for the new
1986 functions for further details.
1990 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1994 * Removed the function names from error messages and deprecated the
1999 * Removed NextStep support and the macro OPENSSL_UNISTD
2003 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2004 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2005 Also removed "export var as function" capability; we do not export
2006 variables, only functions.
2010 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2011 an error and 1 indicating success. In previous versions of OpenSSL this
2012 was a void type. If a key was set longer than the maximum possible this
2017 * Support SM2 signing and verification schemes with X509 certificate.
2021 * Use SHA256 as the default digest for TS query in the `ts` app.
2025 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2029 * Default cipher lists/suites are now available via a function, the
2030 #defines are deprecated.
2034 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2035 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2036 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2040 * Join the directories crypto/x509 and crypto/x509v3
2044 * Added command 'openssl kdf' that uses the EVP_KDF API.
2048 * Added command 'openssl mac' that uses the EVP_MAC API.
2052 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2053 as default directories. Also added the command 'openssl info'
2054 for scripting purposes.
2058 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2063 * Add prediction resistance to the DRBG reseeding process.
2067 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2068 mandated by IEEE Std 1619-2018.
2072 * Added newline escaping functionality to a filename when using openssl dgst.
2073 This output format is to replicate the output format found in the `*sum`
2074 checksum programs. This aims to preserve backward compatibility.
2076 *Matt Eaton, Richard Levitte, and Paul Dale*
2078 * Removed the heartbeat message in DTLS feature, as it has very
2079 little usage and doesn't seem to fulfill a valuable purpose.
2080 The configuration option is now deprecated.
2084 * Changed the output of 'openssl {digestname} < file' to display the
2085 digest name in its output.
2089 * Added a new generic trace API which provides support for enabling
2090 instrumentation through trace output.
2092 *Richard Levitte & Matthias St. Pierre*
2094 * Added build tests for C++. These are generated files that only do one
2095 thing, to include one public OpenSSL head file each. This tests that
2096 the public header files can be usefully included in a C++ application.
2098 This test isn't enabled by default. It can be enabled with the option
2099 'enable-buildtest-c++'.
2103 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2107 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2111 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2115 * Added KMAC to EVP_MAC.
2119 * Added property based algorithm implementation selection framework to
2124 * Added SCA hardening for modular field inversion in EC_GROUP through
2125 a new dedicated field_inv() pointer in EC_METHOD.
2126 This also addresses a leakage affecting conversions from projective
2127 to affine coordinates.
2129 *Billy Bob Brumley, Nicola Tuveri*
2131 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2132 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2133 those algorithms that were already supported through the EVP_PKEY API
2134 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2135 and scrypt are now wrappers that call EVP_KDF.
2139 * Build devcrypto engine as a dynamic engine.
2141 *Eneas U de Queiroz*
2143 * Add keyed BLAKE2 to EVP_MAC.
2147 * Fix a bug in the computation of the endpoint-pair shared secret used
2148 by DTLS over SCTP. This breaks interoperability with older versions
2149 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2150 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2151 interoperability with such broken implementations. However, enabling
2152 this switch breaks interoperability with correct implementations.
2154 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2155 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2159 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2163 * Changed the license to the Apache License v2.0.
2167 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2169 - Major releases (indicated by incrementing the MAJOR release number)
2170 may introduce incompatible API/ABI changes.
2171 - Minor releases (indicated by incrementing the MINOR release number)
2172 may introduce new features but retain API/ABI compatibility.
2173 - Patch releases (indicated by incrementing the PATCH number)
2174 are intended for bug fixes and other improvements of existing
2175 features only (like improving performance or adding documentation)
2176 and retain API/ABI compatibility.
2180 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2184 * Remove the 'dist' target and add a tarball building script. The
2185 'dist' target has fallen out of use, and it shouldn't be
2186 necessary to configure just to create a source distribution.
2190 * Recreate the OS390-Unix config target. It no longer relies on a
2191 special script like it did for OpenSSL pre-1.1.0.
2195 * Instead of having the source directories listed in Configure, add
2196 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2201 * Add GMAC to EVP_MAC.
2205 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2209 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2210 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2211 to facilitate the continued use of MACs through raw private keys in
2212 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2216 * Deprecate ECDH_KDF_X9_62().
2220 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2221 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2222 are retained for backwards compatibility.
2226 * AES-XTS mode now enforces that its two keys are different to mitigate
2227 the attacked described in "Efficient Instantiations of Tweakable
2228 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2229 Details of this attack can be obtained from:
2230 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2234 * Rename the object files, i.e. give them other names than in previous
2235 versions. Their names now include the name of the final product, as
2236 well as its type mnemonic (bin, lib, shlib).
2240 * Added new option for 'openssl list', '-objects', which will display the
2241 list of built in objects, i.e. OIDs with names.
2245 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2246 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2251 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2252 improves application performance by removing data copies and providing
2253 applications with zero-copy system calls such as sendfile and splice.
2257 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2261 * `PKCS12_parse` now maintains the order of the parsed certificates
2262 when outputting them via `*ca` (rather than reversing it).
2266 * Deprecated pthread fork support methods.
2270 * Added support for FFDHE key exchange in TLS 1.3.
2274 * Added a new concept for OpenSSL plugability: providers. This
2275 functionality is designed to replace the ENGINE API and ENGINE
2276 implementations, and to be much more dynamic, allowing provider
2277 authors to introduce new algorithms among other things, as long as
2278 there's an API that supports the algorithm type.
2280 With this concept comes a new core API for interaction between
2281 libcrypto and provider implementations. Public libcrypto functions
2282 that want to use providers do so through this core API.
2284 The main documentation for this core API is found in
2285 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2286 refer to other manuals describing the API specific for supported
2287 algorithm types (also called operations).
2294 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2296 * Avoid loading of a dynamic engine twice.
2300 * Prioritise DANE TLSA issuer certs over peer certs
2304 * Fixed random API for MacOS prior to 10.12
2306 These MacOS versions don't support the CommonCrypto APIs
2310 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2312 * Fixed an SM2 Decryption Buffer Overflow.
2314 In order to decrypt SM2 encrypted data an application is expected to
2315 call the API function EVP_PKEY_decrypt(). Typically an application will
2316 call this function twice. The first time, on entry, the "out" parameter
2317 can be NULL and, on exit, the "outlen" parameter is populated with the
2318 buffer size required to hold the decrypted plaintext. The application
2319 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2320 again, but this time passing a non-NULL value for the "out" parameter.
2322 A bug in the implementation of the SM2 decryption code means that the
2323 calculation of the buffer size required to hold the plaintext returned
2324 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2325 size required by the second call. This can lead to a buffer overflow
2326 when EVP_PKEY_decrypt() is called by the application a second time with
2327 a buffer that is too small.
2329 A malicious attacker who is able present SM2 content for decryption to
2330 an application could cause attacker chosen data to overflow the buffer
2331 by up to a maximum of 62 bytes altering the contents of other data held
2332 after the buffer, possibly changing application behaviour or causing
2333 the application to crash. The location of the buffer is application
2334 dependent but is typically heap allocated.
2339 * Fixed various read buffer overruns processing ASN.1 strings
2341 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2342 structure which contains a buffer holding the string data and a field
2343 holding the buffer length. This contrasts with normal C strings which
2344 are represented as a buffer for the string data which is terminated
2345 with a NUL (0) byte.
2347 Although not a strict requirement, ASN.1 strings that are parsed using
2348 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2349 well as any string whose value has been set with the ASN1_STRING_set()
2350 function will additionally NUL terminate the byte array in the
2351 ASN1_STRING structure.
2353 However, it is possible for applications to directly construct valid
2354 ASN1_STRING structures which do not NUL terminate the byte array by
2355 directly setting the "data" and "length" fields in the ASN1_STRING
2356 array. This can also happen by using the ASN1_STRING_set0() function.
2358 Numerous OpenSSL functions that print ASN.1 data have been found to
2359 assume that the ASN1_STRING byte array will be NUL terminated, even
2360 though this is not guaranteed for strings that have been directly
2361 constructed. Where an application requests an ASN.1 structure to be
2362 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2363 been directly constructed by the application without NUL terminating
2364 the "data" field, then a read buffer overrun can occur.
2366 The same thing can also occur during name constraints processing
2367 of certificates (for example if a certificate has been directly
2368 constructed by the application instead of loading it via the OpenSSL
2369 parsing functions, and the certificate contains non NUL terminated
2370 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2371 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2373 If a malicious actor can cause an application to directly construct an
2374 ASN1_STRING and then process it through one of the affected OpenSSL
2375 functions then this issue could be hit. This might result in a crash
2376 (causing a Denial of Service attack). It could also result in the
2377 disclosure of private memory contents (such as private keys, or
2378 sensitive plaintext).
2383 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2385 * Fixed a problem with verifying a certificate chain when using the
2386 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2387 the certificates present in a certificate chain. It is not set by default.
2389 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2390 the chain that have explicitly encoded elliptic curve parameters was added
2391 as an additional strict check.
2393 An error in the implementation of this check meant that the result of a
2394 previous check to confirm that certificates in the chain are valid CA
2395 certificates was overwritten. This effectively bypasses the check
2396 that non-CA certificates must not be able to issue other certificates.
2398 If a "purpose" has been configured then there is a subsequent opportunity
2399 for checks that the certificate is a valid CA. All of the named "purpose"
2400 values implemented in libcrypto perform this check. Therefore, where
2401 a purpose is set the certificate chain will still be rejected even when the
2402 strict flag has been used. A purpose is set by default in libssl client and
2403 server certificate verification routines, but it can be overridden or
2404 removed by an application.
2406 In order to be affected, an application must explicitly set the
2407 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2408 for the certificate verification or, in the case of TLS client or server
2409 applications, override the default purpose.
2414 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2415 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2416 renegotiation ClientHello omits the signature_algorithms extension (where it
2417 was present in the initial ClientHello), but includes a
2418 signature_algorithms_cert extension then a NULL pointer dereference will
2419 result, leading to a crash and a denial of service attack.
2421 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2422 (which is the default configuration). OpenSSL TLS clients are not impacted by
2426 *Peter Kästle and Samuel Sapalski*
2428 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2430 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2431 create a unique hash value based on the issuer and serial number data
2432 contained within an X509 certificate. However, it was failing to correctly
2433 handle any errors that may occur while parsing the issuer field (which might
2434 occur if the issuer field is maliciously constructed). This may subsequently
2435 result in a NULL pointer deref and a crash leading to a potential denial of
2441 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2442 padding mode to correctly check for rollback attacks. This is considered a
2443 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2448 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2449 functions. Previously they could overflow the output length argument in some
2450 cases where the input length is close to the maximum permissible length for
2451 an integer on the platform. In such cases the return value from the function
2452 call would be 1 (indicating success), but the output length value would be
2453 negative. This could cause applications to behave incorrectly or crash.
2458 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2459 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2460 could be exploited in a side channel attack to recover the password. Since
2461 the attack is local host only this is outside of the current OpenSSL
2462 threat model and therefore no CVE is assigned.
2464 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2469 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2471 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2472 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2473 If an attacker can control both items being compared then this could lead
2474 to a possible denial of service attack. OpenSSL itself uses the
2475 GENERAL_NAME_cmp function for two purposes:
2476 1) Comparing CRL distribution point names between an available CRL and a
2477 CRL distribution point embedded in an X509 certificate
2478 2) When verifying that a timestamp response token signer matches the
2479 timestamp authority name (exposed via the API functions
2480 TS_RESP_verify_response and TS_RESP_verify_token)
2485 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2487 * Certificates with explicit curve parameters are now disallowed in
2488 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2492 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2493 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2494 conversely, silently ignore DTLS protocol version bounds when configuring
2495 TLS-based contexts. The commands can be repeated to set bounds of both
2496 types. The same applies with the corresponding "min_protocol" and
2497 "max_protocol" command-line switches, in case some application uses both TLS
2500 SSL_CTX instances that are created for a fixed protocol version (e.g.
2501 TLSv1_server_method()) also silently ignore version bounds. Previously
2502 attempts to apply bounds to these protocol versions would result in an
2503 error. Now only the "version-flexible" SSL_CTX instances are subject to
2504 limits in configuration files in command-line options.
2508 * Handshake now fails if Extended Master Secret extension is dropped
2513 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2515 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2517 * Fixed segmentation fault in SSL_check_chain()
2518 Server or client applications that call the SSL_check_chain() function
2519 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2520 dereference as a result of incorrect handling of the
2521 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2522 or unrecognised signature algorithm is received from the peer. This could
2523 be exploited by a malicious peer in a Denial of Service attack.
2528 * Added AES consttime code for no-asm configurations
2529 an optional constant time support for AES was added
2530 when building openssl for no-asm.
2531 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2532 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2533 At this time this feature is by default disabled.
2534 It will be enabled by default in 3.0.
2538 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2540 * Revert the change of EOF detection while reading in libssl to avoid
2541 regressions in applications depending on the current way of reporting
2542 the EOF. As the existing method is not fully accurate the change to
2543 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2544 branch and will be present in the 3.0 release.
2548 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2549 when primes for RSA keys are computed.
2550 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2551 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2552 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2553 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2554 This avoids possible fingerprinting of newly generated RSA modules.
2558 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2560 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2561 while reading in libssl then we would report an error back to the
2562 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2563 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2564 therefore give a hint as to what went wrong.
2568 * Check that ed25519 and ed448 are allowed by the security level. Previously
2569 signature algorithms not using an MD were not being checked that they were
2570 allowed by the security level.
2574 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2575 was not quite right. The behaviour was not consistent between resumption
2576 and normal handshakes, and also not quite consistent with historical
2577 behaviour. The behaviour in various scenarios has been clarified and
2578 it has been updated to make it match historical behaviour as closely as
2583 * *[VMS only]* The header files that the VMS compilers include automatically,
2584 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2585 that the C++ compiler doesn't understand. This is a shortcoming in the
2586 compiler, but can be worked around with `__cplusplus` guards.
2588 C++ applications that use OpenSSL libraries must be compiled using the
2589 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2590 functions. Otherwise, only functions with symbols of less than 31
2591 characters can be used, as the linker will not be able to successfully
2592 resolve symbols with longer names.
2596 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2597 The presence of this system service is determined at run-time.
2601 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2606 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2608 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2609 number generator (RNG). This was intended to include protection in the
2610 event of a fork() system call in order to ensure that the parent and child
2611 processes did not share the same RNG state. However, this protection was not
2612 being used in the default case.
2614 A partial mitigation for this issue is that the output from a high
2615 precision timer is mixed into the RNG state so the likelihood of a parent
2616 and child process sharing state is significantly reduced.
2618 If an application already calls OPENSSL_init_crypto() explicitly using
2619 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2622 *Matthias St. Pierre*
2624 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2625 used even when parsing explicit parameters, when loading a encoded key
2626 or calling `EC_GROUP_new_from_ecpkparameters()`/
2627 `EC_GROUP_new_from_ecparameters()`.
2628 This prevents bypass of security hardening and performance gains,
2629 especially for curves with specialized EC_METHODs.
2630 By default, if a key encoded with explicit parameters is loaded and later
2631 encoded, the output is still encoded with explicit parameters, even if
2632 internally a "named" EC_GROUP is used for computation.
2636 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2637 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2638 NULL. After this change, only the cofactor parameter can be NULL. It also
2639 does some minimal sanity checks on the passed order.
2644 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2645 An attack is simple, if the first CMS_recipientInfo is valid but the
2646 second CMS_recipientInfo is chosen ciphertext. If the second
2647 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2648 encryption key will be replaced by garbage, and the message cannot be
2649 decoded, but if the RSA decryption fails, the correct encryption key is
2650 used and the recipient will not notice the attack.
2651 As a work around for this potential attack the length of the decrypted
2652 key must be equal to the cipher default key length, in case the
2653 certificate is not given and all recipientInfo are tried out.
2654 The old behaviour can be re-enabled in the CMS code by setting the
2655 CMS_DEBUG_DECRYPT flag.
2660 * Early start up entropy quality from the DEVRANDOM seed source has been
2661 improved for older Linux systems. The RAND subsystem will wait for
2662 /dev/random to be producing output before seeding from /dev/urandom.
2663 The seeded state is stored for future library initialisations using
2664 a system global shared memory segment. The shared memory identifier
2665 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2666 the desired value. The default identifier is 114.
2670 * Correct the extended master secret constant on EBCDIC systems. Without this
2671 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2672 negotiate EMS will fail. Unfortunately this also means that TLS connections
2673 between EBCDIC systems with this fix, and EBCDIC systems without this
2674 fix will fail if they negotiate EMS.
2678 * Use Windows installation paths in the mingw builds
2680 Mingw isn't a POSIX environment per se, which means that Windows
2681 paths should be used for installation.
2686 * Changed DH_check to accept parameters with order q and 2q subgroups.
2687 With order 2q subgroups the bit 0 of the private key is not secret
2688 but DH_generate_key works around that by clearing bit 0 of the
2689 private key for those. This avoids leaking bit 0 of the private key.
2693 * Significantly reduce secure memory usage by the randomness pools.
2697 * Revert the DEVRANDOM_WAIT feature for Linux systems
2699 The DEVRANDOM_WAIT feature added a select() call to wait for the
2700 /dev/random device to become readable before reading from the
2701 /dev/urandom device.
2703 It turned out that this change had negative side effects on
2704 performance which were not acceptable. After some discussion it
2705 was decided to revert this feature and leave it up to the OS
2706 resp. the platform maintainer to ensure a proper initialization
2707 during early boot time.
2709 *Matthias St. Pierre*
2711 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2713 * Add build tests for C++. These are generated files that only do one
2714 thing, to include one public OpenSSL head file each. This tests that
2715 the public header files can be usefully included in a C++ application.
2717 This test isn't enabled by default. It can be enabled with the option
2718 'enable-buildtest-c++'.
2722 * Enable SHA3 pre-hashing for ECDSA and DSA.
2726 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2727 This changes the size when using the `genpkey` command when no size is given.
2728 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2729 generation commands to use 2048 bits by default.
2733 * Reorganize the manual pages to consistently have RETURN VALUES,
2734 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2735 util/fix-doc-nits accordingly.
2737 *Paul Yang, Joshua Lock*
2739 * Add the missing accessor EVP_PKEY_get0_engine()
2743 * Have commands like `s_client` and `s_server` output the signature scheme
2744 along with other cipher suite parameters when debugging.
2748 * Make OPENSSL_config() error agnostic again.
2752 * Do the error handling in RSA decryption constant time.
2756 * Prevent over long nonces in ChaCha20-Poly1305.
2758 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2759 for every encryption operation. RFC 7539 specifies that the nonce value
2760 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2761 and front pads the nonce with 0 bytes if it is less than 12
2762 bytes. However it also incorrectly allows a nonce to be set of up to 16
2763 bytes. In this case only the last 12 bytes are significant and any
2764 additional leading bytes are ignored.
2766 It is a requirement of using this cipher that nonce values are
2767 unique. Messages encrypted using a reused nonce value are susceptible to
2768 serious confidentiality and integrity attacks. If an application changes
2769 the default nonce length to be longer than 12 bytes and then makes a
2770 change to the leading bytes of the nonce expecting the new value to be a
2771 new unique nonce then such an application could inadvertently encrypt
2772 messages with a reused nonce.
2774 Additionally the ignored bytes in a long nonce are not covered by the
2775 integrity guarantee of this cipher. Any application that relies on the
2776 integrity of these ignored leading bytes of a long nonce may be further
2777 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2778 is safe because no such use sets such a long nonce value. However user
2779 applications that use this cipher directly and set a non-default nonce
2780 length to be longer than 12 bytes may be vulnerable.
2782 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2788 * Add DEVRANDOM_WAIT feature for Linux systems
2790 On older Linux systems where the getrandom() system call is not available,
2791 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2792 Contrary to getrandom(), the /dev/urandom device will not block during
2793 early boot when the kernel CSPRNG has not been seeded yet.
2795 To mitigate this known weakness, use select() to wait for /dev/random to
2796 become readable before reading from /dev/urandom.
2798 * Ensure that SM2 only uses SM3 as digest algorithm
2802 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2804 * Change the info callback signals for the start and end of a post-handshake
2805 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2806 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2807 confused by this and assume that a TLSv1.2 renegotiation has started. This
2808 can break KeyUpdate handling. Instead we no longer signal the start and end
2809 of a post handshake message exchange (although the messages themselves are
2810 still signalled). This could break some applications that were expecting
2811 the old signals. However without this KeyUpdate is not usable for many
2816 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2818 * Timing vulnerability in DSA signature generation
2820 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2821 timing side channel attack. An attacker could use variations in the signing
2822 algorithm to recover the private key.
2824 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2829 * Timing vulnerability in ECDSA signature generation
2831 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2832 timing side channel attack. An attacker could use variations in the signing
2833 algorithm to recover the private key.
2835 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2840 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2841 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2842 of two gigabytes and the error handling improved.
2844 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2845 categorized as a normal bug, not a security issue, because the DRBG reseeds
2846 automatically and is fully functional even without additional randomness
2847 provided by the application.
2849 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2851 * Add a new ClientHello callback. Provides a callback interface that gives
2852 the application the ability to adjust the nascent SSL object at the
2853 earliest stage of ClientHello processing, immediately after extensions have
2854 been collected but before they have been processed. In particular, this
2855 callback can adjust the supported TLS versions in response to the contents
2860 * Add SM2 base algorithm support.
2864 * s390x assembly pack: add (improved) hardware-support for the following
2865 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2866 aes-cfb/cfb8, aes-ecb.
2870 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2871 parameter is no longer accepted, as it leads to a corrupt table. NULL
2872 pem_str is reserved for alias entries only.
2876 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2877 step for prime curves. The new implementation is based on formulae from
2878 differential addition-and-doubling in homogeneous projective coordinates
2879 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2880 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2881 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2882 to work in projective coordinates.
2884 *Billy Bob Brumley, Nicola Tuveri*
2886 * Change generating and checking of primes so that the error rate of not
2887 being prime depends on the intended use based on the size of the input.
2888 For larger primes this will result in more rounds of Miller-Rabin.
2889 The maximal error rate for primes with more than 1080 bits is lowered
2892 *Kurt Roeckx, Annie Yousar*
2894 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2898 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2899 moving between systems, and to avoid confusion when a Windows build is
2900 done with mingw vs with MSVC. For POSIX installs, there's still a
2901 symlink or copy named 'tsget' to avoid that confusion as well.
2905 * Revert blinding in ECDSA sign and instead make problematic addition
2906 length-invariant. Switch even to fixed-length Montgomery multiplication.
2910 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2911 step for binary curves. The new implementation is based on formulae from
2912 differential addition-and-doubling in mixed Lopez-Dahab projective
2913 coordinates, modified to independently blind the operands.
2915 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2917 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2918 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2919 EC_METHODs to implement their own specialized "ladder step", to take
2920 advantage of more favorable coordinate systems or more efficient
2921 differential addition-and-doubling algorithms.
2923 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2925 * Modified the random device based seed sources to keep the relevant
2926 file descriptors open rather than reopening them on each access.
2927 This allows such sources to operate in a chroot() jail without
2928 the associated device nodes being available. This behaviour can be
2929 controlled using RAND_keep_random_devices_open().
2933 * Numerous side-channel attack mitigations have been applied. This may have
2934 performance impacts for some algorithms for the benefit of improved
2935 security. Specific changes are noted in this change log by their respective
2940 * AIX shared library support overhaul. Switch to AIX "natural" way of
2941 handling shared libraries, which means collecting shared objects of
2942 different versions and bitnesses in one common archive. This allows to
2943 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2944 doesn't affect the way 3rd party applications are linked, only how
2945 multi-version installation is managed.
2949 * Make ec_group_do_inverse_ord() more robust and available to other
2950 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2951 mitigations are applied to the fallback BN_mod_inverse().
2952 When using this function rather than BN_mod_inverse() directly, new
2953 EC cryptosystem implementations are then safer-by-default.
2957 * Add coordinate blinding for EC_POINT and implement projective
2958 coordinate blinding for generic prime curves as a countermeasure to
2959 chosen point SCA attacks.
2961 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2963 * Add blinding to ECDSA and DSA signatures to protect against side channel
2964 attacks discovered by Keegan Ryan (NCC Group).
2968 * Enforce checking in the `pkeyutl` command to ensure that the input
2969 length does not exceed the maximum supported digest length when performing
2970 a sign, verify or verifyrecover operation.
2974 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2975 I/O in combination with something like select() or poll() will hang. This
2976 can be turned off again using SSL_CTX_clear_mode().
2977 Many applications do not properly handle non-application data records, and
2978 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2979 around the problems in those applications, but can also break some.
2980 It's recommended to read the manpages about SSL_read(), SSL_write(),
2981 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2982 SSL_CTX_set_read_ahead() again.
2986 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2987 now allow empty (zero character) pass phrases.
2991 * Apply blinding to binary field modular inversion and remove patent
2992 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2996 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2997 binary and prime elliptic curves.
3001 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3002 constant time fixed point multiplication.
3006 * Revise elliptic curve scalar multiplication with timing attack
3007 defenses: ec_wNAF_mul redirects to a constant time implementation
3008 when computing fixed point and variable point multiplication (which
3009 in OpenSSL are mostly used with secret scalars in keygen, sign,
3010 ECDH derive operations).
3011 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3014 * Updated CONTRIBUTING
3018 * Updated DRBG / RAND to request nonce and additional low entropy
3019 randomness from the system.
3021 *Matthias St. Pierre*
3023 * Updated 'openssl rehash' to use OpenSSL consistent default.
3027 * Moved the load of the ssl_conf module to libcrypto, which helps
3028 loading engines that libssl uses before libssl is initialised.
3032 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3036 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3038 *Ingo Schwarze, Rich Salz*
3040 * Added output of accepting IP address and port for 'openssl s_server'
3044 * Added a new API for TLSv1.3 ciphersuites:
3045 SSL_CTX_set_ciphersuites()
3046 SSL_set_ciphersuites()
3050 * Memory allocation failures consistently add an error to the error
3055 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3056 in libcrypto when run as setuid/setgid.
3060 * Load any config file by default when libssl is used.
3064 * Added new public header file <openssl/rand_drbg.h> and documentation
3065 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3067 *Matthias St. Pierre*
3069 * QNX support removed (cannot find contributors to get their approval
3070 for the license change).
3074 * TLSv1.3 replay protection for early data has been implemented. See the
3075 SSL_read_early_data() man page for further details.
3079 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3080 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3081 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3082 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3083 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3084 configuration has been separated out. See the ciphers man page or the
3085 SSL_CTX_set_ciphersuites() man page for more information.
3089 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3090 in responder mode now supports the new "-multi" option, which
3091 spawns the specified number of child processes to handle OCSP
3092 requests. The "-timeout" option now also limits the OCSP
3093 responder's patience to wait to receive the full client request
3094 on a newly accepted connection. Child processes are respawned
3095 as needed, and the CA index file is automatically reloaded
3096 when changed. This makes it possible to run the "ocsp" responder
3097 as a long-running service, making the OpenSSL CA somewhat more
3098 feature-complete. In this mode, most diagnostic messages logged
3099 after entering the event loop are logged via syslog(3) rather than
3104 * Added support for X448 and Ed448. Heavily based on original work by
3109 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3110 objects loaded. This adds the functions OSSL_STORE_expect() and
3111 OSSL_STORE_find() as well as needed tools to construct searches and
3112 get the search data out of them.
3116 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3117 version of OpenSSL should review their configuration settings to ensure
3118 that they are still appropriate for TLSv1.3. For further information see:
3119 <https://wiki.openssl.org/index.php/TLS1.3>
3123 * Grand redesign of the OpenSSL random generator
3125 The default RAND method now utilizes an AES-CTR DRBG according to
3126 NIST standard SP 800-90Ar1. The new random generator is essentially
3127 a port of the default random generator from the OpenSSL FIPS 2.0
3128 object module. It is a hybrid deterministic random bit generator
3129 using an AES-CTR bit stream and which seeds and reseeds itself
3130 automatically using trusted system entropy sources.
3132 Some of its new features are:
3133 - Support for multiple DRBG instances with seed chaining.
3134 - The default RAND method makes use of a DRBG.
3135 - There is a public and private DRBG instance.
3136 - The DRBG instances are fork-safe.
3137 - Keep all global DRBG instances on the secure heap if it is enabled.
3138 - The public and private DRBG instance are per thread for lock free
3141 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3143 * Changed Configure so it only says what it does and doesn't dump
3144 so much data. Instead, ./configdata.pm should be used as a script
3145 to display all sorts of configuration data.
3149 * Added processing of "make variables" to Configure.
3153 * Added SHA512/224 and SHA512/256 algorithm support.
3157 * The last traces of Netware support, first removed in 1.1.0, have
3162 * Get rid of Makefile.shared, and in the process, make the processing
3163 of certain files (rc.obj, or the .def/.map/.opt files produced from
3164 the ordinal files) more visible and hopefully easier to trace and
3165 debug (or make silent).
3169 * Make it possible to have environment variable assignments as
3170 arguments to config / Configure.
3174 * Add multi-prime RSA (RFC 8017) support.
3178 * Add SM3 implemented according to GB/T 32905-2016
3179 *Jack Lloyd <jack.lloyd@ribose.com>,*
3180 *Ronald Tse <ronald.tse@ribose.com>,*
3181 *Erick Borsboom <erick.borsboom@ribose.com>*
3183 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3184 as documented in RFC6066.
3185 Based on a patch from Tomasz Moń
3187 *Filipe Raimundo da Silva*
3189 * Add SM4 implemented according to GB/T 32907-2016.
3190 *Jack Lloyd <jack.lloyd@ribose.com>,*
3191 *Ronald Tse <ronald.tse@ribose.com>,*
3192 *Erick Borsboom <erick.borsboom@ribose.com>*
3194 * Reimplement -newreq-nodes and ERR_error_string_n; the
3195 original author does not agree with the license change.
3199 * Add ARIA AEAD TLS support.
3203 * Some macro definitions to support VS6 have been removed. Visual
3204 Studio 6 has not worked since 1.1.0
3208 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3209 without clearing the errors.
3213 * Add "atfork" functions. If building on a system that without
3214 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3215 requirements. The RAND facility now uses/requires this.
3223 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3224 not possible to disable entirely. However, it's still possible to
3225 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3228 To disable, configure with 'no-ui-console'. 'no-ui' is still
3229 possible to use as an alias. Check at compile time with the
3230 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3231 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3235 * Add a STORE module, which implements a uniform and URI based reader of
3236 stores that can contain keys, certificates, CRLs and numerous other
3237 objects. The main API is loosely based on a few stdio functions,
3238 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3239 OSSL_STORE_error and OSSL_STORE_close.
3240 The implementation uses backends called "loaders" to implement arbitrary
3241 URI schemes. There is one built in "loader" for the 'file' scheme.
3245 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3246 then adjusted to work on FreeBSD 8.4 as well.
3247 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3248 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3252 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3253 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3254 error code calls like this:
3256 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3258 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3259 that can be encoded in C. For the foreseeable future, this will only
3262 *Richard Levitte and Tim Hudson*
3264 * Removed BSD cryptodev engine.
3268 * Add a build target 'build_all_generated', to build all generated files
3269 and only that. This can be used to prepare everything that requires
3270 things like perl for a system that lacks perl and then move everything
3271 to that system and do the rest of the build there.
3275 * In the UI interface, make it possible to duplicate the user data. This
3276 can be used by engines that need to retain the data for a longer time
3277 than just the call where this user data is passed.
3281 * Ignore the '-named_curve auto' value for compatibility of applications
3284 *Tomáš Mráz <tmraz@fedoraproject.org>*
3286 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3287 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3288 alerts across multiple records (some of which could be empty). In practice
3289 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3290 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3291 support this at all. Supporting it adds significant complexity to the
3292 record layer, and its removal is unlikely to cause interoperability
3297 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3298 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3299 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3304 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3305 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3307 *Richard Levitte, Andy Polyakov*
3309 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3314 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3315 platform rather than 'mingw'.
3319 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3320 success if they are asked to add an object which already exists
3321 in the store. This change cascades to other functions which load
3322 certificates and CRLs.
3326 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3327 facilitate stack unwinding even from assembly subroutines.
3331 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3332 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3336 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3337 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3338 which is the minimum version we support.
3342 * Certificate time validation (X509_cmp_time) enforces stricter
3343 compliance with RFC 5280. Fractional seconds and timezone offsets
3344 are no longer allowed.
3348 * Add support for ARIA
3352 * s_client will now send the Server Name Indication (SNI) extension by
3353 default unless the new "-noservername" option is used. The server name is
3354 based on the host provided to the "-connect" option unless overridden by
3355 using "-servername".
3359 * Add support for SipHash
3363 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3364 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3365 prevent issues where no progress is being made and the peer continually
3366 sends unrecognised record types, using up resources processing them.
3370 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3371 using the algorithm defined in
3372 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3376 * Heartbeat support has been removed; the ABI is changed for now.
3378 *Richard Levitte, Rich Salz*
3380 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3384 * The RSA "null" method, which was partially supported to avoid patent
3385 issues, has been replaced to always returns NULL.
3392 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3394 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3395 used even when parsing explicit parameters, when loading a encoded key
3396 or calling `EC_GROUP_new_from_ecpkparameters()`/
3397 `EC_GROUP_new_from_ecparameters()`.
3398 This prevents bypass of security hardening and performance gains,
3399 especially for curves with specialized EC_METHODs.
3400 By default, if a key encoded with explicit parameters is loaded and later
3401 encoded, the output is still encoded with explicit parameters, even if
3402 internally a "named" EC_GROUP is used for computation.
3406 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3407 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3408 NULL. After this change, only the cofactor parameter can be NULL. It also
3409 does some minimal sanity checks on the passed order.
3414 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3415 An attack is simple, if the first CMS_recipientInfo is valid but the
3416 second CMS_recipientInfo is chosen ciphertext. If the second
3417 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3418 encryption key will be replaced by garbage, and the message cannot be
3419 decoded, but if the RSA decryption fails, the correct encryption key is
3420 used and the recipient will not notice the attack.
3421 As a work around for this potential attack the length of the decrypted
3422 key must be equal to the cipher default key length, in case the
3423 certificate is not given and all recipientInfo are tried out.
3424 The old behaviour can be re-enabled in the CMS code by setting the
3425 CMS_DEBUG_DECRYPT flag.
3430 * Use Windows installation paths in the mingw builds
3432 Mingw isn't a POSIX environment per se, which means that Windows
3433 paths should be used for installation.
3438 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3440 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3441 This changes the size when using the `genpkey` command when no size is given.
3442 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3443 generation commands to use 2048 bits by default.
3447 * Prevent over long nonces in ChaCha20-Poly1305.
3449 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3450 for every encryption operation. RFC 7539 specifies that the nonce value
3451 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3452 and front pads the nonce with 0 bytes if it is less than 12
3453 bytes. However it also incorrectly allows a nonce to be set of up to 16
3454 bytes. In this case only the last 12 bytes are significant and any
3455 additional leading bytes are ignored.
3457 It is a requirement of using this cipher that nonce values are
3458 unique. Messages encrypted using a reused nonce value are susceptible to
3459 serious confidentiality and integrity attacks. If an application changes
3460 the default nonce length to be longer than 12 bytes and then makes a
3461 change to the leading bytes of the nonce expecting the new value to be a
3462 new unique nonce then such an application could inadvertently encrypt
3463 messages with a reused nonce.
3465 Additionally the ignored bytes in a long nonce are not covered by the
3466 integrity guarantee of this cipher. Any application that relies on the
3467 integrity of these ignored leading bytes of a long nonce may be further
3468 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3469 is safe because no such use sets such a long nonce value. However user
3470 applications that use this cipher directly and set a non-default nonce
3471 length to be longer than 12 bytes may be vulnerable.
3473 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3479 * Added SCA hardening for modular field inversion in EC_GROUP through
3480 a new dedicated field_inv() pointer in EC_METHOD.
3481 This also addresses a leakage affecting conversions from projective
3482 to affine coordinates.
3484 *Billy Bob Brumley, Nicola Tuveri*
3486 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3487 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3491 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3495 * Remove the 'dist' target and add a tarball building script. The
3496 'dist' target has fallen out of use, and it shouldn't be
3497 necessary to configure just to create a source distribution.
3501 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3503 * Timing vulnerability in DSA signature generation
3505 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3506 timing side channel attack. An attacker could use variations in the signing
3507 algorithm to recover the private key.
3509 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3514 * Timing vulnerability in ECDSA signature generation
3516 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3517 timing side channel attack. An attacker could use variations in the signing
3518 algorithm to recover the private key.
3520 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3525 * Add coordinate blinding for EC_POINT and implement projective
3526 coordinate blinding for generic prime curves as a countermeasure to
3527 chosen point SCA attacks.
3529 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3531 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3533 * Client DoS due to large DH parameter
3535 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3536 malicious server can send a very large prime value to the client. This will
3537 cause the client to spend an unreasonably long period of time generating a
3538 key for this prime resulting in a hang until the client has finished. This
3539 could be exploited in a Denial Of Service attack.
3541 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3546 * Cache timing vulnerability in RSA Key Generation
3548 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3549 a cache timing side channel attack. An attacker with sufficient access to
3550 mount cache timing attacks during the RSA key generation process could
3551 recover the private key.
3553 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3554 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3559 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3560 parameter is no longer accepted, as it leads to a corrupt table. NULL
3561 pem_str is reserved for alias entries only.
3565 * Revert blinding in ECDSA sign and instead make problematic addition
3566 length-invariant. Switch even to fixed-length Montgomery multiplication.
3570 * Change generating and checking of primes so that the error rate of not
3571 being prime depends on the intended use based on the size of the input.
3572 For larger primes this will result in more rounds of Miller-Rabin.
3573 The maximal error rate for primes with more than 1080 bits is lowered
3576 *Kurt Roeckx, Annie Yousar*
3578 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3582 * Add blinding to ECDSA and DSA signatures to protect against side channel
3583 attacks discovered by Keegan Ryan (NCC Group).
3587 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3588 now allow empty (zero character) pass phrases.
3592 * Certificate time validation (X509_cmp_time) enforces stricter
3593 compliance with RFC 5280. Fractional seconds and timezone offsets
3594 are no longer allowed.
3598 * Fixed a text canonicalisation bug in CMS
3600 Where a CMS detached signature is used with text content the text goes
3601 through a canonicalisation process first prior to signing or verifying a
3602 signature. This process strips trailing space at the end of lines, converts
3603 line terminators to CRLF and removes additional trailing line terminators
3604 at the end of a file. A bug in the canonicalisation process meant that
3605 some characters, such as form-feed, were incorrectly treated as whitespace
3606 and removed. This is contrary to the specification (RFC5485). This fix
3607 could mean that detached text data signed with an earlier version of
3608 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3609 signed with a fixed OpenSSL may fail to verify with an earlier version of
3610 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3611 and use the "-binary" flag (for the "cms" command line application) or set
3612 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3616 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3618 * Constructed ASN.1 types with a recursive definition could exceed the stack
3620 Constructed ASN.1 types with a recursive definition (such as can be found
3621 in PKCS7) could eventually exceed the stack given malicious input with
3622 excessive recursion. This could result in a Denial Of Service attack. There
3623 are no such structures used within SSL/TLS that come from untrusted sources
3624 so this is considered safe.
3626 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3632 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3634 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3635 effectively reduced to only comparing the least significant bit of each
3636 byte. This allows an attacker to forge messages that would be considered as
3637 authenticated in an amount of tries lower than that guaranteed by the
3638 security claims of the scheme. The module can only be compiled by the
3639 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3641 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3647 * Add a build target 'build_all_generated', to build all generated files
3648 and only that. This can be used to prepare everything that requires
3649 things like perl for a system that lacks perl and then move everything
3650 to that system and do the rest of the build there.
3654 * Backport SSL_OP_NO_RENGOTIATION
3656 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3657 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3658 changes this is no longer possible in 1.1.0. Therefore, the new
3659 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3660 1.1.0 to provide equivalent functionality.
3662 Note that if an application built against 1.1.0h headers (or above) is run
3663 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3664 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3668 * Removed the OS390-Unix config target. It relied on a script that doesn't
3673 * rsaz_1024_mul_avx2 overflow bug on x86_64
3675 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3676 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3677 Analysis suggests that attacks against RSA and DSA as a result of this
3678 defect would be very difficult to perform and are not believed likely.
3679 Attacks against DH1024 are considered just feasible, because most of the
3680 work necessary to deduce information about a private key may be performed
3681 offline. The amount of resources required for such an attack would be
3682 significant. However, for an attack on TLS to be meaningful, the server
3683 would have to share the DH1024 private key among multiple clients, which is
3684 no longer an option since CVE-2016-0701.
3686 This only affects processors that support the AVX2 but not ADX extensions
3687 like Intel Haswell (4th generation).
3689 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3690 was originally found via the OSS-Fuzz project.
3695 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3697 * bn_sqrx8x_internal carry bug on x86_64
3699 There is a carry propagating bug in the x86_64 Montgomery squaring
3700 procedure. No EC algorithms are affected. Analysis suggests that attacks
3701 against RSA and DSA as a result of this defect would be very difficult to
3702 perform and are not believed likely. Attacks against DH are considered just
3703 feasible (although very difficult) because most of the work necessary to
3704 deduce information about a private key may be performed offline. The amount
3705 of resources required for such an attack would be very significant and
3706 likely only accessible to a limited number of attackers. An attacker would
3707 additionally need online access to an unpatched system using the target
3708 private key in a scenario with persistent DH parameters and a private
3709 key that is shared between multiple clients.
3711 This only affects processors that support the BMI1, BMI2 and ADX extensions
3712 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3714 This issue was reported to OpenSSL by the OSS-Fuzz project.
3719 * Malformed X.509 IPAddressFamily could cause OOB read
3721 If an X.509 certificate has a malformed IPAddressFamily extension,
3722 OpenSSL could do a one-byte buffer overread. The most likely result
3723 would be an erroneous display of the certificate in text format.
3725 This issue was reported to OpenSSL by the OSS-Fuzz project.
3730 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3732 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3733 platform rather than 'mingw'.
3737 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3738 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3739 which is the minimum version we support.
3743 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3745 * Encrypt-Then-Mac renegotiation crash
3747 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3748 negotiated where it was not in the original handshake (or vice-versa) then
3749 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3750 and servers are affected.
3752 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3757 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3759 * Truncated packet could crash via OOB read
3761 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3762 cipher is being used, then a truncated packet can cause that host to
3763 perform an out-of-bounds read, usually resulting in a crash.
3765 This issue was reported to OpenSSL by Robert Święcki of Google.
3770 * Bad (EC)DHE parameters cause a client crash
3772 If a malicious server supplies bad parameters for a DHE or ECDHE key
3773 exchange then this can result in the client attempting to dereference a
3774 NULL pointer leading to a client crash. This could be exploited in a Denial
3777 This issue was reported to OpenSSL by Guido Vranken.
3782 * BN_mod_exp may produce incorrect results on x86_64
3784 There is a carry propagating bug in the x86_64 Montgomery squaring
3785 procedure. No EC algorithms are affected. Analysis suggests that attacks
3786 against RSA and DSA as a result of this defect would be very difficult to
3787 perform and are not believed likely. Attacks against DH are considered just
3788 feasible (although very difficult) because most of the work necessary to
3789 deduce information about a private key may be performed offline. The amount
3790 of resources required for such an attack would be very significant and
3791 likely only accessible to a limited number of attackers. An attacker would
3792 additionally need online access to an unpatched system using the target
3793 private key in a scenario with persistent DH parameters and a private
3794 key that is shared between multiple clients. For example this can occur by
3795 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3796 similar to CVE-2015-3193 but must be treated as a separate problem.
3798 This issue was reported to OpenSSL by the OSS-Fuzz project.
3803 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3805 * ChaCha20/Poly1305 heap-buffer-overflow
3807 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3808 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3809 crash. This issue is not considered to be exploitable beyond a DoS.
3811 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3816 * CMS Null dereference
3818 Applications parsing invalid CMS structures can crash with a NULL pointer
3819 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3820 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3821 structure callback if an attempt is made to free certain invalid encodings.
3822 Only CHOICE structures using a callback which do not handle NULL value are
3825 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3830 * Montgomery multiplication may produce incorrect results
3832 There is a carry propagating bug in the Broadwell-specific Montgomery
3833 multiplication procedure that handles input lengths divisible by, but
3834 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3835 and DH private keys are impossible. This is because the subroutine in
3836 question is not used in operations with the private key itself and an input
3837 of the attacker's direct choice. Otherwise the bug can manifest itself as
3838 transient authentication and key negotiation failures or reproducible
3839 erroneous outcome of public-key operations with specially crafted input.
3840 Among EC algorithms only Brainpool P-512 curves are affected and one
3841 presumably can attack ECDH key negotiation. Impact was not analyzed in
3842 detail, because pre-requisites for attack are considered unlikely. Namely
3843 multiple clients have to choose the curve in question and the server has to
3844 share the private key among them, neither of which is default behaviour.
3845 Even then only clients that chose the curve will be affected.
3847 This issue was publicly reported as transient failures and was not
3848 initially recognized as a security issue. Thanks to Richard Morgan for
3849 providing reproducible case.
3854 * Removed automatic addition of RPATH in shared libraries and executables,
3855 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3859 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3861 * Fix Use After Free for large message sizes
3863 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3864 message larger than approx 16k is received then the underlying buffer to
3865 store the incoming message is reallocated and moved. Unfortunately a
3866 dangling pointer to the old location is left which results in an attempt to
3867 write to the previously freed location. This is likely to result in a
3868 crash, however it could potentially lead to execution of arbitrary code.
3870 This issue only affects OpenSSL 1.1.0a.
3872 This issue was reported to OpenSSL by Robert Święcki.
3877 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3879 * OCSP Status Request extension unbounded memory growth
3881 A malicious client can send an excessively large OCSP Status Request
3882 extension. If that client continually requests renegotiation, sending a
3883 large OCSP Status Request extension each time, then there will be unbounded
3884 memory growth on the server. This will eventually lead to a Denial Of
3885 Service attack through memory exhaustion. Servers with a default
3886 configuration are vulnerable even if they do not support OCSP. Builds using
3887 the "no-ocsp" build time option are not affected.
3889 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3894 * SSL_peek() hang on empty record
3896 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3897 sends an empty record. This could be exploited by a malicious peer in a
3898 Denial Of Service attack.
3900 This issue was reported to OpenSSL by Alex Gaynor.
3905 * Excessive allocation of memory in tls_get_message_header() and
3906 dtls1_preprocess_fragment()
3908 A (D)TLS message includes 3 bytes for its length in the header for the
3909 message. This would allow for messages up to 16Mb in length. Messages of
3910 this length are excessive and OpenSSL includes a check to ensure that a
3911 peer is sending reasonably sized messages in order to avoid too much memory
3912 being consumed to service a connection. A flaw in the logic of version
3913 1.1.0 means that memory for the message is allocated too early, prior to
3914 the excessive message length check. Due to way memory is allocated in
3915 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3916 to service a connection. This could lead to a Denial of Service through
3917 memory exhaustion. However, the excessive message length check still takes
3918 place, and this would cause the connection to immediately fail. Assuming
3919 that the application calls SSL_free() on the failed connection in a timely
3920 manner then the 21Mb of allocated memory will then be immediately freed
3921 again. Therefore, the excessive memory allocation will be transitory in
3922 nature. This then means that there is only a security impact if:
3924 1) The application does not call SSL_free() in a timely manner in the event
3925 that the connection fails
3927 2) The application is working in a constrained environment where there is
3928 very little free memory
3930 3) The attacker initiates multiple connection attempts such that there are
3931 multiple connections in a state where memory has been allocated for the
3932 connection; SSL_free() has not yet been called; and there is insufficient
3933 memory to service the multiple requests.
3935 Except in the instance of (1) above any Denial Of Service is likely to be
3936 transitory because as soon as the connection fails the memory is
3937 subsequently freed again in the SSL_free() call. However there is an
3938 increased risk during this period of application crashes due to the lack of
3939 memory - which would then mean a more serious Denial of Service.
3941 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3942 (CVE-2016-6307 and CVE-2016-6308)
3946 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3947 had to be removed. Primary reason is that vendor assembler can't
3948 assemble our modules with -KPIC flag. As result it, assembly
3949 support, was not even available as option. But its lack means
3950 lack of side-channel resistant code, which is incompatible with
3951 security by todays standards. Fortunately gcc is readily available
3952 prepackaged option, which we firmly point at...
3956 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3958 * Windows command-line tool supports UTF-8 opt-in option for arguments
3959 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3960 (to any value) allows Windows user to access PKCS#12 file generated
3961 with Windows CryptoAPI and protected with non-ASCII password, as well
3962 as files generated under UTF-8 locale on Linux also protected with
3967 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3968 have been disabled by default and removed from DEFAULT, just like RC4.
3969 See the RC4 item below to re-enable both.
3973 * The method for finding the storage location for the Windows RAND seed file
3974 has changed. First we check %RANDFILE%. If that is not set then we check
3975 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3976 all else fails we fall back to C:\.
3980 * The EVP_EncryptUpdate() function has had its return type changed from void
3981 to int. A return of 0 indicates and error while a return of 1 indicates
3986 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3987 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3988 off the constant time implementation for RSA, DSA and DH have been made
3989 no-ops and deprecated.
3993 * Windows RAND implementation was simplified to only get entropy by
3994 calling CryptGenRandom(). Various other RAND-related tickets
3997 *Joseph Wylie Yandle, Rich Salz*
3999 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4000 and `OPENSSL_LH_`, respectively. The old names are available
4001 with API compatibility. They new names are now completely documented.
4005 * Unify TYPE_up_ref(obj) methods signature.
4006 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4007 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4008 int (instead of void) like all others TYPE_up_ref() methods.
4009 So now these methods also check the return value of CRYPTO_atomic_add(),
4010 and the validity of object reference counter.
4012 *fdasilvayy@gmail.com*
4014 * With Windows Visual Studio builds, the .pdb files are installed
4015 alongside the installed libraries and executables. For a static
4016 library installation, ossl_static.pdb is the associate compiler
4017 generated .pdb file to be used when linking programs.
4021 * Remove openssl.spec. Packaging files belong with the packagers.
4025 * Automatic Darwin/OSX configuration has had a refresh, it will now
4026 recognise x86_64 architectures automatically. You can still decide
4027 to build for a different bitness with the environment variable
4028 KERNEL_BITS (can be 32 or 64), for example:
4030 KERNEL_BITS=32 ./config
4034 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4035 256 bit AES and HMAC with SHA256.
4039 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4043 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4047 * To enable users to have their own config files and build file templates,
4048 Configure looks in the directory indicated by the environment variable
4049 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4050 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4051 name and is used as is.
4055 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4056 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4057 X509_CERT_FILE_CTX was removed.
4061 * "shared" builds are now the default. To create only static libraries use
4062 the "no-shared" Configure option.
4066 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4067 All of these option have not worked for some while and are fundamental
4072 * Make various cleanup routines no-ops and mark them as deprecated. Most
4073 global cleanup functions are no longer required because they are handled
4074 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4075 Explicitly de-initing can cause problems (e.g. where a library that uses
4076 OpenSSL de-inits, but an application is still using it). The affected
4077 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4078 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4079 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4080 COMP_zlib_cleanup().
4084 * --strict-warnings no longer enables runtime debugging options
4085 such as REF_DEBUG. Instead, debug options are automatically
4086 enabled with '--debug' builds.
4088 *Andy Polyakov, Emilia Käsper*
4090 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4091 have been moved out of the public header files. New functions for managing
4092 these have been added.
4096 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4097 objects have been moved out of the public header files. New
4098 functions for managing these have been added.
4102 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4103 have been moved out of the public header files. New functions for managing
4104 these have been added.
4108 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4109 moved out of the public header files. New functions for managing these
4114 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4118 * Removed the mk1mf build scripts.
4122 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4123 it is always safe to #include a header now.
4127 * Removed the aged BC-32 config and all its supporting scripts
4131 * Removed support for Ultrix, Netware, and OS/2.
4135 * Add support for HKDF.
4137 *Alessandro Ghedini*
4139 * Add support for blake2b and blake2s
4143 * Added support for "pipelining". Ciphers that have the
4144 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4145 encryptions/decryptions simultaneously. There are currently no built-in
4146 ciphers with this property but the expectation is that engines will be able
4147 to offer it to significantly improve throughput. Support has been extended
4148 into libssl so that multiple records for a single connection can be
4149 processed in one go (for >=TLS 1.1).
4153 * Added the AFALG engine. This is an async capable engine which is able to
4154 offload work to the Linux kernel. In this initial version it only supports
4155 AES128-CBC. The kernel must be version 4.1.0 or greater.
4159 * OpenSSL now uses a new threading API. It is no longer necessary to
4160 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4161 are two supported threading models: pthreads and windows threads. It is
4162 also possible to configure OpenSSL at compile time for "no-threads". The
4163 old threading API should no longer be used. The functions have been
4164 replaced with "no-op" compatibility macros.
4166 *Alessandro Ghedini, Matt Caswell*
4168 * Modify behavior of ALPN to invoke callback after SNI/servername
4169 callback, such that updates to the SSL_CTX affect ALPN.
4173 * Add SSL_CIPHER queries for authentication and key-exchange.
4177 * Changes to the DEFAULT cipherlist:
4178 - Prefer (EC)DHE handshakes over plain RSA.
4179 - Prefer AEAD ciphers over legacy ciphers.
4180 - Prefer ECDSA over RSA when both certificates are available.
4181 - Prefer TLSv1.2 ciphers/PRF.
4182 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4187 * Change the ECC default curve list to be this, in order: x25519,
4188 secp256r1, secp521r1, secp384r1.
4192 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4193 disabled by default. They can be re-enabled using the
4194 enable-weak-ssl-ciphers option to Configure.
4198 * If the server has ALPN configured, but supports no protocols that the
4199 client advertises, send a fatal "no_application_protocol" alert.
4200 This behaviour is SHALL in RFC 7301, though it isn't universally
4201 implemented by other servers.
4205 * Add X25519 support.
4206 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4207 for public and private key encoding using the format documented in
4208 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4209 key generation and key derivation.
4211 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4216 * Deprecate SRP_VBASE_get_by_user.
4217 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4218 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4219 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4220 seed, even if the seed is configured.
4222 Users should use SRP_VBASE_get1_by_user instead. Note that in
4223 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4224 also that even though configuring the SRP seed attempts to hide
4225 invalid usernames by continuing the handshake with fake
4226 credentials, this behaviour is not constant time and no strong
4227 guarantees are made that the handshake is indistinguishable from
4228 that of a valid user.
4232 * Configuration change; it's now possible to build dynamic engines
4233 without having to build shared libraries and vice versa. This
4234 only applies to the engines in `engines/`, those in `crypto/engine/`
4235 will always be built into libcrypto (i.e. "static").
4237 Building dynamic engines is enabled by default; to disable, use
4238 the configuration option "disable-dynamic-engine".
4240 The only requirements for building dynamic engines are the
4241 presence of the DSO module and building with position independent
4242 code, so they will also automatically be disabled if configuring
4243 with "disable-dso" or "disable-pic".
4245 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4246 are also taken away from openssl/opensslconf.h, as they are
4251 * Configuration change; if there is a known flag to compile
4252 position independent code, it will always be applied on the
4253 libcrypto and libssl object files, and never on the application
4254 object files. This means other libraries that use routines from
4255 libcrypto / libssl can be made into shared libraries regardless
4256 of how OpenSSL was configured.
4258 If this isn't desirable, the configuration options "disable-pic"
4259 or "no-pic" can be used to disable the use of PIC. This will
4260 also disable building shared libraries and dynamic engines.
4264 * Removed JPAKE code. It was experimental and has no wide use.
4268 * The INSTALL_PREFIX Makefile variable has been renamed to
4269 DESTDIR. That makes for less confusion on what this variable
4270 is for. Also, the configuration option --install_prefix is
4275 * Heartbeat for TLS has been removed and is disabled by default
4276 for DTLS; configure with enable-heartbeats. Code that uses the
4277 old #define's might need to be updated.
4279 *Emilia Käsper, Rich Salz*
4281 * Rename REF_CHECK to REF_DEBUG.
4285 * New "unified" build system
4287 The "unified" build system is aimed to be a common system for all
4288 platforms we support. With it comes new support for VMS.
4290 This system builds supports building in a different directory tree
4291 than the source tree. It produces one Makefile (for unix family
4292 or lookalikes), or one descrip.mms (for VMS).
4294 The source of information to make the Makefile / descrip.mms is
4295 small files called 'build.info', holding the necessary
4296 information for each directory with source to compile, and a
4297 template in Configurations, like unix-Makefile.tmpl or
4300 With this change, the library names were also renamed on Windows
4301 and on VMS. They now have names that are closer to the standard
4302 on Unix, and include the major version number, and in certain
4303 cases, the architecture they are built for. See "Notes on shared
4304 libraries" in INSTALL.
4306 We rely heavily on the perl module Text::Template.
4310 * Added support for auto-initialisation and de-initialisation of the library.
4311 OpenSSL no longer requires explicit init or deinit routines to be called,
4312 except in certain circumstances. See the OPENSSL_init_crypto() and
4313 OPENSSL_init_ssl() man pages for further information.
4317 * The arguments to the DTLSv1_listen function have changed. Specifically the
4318 "peer" argument is now expected to be a BIO_ADDR object.
4320 * Rewrite of BIO networking library. The BIO library lacked consistent
4321 support of IPv6, and adding it required some more extensive
4322 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4323 which hold all types of addresses and chains of address information.
4324 It also introduces a new API, with functions like BIO_socket,
4325 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4326 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4327 have been adapted accordingly.
4331 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4336 * CRIME protection: disable compression by default, even if OpenSSL is
4337 compiled with zlib enabled. Applications can still enable compression
4338 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4339 using the SSL_CONF library to configure compression.
4343 * The signature of the session callback configured with
4344 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4345 was explicitly marked as `const unsigned char*` instead of
4350 * Always DPURIFY. Remove the use of uninitialized memory in the
4351 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4355 * Removed many obsolete configuration items, including
4356 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4357 MD2_CHAR, MD2_INT, MD2_LONG
4359 IDEA_SHORT, IDEA_LONG
4360 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4362 *Rich Salz, with advice from Andy Polyakov*
4364 * Many BN internals have been moved to an internal header file.
4366 *Rich Salz with help from Andy Polyakov*
4368 * Configuration and writing out the results from it has changed.
4369 Files such as Makefile include/openssl/opensslconf.h and are now
4370 produced through general templates, such as Makefile.in and
4371 crypto/opensslconf.h.in and some help from the perl module
4374 Also, the center of configuration information is no longer
4375 Makefile. Instead, Configure produces a perl module in
4376 configdata.pm which holds most of the config data (in the hash
4377 table %config), the target data that comes from the target
4378 configuration in one of the `Configurations/*.conf` files (in
4383 * To clarify their intended purposes, the Configure options
4384 --prefix and --openssldir change their semantics, and become more
4385 straightforward and less interdependent.
4387 --prefix shall be used exclusively to give the location INSTALLTOP
4388 where programs, scripts, libraries, include files and manuals are
4389 going to be installed. The default is now /usr/local.
4391 --openssldir shall be used exclusively to give the default
4392 location OPENSSLDIR where certificates, private keys, CRLs are
4393 managed. This is also where the default openssl.cnf gets
4395 If the directory given with this option is a relative path, the
4396 values of both the --prefix value and the --openssldir value will
4397 be combined to become OPENSSLDIR.
4398 The default for --openssldir is INSTALLTOP/ssl.
4400 Anyone who uses --openssldir to specify where OpenSSL is to be
4401 installed MUST change to use --prefix instead.
4405 * The GOST engine was out of date and therefore it has been removed. An up
4406 to date GOST engine is now being maintained in an external repository.
4407 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4408 support for GOST ciphersuites (these are only activated if a GOST engine
4413 * EGD is no longer supported by default; use enable-egd when
4416 *Ben Kaduk and Rich Salz*
4418 * The distribution now has Makefile.in files, which are used to
4419 create Makefile's when Configure is run. *Configure must be run
4420 before trying to build now.*
4424 * The return value for SSL_CIPHER_description() for error conditions
4429 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4431 Obtaining and performing DNSSEC validation of TLSA records is
4432 the application's responsibility. The application provides
4433 the TLSA records of its choice to OpenSSL, and these are then
4434 used to authenticate the peer.
4436 The TLSA records need not even come from DNS. They can, for
4437 example, be used to implement local end-entity certificate or
4438 trust-anchor "pinning", where the "pin" data takes the form
4439 of TLSA records, which can augment or replace verification
4440 based on the usual WebPKI public certification authorities.
4444 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4445 continues to support deprecated interfaces in default builds.
4446 However, applications are strongly advised to compile their
4447 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4448 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4449 or the 1.1.0 releases.
4451 In environments in which all applications have been ported to
4452 not use any deprecated interfaces OpenSSL's Configure script
4453 should be used with the --api=1.1.0 option to entirely remove
4454 support for the deprecated features from the library and
4455 unconditionally disable them in the installed headers.
4456 Essentially the same effect can be achieved with the "no-deprecated"
4457 argument to Configure, except that this will always restrict
4458 the build to just the latest API, rather than a fixed API
4461 As applications are ported to future revisions of the API,
4462 they should update their compile-time OPENSSL_API_COMPAT define
4463 accordingly, but in most cases should be able to continue to
4464 compile with later releases.
4466 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4467 0x10000000L and 0x00908000L, respectively. However those
4468 versions did not support the OPENSSL_API_COMPAT feature, and
4469 so applications are not typically tested for explicit support
4470 of just the undeprecated features of either release.
4474 * Add support for setting the minimum and maximum supported protocol.
4475 It can bet set via the SSL_set_min_proto_version() and
4476 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4477 MaxProtocol. It's recommended to use the new APIs to disable
4478 protocols instead of disabling individual protocols using
4479 SSL_set_options() or SSL_CONF's Protocol. This change also
4480 removes support for disabling TLS 1.2 in the OpenSSL TLS
4481 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4485 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4489 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4490 and integrates ECDSA and ECDH functionality into EC. Implementations can
4491 now redirect key generation and no longer need to convert to or from
4494 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4495 include the ec.h header file instead.
4499 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4500 ciphers who are no longer supported and drops support the ephemeral RSA key
4501 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4505 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4506 opaque. For HMAC_CTX, the following constructors and destructors
4509 HMAC_CTX *HMAC_CTX_new(void);
4510 void HMAC_CTX_free(HMAC_CTX *ctx);
4512 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4513 destroy such methods has been added. See EVP_MD_meth_new(3) and
4514 EVP_CIPHER_meth_new(3) for documentation.
4517 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4518 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4519 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4520 an already created structure.
4521 2) For consistency with the majority of our object creators and
4522 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4523 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4524 for deprecated builds.
4528 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4529 cryptographic operations to be performed asynchronously as long as an
4530 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4531 further details. Libssl has also had this capability integrated with the
4532 introduction of the new mode SSL_MODE_ASYNC and associated error
4533 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4534 pages. This work was developed in partnership with Intel Corp.
4538 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4539 always enabled now. If you want to disable the support you should
4540 exclude it using the list of supported ciphers. This also means that the
4541 "-no_ecdhe" option has been removed from s_server.
4545 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4546 SSL_{CTX_}set1_curves() which can set a list.
4550 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4551 curve you want to support using SSL_{CTX_}set1_curves().
4555 * State machine rewrite. The state machine code has been significantly
4556 refactored in order to remove much duplication of code and solve issues
4557 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4558 further details). This change does have some associated API changes.
4559 Notably the SSL_state() function has been removed and replaced by
4560 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4561 SSL_set_state() has been removed altogether. The previous handshake states
4562 defined in ssl.h and ssl3.h have also been removed.
4566 * All instances of the string "ssleay" in the public API were replaced
4567 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4568 Some error codes related to internal RSA_eay API's were renamed.
4572 * The demo files in crypto/threads were moved to demo/threads.
4576 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4579 *Matt Caswell, Rich Salz*
4581 * New ASN.1 embed macro.
4583 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4584 structure is not allocated: it is part of the parent. That is instead of
4592 This reduces memory fragmentation and make it impossible to accidentally
4593 set a mandatory field to NULL.
4595 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4596 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4597 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4602 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4606 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4607 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4608 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4609 DES and RC4 ciphersuites.
4613 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4614 This changes the decoding behaviour for some invalid messages,
4615 though the change is mostly in the more lenient direction, and
4616 legacy behaviour is preserved as much as possible.
4620 * Fix no-stdio build.
4621 *David Woodhouse <David.Woodhouse@intel.com> and also*
4622 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4624 * New testing framework
4625 The testing framework has been largely rewritten and is now using
4626 perl and the perl modules Test::Harness and an extended variant of
4627 Test::More called OpenSSL::Test to do its work. All test scripts in
4628 test/ have been rewritten into test recipes, and all direct calls to
4629 executables in test/Makefile have become individual recipes using the
4630 simplified testing OpenSSL::Test::Simple.
4632 For documentation on our testing modules, do:
4634 perldoc test/testlib/OpenSSL/Test/Simple.pm
4635 perldoc test/testlib/OpenSSL/Test.pm
4639 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4640 are used; the latter aborts on memory leaks (usually checked on exit).
4641 Some undocumented "set malloc, etc., hooks" functions were removed
4642 and others were changed. All are now documented.
4646 * In DSA_generate_parameters_ex, if the provided seed is too short,
4649 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4651 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4652 from RFC4279, RFC4785, RFC5487, RFC5489.
4654 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4655 original RSA_PSK patch.
4659 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4660 era flag was never set throughout the codebase (only read). Also removed
4661 SSL3_FLAGS_POP_BUFFER which was only used if
4662 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4666 * Changed the default name options in the "ca", "crl", "req" and "x509"
4667 to be "oneline" instead of "compat".
4671 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4672 not aware of clients that still exhibit this bug, and the workaround
4673 hasn't been working properly for a while.
4677 * The return type of BIO_number_read() and BIO_number_written() as well as
4678 the corresponding num_read and num_write members in the BIO structure has
4679 changed from unsigned long to uint64_t. On platforms where an unsigned
4680 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4685 * Given the pervasive nature of TLS extensions it is inadvisable to run
4686 OpenSSL without support for them. It also means that maintaining
4687 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4688 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4692 * Removed support for the two export grade static DH ciphersuites
4693 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4694 were newly added (along with a number of other static DH ciphersuites) to
4695 1.0.2. However the two export ones have *never* worked since they were
4696 introduced. It seems strange in any case to be adding new export
4697 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4701 * Version negotiation has been rewritten. In particular SSLv23_method(),
4702 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4703 and turned into macros which simply call the new preferred function names
4704 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4705 should use the new names instead. Also as part of this change the ssl23.h
4706 header file has been removed.
4710 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4711 code and the associated standard is no longer considered fit-for-purpose.
4715 * RT2547 was closed. When generating a private key, try to make the
4716 output file readable only by the owner. This behavior change might
4717 be noticeable when interacting with other software.
4719 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4724 * Added HTTP GET support to the ocsp command.
4728 * Changed default digest for the dgst and enc commands from MD5 to
4733 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4737 * Added support for TLS extended master secret from
4738 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4739 initial patch which was a great help during development.
4743 * All libssl internal structures have been removed from the public header
4744 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4745 now redundant). Users should not attempt to access internal structures
4746 directly. Instead they should use the provided API functions.
4750 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4751 Access to deprecated functions can be re-enabled by running config with
4752 "enable-deprecated". In addition applications wishing to use deprecated
4753 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4754 will, by default, disable some transitive includes that previously existed
4755 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4759 * Added support for OCB mode. OpenSSL has been granted a patent license
4760 compatible with the OpenSSL license for use of OCB. Details are available
4761 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4762 for OCB can be removed by calling config with no-ocb.
4766 * SSLv2 support has been removed. It still supports receiving an SSLv2
4767 compatible client hello.
4771 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4772 done while fixing the error code for the key-too-small case.
4774 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4776 * CA.sh has been removed; use CA.pl instead.
4780 * Removed old DES API.
4784 * Remove various unsupported platforms:
4790 Sinix/ReliantUNIX RM400
4795 16-bit platforms such as WIN16
4799 * Clean up OPENSSL_NO_xxx #define's
4800 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4801 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4802 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4803 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4804 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4805 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4806 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4807 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4808 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4809 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4813 * Cleaned up dead code
4814 Remove all but one '#ifdef undef' which is to be looked at.
4818 * Clean up calling of xxx_free routines.
4819 Just like free(), fix most of the xxx_free routines to accept
4820 NULL. Remove the non-null checks from callers. Save much code.
4824 * Add secure heap for storage of private keys (when possible).
4825 Add BIO_s_secmem(), CBIGNUM, etc.
4826 Contributed by Akamai Technologies under our Corporate CLA.
4830 * Experimental support for a new, fast, unbiased prime candidate generator,
4831 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4833 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4835 * New output format NSS in the sess_id command line tool. This allows
4836 exporting the session id and the master key in NSS keylog format.
4838 *Martin Kaiser <martin@kaiser.cx>*
4840 * Harmonize version and its documentation. -f flag is used to display
4843 *mancha <mancha1@zoho.com>*
4845 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4846 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4848 *mancha <mancha1@zoho.com>*
4850 * Fix some double frees. These are not thought to be exploitable.
4852 *mancha <mancha1@zoho.com>*
4854 * A missing bounds check in the handling of the TLS heartbeat extension
4855 can be used to reveal up to 64k of memory to a connected client or
4858 Thanks for Neel Mehta of Google Security for discovering this bug and to
4859 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4860 preparing the fix ([CVE-2014-0160])
4862 *Adam Langley, Bodo Moeller*
4864 * Fix for the attack described in the paper "Recovering OpenSSL
4865 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4866 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4867 <http://eprint.iacr.org/2014/140>
4869 Thanks to Yuval Yarom and Naomi Benger for discovering this
4870 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4872 *Yuval Yarom and Naomi Benger*
4874 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4875 this fixes a limitation in previous versions of OpenSSL.
4879 * Experimental encrypt-then-mac support.
4881 Experimental support for encrypt then mac from
4882 draft-gutmann-tls-encrypt-then-mac-02.txt
4884 To enable it set the appropriate extension number (0x42 for the test
4885 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4887 For non-compliant peers (i.e. just about everything) this should have no
4890 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4894 * Add EVP support for key wrapping algorithms, to avoid problems with
4895 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4896 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4897 algorithms and include tests cases.
4901 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4906 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4907 MGF1 digest and OAEP label.
4911 * Make openssl verify return errors.
4913 *Chris Palmer <palmer@google.com> and Ben Laurie*
4915 * New function ASN1_TIME_diff to calculate the difference between two
4916 ASN1_TIME structures or one structure and the current time.
4920 * Update fips_test_suite to support multiple command line options. New
4921 test to induce all self test errors in sequence and check expected
4926 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4927 sign or verify all in one operation.
4931 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4932 test programs and fips_test_suite. Includes functionality to parse
4933 the minimal script output of fipsalgest.pl directly.
4937 * Add authorisation parameter to FIPS_module_mode_set().
4941 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4945 * Use separate DRBG fields for internal and external flags. New function
4946 FIPS_drbg_health_check() to perform on demand health checking. Add
4947 generation tests to fips_test_suite with reduced health check interval to
4948 demonstrate periodic health checking. Add "nodh" option to
4949 fips_test_suite to skip very slow DH test.
4953 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4958 * More extensive health check for DRBG checking many more failure modes.
4959 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4960 combination: call this in fips_test_suite.
4964 * Add support for canonical generation of DSA parameter 'g'. See
4967 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4968 POST to handle HMAC cases.
4972 * Add functions FIPS_module_version() and FIPS_module_version_text()
4973 to return numerical and string versions of the FIPS module number.
4977 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4978 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4979 outside the validated module in the FIPS capable OpenSSL.
4983 * Minor change to DRBG entropy callback semantics. In some cases
4984 there is no multiple of the block length between min_len and
4985 max_len. Allow the callback to return more than max_len bytes
4986 of entropy but discard any extra: it is the callback's responsibility
4987 to ensure that the extra data discarded does not impact the
4988 requested amount of entropy.
4992 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4993 information in FIPS186-3, SP800-57 and SP800-131A.
4997 * CCM support via EVP. Interface is very similar to GCM case except we
4998 must supply all data in one chunk (i.e. no update, final) and the
4999 message length must be supplied if AAD is used. Add algorithm test
5004 * Initial version of POST overhaul. Add POST callback to allow the status
5005 of POST to be monitored and/or failures induced. Modify fips_test_suite
5006 to use callback. Always run all selftests even if one fails.
5010 * XTS support including algorithm test driver in the fips_gcmtest program.
5011 Note: this does increase the maximum key length from 32 to 64 bytes but
5012 there should be no binary compatibility issues as existing applications
5013 will never use XTS mode.
5017 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5018 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5019 performs algorithm blocking for unapproved PRNG types. Also do not
5020 set PRNG type in FIPS_mode_set(): leave this to the application.
5021 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5022 the standard OpenSSL PRNG: set additional data to a date time vector.
5026 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5027 This shouldn't present any incompatibility problems because applications
5028 shouldn't be using these directly and any that are will need to rethink
5029 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5033 * Extensive self tests and health checking required by SP800-90 DRBG.
5034 Remove strength parameter from FIPS_drbg_instantiate and always
5035 instantiate at maximum supported strength.
5039 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5043 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5047 * New function DH_compute_key_padded() to compute a DH key and pad with
5048 leading zeroes if needed: this complies with SP800-56A et al.
5052 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5053 anything, incomplete, subject to change and largely untested at present.
5057 * Modify fipscanisteronly build option to only build the necessary object
5058 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5062 * Add experimental option FIPSSYMS to give all symbols in
5063 fipscanister.o and FIPS or fips prefix. This will avoid
5064 conflicts with future versions of OpenSSL. Add perl script
5065 util/fipsas.pl to preprocess assembly language source files
5066 and rename any affected symbols.
5070 * Add selftest checks and algorithm block of non-fips algorithms in
5071 FIPS mode. Remove DES2 from selftests.
5075 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5076 return internal method without any ENGINE dependencies. Add new
5077 tiny fips sign and verify functions.
5081 * New build option no-ec2m to disable characteristic 2 code.
5085 * New build option "fipscanisteronly". This only builds fipscanister.o
5086 and (currently) associated fips utilities. Uses the file Makefile.fips
5087 instead of Makefile.org as the prototype.
5091 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5092 Update fips_gcmtest to use IV generator.
5096 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5097 setting output buffer to NULL. The `*Final` function must be
5098 called although it will not retrieve any additional data. The tag
5099 can be set or retrieved with a ctrl. The IV length is by default 12
5100 bytes (96 bits) but can be set to an alternative value. If the IV
5101 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5106 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5107 underlying do_cipher function handles all cipher semantics itself
5108 including padding and finalisation. This is useful if (for example)
5109 an ENGINE cipher handles block padding itself. The behaviour of
5110 do_cipher is subtly changed if this flag is set: the return value
5111 is the number of characters written to the output buffer (zero is
5112 no longer an error code) or a negative error code. Also if the
5113 input buffer is NULL and length 0 finalisation should be performed.
5117 * If a candidate issuer certificate is already part of the constructed
5118 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5122 * Improve forward-security support: add functions
5124 void SSL_CTX_set_not_resumable_session_callback(
5125 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5126 void SSL_set_not_resumable_session_callback(
5127 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5129 for use by SSL/TLS servers; the callback function will be called whenever a
5130 new session is created, and gets to decide whether the session may be
5131 cached to make it resumable (return 0) or not (return 1). (As by the
5132 SSL/TLS protocol specifications, the session_id sent by the server will be
5133 empty to indicate that the session is not resumable; also, the server will
5134 not generate RFC 4507 (RFC 5077) session tickets.)
5136 A simple reasonable callback implementation is to return is_forward_secure.
5137 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5138 by the SSL/TLS server library, indicating whether it can provide forward
5141 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5143 * New -verify_name option in command line utilities to set verification
5148 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5149 Add CMAC pkey methods.
5153 * Experimental renegotiation in s_server -www mode. If the client
5154 browses /reneg connection is renegotiated. If /renegcert it is
5155 renegotiated requesting a certificate.
5159 * Add an "external" session cache for debugging purposes to s_server. This
5160 should help trace issues which normally are only apparent in deployed
5161 multi-process servers.
5165 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5166 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5167 BIO_set_cipher() and some obscure PEM functions were changed so they
5168 can now return an error. The RAND changes required a change to the
5169 RAND_METHOD structure.
5173 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5174 a gcc attribute to warn if the result of a function is ignored. This
5175 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5176 whose return value is often ignored.
5180 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5181 These allow SCTs (signed certificate timestamps) to be requested and
5182 validated when establishing a connection.
5184 *Rob Percival <robpercival@google.com>*
5189 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5191 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5192 used even when parsing explicit parameters, when loading a encoded key
5193 or calling `EC_GROUP_new_from_ecpkparameters()`/
5194 `EC_GROUP_new_from_ecparameters()`.
5195 This prevents bypass of security hardening and performance gains,
5196 especially for curves with specialized EC_METHODs.
5197 By default, if a key encoded with explicit parameters is loaded and later
5198 encoded, the output is still encoded with explicit parameters, even if
5199 internally a "named" EC_GROUP is used for computation.
5203 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5204 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5205 NULL. After this change, only the cofactor parameter can be NULL. It also
5206 does some minimal sanity checks on the passed order.
5211 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5212 An attack is simple, if the first CMS_recipientInfo is valid but the
5213 second CMS_recipientInfo is chosen ciphertext. If the second
5214 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5215 encryption key will be replaced by garbage, and the message cannot be
5216 decoded, but if the RSA decryption fails, the correct encryption key is
5217 used and the recipient will not notice the attack.
5218 As a work around for this potential attack the length of the decrypted
5219 key must be equal to the cipher default key length, in case the
5220 certificate is not given and all recipientInfo are tried out.
5221 The old behaviour can be re-enabled in the CMS code by setting the
5222 CMS_DEBUG_DECRYPT flag.
5227 * Document issue with installation paths in diverse Windows builds
5229 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5230 binaries and run-time config file.
5235 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5237 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5238 This changes the size when using the `genpkey` command when no size is given.
5239 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5240 generation commands to use 2048 bits by default.
5244 * Add FIPS support for Android Arm 64-bit
5246 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5247 Module in Version 2.0.10. For some reason, the corresponding target
5248 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5249 built with FIPS support on Android Arm 64-bit. This omission has been
5252 *Matthias St. Pierre*
5254 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5256 * 0-byte record padding oracle
5258 If an application encounters a fatal protocol error and then calls
5259 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5260 then OpenSSL can respond differently to the calling application if a 0 byte
5261 record is received with invalid padding compared to if a 0 byte record is
5262 received with an invalid MAC. If the application then behaves differently
5263 based on that in a way that is detectable to the remote peer, then this
5264 amounts to a padding oracle that could be used to decrypt data.
5266 In order for this to be exploitable "non-stitched" ciphersuites must be in
5267 use. Stitched ciphersuites are optimised implementations of certain
5268 commonly used ciphersuites. Also the application must call SSL_shutdown()
5269 twice even if a protocol error has occurred (applications should not do
5270 this but some do anyway).
5272 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5273 Aviram, with additional investigation by Steven Collison and Andrew
5274 Hourselt. It was reported to OpenSSL on 10th December 2018.
5279 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5283 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5285 * Microarchitecture timing vulnerability in ECC scalar multiplication
5287 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5288 shown to be vulnerable to a microarchitecture timing side channel attack.
5289 An attacker with sufficient access to mount local timing attacks during
5290 ECDSA signature generation could recover the private key.
5292 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5293 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5299 * Timing vulnerability in DSA signature generation
5301 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5302 timing side channel attack. An attacker could use variations in the signing
5303 algorithm to recover the private key.
5305 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5310 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5311 Module, accidentally introduced while backporting security fixes from the
5312 development branch and hindering the use of ECC in FIPS mode.
5316 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5318 * Client DoS due to large DH parameter
5320 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5321 malicious server can send a very large prime value to the client. This will
5322 cause the client to spend an unreasonably long period of time generating a
5323 key for this prime resulting in a hang until the client has finished. This
5324 could be exploited in a Denial Of Service attack.
5326 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5331 * Cache timing vulnerability in RSA Key Generation
5333 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5334 a cache timing side channel attack. An attacker with sufficient access to
5335 mount cache timing attacks during the RSA key generation process could
5336 recover the private key.
5338 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5339 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5344 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5345 parameter is no longer accepted, as it leads to a corrupt table. NULL
5346 pem_str is reserved for alias entries only.
5350 * Revert blinding in ECDSA sign and instead make problematic addition
5351 length-invariant. Switch even to fixed-length Montgomery multiplication.
5355 * Change generating and checking of primes so that the error rate of not
5356 being prime depends on the intended use based on the size of the input.
5357 For larger primes this will result in more rounds of Miller-Rabin.
5358 The maximal error rate for primes with more than 1080 bits is lowered
5361 *Kurt Roeckx, Annie Yousar*
5363 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5367 * Add blinding to ECDSA and DSA signatures to protect against side channel
5368 attacks discovered by Keegan Ryan (NCC Group).
5372 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5373 now allow empty (zero character) pass phrases.
5377 * Certificate time validation (X509_cmp_time) enforces stricter
5378 compliance with RFC 5280. Fractional seconds and timezone offsets
5379 are no longer allowed.
5383 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5385 * Constructed ASN.1 types with a recursive definition could exceed the stack
5387 Constructed ASN.1 types with a recursive definition (such as can be found
5388 in PKCS7) could eventually exceed the stack given malicious input with
5389 excessive recursion. This could result in a Denial Of Service attack. There
5390 are no such structures used within SSL/TLS that come from untrusted sources
5391 so this is considered safe.
5393 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5399 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5401 * Read/write after SSL object in error state
5403 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5404 mechanism. The intent was that if a fatal error occurred during a handshake
5405 then OpenSSL would move into the error state and would immediately fail if
5406 you attempted to continue the handshake. This works as designed for the
5407 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5408 SSL_connect()), however due to a bug it does not work correctly if
5409 SSL_read() or SSL_write() is called directly. In that scenario, if the
5410 handshake fails then a fatal error will be returned in the initial function
5411 call. If SSL_read()/SSL_write() is subsequently called by the application
5412 for the same SSL object then it will succeed and the data is passed without
5413 being decrypted/encrypted directly from the SSL/TLS record layer.
5415 In order to exploit this issue an application bug would have to be present
5416 that resulted in a call to SSL_read()/SSL_write() being issued after having
5417 already received a fatal error.
5419 This issue was reported to OpenSSL by David Benjamin (Google).
5424 * rsaz_1024_mul_avx2 overflow bug on x86_64
5426 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5427 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5428 Analysis suggests that attacks against RSA and DSA as a result of this
5429 defect would be very difficult to perform and are not believed likely.
5430 Attacks against DH1024 are considered just feasible, because most of the
5431 work necessary to deduce information about a private key may be performed
5432 offline. The amount of resources required for such an attack would be
5433 significant. However, for an attack on TLS to be meaningful, the server
5434 would have to share the DH1024 private key among multiple clients, which is
5435 no longer an option since CVE-2016-0701.
5437 This only affects processors that support the AVX2 but not ADX extensions
5438 like Intel Haswell (4th generation).
5440 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5441 was originally found via the OSS-Fuzz project.
5446 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5448 * bn_sqrx8x_internal carry bug on x86_64
5450 There is a carry propagating bug in the x86_64 Montgomery squaring
5451 procedure. No EC algorithms are affected. Analysis suggests that attacks
5452 against RSA and DSA as a result of this defect would be very difficult to
5453 perform and are not believed likely. Attacks against DH are considered just
5454 feasible (although very difficult) because most of the work necessary to
5455 deduce information about a private key may be performed offline. The amount
5456 of resources required for such an attack would be very significant and
5457 likely only accessible to a limited number of attackers. An attacker would
5458 additionally need online access to an unpatched system using the target
5459 private key in a scenario with persistent DH parameters and a private
5460 key that is shared between multiple clients.
5462 This only affects processors that support the BMI1, BMI2 and ADX extensions
5463 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5465 This issue was reported to OpenSSL by the OSS-Fuzz project.
5470 * Malformed X.509 IPAddressFamily could cause OOB read
5472 If an X.509 certificate has a malformed IPAddressFamily extension,
5473 OpenSSL could do a one-byte buffer overread. The most likely result
5474 would be an erroneous display of the certificate in text format.
5476 This issue was reported to OpenSSL by the OSS-Fuzz project.
5480 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5482 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5483 platform rather than 'mingw'.
5487 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5489 * Truncated packet could crash via OOB read
5491 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5492 cipher is being used, then a truncated packet can cause that host to
5493 perform an out-of-bounds read, usually resulting in a crash.
5495 This issue was reported to OpenSSL by Robert Święcki of Google.
5500 * BN_mod_exp may produce incorrect results on x86_64
5502 There is a carry propagating bug in the x86_64 Montgomery squaring
5503 procedure. No EC algorithms are affected. Analysis suggests that attacks
5504 against RSA and DSA as a result of this defect would be very difficult to
5505 perform and are not believed likely. Attacks against DH are considered just
5506 feasible (although very difficult) because most of the work necessary to
5507 deduce information about a private key may be performed offline. The amount
5508 of resources required for such an attack would be very significant and
5509 likely only accessible to a limited number of attackers. An attacker would
5510 additionally need online access to an unpatched system using the target
5511 private key in a scenario with persistent DH parameters and a private
5512 key that is shared between multiple clients. For example this can occur by
5513 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5514 similar to CVE-2015-3193 but must be treated as a separate problem.
5516 This issue was reported to OpenSSL by the OSS-Fuzz project.
5521 * Montgomery multiplication may produce incorrect results
5523 There is a carry propagating bug in the Broadwell-specific Montgomery
5524 multiplication procedure that handles input lengths divisible by, but
5525 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5526 and DH private keys are impossible. This is because the subroutine in
5527 question is not used in operations with the private key itself and an input
5528 of the attacker's direct choice. Otherwise the bug can manifest itself as
5529 transient authentication and key negotiation failures or reproducible
5530 erroneous outcome of public-key operations with specially crafted input.
5531 Among EC algorithms only Brainpool P-512 curves are affected and one
5532 presumably can attack ECDH key negotiation. Impact was not analyzed in
5533 detail, because pre-requisites for attack are considered unlikely. Namely
5534 multiple clients have to choose the curve in question and the server has to
5535 share the private key among them, neither of which is default behaviour.
5536 Even then only clients that chose the curve will be affected.
5538 This issue was publicly reported as transient failures and was not
5539 initially recognized as a security issue. Thanks to Richard Morgan for
5540 providing reproducible case.
5545 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5546 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5547 prevent issues where no progress is being made and the peer continually
5548 sends unrecognised record types, using up resources processing them.
5552 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5554 * Missing CRL sanity check
5556 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5557 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5558 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5560 This issue only affects the OpenSSL 1.0.2i
5565 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5567 * OCSP Status Request extension unbounded memory growth
5569 A malicious client can send an excessively large OCSP Status Request
5570 extension. If that client continually requests renegotiation, sending a
5571 large OCSP Status Request extension each time, then there will be unbounded
5572 memory growth on the server. This will eventually lead to a Denial Of
5573 Service attack through memory exhaustion. Servers with a default
5574 configuration are vulnerable even if they do not support OCSP. Builds using
5575 the "no-ocsp" build time option are not affected.
5577 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5582 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5585 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5591 * OOB write in MDC2_Update()
5593 An overflow can occur in MDC2_Update() either if called directly or
5594 through the EVP_DigestUpdate() function using MDC2. If an attacker
5595 is able to supply very large amounts of input data after a previous
5596 call to EVP_EncryptUpdate() with a partial block then a length check
5597 can overflow resulting in a heap corruption.
5599 The amount of data needed is comparable to SIZE_MAX which is impractical
5602 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5607 * Malformed SHA512 ticket DoS
5609 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5610 DoS attack where a malformed ticket will result in an OOB read which will
5613 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5614 a custom server callback and ticket lookup mechanism.
5616 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5621 * OOB write in BN_bn2dec()
5623 The function BN_bn2dec() does not check the return value of BN_div_word().
5624 This can cause an OOB write if an application uses this function with an
5625 overly large BIGNUM. This could be a problem if an overly large certificate
5626 or CRL is printed out from an untrusted source. TLS is not affected because
5627 record limits will reject an oversized certificate before it is parsed.
5629 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5634 * OOB read in TS_OBJ_print_bio()
5636 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5637 the total length the OID text representation would use and not the amount
5638 of data written. This will result in OOB reads when large OIDs are
5641 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5646 * Pointer arithmetic undefined behaviour
5648 Avoid some undefined pointer arithmetic
5650 A common idiom in the codebase is to check limits in the following manner:
5653 Where "p" points to some malloc'd data of SIZE bytes and
5656 "len" here could be from some externally supplied data (e.g. from a TLS
5659 The rules of C pointer arithmetic are such that "p + len" is only well
5660 defined where len <= SIZE. Therefore the above idiom is actually
5661 undefined behaviour.
5663 For example this could cause problems if some malloc implementation
5664 provides an address for "p" such that "p + len" actually overflows for
5665 values of len that are too big and therefore p + len < limit.
5667 This issue was reported to OpenSSL by Guido Vranken
5672 * Constant time flag not preserved in DSA signing
5674 Operations in the DSA signing algorithm should run in constant time in
5675 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5676 implementation means that a non-constant time codepath is followed for
5677 certain operations. This has been demonstrated through a cache-timing
5678 attack to be sufficient for an attacker to recover the private DSA key.
5680 This issue was reported by César Pereida (Aalto University), Billy Brumley
5681 (Tampere University of Technology), and Yuval Yarom (The University of
5682 Adelaide and NICTA).
5687 * DTLS buffered message DoS
5689 In a DTLS connection where handshake messages are delivered out-of-order
5690 those messages that OpenSSL is not yet ready to process will be buffered
5691 for later use. Under certain circumstances, a flaw in the logic means that
5692 those messages do not get removed from the buffer even though the handshake
5693 has been completed. An attacker could force up to approx. 15 messages to
5694 remain in the buffer when they are no longer required. These messages will
5695 be cleared when the DTLS connection is closed. The default maximum size for
5696 a message is 100k. Therefore, the attacker could force an additional 1500k
5697 to be consumed per connection. By opening many simultaneous connections an
5698 attacker could cause a DoS attack through memory exhaustion.
5700 This issue was reported to OpenSSL by Quan Luo.
5705 * DTLS replay protection DoS
5707 A flaw in the DTLS replay attack protection mechanism means that records
5708 that arrive for future epochs update the replay protection "window" before
5709 the MAC for the record has been validated. This could be exploited by an
5710 attacker by sending a record for the next epoch (which does not have to
5711 decrypt or have a valid MAC), with a very large sequence number. This means
5712 that all subsequent legitimate packets are dropped causing a denial of
5713 service for a specific DTLS connection.
5715 This issue was reported to OpenSSL by the OCAP audit team.
5720 * Certificate message OOB reads
5722 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5723 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5724 theoretical DoS risk but this has not been observed in practice on common
5727 The messages affected are client certificate, client certificate request
5728 and server certificate. As a result the attack can only be performed
5729 against a client or a server which enables client authentication.
5731 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5736 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5738 * Prevent padding oracle in AES-NI CBC MAC check
5740 A MITM attacker can use a padding oracle attack to decrypt traffic
5741 when the connection uses an AES CBC cipher and the server support
5744 This issue was introduced as part of the fix for Lucky 13 padding
5745 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5746 constant time by making sure that always the same bytes are read and
5747 compared against either the MAC or padding bytes. But it no longer
5748 checked that there was enough data to have both the MAC and padding
5751 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5755 * Fix EVP_EncodeUpdate overflow
5757 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5758 Base64 encoding of binary data. If an attacker is able to supply very large
5759 amounts of input data then a length check can overflow resulting in a heap
5762 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5763 the `PEM_write_bio*` family of functions. These are mainly used within the
5764 OpenSSL command line applications, so any application which processes data
5765 from an untrusted source and outputs it as a PEM file should be considered
5766 vulnerable to this issue. User applications that call these APIs directly
5767 with large amounts of untrusted data may also be vulnerable.
5769 This issue was reported by Guido Vranken.
5774 * Fix EVP_EncryptUpdate overflow
5776 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5777 is able to supply very large amounts of input data after a previous call to
5778 EVP_EncryptUpdate() with a partial block then a length check can overflow
5779 resulting in a heap corruption. Following an analysis of all OpenSSL
5780 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5781 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5782 the first called function after an EVP_EncryptInit(), and therefore that
5783 specific call must be safe. The second form is where the length passed to
5784 EVP_EncryptUpdate() can be seen from the code to be some small value and
5785 therefore there is no possibility of an overflow. Since all instances are
5786 one of these two forms, it is believed that there can be no overflows in
5787 internal code due to this problem. It should be noted that
5788 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5789 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5790 of these calls have also been analysed too and it is believed there are no
5791 instances in internal usage where an overflow could occur.
5793 This issue was reported by Guido Vranken.
5798 * Prevent ASN.1 BIO excessive memory allocation
5800 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5801 a short invalid encoding can cause allocation of large amounts of memory
5802 potentially consuming excessive resources or exhausting memory.
5804 Any application parsing untrusted data through d2i BIO functions is
5805 affected. The memory based functions such as d2i_X509() are *not* affected.
5806 Since the memory based functions are used by the TLS library, TLS
5807 applications are not affected.
5809 This issue was reported by Brian Carpenter.
5816 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5817 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5818 in arbitrary stack data being returned in the buffer.
5820 This issue was reported by Guido Vranken.
5825 * Modify behavior of ALPN to invoke callback after SNI/servername
5826 callback, such that updates to the SSL_CTX affect ALPN.
5830 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5835 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5836 methods are enabled and ssl2 is disabled the methods return NULL.
5840 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5842 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5843 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5844 provide any "EXPORT" or "LOW" strength ciphers.
5848 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5849 is by default disabled at build-time. Builds that are not configured with
5850 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5851 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5852 will need to explicitly call either of:
5854 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5856 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5858 as appropriate. Even if either of those is used, or the application
5859 explicitly uses the version-specific SSLv2_method() or its client and
5860 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5861 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5862 ciphers, and SSLv2 56-bit DES are no longer available.
5867 * Fix a double-free in DSA code
5869 A double free bug was discovered when OpenSSL parses malformed DSA private
5870 keys and could lead to a DoS attack or memory corruption for applications
5871 that receive DSA private keys from untrusted sources. This scenario is
5874 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5880 * Disable SRP fake user seed to address a server memory leak.
5882 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5884 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5885 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5886 was changed to ignore the "fake user" SRP seed, even if the seed
5889 Users should use SRP_VBASE_get1_by_user instead. Note that in
5890 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5891 also that even though configuring the SRP seed attempts to hide
5892 invalid usernames by continuing the handshake with fake
5893 credentials, this behaviour is not constant time and no strong
5894 guarantees are made that the handshake is indistinguishable from
5895 that of a valid user.
5900 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5902 In the BN_hex2bn function the number of hex digits is calculated using an
5903 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5904 large values of `i` this can result in `bn_expand` not allocating any
5905 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5906 field as NULL leading to a subsequent NULL ptr deref. For very large values
5907 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5908 In this case memory is allocated to the internal BIGNUM data field, but it
5909 is insufficiently sized leading to heap corruption. A similar issue exists
5910 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5911 is ever called by user applications with very large untrusted hex/dec data.
5912 This is anticipated to be a rare occurrence.
5914 All OpenSSL internal usage of these functions use data that is not expected
5915 to be untrusted, e.g. config file data or application command line
5916 arguments. If user developed applications generate config file data based
5917 on untrusted data then it is possible that this could also lead to security
5918 consequences. This is also anticipated to be rare.
5920 This issue was reported to OpenSSL by Guido Vranken.
5925 * Fix memory issues in `BIO_*printf` functions
5927 The internal `fmtstr` function used in processing a "%s" format string in
5928 the `BIO_*printf` functions could overflow while calculating the length of a
5929 string and cause an OOB read when printing very long strings.
5931 Additionally the internal `doapr_outch` function can attempt to write to an
5932 OOB memory location (at an offset from the NULL pointer) in the event of a
5933 memory allocation failure. In 1.0.2 and below this could be caused where
5934 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5935 could be in processing a very long "%s" format string. Memory leaks can
5938 The first issue may mask the second issue dependent on compiler behaviour.
5939 These problems could enable attacks where large amounts of untrusted data
5940 is passed to the `BIO_*printf` functions. If applications use these functions
5941 in this way then they could be vulnerable. OpenSSL itself uses these
5942 functions when printing out human-readable dumps of ASN.1 data. Therefore
5943 applications that print this data could be vulnerable if the data is from
5944 untrusted sources. OpenSSL command line applications could also be
5945 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5946 as command line arguments.
5948 Libssl is not considered directly vulnerable. Additionally certificates etc
5949 received via remote connections via libssl are also unlikely to be able to
5950 trigger these issues because of message size limits enforced within libssl.
5952 This issue was reported to OpenSSL Guido Vranken.
5957 * Side channel attack on modular exponentiation
5959 A side-channel attack was found which makes use of cache-bank conflicts on
5960 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5961 of RSA keys. The ability to exploit this issue is limited as it relies on
5962 an attacker who has control of code in a thread running on the same
5963 hyper-threaded core as the victim thread which is performing decryptions.
5965 This issue was reported to OpenSSL by Yuval Yarom, The University of
5966 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5967 Nadia Heninger, University of Pennsylvania with more information at
5968 <http://cachebleed.info>.
5973 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5974 if no keysize is specified with default_bits. This fixes an
5975 omission in an earlier change that changed all RSA/DSA key generation
5976 commands to use 2048 bits by default.
5980 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5982 * DH small subgroups
5984 Historically OpenSSL only ever generated DH parameters based on "safe"
5985 primes. More recently (in version 1.0.2) support was provided for
5986 generating X9.42 style parameter files such as those required for RFC 5114
5987 support. The primes used in such files may not be "safe". Where an
5988 application is using DH configured with parameters based on primes that are
5989 not "safe" then an attacker could use this fact to find a peer's private
5990 DH exponent. This attack requires that the attacker complete multiple
5991 handshakes in which the peer uses the same private DH exponent. For example
5992 this could be used to discover a TLS server's private DH exponent if it's
5993 reusing the private DH exponent or it's using a static DH ciphersuite.
5995 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5996 TLS. It is not on by default. If the option is not set then the server
5997 reuses the same private DH exponent for the life of the server process and
5998 would be vulnerable to this attack. It is believed that many popular
5999 applications do set this option and would therefore not be at risk.
6001 The fix for this issue adds an additional check where a "q" parameter is
6002 available (as is the case in X9.42 based parameters). This detects the
6003 only known attack, and is the only possible defense for static DH
6004 ciphersuites. This could have some performance impact.
6006 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6007 default and cannot be disabled. This could have some performance impact.
6009 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6014 * SSLv2 doesn't block disabled ciphers
6016 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6017 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6018 been disabled, provided that the SSLv2 protocol was not also disabled via
6021 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6022 and Sebastian Schinzel.
6027 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6029 * BN_mod_exp may produce incorrect results on x86_64
6031 There is a carry propagating bug in the x86_64 Montgomery squaring
6032 procedure. No EC algorithms are affected. Analysis suggests that attacks
6033 against RSA and DSA as a result of this defect would be very difficult to
6034 perform and are not believed likely. Attacks against DH are considered just
6035 feasible (although very difficult) because most of the work necessary to
6036 deduce information about a private key may be performed offline. The amount
6037 of resources required for such an attack would be very significant and
6038 likely only accessible to a limited number of attackers. An attacker would
6039 additionally need online access to an unpatched system using the target
6040 private key in a scenario with persistent DH parameters and a private
6041 key that is shared between multiple clients. For example this can occur by
6042 default in OpenSSL DHE based SSL/TLS ciphersuites.
6044 This issue was reported to OpenSSL by Hanno Böck.
6049 * Certificate verify crash with missing PSS parameter
6051 The signature verification routines will crash with a NULL pointer
6052 dereference if presented with an ASN.1 signature using the RSA PSS
6053 algorithm and absent mask generation function parameter. Since these
6054 routines are used to verify certificate signature algorithms this can be
6055 used to crash any certificate verification operation and exploited in a
6056 DoS attack. Any application which performs certificate verification is
6057 vulnerable including OpenSSL clients and servers which enable client
6060 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6065 * X509_ATTRIBUTE memory leak
6067 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6068 memory. This structure is used by the PKCS#7 and CMS routines so any
6069 application which reads PKCS#7 or CMS data from untrusted sources is
6070 affected. SSL/TLS is not affected.
6072 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6078 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6079 This changes the decoding behaviour for some invalid messages,
6080 though the change is mostly in the more lenient direction, and
6081 legacy behaviour is preserved as much as possible.
6085 * In DSA_generate_parameters_ex, if the provided seed is too short,
6088 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6090 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6092 * Alternate chains certificate forgery
6094 During certificate verification, OpenSSL will attempt to find an
6095 alternative certificate chain if the first attempt to build such a chain
6096 fails. An error in the implementation of this logic can mean that an
6097 attacker could cause certain checks on untrusted certificates to be
6098 bypassed, such as the CA flag, enabling them to use a valid leaf
6099 certificate to act as a CA and "issue" an invalid certificate.
6101 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6106 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6108 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6109 incompatibility in the handling of HMAC. The previous ABI has now been
6114 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6116 * Malformed ECParameters causes infinite loop
6118 When processing an ECParameters structure OpenSSL enters an infinite loop
6119 if the curve specified is over a specially malformed binary polynomial
6122 This can be used to perform denial of service against any
6123 system which processes public keys, certificate requests or
6124 certificates. This includes TLS clients and TLS servers with
6125 client authentication enabled.
6127 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6132 * Exploitable out-of-bounds read in X509_cmp_time
6134 X509_cmp_time does not properly check the length of the ASN1_TIME
6135 string and can read a few bytes out of bounds. In addition,
6136 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6139 An attacker can use this to craft malformed certificates and CRLs of
6140 various sizes and potentially cause a segmentation fault, resulting in
6141 a DoS on applications that verify certificates or CRLs. TLS clients
6142 that verify CRLs are affected. TLS clients and servers with client
6143 authentication enabled may be affected if they use custom verification
6146 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6147 independently by Hanno Böck.
6152 * PKCS7 crash with missing EnvelopedContent
6154 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6155 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6156 with missing content and trigger a NULL pointer dereference on parsing.
6158 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6159 structures from untrusted sources are affected. OpenSSL clients and
6160 servers are not affected.
6162 This issue was reported to OpenSSL by Michal Zalewski (Google).
6167 * CMS verify infinite loop with unknown hash function
6169 When verifying a signedData message the CMS code can enter an infinite loop
6170 if presented with an unknown hash function OID. This can be used to perform
6171 denial of service against any system which verifies signedData messages using
6173 This issue was reported to OpenSSL by Johannes Bauer.
6178 * Race condition handling NewSessionTicket
6180 If a NewSessionTicket is received by a multi-threaded client when attempting to
6181 reuse a previous ticket then a race condition can occur potentially leading to
6182 a double free of the ticket data.
6187 * Only support 256-bit or stronger elliptic curves with the
6188 'ecdh_auto' setting (server) or by default (client). Of supported
6189 curves, prefer P-256 (both).
6193 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6195 * ClientHello sigalgs DoS fix
6197 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6198 invalid signature algorithms extension a NULL pointer dereference will
6199 occur. This can be exploited in a DoS attack against the server.
6201 This issue was was reported to OpenSSL by David Ramos of Stanford
6205 *Stephen Henson and Matt Caswell*
6207 * Multiblock corrupted pointer fix
6209 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6210 feature only applies on 64 bit x86 architecture platforms that support AES
6211 NI instructions. A defect in the implementation of "multiblock" can cause
6212 OpenSSL's internal write buffer to become incorrectly set to NULL when
6213 using non-blocking IO. Typically, when the user application is using a
6214 socket BIO for writing, this will only result in a failed connection.
6215 However if some other BIO is used then it is likely that a segmentation
6216 fault will be triggered, thus enabling a potential DoS attack.
6218 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6223 * Segmentation fault in DTLSv1_listen fix
6225 The DTLSv1_listen function is intended to be stateless and processes the
6226 initial ClientHello from many peers. It is common for user code to loop
6227 over the call to DTLSv1_listen until a valid ClientHello is received with
6228 an associated cookie. A defect in the implementation of DTLSv1_listen means
6229 that state is preserved in the SSL object from one invocation to the next
6230 that can lead to a segmentation fault. Errors processing the initial
6231 ClientHello can trigger this scenario. An example of such an error could be
6232 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6235 This issue was reported to OpenSSL by Per Allansson.
6240 * Segmentation fault in ASN1_TYPE_cmp fix
6242 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6243 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6244 certificate signature algorithm consistency this can be used to crash any
6245 certificate verification operation and exploited in a DoS attack. Any
6246 application which performs certificate verification is vulnerable including
6247 OpenSSL clients and servers which enable client authentication.
6252 * Segmentation fault for invalid PSS parameters fix
6254 The signature verification routines will crash with a NULL pointer
6255 dereference if presented with an ASN.1 signature using the RSA PSS
6256 algorithm and invalid parameters. Since these routines are used to verify
6257 certificate signature algorithms this can be used to crash any
6258 certificate verification operation and exploited in a DoS attack. Any
6259 application which performs certificate verification is vulnerable including
6260 OpenSSL clients and servers which enable client authentication.
6262 This issue was was reported to OpenSSL by Brian Carpenter.
6267 * ASN.1 structure reuse memory corruption fix
6269 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6270 memory corruption via an invalid write. Such reuse is and has been
6271 strongly discouraged and is believed to be rare.
6273 Applications that parse structures containing CHOICE or ANY DEFINED BY
6274 components may be affected. Certificate parsing (d2i_X509 and related
6275 functions) are however not affected. OpenSSL clients and servers are
6281 * PKCS7 NULL pointer dereferences fix
6283 The PKCS#7 parsing code does not handle missing outer ContentInfo
6284 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6285 missing content and trigger a NULL pointer dereference on parsing.
6287 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6288 otherwise parse PKCS#7 structures from untrusted sources are
6289 affected. OpenSSL clients and servers are not affected.
6291 This issue was reported to OpenSSL by Michal Zalewski (Google).
6296 * DoS via reachable assert in SSLv2 servers fix
6298 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6299 servers that both support SSLv2 and enable export cipher suites by sending
6300 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6302 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6303 (OpenSSL development team).
6308 * Empty CKE with client auth and DHE fix
6310 If client auth is used then a server can seg fault in the event of a DHE
6311 ciphersuite being selected and a zero length ClientKeyExchange message
6312 being sent by the client. This could be exploited in a DoS attack.
6317 * Handshake with unseeded PRNG fix
6319 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6320 with an unseeded PRNG. The conditions are:
6321 - The client is on a platform where the PRNG has not been seeded
6322 automatically, and the user has not seeded manually
6323 - A protocol specific client method version has been used (i.e. not
6324 SSL_client_methodv23)
6325 - A ciphersuite is used that does not require additional random data from
6326 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6328 If the handshake succeeds then the client random that has been used will
6329 have been generated from a PRNG with insufficient entropy and therefore the
6330 output may be predictable.
6332 For example using the following command with an unseeded openssl will
6333 succeed on an unpatched platform:
6335 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6340 * Use After Free following d2i_ECPrivatekey error fix
6342 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6343 could cause a use after free condition. This, in turn, could cause a double
6344 free in several private key parsing functions (such as d2i_PrivateKey
6345 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6346 for applications that receive EC private keys from untrusted
6347 sources. This scenario is considered rare.
6349 This issue was discovered by the BoringSSL project and fixed in their
6355 * X509_to_X509_REQ NULL pointer deref fix
6357 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6358 the certificate key is invalid. This function is rarely used in practice.
6360 This issue was discovered by Brian Carpenter.
6365 * Removed the export ciphers from the DEFAULT ciphers
6369 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6371 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6372 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6373 So far those who have to target multiple platforms would compromise
6374 and argue that binary targeting say ARMv5 would still execute on
6375 ARMv8. "Universal" build resolves this compromise by providing
6376 near-optimal performance even on newer platforms.
6380 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6381 (other platforms pending).
6383 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6385 * Add support for the SignedCertificateTimestampList certificate and
6386 OCSP response extensions from RFC6962.
6390 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6391 for corner cases. (Certain input points at infinity could lead to
6392 bogus results, with non-infinity inputs mapped to infinity too.)
6396 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6397 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6398 common cases are optimized and there still is room for further
6399 improvements. Vector Permutation AES for Altivec is also added.
6403 * Add support for little-endian ppc64 Linux target.
6405 *Marcelo Cerri (IBM)*
6407 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6408 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6409 are optimized and there still is room for further improvements.
6410 Both 32- and 64-bit modes are supported.
6412 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6414 * Improved ARMv7 NEON support.
6418 * Support for SPARC Architecture 2011 crypto extensions, first
6419 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6420 SHA256/512, MD5, GHASH and modular exponentiation.
6422 *Andy Polyakov, David Miller*
6424 * Accelerated modular exponentiation for Intel processors, a.k.a.
6427 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6429 * Support for new and upcoming Intel processors, including AVX2,
6430 BMI and SHA ISA extensions. This includes additional "stitched"
6431 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6434 This work was sponsored by Intel Corp.
6438 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6439 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6440 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6444 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6445 this fixes a limitation in previous versions of OpenSSL.
6449 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6450 MGF1 digest and OAEP label.
6454 * Add EVP support for key wrapping algorithms, to avoid problems with
6455 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6456 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6457 algorithms and include tests cases.
6461 * Add functions to allocate and set the fields of an ECDSA_METHOD
6464 *Douglas E. Engert, Steve Henson*
6466 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6467 difference in days and seconds between two tm or ASN1_TIME structures.
6471 * Add -rev test option to s_server to just reverse order of characters
6472 received by client and send back to server. Also prints an abbreviated
6473 summary of the connection parameters.
6477 * New option -brief for s_client and s_server to print out a brief summary
6478 of connection parameters.
6482 * Add callbacks for arbitrary TLS extensions.
6484 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6486 * New option -crl_download in several openssl utilities to download CRLs
6487 from CRLDP extension in certificates.
6491 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6495 * New function X509_CRL_diff to generate a delta CRL from the difference
6496 of two full CRLs. Add support to "crl" utility.
6500 * New functions to set lookup_crls function and to retrieve
6501 X509_STORE from X509_STORE_CTX.
6505 * Print out deprecated issuer and subject unique ID fields in
6510 * Extend OCSP I/O functions so they can be used for simple general purpose
6511 HTTP as well as OCSP. New wrapper function which can be used to download
6512 CRLs using the OCSP API.
6516 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6520 * `SSL_CONF*` functions. These provide a common framework for application
6521 configuration using configuration files or command lines.
6525 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6526 message callback and prints the results. Needs compile time option
6527 "enable-ssl-trace". New options to s_client and s_server to enable
6532 * New ctrl and macro to retrieve supported points extensions.
6533 Print out extension in s_server and s_client.
6537 * New functions to retrieve certificate signature and signature
6542 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6547 * New Suite B modes for TLS code. These use and enforce the requirements
6548 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6549 only use Suite B curves. The Suite B modes can be set by using the
6550 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6554 * New chain verification flags for Suite B levels of security. Check
6555 algorithms are acceptable when flags are set in X509_verify_cert.
6559 * Make tls1_check_chain return a set of flags indicating checks passed
6560 by a certificate chain. Add additional tests to handle client
6561 certificates: checks for matching certificate type and issuer name
6566 * If an attempt is made to use a signature algorithm not in the peer
6567 preference list abort the handshake. If client has no suitable
6568 signature algorithms in response to a certificate request do not
6569 use the certificate.
6573 * If server EC tmp key is not in client preference list abort handshake.
6577 * Add support for certificate stores in CERT structure. This makes it
6578 possible to have different stores per SSL structure or one store in
6579 the parent SSL_CTX. Include distinct stores for certificate chain
6580 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6581 to build and store a certificate chain in CERT structure: returning
6582 an error if the chain cannot be built: this will allow applications
6583 to test if a chain is correctly configured.
6585 Note: if the CERT based stores are not set then the parent SSL_CTX
6586 store is used to retain compatibility with existing behaviour.
6590 * New function ssl_set_client_disabled to set a ciphersuite disabled
6591 mask based on the current session, check mask when sending client
6592 hello and checking the requested ciphersuite.
6596 * New ctrls to retrieve and set certificate types in a certificate
6597 request message. Print out received values in s_client. If certificate
6598 types is not set with custom values set sensible values based on
6599 supported signature algorithms.
6603 * Support for distinct client and server supported signature algorithms.
6607 * Add certificate callback. If set this is called whenever a certificate
6608 is required by client or server. An application can decide which
6609 certificate chain to present based on arbitrary criteria: for example
6610 supported signature algorithms. Add very simple example to s_server.
6611 This fixes many of the problems and restrictions of the existing client
6612 certificate callback: for example you can now clear an existing
6613 certificate and specify the whole chain.
6617 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6618 the certificate can be used for (if anything). Set valid_flags field
6619 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6620 to have similar checks in it.
6622 Add new "cert_flags" field to CERT structure and include a "strict mode".
6623 This enforces some TLS certificate requirements (such as only permitting
6624 certificate signature algorithms contained in the supported algorithms
6625 extension) which some implementations ignore: this option should be used
6626 with caution as it could cause interoperability issues.
6630 * Update and tidy signature algorithm extension processing. Work out
6631 shared signature algorithms based on preferences and peer algorithms
6632 and print them out in s_client and s_server. Abort handshake if no
6633 shared signature algorithms.
6637 * Add new functions to allow customised supported signature algorithms
6638 for SSL and SSL_CTX structures. Add options to s_client and s_server
6643 * New function SSL_certs_clear() to delete all references to certificates
6644 from an SSL structure. Before this once a certificate had been added
6645 it couldn't be removed.
6649 * Integrate hostname, email address and IP address checking with certificate
6650 verification. New verify options supporting checking in openssl utility.
6654 * Fixes and wildcard matching support to hostname and email checking
6655 functions. Add manual page.
6657 *Florian Weimer (Red Hat Product Security Team)*
6659 * New functions to check a hostname email or IP address against a
6660 certificate. Add options x509 utility to print results of checks against
6665 * Fix OCSP checking.
6667 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6669 * Initial experimental support for explicitly trusted non-root CAs.
6670 OpenSSL still tries to build a complete chain to a root but if an
6671 intermediate CA has a trust setting included that is used. The first
6672 setting is used: whether to trust (e.g., -addtrust option to the x509
6677 * Add -trusted_first option which attempts to find certificates in the
6678 trusted store even if an untrusted chain is also supplied.
6682 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6683 platform support for Linux and Android.
6687 * Support for linux-x32, ILP32 environment in x86_64 framework.
6691 * Experimental multi-implementation support for FIPS capable OpenSSL.
6692 When in FIPS mode the approved implementations are used as normal,
6693 when not in FIPS mode the internal unapproved versions are used instead.
6694 This means that the FIPS capable OpenSSL isn't forced to use the
6695 (often lower performance) FIPS implementations outside FIPS mode.
6699 * Transparently support X9.42 DH parameters when calling
6700 PEM_read_bio_DHparameters. This means existing applications can handle
6701 the new parameter format automatically.
6705 * Initial experimental support for X9.42 DH parameter format: mainly
6706 to support use of 'q' parameter for RFC5114 parameters.
6710 * Add DH parameters from RFC5114 including test data to dhtest.
6714 * Support for automatic EC temporary key parameter selection. If enabled
6715 the most preferred EC parameters are automatically used instead of
6716 hardcoded fixed parameters. Now a server just has to call:
6717 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6718 support ECDH and use the most appropriate parameters.
6722 * Enhance and tidy EC curve and point format TLS extension code. Use
6723 static structures instead of allocation if default values are used.
6724 New ctrls to set curves we wish to support and to retrieve shared curves.
6725 Print out shared curves in s_server. New options to s_server and s_client
6726 to set list of supported curves.
6730 * New ctrls to retrieve supported signature algorithms and
6731 supported curve values as an array of NIDs. Extend openssl utility
6732 to print out received values.
6736 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6737 between NIDs and the more common NIST names such as "P-256". Enhance
6738 ecparam utility and ECC method to recognise the NIST names for curves.
6742 * Enhance SSL/TLS certificate chain handling to support different
6743 chains for each certificate instead of one chain in the parent SSL_CTX.
6747 * Support for fixed DH ciphersuite client authentication: where both
6748 server and client use DH certificates with common parameters.
6752 * Support for fixed DH ciphersuites: those requiring DH server
6757 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6759 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6760 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6761 X509_CINF_get_signature were reverted post internal team review.
6766 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6768 * OCSP Status Request extension unbounded memory growth
6770 A malicious client can send an excessively large OCSP Status Request
6771 extension. If that client continually requests renegotiation, sending a
6772 large OCSP Status Request extension each time, then there will be unbounded
6773 memory growth on the server. This will eventually lead to a Denial Of
6774 Service attack through memory exhaustion. Servers with a default
6775 configuration are vulnerable even if they do not support OCSP. Builds using
6776 the "no-ocsp" build time option are not affected.
6778 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6783 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6786 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6792 * OOB write in MDC2_Update()
6794 An overflow can occur in MDC2_Update() either if called directly or
6795 through the EVP_DigestUpdate() function using MDC2. If an attacker
6796 is able to supply very large amounts of input data after a previous
6797 call to EVP_EncryptUpdate() with a partial block then a length check
6798 can overflow resulting in a heap corruption.
6800 The amount of data needed is comparable to SIZE_MAX which is impractical
6803 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6808 * Malformed SHA512 ticket DoS
6810 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6811 DoS attack where a malformed ticket will result in an OOB read which will
6814 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6815 a custom server callback and ticket lookup mechanism.
6817 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6822 * OOB write in BN_bn2dec()
6824 The function BN_bn2dec() does not check the return value of BN_div_word().
6825 This can cause an OOB write if an application uses this function with an
6826 overly large BIGNUM. This could be a problem if an overly large certificate
6827 or CRL is printed out from an untrusted source. TLS is not affected because
6828 record limits will reject an oversized certificate before it is parsed.
6830 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6835 * OOB read in TS_OBJ_print_bio()
6837 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6838 the total length the OID text representation would use and not the amount
6839 of data written. This will result in OOB reads when large OIDs are
6842 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6847 * Pointer arithmetic undefined behaviour
6849 Avoid some undefined pointer arithmetic
6851 A common idiom in the codebase is to check limits in the following manner:
6854 Where "p" points to some malloc'd data of SIZE bytes and
6857 "len" here could be from some externally supplied data (e.g. from a TLS
6860 The rules of C pointer arithmetic are such that "p + len" is only well
6861 defined where len <= SIZE. Therefore, the above idiom is actually
6862 undefined behaviour.
6864 For example this could cause problems if some malloc implementation
6865 provides an address for "p" such that "p + len" actually overflows for
6866 values of len that are too big and therefore p + len < limit.
6868 This issue was reported to OpenSSL by Guido Vranken
6873 * Constant time flag not preserved in DSA signing
6875 Operations in the DSA signing algorithm should run in constant time in
6876 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6877 implementation means that a non-constant time codepath is followed for
6878 certain operations. This has been demonstrated through a cache-timing
6879 attack to be sufficient for an attacker to recover the private DSA key.
6881 This issue was reported by César Pereida (Aalto University), Billy Brumley
6882 (Tampere University of Technology), and Yuval Yarom (The University of
6883 Adelaide and NICTA).
6888 * DTLS buffered message DoS
6890 In a DTLS connection where handshake messages are delivered out-of-order
6891 those messages that OpenSSL is not yet ready to process will be buffered
6892 for later use. Under certain circumstances, a flaw in the logic means that
6893 those messages do not get removed from the buffer even though the handshake
6894 has been completed. An attacker could force up to approx. 15 messages to
6895 remain in the buffer when they are no longer required. These messages will
6896 be cleared when the DTLS connection is closed. The default maximum size for
6897 a message is 100k. Therefore, the attacker could force an additional 1500k
6898 to be consumed per connection. By opening many simultaneous connections an
6899 attacker could cause a DoS attack through memory exhaustion.
6901 This issue was reported to OpenSSL by Quan Luo.
6906 * DTLS replay protection DoS
6908 A flaw in the DTLS replay attack protection mechanism means that records
6909 that arrive for future epochs update the replay protection "window" before
6910 the MAC for the record has been validated. This could be exploited by an
6911 attacker by sending a record for the next epoch (which does not have to
6912 decrypt or have a valid MAC), with a very large sequence number. This means
6913 that all subsequent legitimate packets are dropped causing a denial of
6914 service for a specific DTLS connection.
6916 This issue was reported to OpenSSL by the OCAP audit team.
6921 * Certificate message OOB reads
6923 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6924 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6925 theoretical DoS risk but this has not been observed in practice on common
6928 The messages affected are client certificate, client certificate request
6929 and server certificate. As a result the attack can only be performed
6930 against a client or a server which enables client authentication.
6932 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6937 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6939 * Prevent padding oracle in AES-NI CBC MAC check
6941 A MITM attacker can use a padding oracle attack to decrypt traffic
6942 when the connection uses an AES CBC cipher and the server support
6945 This issue was introduced as part of the fix for Lucky 13 padding
6946 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6947 constant time by making sure that always the same bytes are read and
6948 compared against either the MAC or padding bytes. But it no longer
6949 checked that there was enough data to have both the MAC and padding
6952 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6957 * Fix EVP_EncodeUpdate overflow
6959 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6960 Base64 encoding of binary data. If an attacker is able to supply very large
6961 amounts of input data then a length check can overflow resulting in a heap
6964 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6965 the `PEM_write_bio*` family of functions. These are mainly used within the
6966 OpenSSL command line applications, so any application which processes data
6967 from an untrusted source and outputs it as a PEM file should be considered
6968 vulnerable to this issue. User applications that call these APIs directly
6969 with large amounts of untrusted data may also be vulnerable.
6971 This issue was reported by Guido Vranken.
6976 * Fix EVP_EncryptUpdate overflow
6978 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6979 is able to supply very large amounts of input data after a previous call to
6980 EVP_EncryptUpdate() with a partial block then a length check can overflow
6981 resulting in a heap corruption. Following an analysis of all OpenSSL
6982 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6983 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6984 the first called function after an EVP_EncryptInit(), and therefore that
6985 specific call must be safe. The second form is where the length passed to
6986 EVP_EncryptUpdate() can be seen from the code to be some small value and
6987 therefore there is no possibility of an overflow. Since all instances are
6988 one of these two forms, it is believed that there can be no overflows in
6989 internal code due to this problem. It should be noted that
6990 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6991 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6992 of these calls have also been analysed too and it is believed there are no
6993 instances in internal usage where an overflow could occur.
6995 This issue was reported by Guido Vranken.
7000 * Prevent ASN.1 BIO excessive memory allocation
7002 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7003 a short invalid encoding can casuse allocation of large amounts of memory
7004 potentially consuming excessive resources or exhausting memory.
7006 Any application parsing untrusted data through d2i BIO functions is
7007 affected. The memory based functions such as d2i_X509() are *not* affected.
7008 Since the memory based functions are used by the TLS library, TLS
7009 applications are not affected.
7011 This issue was reported by Brian Carpenter.
7018 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7019 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7020 in arbitrary stack data being returned in the buffer.
7022 This issue was reported by Guido Vranken.
7027 * Modify behavior of ALPN to invoke callback after SNI/servername
7028 callback, such that updates to the SSL_CTX affect ALPN.
7032 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7037 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7038 methods are enabled and ssl2 is disabled the methods return NULL.
7042 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7044 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7045 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7046 provide any "EXPORT" or "LOW" strength ciphers.
7050 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7051 is by default disabled at build-time. Builds that are not configured with
7052 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7053 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7054 will need to explicitly call either of:
7056 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7058 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7060 as appropriate. Even if either of those is used, or the application
7061 explicitly uses the version-specific SSLv2_method() or its client and
7062 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7063 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7064 ciphers, and SSLv2 56-bit DES are no longer available.
7069 * Fix a double-free in DSA code
7071 A double free bug was discovered when OpenSSL parses malformed DSA private
7072 keys and could lead to a DoS attack or memory corruption for applications
7073 that receive DSA private keys from untrusted sources. This scenario is
7076 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7082 * Disable SRP fake user seed to address a server memory leak.
7084 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7086 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7087 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7088 was changed to ignore the "fake user" SRP seed, even if the seed
7091 Users should use SRP_VBASE_get1_by_user instead. Note that in
7092 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7093 also that even though configuring the SRP seed attempts to hide
7094 invalid usernames by continuing the handshake with fake
7095 credentials, this behaviour is not constant time and no strong
7096 guarantees are made that the handshake is indistinguishable from
7097 that of a valid user.
7102 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7104 In the BN_hex2bn function the number of hex digits is calculated using an
7105 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7106 large values of `i` this can result in `bn_expand` not allocating any
7107 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7108 field as NULL leading to a subsequent NULL ptr deref. For very large values
7109 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7110 In this case memory is allocated to the internal BIGNUM data field, but it
7111 is insufficiently sized leading to heap corruption. A similar issue exists
7112 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7113 is ever called by user applications with very large untrusted hex/dec data.
7114 This is anticipated to be a rare occurrence.
7116 All OpenSSL internal usage of these functions use data that is not expected
7117 to be untrusted, e.g. config file data or application command line
7118 arguments. If user developed applications generate config file data based
7119 on untrusted data then it is possible that this could also lead to security
7120 consequences. This is also anticipated to be rare.
7122 This issue was reported to OpenSSL by Guido Vranken.
7127 * Fix memory issues in `BIO_*printf` functions
7129 The internal `fmtstr` function used in processing a "%s" format string in
7130 the `BIO_*printf` functions could overflow while calculating the length of a
7131 string and cause an OOB read when printing very long strings.
7133 Additionally the internal `doapr_outch` function can attempt to write to an
7134 OOB memory location (at an offset from the NULL pointer) in the event of a
7135 memory allocation failure. In 1.0.2 and below this could be caused where
7136 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7137 could be in processing a very long "%s" format string. Memory leaks can
7140 The first issue may mask the second issue dependent on compiler behaviour.
7141 These problems could enable attacks where large amounts of untrusted data
7142 is passed to the `BIO_*printf` functions. If applications use these functions
7143 in this way then they could be vulnerable. OpenSSL itself uses these
7144 functions when printing out human-readable dumps of ASN.1 data. Therefore
7145 applications that print this data could be vulnerable if the data is from
7146 untrusted sources. OpenSSL command line applications could also be
7147 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7148 as command line arguments.
7150 Libssl is not considered directly vulnerable. Additionally certificates etc
7151 received via remote connections via libssl are also unlikely to be able to
7152 trigger these issues because of message size limits enforced within libssl.
7154 This issue was reported to OpenSSL Guido Vranken.
7159 * Side channel attack on modular exponentiation
7161 A side-channel attack was found which makes use of cache-bank conflicts on
7162 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7163 of RSA keys. The ability to exploit this issue is limited as it relies on
7164 an attacker who has control of code in a thread running on the same
7165 hyper-threaded core as the victim thread which is performing decryptions.
7167 This issue was reported to OpenSSL by Yuval Yarom, The University of
7168 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7169 Nadia Heninger, University of Pennsylvania with more information at
7170 <http://cachebleed.info>.
7175 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7176 if no keysize is specified with default_bits. This fixes an
7177 omission in an earlier change that changed all RSA/DSA key generation
7178 commands to use 2048 bits by default.
7182 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7184 * Protection for DH small subgroup attacks
7186 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7187 switched on by default and cannot be disabled. This could have some
7192 * SSLv2 doesn't block disabled ciphers
7194 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7195 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7196 been disabled, provided that the SSLv2 protocol was not also disabled via
7199 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7200 and Sebastian Schinzel.
7205 * Reject DH handshakes with parameters shorter than 1024 bits.
7209 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7211 * Certificate verify crash with missing PSS parameter
7213 The signature verification routines will crash with a NULL pointer
7214 dereference if presented with an ASN.1 signature using the RSA PSS
7215 algorithm and absent mask generation function parameter. Since these
7216 routines are used to verify certificate signature algorithms this can be
7217 used to crash any certificate verification operation and exploited in a
7218 DoS attack. Any application which performs certificate verification is
7219 vulnerable including OpenSSL clients and servers which enable client
7222 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7227 * X509_ATTRIBUTE memory leak
7229 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7230 memory. This structure is used by the PKCS#7 and CMS routines so any
7231 application which reads PKCS#7 or CMS data from untrusted sources is
7232 affected. SSL/TLS is not affected.
7234 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7240 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7241 This changes the decoding behaviour for some invalid messages,
7242 though the change is mostly in the more lenient direction, and
7243 legacy behaviour is preserved as much as possible.
7247 * In DSA_generate_parameters_ex, if the provided seed is too short,
7248 use a random seed, as already documented.
7250 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7252 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7254 * Alternate chains certificate forgery
7256 During certificate verfification, OpenSSL will attempt to find an
7257 alternative certificate chain if the first attempt to build such a chain
7258 fails. An error in the implementation of this logic can mean that an
7259 attacker could cause certain checks on untrusted certificates to be
7260 bypassed, such as the CA flag, enabling them to use a valid leaf
7261 certificate to act as a CA and "issue" an invalid certificate.
7263 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7269 * Race condition handling PSK identify hint
7271 If PSK identity hints are received by a multi-threaded client then
7272 the values are wrongly updated in the parent SSL_CTX structure. This can
7273 result in a race condition potentially leading to a double free of the
7279 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7281 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7282 incompatibility in the handling of HMAC. The previous ABI has now been
7285 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7287 * Malformed ECParameters causes infinite loop
7289 When processing an ECParameters structure OpenSSL enters an infinite loop
7290 if the curve specified is over a specially malformed binary polynomial
7293 This can be used to perform denial of service against any
7294 system which processes public keys, certificate requests or
7295 certificates. This includes TLS clients and TLS servers with
7296 client authentication enabled.
7298 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7303 * Exploitable out-of-bounds read in X509_cmp_time
7305 X509_cmp_time does not properly check the length of the ASN1_TIME
7306 string and can read a few bytes out of bounds. In addition,
7307 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7310 An attacker can use this to craft malformed certificates and CRLs of
7311 various sizes and potentially cause a segmentation fault, resulting in
7312 a DoS on applications that verify certificates or CRLs. TLS clients
7313 that verify CRLs are affected. TLS clients and servers with client
7314 authentication enabled may be affected if they use custom verification
7317 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7318 independently by Hanno Böck.
7323 * PKCS7 crash with missing EnvelopedContent
7325 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7326 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7327 with missing content and trigger a NULL pointer dereference on parsing.
7329 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7330 structures from untrusted sources are affected. OpenSSL clients and
7331 servers are not affected.
7333 This issue was reported to OpenSSL by Michal Zalewski (Google).
7338 * CMS verify infinite loop with unknown hash function
7340 When verifying a signedData message the CMS code can enter an infinite loop
7341 if presented with an unknown hash function OID. This can be used to perform
7342 denial of service against any system which verifies signedData messages using
7344 This issue was reported to OpenSSL by Johannes Bauer.
7349 * Race condition handling NewSessionTicket
7351 If a NewSessionTicket is received by a multi-threaded client when attempting to
7352 reuse a previous ticket then a race condition can occur potentially leading to
7353 a double free of the ticket data.
7358 * Reject DH handshakes with parameters shorter than 768 bits.
7360 *Kurt Roeckx and Emilia Kasper*
7362 * dhparam: generate 2048-bit parameters by default.
7364 *Kurt Roeckx and Emilia Kasper*
7366 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7368 * Segmentation fault in ASN1_TYPE_cmp fix
7370 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7371 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7372 certificate signature algorithm consistency this can be used to crash any
7373 certificate verification operation and exploited in a DoS attack. Any
7374 application which performs certificate verification is vulnerable including
7375 OpenSSL clients and servers which enable client authentication.
7380 * ASN.1 structure reuse memory corruption fix
7382 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7383 memory corruption via an invalid write. Such reuse is and has been
7384 strongly discouraged and is believed to be rare.
7386 Applications that parse structures containing CHOICE or ANY DEFINED BY
7387 components may be affected. Certificate parsing (d2i_X509 and related
7388 functions) are however not affected. OpenSSL clients and servers are
7394 * PKCS7 NULL pointer dereferences fix
7396 The PKCS#7 parsing code does not handle missing outer ContentInfo
7397 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7398 missing content and trigger a NULL pointer dereference on parsing.
7400 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7401 otherwise parse PKCS#7 structures from untrusted sources are
7402 affected. OpenSSL clients and servers are not affected.
7404 This issue was reported to OpenSSL by Michal Zalewski (Google).
7409 * DoS via reachable assert in SSLv2 servers fix
7411 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7412 servers that both support SSLv2 and enable export cipher suites by sending
7413 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7415 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7416 (OpenSSL development team).
7421 * Use After Free following d2i_ECPrivatekey error fix
7423 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7424 could cause a use after free condition. This, in turn, could cause a double
7425 free in several private key parsing functions (such as d2i_PrivateKey
7426 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7427 for applications that receive EC private keys from untrusted
7428 sources. This scenario is considered rare.
7430 This issue was discovered by the BoringSSL project and fixed in their
7436 * X509_to_X509_REQ NULL pointer deref fix
7438 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7439 the certificate key is invalid. This function is rarely used in practice.
7441 This issue was discovered by Brian Carpenter.
7446 * Removed the export ciphers from the DEFAULT ciphers
7450 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7452 * Build fixes for the Windows and OpenVMS platforms
7454 *Matt Caswell and Richard Levitte*
7456 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7458 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7459 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7460 dereference. This could lead to a Denial Of Service attack. Thanks to
7461 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7466 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7467 dtls1_buffer_record function under certain conditions. In particular this
7468 could occur if an attacker sent repeated DTLS records with the same
7469 sequence number but for the next epoch. The memory leak could be exploited
7470 by an attacker in a Denial of Service attack through memory exhaustion.
7471 Thanks to Chris Mueller for reporting this issue.
7476 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7477 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7478 method would be set to NULL which could later result in a NULL pointer
7479 dereference. Thanks to Frank Schmirler for reporting this issue.
7484 * Abort handshake if server key exchange message is omitted for ephemeral
7487 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7488 reporting this issue.
7493 * Remove non-export ephemeral RSA code on client and server. This code
7494 violated the TLS standard by allowing the use of temporary RSA keys in
7495 non-export ciphersuites and could be used by a server to effectively
7496 downgrade the RSA key length used to a value smaller than the server
7497 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7498 INRIA or reporting this issue.
7503 * Fixed issue where DH client certificates are accepted without verification.
7504 An OpenSSL server will accept a DH certificate for client authentication
7505 without the certificate verify message. This effectively allows a client to
7506 authenticate without the use of a private key. This only affects servers
7507 which trust a client certificate authority which issues certificates
7508 containing DH keys: these are extremely rare and hardly ever encountered.
7509 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7515 * Ensure that the session ID context of an SSL is updated when its
7516 SSL_CTX is updated via SSL_set_SSL_CTX.
7518 The session ID context is typically set from the parent SSL_CTX,
7519 and can vary with the CTX.
7523 * Fix various certificate fingerprint issues.
7525 By using non-DER or invalid encodings outside the signed portion of a
7526 certificate the fingerprint can be changed without breaking the signature.
7527 Although no details of the signed portion of the certificate can be changed
7528 this can cause problems with some applications: e.g. those using the
7529 certificate fingerprint for blacklists.
7531 1. Reject signatures with non zero unused bits.
7533 If the BIT STRING containing the signature has non zero unused bits reject
7534 the signature. All current signature algorithms require zero unused bits.
7536 2. Check certificate algorithm consistency.
7538 Check the AlgorithmIdentifier inside TBS matches the one in the
7539 certificate signature. NB: this will result in signature failure
7540 errors for some broken certificates.
7542 Thanks to Konrad Kraszewski from Google for reporting this issue.
7544 3. Check DSA/ECDSA signatures use DER.
7546 Re-encode DSA/ECDSA signatures and compare with the original received
7547 signature. Return an error if there is a mismatch.
7549 This will reject various cases including garbage after signature
7550 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7551 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7552 (negative or with leading zeroes).
7554 Further analysis was conducted and fixes were developed by Stephen Henson
7555 of the OpenSSL core team.
7561 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7562 results on some platforms, including x86_64. This bug occurs at random
7563 with a very low probability, and is not known to be exploitable in any
7564 way, though its exact impact is difficult to determine. Thanks to Pieter
7565 Wuille (Blockstream) who reported this issue and also suggested an initial
7566 fix. Further analysis was conducted by the OpenSSL development team and
7567 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7568 the OpenSSL core team.
7573 * Do not resume sessions on the server if the negotiated protocol
7574 version does not match the session's version. Resuming with a different
7575 version, while not strictly forbidden by the RFC, is of questionable
7576 sanity and breaks all known clients.
7578 *David Benjamin, Emilia Käsper*
7580 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7581 early CCS messages during renegotiation. (Note that because
7582 renegotiation is encrypted, this early CCS was not exploitable.)
7586 * Tighten client-side session ticket handling during renegotiation:
7587 ensure that the client only accepts a session ticket if the server sends
7588 the extension anew in the ServerHello. Previously, a TLS client would
7589 reuse the old extension state and thus accept a session ticket if one was
7590 announced in the initial ServerHello.
7592 Similarly, ensure that the client requires a session ticket if one
7593 was advertised in the ServerHello. Previously, a TLS client would
7594 ignore a missing NewSessionTicket message.
7598 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7602 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7603 sends a carefully crafted handshake message, to cause OpenSSL to fail
7604 to free up to 64k of memory causing a memory leak. This could be
7605 exploited in a Denial Of Service attack. This issue affects OpenSSL
7606 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7607 whether SRTP is used or configured. Implementations of OpenSSL that
7608 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7610 The fix was developed by the OpenSSL team.
7615 * Session Ticket Memory Leak.
7617 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7618 integrity of that ticket is first verified. In the event of a session
7619 ticket integrity check failing, OpenSSL will fail to free memory
7620 causing a memory leak. By sending a large number of invalid session
7621 tickets an attacker could exploit this issue in a Denial Of Service
7627 * Build option no-ssl3 is incomplete.
7629 When OpenSSL is configured with "no-ssl3" as a build option, servers
7630 could accept and complete an SSL 3.0 handshake, and clients could be
7631 configured to send them.
7634 *Akamai and the OpenSSL team*
7636 * Add support for TLS_FALLBACK_SCSV.
7637 Client applications doing fallback retries should call
7638 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7641 *Adam Langley, Bodo Moeller*
7643 * Add additional DigestInfo checks.
7645 Re-encode DigestInto in DER and check against the original when
7646 verifying RSA signature: this will reject any improperly encoded
7647 DigestInfo structures.
7649 Note: this is a precautionary measure and no attacks are currently known.
7653 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7655 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7656 SRP code can be overrun an internal buffer. Add sanity check that
7657 g, A, B < N to SRP code.
7659 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7660 Group for discovering this issue.
7665 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7666 TLS 1.0 instead of higher protocol versions when the ClientHello message
7667 is badly fragmented. This allows a man-in-the-middle attacker to force a
7668 downgrade to TLS 1.0 even if both the server and the client support a
7669 higher protocol version, by modifying the client's TLS records.
7671 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7672 researching this issue.
7677 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7678 to a denial of service attack. A malicious server can crash the client
7679 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7680 ciphersuite and sending carefully crafted handshake messages.
7682 Thanks to Felix Gröbert (Google) for discovering and researching this
7688 * By sending carefully crafted DTLS packets an attacker could cause openssl
7689 to leak memory. This can be exploited through a Denial of Service attack.
7690 Thanks to Adam Langley for discovering and researching this issue.
7695 * An attacker can force openssl to consume large amounts of memory whilst
7696 processing DTLS handshake messages. This can be exploited through a
7697 Denial of Service attack.
7698 Thanks to Adam Langley for discovering and researching this issue.
7703 * An attacker can force an error condition which causes openssl to crash
7704 whilst processing DTLS packets due to memory being freed twice. This
7705 can be exploited through a Denial of Service attack.
7706 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7712 * If a multithreaded client connects to a malicious server using a resumed
7713 session and the server sends an ec point format extension it could write
7714 up to 255 bytes to freed memory.
7716 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7722 * A malicious server can crash an OpenSSL client with a null pointer
7723 dereference (read) by specifying an SRP ciphersuite even though it was not
7724 properly negotiated with the client. This can be exploited through a
7725 Denial of Service attack.
7727 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7728 discovering and researching this issue.
7733 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7734 X509_name_oneline, X509_name_print_ex et al. to leak some information
7735 from the stack. Applications may be affected if they echo pretty printing
7736 output to the attacker.
7738 Thanks to Ivan Fratric (Google) for discovering this issue.
7741 *Emilia Käsper, and Steve Henson*
7743 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7744 for corner cases. (Certain input points at infinity could lead to
7745 bogus results, with non-infinity inputs mapped to infinity too.)
7749 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7751 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7752 handshake can force the use of weak keying material in OpenSSL
7753 SSL/TLS clients and servers.
7755 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7756 researching this issue. ([CVE-2014-0224])
7758 *KIKUCHI Masashi, Steve Henson*
7760 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7761 OpenSSL DTLS client the code can be made to recurse eventually crashing
7764 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7767 *Imre Rad, Steve Henson*
7769 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7770 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7771 client or server. This is potentially exploitable to run arbitrary
7772 code on a vulnerable client or server.
7774 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7776 *Jüri Aedla, Steve Henson*
7778 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7779 are subject to a denial of service attack.
7781 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7782 this issue. ([CVE-2014-3470])
7784 *Felix Gröbert, Ivan Fratric, Steve Henson*
7786 * Harmonize version and its documentation. -f flag is used to display
7789 *mancha <mancha1@zoho.com>*
7791 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7792 in i2d_ECPrivateKey.
7794 *mancha <mancha1@zoho.com>*
7796 * Fix some double frees. These are not thought to be exploitable.
7798 *mancha <mancha1@zoho.com>*
7800 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7802 * A missing bounds check in the handling of the TLS heartbeat extension
7803 can be used to reveal up to 64k of memory to a connected client or
7806 Thanks for Neel Mehta of Google Security for discovering this bug and to
7807 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7808 preparing the fix ([CVE-2014-0160])
7810 *Adam Langley, Bodo Moeller*
7812 * Fix for the attack described in the paper "Recovering OpenSSL
7813 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7814 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7815 <http://eprint.iacr.org/2014/140>
7817 Thanks to Yuval Yarom and Naomi Benger for discovering this
7818 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7820 *Yuval Yarom and Naomi Benger*
7822 * TLS pad extension: draft-agl-tls-padding-03
7824 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7825 TLS client Hello record length value would otherwise be > 255 and
7826 less that 512 pad with a dummy extension containing zeroes so it
7827 is at least 512 bytes long.
7829 *Adam Langley, Steve Henson*
7831 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7833 * Fix for TLS record tampering bug. A carefully crafted invalid
7834 handshake could crash OpenSSL with a NULL pointer exception.
7835 Thanks to Anton Johansson for reporting this issues.
7838 * Keep original DTLS digest and encryption contexts in retransmission
7839 structures so we can use the previous session parameters if they need
7840 to be resent. ([CVE-2013-6450])
7844 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7845 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7846 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7847 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7848 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7849 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7851 *Rob Stradling, Adam Langley*
7853 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7855 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7856 supporting platforms or when small records were transferred.
7858 *Andy Polyakov, Steve Henson*
7860 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7862 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7864 This addresses the flaw in CBC record processing discovered by
7865 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7866 at: <http://www.isg.rhul.ac.uk/tls/>
7868 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7869 Security Group at Royal Holloway, University of London
7870 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7871 Emilia Käsper for the initial patch.
7874 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7876 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7877 ciphersuites which can be exploited in a denial of service attack.
7878 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7879 and detecting this bug and to Wolfgang Ettlinger
7880 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7885 * Return an error when checking OCSP signatures when key is NULL.
7886 This fixes a DoS attack. ([CVE-2013-0166])
7890 * Make openssl verify return errors.
7892 *Chris Palmer <palmer@google.com> and Ben Laurie*
7894 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7895 the right response is stapled. Also change SSL_get_certificate()
7896 so it returns the certificate actually sent.
7897 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7899 *Rob Stradling <rob.stradling@comodo.com>*
7901 * Fix possible deadlock when decoding public keys.
7905 * Don't use TLS 1.0 record version number in initial client hello
7910 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7912 * Sanity check record length before skipping explicit IV in TLS
7913 1.2, 1.1 and DTLS to fix DoS attack.
7915 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7916 fuzzing as a service testing platform.
7921 * Initialise tkeylen properly when encrypting CMS messages.
7922 Thanks to Solar Designer of Openwall for reporting this issue.
7926 * In FIPS mode don't try to use composite ciphers as they are not
7931 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7933 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7934 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7935 mean any application compiled against OpenSSL 1.0.0 headers setting
7936 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7937 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7938 0x10000000L Any application which was previously compiled against
7939 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7940 will need to be recompiled as a result. Letting be results in
7941 inability to disable specifically TLS 1.1 and in client context,
7942 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7946 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7947 disable just protocol X, but all protocols above X *if* there are
7948 protocols *below* X still enabled. In more practical terms it means
7949 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7950 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7951 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7956 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7958 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7959 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7960 in CRYPTO_realloc_clean.
7962 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7963 issue and to Adam Langley <agl@chromium.org> for fixing it.
7966 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7968 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7972 * Workarounds for some broken servers that "hang" if a client hello
7973 record length exceeds 255 bytes.
7975 1. Do not use record version number > TLS 1.0 in initial client
7976 hello: some (but not all) hanging servers will now work.
7977 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7978 the number of ciphers sent in the client hello. This should be
7979 set to an even number, such as 50, for example by passing:
7980 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7981 Most broken servers should now work.
7982 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7983 TLS 1.2 client support entirely.
7987 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7991 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7993 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7994 STRING form instead of a DigestInfo.
7998 * The format used for MDC2 RSA signatures is inconsistent between EVP
7999 and the RSA_sign/RSA_verify functions. This was made more apparent when
8000 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8001 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8002 the correct format in RSA_verify so both forms transparently work.
8006 * Some servers which support TLS 1.0 can choke if we initially indicate
8007 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8008 encrypted premaster secret. As a workaround use the maximum permitted
8009 client version in client hello, this should keep such servers happy
8010 and still work with previous versions of OpenSSL.
8014 * Add support for TLS/DTLS heartbeats.
8016 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8018 * Add support for SCTP.
8020 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8022 * Improved PRNG seeding for VOS.
8024 *Paul Green <Paul.Green@stratus.com>*
8026 * Extensive assembler packs updates, most notably:
8028 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8029 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8030 - x86_64: bit-sliced AES implementation;
8031 - ARM: NEON support, contemporary platforms optimizations;
8032 - s390x: z196 support;
8033 - `*`: GHASH and GF(2^m) multiplication implementations;
8037 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8038 (removal of unnecessary code)
8040 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8042 * Add TLS key material exporter from RFC 5705.
8046 * Add DTLS-SRTP negotiation from RFC 5764.
8050 * Add Next Protocol Negotiation,
8051 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8052 disabled with a no-npn flag to config or Configure. Code donated
8055 *Adam Langley <agl@google.com> and Ben Laurie*
8057 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8058 NIST-P256, NIST-P521, with constant-time single point multiplication on
8059 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8060 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8061 Code made available under Apache License version 2.0.
8063 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8064 line to include this in your build of OpenSSL, and run "make depend" (or
8065 "make update"). This enables the following EC_METHODs:
8067 EC_GFp_nistp224_method()
8068 EC_GFp_nistp256_method()
8069 EC_GFp_nistp521_method()
8071 EC_GROUP_new_by_curve_name() will automatically use these (while
8072 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8075 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8077 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8078 all platforms. Move ssize_t definition from e_os.h to the public
8079 header file e_os2.h as it now appears in public header file cms.h
8083 * New -sigopt option to the ca, req and x509 utilities. Additional
8084 signature parameters can be passed using this option and in
8089 * Add RSA PSS signing function. This will generate and set the
8090 appropriate AlgorithmIdentifiers for PSS based on those in the
8091 corresponding EVP_MD_CTX structure. No application support yet.
8095 * Support for companion algorithm specific ASN1 signing routines.
8096 New function ASN1_item_sign_ctx() signs a pre-initialised
8097 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8098 the appropriate parameters.
8102 * Add new algorithm specific ASN1 verification initialisation function
8103 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8104 handling will be the same no matter what EVP_PKEY_METHOD is used.
8105 Add a PSS handler to support verification of PSS signatures: checked
8106 against a number of sample certificates.
8110 * Add signature printing for PSS. Add PSS OIDs.
8112 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8114 * Add algorithm specific signature printing. An individual ASN1 method
8115 can now print out signatures instead of the standard hex dump.
8117 More complex signatures (e.g. PSS) can print out more meaningful
8118 information. Include DSA version that prints out the signature
8123 * Password based recipient info support for CMS library: implementing
8128 * Split password based encryption into PBES2 and PBKDF2 functions. This
8129 neatly separates the code into cipher and PBE sections and is required
8130 for some algorithms that split PBES2 into separate pieces (such as
8131 password based CMS).
8135 * Session-handling fixes:
8136 - Fix handling of connections that are resuming with a session ID,
8137 but also support Session Tickets.
8138 - Fix a bug that suppressed issuing of a new ticket if the client
8139 presented a ticket with an expired session.
8140 - Try to set the ticket lifetime hint to something reasonable.
8141 - Make tickets shorter by excluding irrelevant information.
8142 - On the client side, don't ignore renewed tickets.
8144 *Adam Langley, Bodo Moeller (Google)*
8146 * Fix PSK session representation.
8150 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8152 This work was sponsored by Intel.
8156 * Add GCM support to TLS library. Some custom code is needed to split
8157 the IV between the fixed (from PRF) and explicit (from TLS record)
8158 portions. This adds all GCM ciphersuites supported by RFC5288 and
8159 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8160 add a special AESGCM string for GCM only.
8164 * Expand range of ctrls for AES GCM. Permit setting invocation
8165 field on decrypt and retrieval of invocation field only on encrypt.
8169 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8170 As required by RFC5289 these ciphersuites cannot be used if for
8171 versions of TLS earlier than 1.2.
8175 * For FIPS capable OpenSSL interpret a NULL default public key method
8176 as unset and return the appropriate default but do *not* set the default.
8177 This means we can return the appropriate method in applications that
8178 switch between FIPS and non-FIPS modes.
8182 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8183 ENGINE is used then we cannot handle that in the FIPS module so we
8184 keep original code iff non-FIPS operations are allowed.
8188 * Add -attime option to openssl utilities.
8190 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8192 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8196 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8197 FIPS EC methods unconditionally for now.
8201 * New build option no-ec2m to disable characteristic 2 code.
8205 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8206 all cases can be covered as some introduce binary incompatibilities.
8210 * Redirect RSA operations to FIPS module including keygen,
8211 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8215 * Add similar low-level API blocking to ciphers.
8219 * low-level digest APIs are not approved in FIPS mode: any attempt
8220 to use these will cause a fatal error. Applications that *really* want
8221 to use them can use the `private_*` version instead.
8225 * Redirect cipher operations to FIPS module for FIPS builds.
8229 * Redirect digest operations to FIPS module for FIPS builds.
8233 * Update build system to add "fips" flag which will link in fipscanister.o
8234 for static and shared library builds embedding a signature if needed.
8238 * Output TLS supported curves in preference order instead of numerical
8239 order. This is currently hardcoded for the highest order curves first.
8240 This should be configurable so applications can judge speed vs strength.
8244 * Add TLS v1.2 server support for client authentication.
8248 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8253 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8254 FIPS modules versions.
8258 * Add TLS v1.2 client side support for client authentication. Keep cache
8259 of handshake records longer as we don't know the hash algorithm to use
8260 until after the certificate request message is received.
8264 * Initial TLS v1.2 client support. Add a default signature algorithms
8265 extension including all the algorithms we support. Parse new signature
8266 format in client key exchange. Relax some ECC signing restrictions for
8267 TLS v1.2 as indicated in RFC5246.
8271 * Add server support for TLS v1.2 signature algorithms extension. Switch
8272 to new signature format when needed using client digest preference.
8273 All server ciphersuites should now work correctly in TLS v1.2. No client
8274 support yet and no support for client certificates.
8278 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8279 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8280 ciphersuites. At present only RSA key exchange ciphersuites work with
8281 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8282 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8283 and version checking.
8287 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8288 with this defined it will not be affected by any changes to ssl internal
8289 structures. Add several utility functions to allow openssl application
8290 to work with OPENSSL_NO_SSL_INTERN defined.
8294 * A long standing patch to add support for SRP from EdelWeb (Peter
8295 Sylvester and Christophe Renou) was integrated.
8296 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8297 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8300 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8304 * Permit abbreviated handshakes when renegotiating using the function
8305 SSL_renegotiate_abbreviated().
8307 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8309 * Add call to ENGINE_register_all_complete() to
8310 ENGINE_load_builtin_engines(), so some implementations get used
8311 automatically instead of needing explicit application support.
8315 * Add support for TLS key exporter as described in RFC5705.
8317 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8319 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8320 a few changes are required:
8322 Add SSL_OP_NO_TLSv1_1 flag.
8323 Add TLSv1_1 methods.
8324 Update version checking logic to handle version 1.1.
8325 Add explicit IV handling (ported from DTLS code).
8326 Add command line options to s_client/s_server.
8333 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8335 * X509_ATTRIBUTE memory leak
8337 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8338 memory. This structure is used by the PKCS#7 and CMS routines so any
8339 application which reads PKCS#7 or CMS data from untrusted sources is
8340 affected. SSL/TLS is not affected.
8342 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8348 * Race condition handling PSK identify hint
8350 If PSK identity hints are received by a multi-threaded client then
8351 the values are wrongly updated in the parent SSL_CTX structure. This can
8352 result in a race condition potentially leading to a double free of the
8358 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8360 * Malformed ECParameters causes infinite loop
8362 When processing an ECParameters structure OpenSSL enters an infinite loop
8363 if the curve specified is over a specially malformed binary polynomial
8366 This can be used to perform denial of service against any
8367 system which processes public keys, certificate requests or
8368 certificates. This includes TLS clients and TLS servers with
8369 client authentication enabled.
8371 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8376 * Exploitable out-of-bounds read in X509_cmp_time
8378 X509_cmp_time does not properly check the length of the ASN1_TIME
8379 string and can read a few bytes out of bounds. In addition,
8380 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8383 An attacker can use this to craft malformed certificates and CRLs of
8384 various sizes and potentially cause a segmentation fault, resulting in
8385 a DoS on applications that verify certificates or CRLs. TLS clients
8386 that verify CRLs are affected. TLS clients and servers with client
8387 authentication enabled may be affected if they use custom verification
8390 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8391 independently by Hanno Böck.
8396 * PKCS7 crash with missing EnvelopedContent
8398 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8399 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8400 with missing content and trigger a NULL pointer dereference on parsing.
8402 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8403 structures from untrusted sources are affected. OpenSSL clients and
8404 servers are not affected.
8406 This issue was reported to OpenSSL by Michal Zalewski (Google).
8411 * CMS verify infinite loop with unknown hash function
8413 When verifying a signedData message the CMS code can enter an infinite loop
8414 if presented with an unknown hash function OID. This can be used to perform
8415 denial of service against any system which verifies signedData messages using
8417 This issue was reported to OpenSSL by Johannes Bauer.
8422 * Race condition handling NewSessionTicket
8424 If a NewSessionTicket is received by a multi-threaded client when attempting to
8425 reuse a previous ticket then a race condition can occur potentially leading to
8426 a double free of the ticket data.
8431 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8433 * Segmentation fault in ASN1_TYPE_cmp fix
8435 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8436 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8437 certificate signature algorithm consistency this can be used to crash any
8438 certificate verification operation and exploited in a DoS attack. Any
8439 application which performs certificate verification is vulnerable including
8440 OpenSSL clients and servers which enable client authentication.
8445 * ASN.1 structure reuse memory corruption fix
8447 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8448 memory corruption via an invalid write. Such reuse is and has been
8449 strongly discouraged and is believed to be rare.
8451 Applications that parse structures containing CHOICE or ANY DEFINED BY
8452 components may be affected. Certificate parsing (d2i_X509 and related
8453 functions) are however not affected. OpenSSL clients and servers are
8459 * PKCS7 NULL pointer dereferences fix
8461 The PKCS#7 parsing code does not handle missing outer ContentInfo
8462 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8463 missing content and trigger a NULL pointer dereference on parsing.
8465 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8466 otherwise parse PKCS#7 structures from untrusted sources are
8467 affected. OpenSSL clients and servers are not affected.
8469 This issue was reported to OpenSSL by Michal Zalewski (Google).
8474 * DoS via reachable assert in SSLv2 servers fix
8476 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8477 servers that both support SSLv2 and enable export cipher suites by sending
8478 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8480 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8481 (OpenSSL development team).
8486 * Use After Free following d2i_ECPrivatekey error fix
8488 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8489 could cause a use after free condition. This, in turn, could cause a double
8490 free in several private key parsing functions (such as d2i_PrivateKey
8491 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8492 for applications that receive EC private keys from untrusted
8493 sources. This scenario is considered rare.
8495 This issue was discovered by the BoringSSL project and fixed in their
8501 * X509_to_X509_REQ NULL pointer deref fix
8503 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8504 the certificate key is invalid. This function is rarely used in practice.
8506 This issue was discovered by Brian Carpenter.
8511 * Removed the export ciphers from the DEFAULT ciphers
8515 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8517 * Build fixes for the Windows and OpenVMS platforms
8519 *Matt Caswell and Richard Levitte*
8521 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8523 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8524 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8525 dereference. This could lead to a Denial Of Service attack. Thanks to
8526 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8531 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8532 dtls1_buffer_record function under certain conditions. In particular this
8533 could occur if an attacker sent repeated DTLS records with the same
8534 sequence number but for the next epoch. The memory leak could be exploited
8535 by an attacker in a Denial of Service attack through memory exhaustion.
8536 Thanks to Chris Mueller for reporting this issue.
8541 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8542 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8543 method would be set to NULL which could later result in a NULL pointer
8544 dereference. Thanks to Frank Schmirler for reporting this issue.
8549 * Abort handshake if server key exchange message is omitted for ephemeral
8552 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8553 reporting this issue.
8558 * Remove non-export ephemeral RSA code on client and server. This code
8559 violated the TLS standard by allowing the use of temporary RSA keys in
8560 non-export ciphersuites and could be used by a server to effectively
8561 downgrade the RSA key length used to a value smaller than the server
8562 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8563 INRIA or reporting this issue.
8568 * Fixed issue where DH client certificates are accepted without verification.
8569 An OpenSSL server will accept a DH certificate for client authentication
8570 without the certificate verify message. This effectively allows a client to
8571 authenticate without the use of a private key. This only affects servers
8572 which trust a client certificate authority which issues certificates
8573 containing DH keys: these are extremely rare and hardly ever encountered.
8574 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8580 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8581 results on some platforms, including x86_64. This bug occurs at random
8582 with a very low probability, and is not known to be exploitable in any
8583 way, though its exact impact is difficult to determine. Thanks to Pieter
8584 Wuille (Blockstream) who reported this issue and also suggested an initial
8585 fix. Further analysis was conducted by the OpenSSL development team and
8586 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8587 the OpenSSL core team.
8592 * Fix various certificate fingerprint issues.
8594 By using non-DER or invalid encodings outside the signed portion of a
8595 certificate the fingerprint can be changed without breaking the signature.
8596 Although no details of the signed portion of the certificate can be changed
8597 this can cause problems with some applications: e.g. those using the
8598 certificate fingerprint for blacklists.
8600 1. Reject signatures with non zero unused bits.
8602 If the BIT STRING containing the signature has non zero unused bits reject
8603 the signature. All current signature algorithms require zero unused bits.
8605 2. Check certificate algorithm consistency.
8607 Check the AlgorithmIdentifier inside TBS matches the one in the
8608 certificate signature. NB: this will result in signature failure
8609 errors for some broken certificates.
8611 Thanks to Konrad Kraszewski from Google for reporting this issue.
8613 3. Check DSA/ECDSA signatures use DER.
8615 Reencode DSA/ECDSA signatures and compare with the original received
8616 signature. Return an error if there is a mismatch.
8618 This will reject various cases including garbage after signature
8619 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8620 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8621 (negative or with leading zeroes).
8623 Further analysis was conducted and fixes were developed by Stephen Henson
8624 of the OpenSSL core team.
8630 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8632 * Session Ticket Memory Leak.
8634 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8635 integrity of that ticket is first verified. In the event of a session
8636 ticket integrity check failing, OpenSSL will fail to free memory
8637 causing a memory leak. By sending a large number of invalid session
8638 tickets an attacker could exploit this issue in a Denial Of Service
8644 * Build option no-ssl3 is incomplete.
8646 When OpenSSL is configured with "no-ssl3" as a build option, servers
8647 could accept and complete an SSL 3.0 handshake, and clients could be
8648 configured to send them.
8651 *Akamai and the OpenSSL team*
8653 * Add support for TLS_FALLBACK_SCSV.
8654 Client applications doing fallback retries should call
8655 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8658 *Adam Langley, Bodo Moeller*
8660 * Add additional DigestInfo checks.
8662 Reencode DigestInto in DER and check against the original when
8663 verifying RSA signature: this will reject any improperly encoded
8664 DigestInfo structures.
8666 Note: this is a precautionary measure and no attacks are currently known.
8670 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8672 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8673 to a denial of service attack. A malicious server can crash the client
8674 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8675 ciphersuite and sending carefully crafted handshake messages.
8677 Thanks to Felix Gröbert (Google) for discovering and researching this
8683 * By sending carefully crafted DTLS packets an attacker could cause openssl
8684 to leak memory. This can be exploited through a Denial of Service attack.
8685 Thanks to Adam Langley for discovering and researching this issue.
8690 * An attacker can force openssl to consume large amounts of memory whilst
8691 processing DTLS handshake messages. This can be exploited through a
8692 Denial of Service attack.
8693 Thanks to Adam Langley for discovering and researching this issue.
8698 * An attacker can force an error condition which causes openssl to crash
8699 whilst processing DTLS packets due to memory being freed twice. This
8700 can be exploited through a Denial of Service attack.
8701 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8707 * If a multithreaded client connects to a malicious server using a resumed
8708 session and the server sends an ec point format extension it could write
8709 up to 255 bytes to freed memory.
8711 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8717 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8718 X509_name_oneline, X509_name_print_ex et al. to leak some information
8719 from the stack. Applications may be affected if they echo pretty printing
8720 output to the attacker.
8722 Thanks to Ivan Fratric (Google) for discovering this issue.
8725 *Emilia Käsper, and Steve Henson*
8727 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8728 for corner cases. (Certain input points at infinity could lead to
8729 bogus results, with non-infinity inputs mapped to infinity too.)
8733 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8735 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8736 handshake can force the use of weak keying material in OpenSSL
8737 SSL/TLS clients and servers.
8739 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8740 researching this issue. ([CVE-2014-0224])
8742 *KIKUCHI Masashi, Steve Henson*
8744 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8745 OpenSSL DTLS client the code can be made to recurse eventually crashing
8748 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8751 *Imre Rad, Steve Henson*
8753 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8754 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8755 client or server. This is potentially exploitable to run arbitrary
8756 code on a vulnerable client or server.
8758 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8760 *Jüri Aedla, Steve Henson*
8762 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8763 are subject to a denial of service attack.
8765 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8766 this issue. ([CVE-2014-3470])
8768 *Felix Gröbert, Ivan Fratric, Steve Henson*
8770 * Harmonize version and its documentation. -f flag is used to display
8773 *mancha <mancha1@zoho.com>*
8775 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8776 in i2d_ECPrivateKey.
8778 *mancha <mancha1@zoho.com>*
8780 * Fix some double frees. These are not thought to be exploitable.
8782 *mancha <mancha1@zoho.com>*
8784 * Fix for the attack described in the paper "Recovering OpenSSL
8785 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8786 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8787 <http://eprint.iacr.org/2014/140>
8789 Thanks to Yuval Yarom and Naomi Benger for discovering this
8790 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8792 *Yuval Yarom and Naomi Benger*
8794 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8796 * Keep original DTLS digest and encryption contexts in retransmission
8797 structures so we can use the previous session parameters if they need
8798 to be resent. ([CVE-2013-6450])
8802 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8803 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8804 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8805 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8806 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8807 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8809 *Rob Stradling, Adam Langley*
8811 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8813 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8815 This addresses the flaw in CBC record processing discovered by
8816 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8817 at: <http://www.isg.rhul.ac.uk/tls/>
8819 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8820 Security Group at Royal Holloway, University of London
8821 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8822 Emilia Käsper for the initial patch.
8825 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8827 * Return an error when checking OCSP signatures when key is NULL.
8828 This fixes a DoS attack. ([CVE-2013-0166])
8832 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8833 the right response is stapled. Also change SSL_get_certificate()
8834 so it returns the certificate actually sent.
8835 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8836 (This is a backport)
8838 *Rob Stradling <rob.stradling@comodo.com>*
8840 * Fix possible deadlock when decoding public keys.
8844 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8846 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8849 * Sanity check record length before skipping explicit IV in DTLS
8852 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8853 fuzzing as a service testing platform.
8858 * Initialise tkeylen properly when encrypting CMS messages.
8859 Thanks to Solar Designer of Openwall for reporting this issue.
8863 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8865 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8866 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8867 in CRYPTO_realloc_clean.
8869 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8870 issue and to Adam Langley <agl@chromium.org> for fixing it.
8873 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8875 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8877 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8878 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8879 content decryption and always return the same error. Note: this attack
8880 needs on average 2^20 messages so it only affects automated senders. The
8881 old behaviour can be re-enabled in the CMS code by setting the
8882 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8883 an MMA defence is not necessary.
8884 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8885 this issue. ([CVE-2012-0884])
8889 * Fix CVE-2011-4619: make sure we really are receiving a
8890 client hello before rejecting multiple SGC restarts. Thanks to
8891 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8895 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8897 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8898 Thanks to Antonio Martin, Enterprise Secure Access Research and
8899 Development, Cisco Systems, Inc. for discovering this bug and
8900 preparing a fix. ([CVE-2012-0050])
8904 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8906 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8907 of the Vaudenay padding oracle attack on CBC mode encryption
8908 which enables an efficient plaintext recovery attack against
8909 the OpenSSL implementation of DTLS. Their attack exploits timing
8910 differences arising during decryption processing. A research
8911 paper describing this attack can be found at:
8912 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8913 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8914 Security Group at Royal Holloway, University of London
8915 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8916 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8917 for preparing the fix. ([CVE-2011-4108])
8919 *Robin Seggelmann, Michael Tuexen*
8921 * Clear bytes used for block padding of SSL 3.0 records.
8924 *Adam Langley (Google)*
8926 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8927 Kadianakis <desnacked@gmail.com> for discovering this issue and
8928 Adam Langley for preparing the fix. ([CVE-2011-4619])
8930 *Adam Langley (Google)*
8932 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8934 *Andrey Kulikov <amdeich@gmail.com>*
8936 * Prevent malformed RFC3779 data triggering an assertion failure.
8937 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8938 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8940 *Rob Austein <sra@hactrn.net>*
8942 * Improved PRNG seeding for VOS.
8944 *Paul Green <Paul.Green@stratus.com>*
8946 * Fix ssl_ciph.c set-up race.
8948 *Adam Langley (Google)*
8950 * Fix spurious failures in ecdsatest.c.
8952 *Emilia Käsper (Google)*
8954 * Fix the BIO_f_buffer() implementation (which was mixing different
8955 interpretations of the `..._len` fields).
8957 *Adam Langley (Google)*
8959 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8960 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8961 threads won't reuse the same blinding coefficients.
8963 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8964 lock to call BN_BLINDING_invert_ex, and avoids one use of
8965 BN_BLINDING_update for each BN_BLINDING structure (previously,
8966 the last update always remained unused).
8968 *Emilia Käsper (Google)*
8970 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8972 *Bob Buckholz (Google)*
8974 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8976 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8977 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8979 *Kaspar Brand <ossl@velox.ch>*
8981 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8982 for multi-threaded use of ECDH. ([CVE-2011-3210])
8984 *Adam Langley (Google)*
8986 * Fix x509_name_ex_d2i memory leak on bad inputs.
8990 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8991 signature public key algorithm by using OID xref utilities instead.
8992 Before this you could only use some ECC ciphersuites with SHA1 only.
8996 * Add protection against ECDSA timing attacks as mentioned in the paper
8997 by Billy Bob Brumley and Nicola Tuveri, see:
8998 <http://eprint.iacr.org/2011/232.pdf>
9000 *Billy Bob Brumley and Nicola Tuveri*
9002 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9004 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9006 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9008 * Fix bug in string printing code: if *any* escaping is enabled we must
9009 escape the escape character (backslash) or the resulting string is
9014 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9016 * Disable code workaround for ancient and obsolete Netscape browsers
9017 and servers: an attacker can use it in a ciphersuite downgrade attack.
9018 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9022 * Fixed J-PAKE implementation error, originally discovered by
9023 Sebastien Martini, further info and confirmation from Stefan
9024 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9028 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9030 * Fix extension code to avoid race conditions which can result in a buffer
9031 overrun vulnerability: resumed sessions must not be modified as they can
9032 be shared by multiple threads. CVE-2010-3864
9036 * Fix WIN32 build system to correctly link an ENGINE directory into
9041 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9043 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9046 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9048 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9050 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9051 context. The operation can be customised via the ctrl mechanism in
9052 case ENGINEs want to include additional functionality.
9056 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9060 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9061 output hashes compatible with older versions of OpenSSL.
9063 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9065 * Fix compression algorithm handling: if resuming a session use the
9066 compression algorithm of the resumed session instead of determining
9067 it from client hello again. Don't allow server to change algorithm.
9071 * Add load_crls() function to commands tidying load_certs() too. Add option
9072 to verify utility to allow additional CRLs to be included.
9076 * Update OCSP request code to permit adding custom headers to the request:
9077 some responders need this.
9081 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9084 *Julia Lawall <julia@diku.dk>*
9086 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9087 needlessly dereferenced structures, used obsolete functions and
9088 didn't handle all updated verify codes correctly.
9092 * Disable MD2 in the default configuration.
9096 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9097 indicate the initial BIO being pushed or popped. This makes it possible
9098 to determine whether the BIO is the one explicitly called or as a result
9099 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9100 it handles reference counts correctly and doesn't zero out the I/O bio
9101 when it is not being explicitly popped. WARNING: applications which
9102 included workarounds for the old buggy behaviour will need to be modified
9103 or they could free up already freed BIOs.
9107 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9108 renaming to all platforms (within the 0.9.8 branch, this was
9109 done conditionally on Netware platforms to avoid a name clash).
9111 *Guenter <lists@gknw.net>*
9113 * Add ECDHE and PSK support to DTLS.
9115 *Michael Tuexen <tuexen@fh-muenster.de>*
9117 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9122 * Add "missing" function EVP_MD_flags() (without this the only way to
9123 retrieve a digest flags is by accessing the structure directly. Update
9124 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9125 or cipher is registered as in the "from" argument. Print out all
9126 registered digests in the dgst usage message instead of manually
9127 attempting to work them out.
9131 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9132 this allows the use of compression and extensions. Change default cipher
9133 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9134 by default unless an application cipher string requests it.
9138 * Alter match criteria in PKCS12_parse(). It used to try to use local
9139 key ids to find matching certificates and keys but some PKCS#12 files
9140 don't follow the (somewhat unwritten) rules and this strategy fails.
9141 Now just gather all certificates together and the first private key
9142 then look for the first certificate that matches the key.
9146 * Support use of registered digest and cipher names for dgst and cipher
9147 commands instead of having to add each one as a special case. So now
9154 openssl dgst -sha256 foo
9156 and this works for ENGINE based algorithms too.
9160 * Update Gost ENGINE to support parameter files.
9162 *Victor B. Wagner <vitus@cryptocom.ru>*
9164 * Support GeneralizedTime in ca utility.
9166 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9168 * Enhance the hash format used for certificate directory links. The new
9169 form uses the canonical encoding (meaning equivalent names will work
9170 even if they aren't identical) and uses SHA1 instead of MD5. This form
9171 is incompatible with the older format and as a result c_rehash should
9172 be used to rebuild symbolic links.
9176 * Make PKCS#8 the default write format for private keys, replacing the
9177 traditional format. This form is standardised, more secure and doesn't
9178 include an implicit MD5 dependency.
9182 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9183 committed to OpenSSL should pass this lot as a minimum.
9187 * Add session ticket override functionality for use by EAP-FAST.
9189 *Jouni Malinen <j@w1.fi>*
9191 * Modify HMAC functions to return a value. Since these can be implemented
9192 in an ENGINE errors can occur.
9196 * Type-checked OBJ_bsearch_ex.
9200 * Type-checked OBJ_bsearch. Also some constification necessitated
9201 by type-checking. Still to come: TXT_DB, bsearch(?),
9202 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9207 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9208 seconds to a tm structure directly, instead of going through OS
9209 specific date routines. This avoids any issues with OS routines such
9210 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9211 and X509_time_adj_ex() to cover the extended range. The existing
9212 X509_time_adj() is still usable and will no longer have any date issues.
9216 * Delta CRL support. New use deltas option which will attempt to locate
9217 and search any appropriate delta CRLs available.
9219 This work was sponsored by Google.
9223 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9224 code and add additional score elements. Validate alternate CRL paths
9225 as part of the CRL checking and indicate a new error "CRL path validation
9226 error" in this case. Applications wanting additional details can use
9227 the verify callback and check the new "parent" field. If this is not
9228 NULL CRL path validation is taking place. Existing applications won't
9229 see this because it requires extended CRL support which is off by
9232 This work was sponsored by Google.
9236 * Support for freshest CRL extension.
9238 This work was sponsored by Google.
9242 * Initial indirect CRL support. Currently only supported in the CRLs
9243 passed directly and not via lookup. Process certificate issuer
9244 CRL entry extension and lookup CRL entries by bother issuer name
9245 and serial number. Check and process CRL issuer entry in IDP extension.
9247 This work was sponsored by Google.
9251 * Add support for distinct certificate and CRL paths. The CRL issuer
9252 certificate is validated separately in this case. Only enabled if
9253 an extended CRL support flag is set: this flag will enable additional
9254 CRL functionality in future.
9256 This work was sponsored by Google.
9260 * Add support for policy mappings extension.
9262 This work was sponsored by Google.
9266 * Fixes to pathlength constraint, self issued certificate handling,
9267 policy processing to align with RFC3280 and PKITS tests.
9269 This work was sponsored by Google.
9273 * Support for name constraints certificate extension. DN, email, DNS
9274 and URI types are currently supported.
9276 This work was sponsored by Google.
9280 * To cater for systems that provide a pointer-based thread ID rather
9281 than numeric, deprecate the current numeric thread ID mechanism and
9282 replace it with a structure and associated callback type. This
9283 mechanism allows a numeric "hash" to be extracted from a thread ID in
9284 either case, and on platforms where pointers are larger than 'long',
9285 mixing is done to help ensure the numeric 'hash' is usable even if it
9286 can't be guaranteed unique. The default mechanism is to use "&errno"
9287 as a pointer-based thread ID to distinguish between threads.
9289 Applications that want to provide their own thread IDs should now use
9290 CRYPTO_THREADID_set_callback() to register a callback that will call
9291 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9293 Note that ERR_remove_state() is now deprecated, because it is tied
9294 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9295 to free the current thread's error state should be replaced by
9296 ERR_remove_thread_state(NULL).
9298 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9299 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9300 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9301 application was previously providing a numeric thread callback that
9302 was inappropriate for distinguishing threads, then uniqueness might
9303 have been obtained with &errno that happened immediately in the
9304 intermediate development versions of OpenSSL; this is no longer the
9305 case, the numeric thread callback will now override the automatic use
9308 *Geoff Thorpe, with help from Bodo Moeller*
9310 * Initial support for different CRL issuing certificates. This covers a
9311 simple case where the self issued certificates in the chain exist and
9312 the real CRL issuer is higher in the existing chain.
9314 This work was sponsored by Google.
9318 * Removed effectively defunct crypto/store from the build.
9322 * Revamp of STACK to provide stronger type-checking. Still to come:
9323 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9324 ASN1_STRING, CONF_VALUE.
9328 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9329 RAM on SSL connections. This option can save about 34k per idle SSL.
9333 * Revamp of LHASH to provide stronger type-checking. Still to come:
9334 STACK, TXT_DB, bsearch, qsort.
9338 * Initial support for Cryptographic Message Syntax (aka CMS) based
9339 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9340 support for data, signedData, compressedData, digestedData and
9341 encryptedData, envelopedData types included. Scripts to check against
9342 RFC4134 examples draft and interop and consistency checks of many
9343 content types and variants.
9347 * Add options to enc utility to support use of zlib compression BIO.
9351 * Extend mk1mf to support importing of options and assembly language
9352 files from Configure script, currently only included in VC-WIN32.
9353 The assembly language rules can now optionally generate the source
9354 files from the associated perl scripts.
9358 * Implement remaining functionality needed to support GOST ciphersuites.
9359 Interop testing has been performed using CryptoPro implementations.
9361 *Victor B. Wagner <vitus@cryptocom.ru>*
9363 * s390x assembler pack.
9367 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9372 * Implement Opaque PRF Input TLS extension as specified in
9373 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9374 official specification yet and no extension type assignment by
9375 IANA exists, this extension (for now) will have to be explicitly
9376 enabled when building OpenSSL by providing the extension number
9377 to use. For example, specify an option
9379 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9381 to the "config" or "Configure" script to enable the extension,
9382 assuming extension number 0x9527 (which is a completely arbitrary
9383 and unofficial assignment based on the MD5 hash of the Internet
9384 Draft). Note that by doing so, you potentially lose
9385 interoperability with other TLS implementations since these might
9386 be using the same extension number for other purposes.
9388 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9389 opaque PRF input value to use in the handshake. This will create
9390 an internal copy of the length-'len' string at 'src', and will
9391 return non-zero for success.
9393 To get more control and flexibility, provide a callback function
9396 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9397 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9401 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9404 Callback function 'cb' will be called in handshakes, and is
9405 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9406 Argument 'arg' is for application purposes (the value as given to
9407 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9408 be provided to the callback function). The callback function
9409 has to return non-zero to report success: usually 1 to use opaque
9410 PRF input just if possible, or 2 to enforce use of the opaque PRF
9411 input. In the latter case, the library will abort the handshake
9412 if opaque PRF input is not successfully negotiated.
9414 Arguments 'peerinput' and 'len' given to the callback function
9415 will always be NULL and 0 in the case of a client. A server will
9416 see the client's opaque PRF input through these variables if
9417 available (NULL and 0 otherwise). Note that if the server
9418 provides an opaque PRF input, the length must be the same as the
9419 length of the client's opaque PRF input.
9421 Note that the callback function will only be called when creating
9422 a new session (session resumption can resume whatever was
9423 previously negotiated), and will not be called in SSL 2.0
9424 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9425 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9426 for applications that need to enforce opaque PRF input.
9430 * Update ssl code to support digests other than SHA1+MD5 for handshake
9433 *Victor B. Wagner <vitus@cryptocom.ru>*
9435 * Add RFC4507 support to OpenSSL. This includes the corrections in
9436 RFC4507bis. The encrypted ticket format is an encrypted encoded
9437 SSL_SESSION structure, that way new session features are automatically
9440 If a client application caches session in an SSL_SESSION structure
9441 support is transparent because tickets are now stored in the encoded
9444 The SSL_CTX structure automatically generates keys for ticket
9445 protection in servers so again support should be possible
9446 with no application modification.
9448 If a client or server wishes to disable RFC4507 support then the option
9449 SSL_OP_NO_TICKET can be set.
9451 Add a TLS extension debugging callback to allow the contents of any client
9452 or server extensions to be examined.
9454 This work was sponsored by Google.
9458 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9459 OpenSSL should now compile cleanly on gcc 4.2
9461 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9463 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9464 support including streaming MAC support: this is required for GOST
9465 ciphersuite support.
9467 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9469 * Add option -stream to use PKCS#7 streaming in smime utility. New
9470 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9471 to output in BER and PEM format.
9475 * Experimental support for use of HMAC via EVP_PKEY interface. This
9476 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9477 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9478 ENGINE support for HMAC keys which are unextractable. New -mac and
9479 -macopt options to dgst utility.
9483 * New option -sigopt to dgst utility. Update dgst to use
9484 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9485 alternative signing parameters such as X9.31 or PSS in the dgst
9490 * Change ssl_cipher_apply_rule(), the internal function that does
9491 the work each time a ciphersuite string requests enabling
9492 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9493 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9494 the order of disabled ciphersuites such that those ciphersuites
9495 that most recently went from enabled to disabled not only stay
9496 in order with respect to each other, but also have higher priority
9497 than other disabled ciphersuites the next time ciphersuites are
9500 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9501 the same ciphersuites as with "HIGH" alone, but in a specific
9502 order where the PSK ciphersuites come first (since they are the
9503 most recently disabled ciphersuites when "HIGH" is parsed).
9505 Also, change ssl_create_cipher_list() (using this new
9506 functionality) such that between otherwise identical
9507 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9512 * Change ssl_create_cipher_list() so that it automatically
9513 arranges the ciphersuites in reasonable order before starting
9514 to process the rule string. Thus, the definition for "DEFAULT"
9515 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9516 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9517 This makes it much easier to arrive at a reasonable default order
9518 in applications for which anonymous ciphers are OK (meaning
9519 that you can't actually use DEFAULT).
9521 *Bodo Moeller; suggested by Victor Duchovni*
9523 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9524 processing) into multiple integers instead of setting
9525 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9526 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9527 (These masks as well as the individual bit definitions are hidden
9528 away into the non-exported interface ssl/ssl_locl.h, so this
9529 change to the definition of the SSL_CIPHER structure shouldn't
9530 affect applications.) This give us more bits for each of these
9531 categories, so there is no longer a need to coagulate AES128 and
9532 AES256 into a single algorithm bit, and to coagulate Camellia128
9533 and Camellia256 into a single algorithm bit, which has led to all
9536 Thus, among other things, the kludge introduced in 0.9.7m and
9537 0.9.8e for masking out AES256 independently of AES128 or masking
9538 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9540 With the change, we also introduce new ciphersuite aliases that
9541 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9546 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9547 Use the leftmost N bytes of the signature input if the input is
9548 larger than the prime q (with N being the size in bytes of q).
9552 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9553 it yet and it is largely untested.
9557 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9561 * Initial incomplete changes to avoid need for function casts in OpenSSL
9562 some compilers (gcc 4.2 and later) reject their use. Safestack is
9563 reimplemented. Update ASN1 to avoid use of legacy functions.
9567 * Win32/64 targets are linked with Winsock2.
9571 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9572 to external functions. This can be used to increase CRL handling
9573 efficiency especially when CRLs are very large by (for example) storing
9574 the CRL revoked certificates in a database.
9578 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9579 new CRLs added to a directory can be used. New command line option
9580 -verify_return_error to s_client and s_server. This causes real errors
9581 to be returned by the verify callback instead of carrying on no matter
9582 what. This reflects the way a "real world" verify callback would behave.
9586 * GOST engine, supporting several GOST algorithms and public key formats.
9587 Kindly donated by Cryptocom.
9591 * Partial support for Issuing Distribution Point CRL extension. CRLs
9592 partitioned by DP are handled but no indirect CRL or reason partitioning
9593 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9594 selected via a scoring technique which handles IDP and AKID in CRLs.
9598 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9599 will ultimately be used for all verify operations: this will remove the
9600 X509_STORE dependency on certificate verification and allow alternative
9601 lookup methods. X509_STORE based implementations of these two callbacks.
9605 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9606 Modify get_crl() to find a valid (unexpired) CRL if possible.
9610 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9611 this would be called X509_CRL_cmp() but that name is already used by
9612 a function that just compares CRL issuer names. Cache several CRL
9613 extensions in X509_CRL structure and cache CRLDP in X509.
9617 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9618 this maps equivalent X509_NAME structures into a consistent structure.
9619 Name comparison can then be performed rapidly using memcmp().
9623 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9628 * Allow digests to supply their own micalg string for S/MIME type using
9629 the ctrl EVP_MD_CTRL_MICALG.
9633 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9634 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9635 ctrl. It can then customise the structure before and/or after signing
9640 * New function OBJ_add_sigid() to allow application defined signature OIDs
9641 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9642 to free up any added signature OIDs.
9646 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9647 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9648 digest and cipher tables. New options added to openssl utility:
9649 list-message-digest-algorithms and list-cipher-algorithms.
9653 * Change the array representation of binary polynomials: the list
9654 of degrees of non-zero coefficients is now terminated with -1.
9655 Previously it was terminated with 0, which was also part of the
9656 value; thus, the array representation was not applicable to
9657 polynomials where t^0 has coefficient zero. This change makes
9658 the array representation useful in a more general context.
9662 * Various modifications and fixes to SSL/TLS cipher string
9663 handling. For ECC, the code now distinguishes between fixed ECDH
9664 with RSA certificates on the one hand and with ECDSA certificates
9665 on the other hand, since these are separate ciphersuites. The
9666 unused code for Fortezza ciphersuites has been removed.
9668 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9669 (not "ECDHE"). For consistency with the code for DH
9670 certificates, use of ECDH certificates is now considered ECDH
9671 authentication, not RSA or ECDSA authentication (the latter is
9672 merely the CA's signing algorithm and not actively used in the
9675 The temporary ciphersuite alias "ECCdraft" is no longer
9676 available, and ECC ciphersuites are no longer excluded from "ALL"
9677 and "DEFAULT". The following aliases now exist for RFC 4492
9678 ciphersuites, most of these by analogy with the DH case:
9680 kECDHr - ECDH cert, signed with RSA
9681 kECDHe - ECDH cert, signed with ECDSA
9682 kECDH - ECDH cert (signed with either RSA or ECDSA)
9683 kEECDH - ephemeral ECDH
9684 ECDH - ECDH cert or ephemeral ECDH
9690 AECDH - anonymous ECDH
9691 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9695 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9696 Use correct micalg parameters depending on digest(s) in signed message.
9700 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9701 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9705 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9706 an engine to register a method. Add ENGINE lookups for methods and
9707 functional reference processing.
9711 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9712 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9717 * New -resign option to smime utility. This adds one or more signers
9718 to an existing PKCS#7 signedData structure. Also -md option to use an
9719 alternative message digest algorithm for signing.
9723 * Tidy up PKCS#7 routines and add new functions to make it easier to
9724 create PKCS7 structures containing multiple signers. Update smime
9725 application to support multiple signers.
9729 * New -macalg option to pkcs12 utility to allow setting of an alternative
9734 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9735 Reorganize PBE internals to lookup from a static table using NIDs,
9736 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9737 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9738 PRF which will be automatically used with PBES2.
9742 * Replace the algorithm specific calls to generate keys in "req" with the
9747 * Update PKCS#7 enveloped data routines to use new API. This is now
9748 supported by any public key method supporting the encrypt operation. A
9749 ctrl is added to allow the public key algorithm to examine or modify
9750 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9755 * Add a ctrl to asn1 method to allow a public key algorithm to express
9756 a default digest type to use. In most cases this will be SHA1 but some
9757 algorithms (such as GOST) need to specify an alternative digest. The
9758 return value indicates how strong the preference is 1 means optional and
9759 2 is mandatory (that is it is the only supported type). Modify
9760 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9761 use the default md. Update openssl utilities to use the default digest
9762 type for signing if it is not explicitly indicated.
9766 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9767 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9768 signing method from the key type. This effectively removes the link
9769 between digests and public key types.
9773 * Add an OID cross reference table and utility functions. Its purpose is to
9774 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9775 rsaEncryption. This will allow some of the algorithm specific hackery
9776 needed to use the correct OID to be removed.
9780 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9781 structures for PKCS7_sign(). They are now set up by the relevant public
9786 * Add provisional EC pkey method with support for ECDSA and ECDH.
9790 * Add support for key derivation (agreement) in the API, DH method and
9795 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9796 public and private key formats. As a side effect these add additional
9797 command line functionality not previously available: DSA signatures can be
9798 generated and verified using pkeyutl and DH key support and generation in
9805 *Oliver Tappe <zooey@hirschkaefer.de>*
9807 * New make target "install_html_docs" installs HTML renditions of the
9810 *Oliver Tappe <zooey@hirschkaefer.de>*
9812 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9813 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9814 support key and parameter generation and add initial key generation
9815 functionality for RSA.
9819 * Add functions for main EVP_PKEY_method operations. The undocumented
9820 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9821 `EVP_PKEY_{encrypt,decrypt}_old`.
9825 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9826 key API, doesn't do much yet.
9830 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9831 public key algorithms. New option to openssl utility:
9832 "list-public-key-algorithms" to print out info.
9836 * Implement the Supported Elliptic Curves Extension for
9837 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9841 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9842 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9846 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9847 utilities such as rsa, dsa, dsaparam etc except they process any key
9852 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9853 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9854 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9859 * Initial support for pluggable public key ASN1.
9860 De-spaghettify the public key ASN1 handling. Move public and private
9861 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9862 algorithm specific handling to a single module within the relevant
9863 algorithm directory. Add functions to allow (near) opaque processing
9864 of public and private key structures.
9868 * Implement the Supported Point Formats Extension for
9869 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9873 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9874 for the psk identity [hint] and the psk callback functions to the
9875 SSL_SESSION, SSL and SSL_CTX structure.
9878 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9882 SSL_CTX_use_psk_identity_hint
9883 SSL_get_psk_identity_hint
9884 SSL_get_psk_identity
9885 SSL_use_psk_identity_hint
9887 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9889 * Add RFC 3161 compliant time stamp request creation, response generation
9890 and response verification functionality.
9892 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9894 * Add initial support for TLS extensions, specifically for the server_name
9895 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9896 have new members for a hostname. The SSL data structure has an
9897 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9898 stored in that context to allow for session resumption, even after the
9899 SSL has been switched to a new SSL_CTX in reaction to a client's
9900 server_name extension.
9902 New functions (subject to change):
9904 SSL_get_servername()
9905 SSL_get_servername_type()
9908 New CTRL codes and macros (subject to change):
9910 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9911 - SSL_CTX_set_tlsext_servername_callback()
9912 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9913 - SSL_CTX_set_tlsext_servername_arg()
9914 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9916 openssl s_client has a new '-servername ...' option.
9918 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9919 '-key2 ...', '-servername_fatal' (subject to change). This allows
9920 testing the HostName extension for a specific single hostname ('-cert'
9921 and '-key' remain fallbacks for handshakes without HostName
9922 negotiation). If the unrecognized_name alert has to be sent, this by
9923 default is a warning; it becomes fatal with the '-servername_fatal'
9926 *Peter Sylvester, Remy Allais, Christophe Renou*
9928 * Whirlpool hash implementation is added.
9932 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9933 bn(64,32). Because of instruction set limitations it doesn't have
9934 any negative impact on performance. This was done mostly in order
9935 to make it possible to share assembler modules, such as bn_mul_mont
9936 implementations, between 32- and 64-bit builds without hassle.
9940 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9941 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9946 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9947 dedicated Montgomery multiplication procedure, is introduced.
9948 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9949 "64-bit" performance on certain 32-bit targets.
9953 * New option SSL_OP_NO_COMP to disable use of compression selectively
9954 in SSL structures. New SSL ctrl to set maximum send fragment size.
9955 Save memory by setting the I/O buffer sizes dynamically instead of
9956 using the maximum available value.
9960 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9961 in addition to the text details.
9965 * Very, very preliminary EXPERIMENTAL support for printing of general
9966 ASN1 structures. This currently produces rather ugly output and doesn't
9967 handle several customised structures at all.
9971 * Integrated support for PVK file format and some related formats such
9972 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9973 these in the 'rsa' and 'dsa' utilities.
9977 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9981 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9982 place for the (very old) "NETSCAPE" format certificates which are now
9983 handled using new ASN1 code equivalents.
9987 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9988 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9989 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9993 * Modify CRL distribution points extension code to print out previously
9994 unsupported fields. Enhance extension setting code to allow setting of
9999 * Add print and set support for Issuing Distribution Point CRL extension.
10003 * Change 'Configure' script to enable Camellia by default.
10010 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10012 * When rejecting SSL/TLS records due to an incorrect version number, never
10013 update s->server with a new major version number. As of
10014 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10015 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10016 the previous behavior could result in a read attempt at NULL when
10017 receiving specific incorrect SSL/TLS records once record payload
10018 protection is active. ([CVE-2010-0740])
10020 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10022 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10023 could be crashed if the relevant tables were not present (e.g. chrooted).
10025 *Tomas Hoger <thoger@redhat.com>*
10027 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10029 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10031 *Martin Olsson, Neel Mehta*
10033 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10034 accommodate for stack sorting, always a write lock!).
10038 * On some versions of WIN32 Heap32Next is very slow. This can cause
10039 excessive delays in the RAND_poll(): over a minute. As a workaround
10040 include a time check in the inner Heap32Next loop too.
10044 * The code that handled flushing of data in SSL/TLS originally used the
10045 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10046 the problem outlined in PR#1949. The fix suggested there however can
10047 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10048 of Apache). So instead simplify the code to flush unconditionally.
10049 This should be fine since flushing with no data to flush is a no op.
10053 * Handle TLS versions 2.0 and later properly and correctly use the
10054 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10055 off ancient servers have a habit of sticking around for a while...
10059 * Modify compression code so it frees up structures without using the
10060 ex_data callbacks. This works around a problem where some applications
10061 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10062 restarting) then use compression (e.g. SSL with compression) later.
10063 This results in significant per-connection memory leaks and
10064 has caused some security issues including CVE-2008-1678 and
10069 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10070 change when encrypting or decrypting.
10074 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10075 connect and renegotiate with servers which do not support RI.
10076 Until RI is more widely deployed this option is enabled by default.
10080 * Add "missing" ssl ctrls to clear options and mode.
10084 * If client attempts to renegotiate and doesn't support RI respond with
10085 a no_renegotiation alert as required by RFC5746. Some renegotiating
10086 TLS clients will continue a connection gracefully when they receive
10087 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10088 waiting for a server hello which it will never receive. Now we treat a
10089 received no_renegotiation alert as a fatal error. This is because
10090 applications requesting a renegotiation might well expect it to succeed
10091 and would have no code in place to handle the server denying it so the
10092 only safe thing to do is to terminate the connection.
10096 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10097 peer supports secure renegotiation and 0 otherwise. Print out peer
10098 renegotiation support in s_client/s_server.
10102 * Replace the highly broken and deprecated SPKAC certification method with
10103 the updated NID creation version. This should correctly handle UTF8.
10107 * Implement RFC5746. Re-enable renegotiation but require the extension
10108 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10109 turns out to be a bad idea. It has been replaced by
10110 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10111 SSL_CTX_set_options(). This is really not recommended unless you
10112 know what you are doing.
10114 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10116 * Fixes to stateless session resumption handling. Use initial_ctx when
10117 issuing and attempting to decrypt tickets in case it has changed during
10118 servername handling. Use a non-zero length session ID when attempting
10119 stateless session resumption: this makes it possible to determine if
10120 a resumption has occurred immediately after receiving server hello
10121 (several places in OpenSSL subtly assume this) instead of later in
10126 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10127 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10128 fixes for a few places where the return code is not checked
10131 *Julia Lawall <julia@diku.dk>*
10133 * Add --strict-warnings option to Configure script to include devteam
10134 warnings in other configurations.
10138 * Add support for --libdir option and LIBDIR variable in makefiles. This
10139 makes it possible to install openssl libraries in locations which
10140 have names other than "lib", for example "/usr/lib64" which some
10143 *Steve Henson, based on patch from Jeremy Utley*
10145 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10146 X690 8.9.12 and can produce some misleading textual output of OIDs.
10148 *Steve Henson, reported by Dan Kaminsky*
10150 * Delete MD2 from algorithm tables. This follows the recommendation in
10151 several standards that it is not used in new applications due to
10152 several cryptographic weaknesses. For binary compatibility reasons
10153 the MD2 API is still compiled in by default.
10157 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10162 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10163 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10166 *Guenter <lists@gknw.net>*
10168 * Fix the server certificate chain building code to use X509_verify_cert(),
10169 it used to have an ad-hoc builder which was unable to cope with anything
10170 other than a simple chain.
10172 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10174 * Don't check self signed certificate signatures in X509_verify_cert()
10175 by default (a flag can override this): it just wastes time without
10176 adding any security. As a useful side effect self signed root CAs
10177 with non-FIPS digests are now usable in FIPS mode.
10181 * In dtls1_process_out_of_seq_message() the check if the current message
10182 is already buffered was missing. For every new message was memory
10183 allocated, allowing an attacker to perform an denial of service attack
10184 with sending out of seq handshake messages until there is no memory
10185 left. Additionally every future message was buffered, even if the
10186 sequence number made no sense and would be part of another handshake.
10187 So only messages with sequence numbers less than 10 in advance will be
10188 buffered. ([CVE-2009-1378])
10190 *Robin Seggelmann, discovered by Daniel Mentz*
10192 * Records are buffered if they arrive with a future epoch to be
10193 processed after finishing the corresponding handshake. There is
10194 currently no limitation to this buffer allowing an attacker to perform
10195 a DOS attack with sending records with future epochs until there is no
10196 memory left. This patch adds the pqueue_size() function to determine
10197 the size of a buffer and limits the record buffer to 100 entries.
10200 *Robin Seggelmann, discovered by Daniel Mentz*
10202 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10203 parent structure is freed. ([CVE-2009-1379])
10207 * Handle non-blocking I/O properly in SSL_shutdown() call.
10209 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10211 * Add `2.5.4.*` OIDs
10213 *Ilya O. <vrghost@gmail.com>*
10215 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10217 * Disable renegotiation completely - this fixes a severe security
10218 problem ([CVE-2009-3555]) at the cost of breaking all
10219 renegotiation. Renegotiation can be re-enabled by setting
10220 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10221 run-time. This is really not recommended unless you know what
10226 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10228 * Don't set val to NULL when freeing up structures, it is freed up by
10229 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10230 zeroing past the valid field. ([CVE-2009-0789])
10232 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10234 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10235 checked correctly. This would allow some invalid signed attributes to
10236 appear to verify correctly. ([CVE-2009-0591])
10238 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10240 * Reject UniversalString and BMPString types with invalid lengths. This
10241 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10242 a legal length. ([CVE-2009-0590])
10246 * Set S/MIME signing as the default purpose rather than setting it
10247 unconditionally. This allows applications to override it at the store
10252 * Permit restricted recursion of ASN1 strings. This is needed in practice
10253 to handle some structures.
10257 * Improve efficiency of mem_gets: don't search whole buffer each time
10260 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10262 * New -hex option for openssl rand.
10266 * Print out UTF8String and NumericString when parsing ASN1.
10270 * Support NumericString type for name components.
10274 * Allow CC in the environment to override the automatically chosen
10275 compiler. Note that nothing is done to ensure flags work with the
10280 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10282 * Properly check EVP_VerifyFinal() and similar return values
10285 *Ben Laurie, Bodo Moeller, Google Security Team*
10287 * Enable TLS extensions by default.
10291 * Allow the CHIL engine to be loaded, whether the application is
10292 multithreaded or not. (This does not release the developer from the
10293 obligation to set up the dynamic locking callbacks.)
10295 *Sander Temme <sander@temme.net>*
10297 * Use correct exit code if there is an error in dgst command.
10299 *Steve Henson; problem pointed out by Roland Dirlewanger*
10301 * Tweak Configure so that you need to say "experimental-jpake" to enable
10302 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10306 * Add experimental JPAKE support, including demo authentication in
10307 s_client and s_server.
10311 * Set the comparison function in v3_addr_canonize().
10313 *Rob Austein <sra@hactrn.net>*
10315 * Add support for XMPP STARTTLS in s_client.
10317 *Philip Paeps <philip@freebsd.org>*
10319 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10320 to ensure that even with this option, only ciphersuites in the
10321 server's preference list will be accepted. (Note that the option
10322 applies only when resuming a session, so the earlier behavior was
10323 just about the algorithm choice for symmetric cryptography.)
10327 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10329 * Fix NULL pointer dereference if a DTLS server received
10330 ChangeCipherSpec as first record ([CVE-2009-1386]).
10334 * Fix a state transition in s3_srvr.c and d1_srvr.c
10335 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10337 *Nagendra Modadugu*
10339 * The fix in 0.9.8c that supposedly got rid of unsafe
10340 double-checked locking was incomplete for RSA blinding,
10341 addressing just one layer of what turns out to have been
10342 doubly unsafe triple-checked locking.
10344 So now fix this for real by retiring the MONT_HELPER macro
10345 in crypto/rsa/rsa_eay.c.
10347 *Bodo Moeller; problem pointed out by Marius Schilder*
10349 * Various precautionary measures:
10351 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10353 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10354 (NB: This would require knowledge of the secret session ticket key
10355 to exploit, in which case you'd be SOL either way.)
10357 - Change bn_nist.c so that it will properly handle input BIGNUMs
10358 outside the expected range.
10360 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10363 *Neel Mehta, Bodo Moeller*
10365 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10366 the load fails. Useful for distros.
10368 *Ben Laurie and the FreeBSD team*
10370 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10374 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10378 * Expand ENGINE to support engine supplied SSL client certificate functions.
10380 This work was sponsored by Logica.
10384 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10385 keystores. Support for SSL/TLS client authentication too.
10386 Not compiled unless enable-capieng specified to Configure.
10388 This work was sponsored by Logica.
10392 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10393 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10394 attribute creation routines such as certificate requests and PKCS#12
10399 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10401 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10402 handshake which could lead to a client crash as found using the
10403 Codenomicon TLS test suite ([CVE-2008-1672])
10405 *Steve Henson, Mark Cox*
10407 * Fix double free in TLS server name extensions which could lead to
10408 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10412 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10414 Clear the error queue to ensure that error entries left from
10415 older function calls do not interfere with the correct operation.
10417 *Lutz Jaenicke, Erik de Castro Lopo*
10419 * Remove root CA certificates of commercial CAs:
10421 The OpenSSL project does not recommend any specific CA and does not
10422 have any policy with respect to including or excluding any CA.
10423 Therefore, it does not make any sense to ship an arbitrary selection
10424 of root CA certificates with the OpenSSL software.
10428 * RSA OAEP patches to fix two separate invalid memory reads.
10429 The first one involves inputs when 'lzero' is greater than
10430 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10431 before the beginning of from). The second one involves inputs where
10432 the 'db' section contains nothing but zeroes (there is a one-byte
10433 invalid read after the end of 'db').
10435 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10437 * Partial backport from 0.9.9-dev:
10439 Introduce bn_mul_mont (dedicated Montgomery multiplication
10440 procedure) as a candidate for BIGNUM assembler implementation.
10441 While 0.9.9-dev uses assembler for various architectures, only
10442 x86_64 is available by default here in the 0.9.8 branch, and
10443 32-bit x86 is available through a compile-time setting.
10445 To try the 32-bit x86 assembler implementation, use Configure
10446 option "enable-montasm" (which exists only for this backport).
10448 As "enable-montasm" for 32-bit x86 disclaims code stability
10449 anyway, in this constellation we activate additional code
10450 backported from 0.9.9-dev for further performance improvements,
10451 namely BN_from_montgomery_word. (To enable this otherwise,
10452 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10454 *Andy Polyakov (backport partially by Bodo Moeller)*
10456 * Add TLS session ticket callback. This allows an application to set
10457 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10458 values. This is useful for key rollover for example where several key
10459 sets may exist with different names.
10463 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10464 This was broken until now in 0.9.8 releases, such that the only way
10465 a registered ENGINE could be used (assuming it initialises
10466 successfully on the host) was to explicitly set it as the default
10467 for the relevant algorithms. This is in contradiction with 0.9.7
10468 behaviour and the documentation. With this fix, when an ENGINE is
10469 registered into a given algorithm's table of implementations, the
10470 'uptodate' flag is reset so that auto-discovery will be used next
10471 time a new context for that algorithm attempts to select an
10474 *Ian Lister (tweaked by Geoff Thorpe)*
10476 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10477 implementation in the following ways:
10479 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10482 Lack of BER streaming support means one pass streaming processing is
10483 only supported if data is detached: setting the streaming flag is
10484 ignored for embedded content.
10486 CMS support is disabled by default and must be explicitly enabled
10487 with the enable-cms configuration option.
10491 * Update the GMP engine glue to do direct copies between BIGNUM and
10492 mpz_t when openssl and GMP use the same limb size. Otherwise the
10493 existing "conversion via a text string export" trick is still used.
10495 *Paul Sheer <paulsheer@gmail.com>*
10497 * Zlib compression BIO. This is a filter BIO which compressed and
10498 uncompresses any data passed through it.
10502 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10503 RFC3394 compatible AES key wrapping.
10507 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10508 sets string data without copying. X509_ALGOR_set0() and
10509 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10510 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10511 from an X509_ATTRIBUTE structure optionally checking it occurs only
10512 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10517 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10518 to get the expected BN_FLG_CONSTTIME behavior.
10520 *Bodo Moeller (Google)*
10524 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10525 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10526 - added some more tests to do_tests.pl
10527 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10528 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10529 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10530 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10531 - various changes to netware.pl to enable gcc-cross builds on Win32
10533 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10534 - various changes to fix missing prototype warnings
10535 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10536 - added AES, WHIRLPOOL and CPUID assembler code to build files
10537 - added missing AES assembler make rules to mk1mf.pl
10538 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10540 *Guenter Knauf <eflash@gmx.net>*
10542 * Implement certificate status request TLS extension defined in RFC3546.
10543 A client can set the appropriate parameters and receive the encoded
10544 OCSP response via a callback. A server can query the supplied parameters
10545 and set the encoded OCSP response in the callback. Add simplified examples
10546 to s_client and s_server.
10550 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10552 * Fix various bugs:
10553 + Binary incompatibility of ssl_ctx_st structure
10554 + DTLS interoperation with non-compliant servers
10555 + Don't call get_session_cb() without proposed session
10556 + Fix ia64 assembler code
10558 *Andy Polyakov, Steve Henson*
10560 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10562 * DTLS Handshake overhaul. There were longstanding issues with
10563 OpenSSL DTLS implementation, which were making it impossible for
10564 RFC 4347 compliant client to communicate with OpenSSL server.
10565 Unfortunately just fixing these incompatibilities would "cut off"
10566 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10567 server keeps tolerating non RFC compliant syntax. The opposite is
10568 not true, 0.9.8f client can not communicate with earlier server.
10569 This update even addresses CVE-2007-4995.
10573 * Changes to avoid need for function casts in OpenSSL: some compilers
10574 (gcc 4.2 and later) reject their use.
10575 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10578 * Add RFC4507 support to OpenSSL. This includes the corrections in
10579 RFC4507bis. The encrypted ticket format is an encrypted encoded
10580 SSL_SESSION structure, that way new session features are automatically
10583 If a client application caches session in an SSL_SESSION structure
10584 support is transparent because tickets are now stored in the encoded
10587 The SSL_CTX structure automatically generates keys for ticket
10588 protection in servers so again support should be possible
10589 with no application modification.
10591 If a client or server wishes to disable RFC4507 support then the option
10592 SSL_OP_NO_TICKET can be set.
10594 Add a TLS extension debugging callback to allow the contents of any client
10595 or server extensions to be examined.
10597 This work was sponsored by Google.
10601 * Add initial support for TLS extensions, specifically for the server_name
10602 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10603 have new members for a hostname. The SSL data structure has an
10604 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10605 stored in that context to allow for session resumption, even after the
10606 SSL has been switched to a new SSL_CTX in reaction to a client's
10607 server_name extension.
10609 New functions (subject to change):
10611 SSL_get_servername()
10612 SSL_get_servername_type()
10615 New CTRL codes and macros (subject to change):
10617 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10618 - SSL_CTX_set_tlsext_servername_callback()
10619 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10620 - SSL_CTX_set_tlsext_servername_arg()
10621 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10623 openssl s_client has a new '-servername ...' option.
10625 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10626 '-key2 ...', '-servername_fatal' (subject to change). This allows
10627 testing the HostName extension for a specific single hostname ('-cert'
10628 and '-key' remain fallbacks for handshakes without HostName
10629 negotiation). If the unrecognized_name alert has to be sent, this by
10630 default is a warning; it becomes fatal with the '-servername_fatal'
10633 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10635 * Add AES and SSE2 assembly language support to VC++ build.
10639 * Mitigate attack on final subtraction in Montgomery reduction.
10643 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10644 (which previously caused an internal error).
10648 * Squeeze another 10% out of IGE mode when in != out.
10652 * AES IGE mode speedup.
10654 *Dean Gaudet (Google)*
10656 * Add the Korean symmetric 128-bit cipher SEED (see
10657 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10658 add SEED ciphersuites from RFC 4162:
10660 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10661 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10662 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10663 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10665 To minimize changes between patchlevels in the OpenSSL 0.9.8
10666 series, SEED remains excluded from compilation unless OpenSSL
10667 is configured with 'enable-seed'.
10669 *KISA, Bodo Moeller*
10671 * Mitigate branch prediction attacks, which can be practical if a
10672 single processor is shared, allowing a spy process to extract
10673 information. For detailed background information, see
10674 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10675 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10676 and Necessary Software Countermeasures"). The core of the change
10677 are new versions BN_div_no_branch() and
10678 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10679 respectively, which are slower, but avoid the security-relevant
10680 conditional branches. These are automatically called by BN_div()
10681 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10682 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10683 remove a conditional branch.
10685 BN_FLG_CONSTTIME is the new name for the previous
10686 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10687 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10688 in the exponent causes BN_mod_exp_mont() to use the alternative
10689 implementation in BN_mod_exp_mont_consttime().) The old name
10690 remains as a deprecated alias.
10692 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10693 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10694 constant-time implementations for more than just exponentiation.
10695 Here too the old name is kept as a deprecated alias.
10697 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10698 the BN_BLINDING structure gets an independent copy of the
10699 modulus. This means that the previous `BIGNUM *m` argument to
10700 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10701 essentially becomes `const BIGNUM *m`, although we can't actually
10702 change this in the header file before 0.9.9. It allows
10703 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10704 enable BN_FLG_CONSTTIME.
10706 *Matthew D Wood (Intel Corp)*
10708 * In the SSL/TLS server implementation, be strict about session ID
10709 context matching (which matters if an application uses a single
10710 external cache for different purposes). Previously,
10711 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10712 set. This did ensure strict client verification, but meant that,
10713 with applications using a single external cache for quite
10714 different requirements, clients could circumvent ciphersuite
10715 restrictions for a given session ID context by starting a session
10716 in a different context.
10720 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10721 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10722 authentication-only ciphersuites.
10726 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10727 not complete and could lead to a possible single byte overflow
10728 ([CVE-2007-5135]) [Ben Laurie]
10730 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10732 * Since AES128 and AES256 (and similarly Camellia128 and
10733 Camellia256) share a single mask bit in the logic of
10734 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10735 kludge to work properly if AES128 is available and AES256 isn't
10736 (or if Camellia128 is available and Camellia256 isn't).
10740 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10741 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10742 When a point or a seed is encoded in a BIT STRING, we need to
10743 prevent the removal of trailing zero bits to get the proper DER
10744 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10745 of a NamedBitList, for which trailing 0 bits need to be removed.)
10749 * Have SSL/TLS server implementation tolerate "mismatched" record
10750 protocol version while receiving ClientHello even if the
10751 ClientHello is fragmented. (The server can't insist on the
10752 particular protocol version it has chosen before the ServerHello
10753 message has informed the client about his choice.)
10757 * Add RFC 3779 support.
10759 *Rob Austein for ARIN, Ben Laurie*
10761 * Load error codes if they are not already present instead of using a
10762 static variable. This allows them to be cleanly unloaded and reloaded.
10763 Improve header file function name parsing.
10767 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10768 or CAPABILITY handshake as required by RFCs.
10770 *Goetz Babin-Ebell*
10772 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10774 * Introduce limits to prevent malicious keys being able to
10775 cause a denial of service. ([CVE-2006-2940])
10777 *Steve Henson, Bodo Moeller*
10779 * Fix ASN.1 parsing of certain invalid structures that can result
10780 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10782 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10783 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10785 * Fix SSL client code which could crash if connecting to a
10786 malicious SSLv2 server. ([CVE-2006-4343])
10788 *Tavis Ormandy and Will Drewry, Google Security Team*
10790 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10791 match only those. Before that, "AES256-SHA" would be interpreted
10792 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10793 the same strength classification in 0.9.7h) as we currently only
10794 have a single AES bit in the ciphersuite description bitmap.
10795 That change, however, also applied to ciphersuite strings such as
10796 "RC4-MD5" that intentionally matched multiple ciphersuites --
10797 namely, SSL 2.0 ciphersuites in addition to the more common ones
10798 from SSL 3.0/TLS 1.0.
10800 So we change the selection algorithm again: Naming an explicit
10801 ciphersuite selects this one ciphersuite, and any other similar
10802 ciphersuite (same bitmap) from *other* protocol versions.
10803 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10804 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10806 Since SSL 2.0 does not have any ciphersuites for which the
10807 128/256 bit distinction would be relevant, this works for now.
10808 The proper fix will be to use different bits for AES128 and
10809 AES256, which would have avoided the problems from the beginning;
10810 however, bits are scarce, so we can only do this in a new release
10811 (not just a patchlevel) when we can change the SSL_CIPHER
10812 definition to split the single 'unsigned long mask' bitmap into
10813 multiple values to extend the available space.
10817 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10819 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10820 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10822 * Add AES IGE and biIGE modes.
10826 * Change the Unix randomness entropy gathering to use poll() when
10827 possible instead of select(), since the latter has some
10828 undesirable limitations.
10830 *Darryl Miles via Richard Levitte and Bodo Moeller*
10832 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10833 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10834 cannot be implicitly activated as part of, e.g., the "AES" alias.
10835 However, please upgrade to OpenSSL 0.9.9[-dev] for
10836 non-experimental use of the ECC ciphersuites to get TLS extension
10837 support, which is required for curve and point format negotiation
10838 to avoid potential handshake problems.
10842 * Disable rogue ciphersuites:
10844 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10845 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10846 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10848 The latter two were purportedly from
10849 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10852 Also deactivate the remaining ciphersuites from
10853 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10854 unofficial, and the ID has long expired.
10858 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10859 dual-core machines) and other potential thread-safety issues.
10863 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10864 versions), which is now available for royalty-free use
10865 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10866 Also, add Camellia TLS ciphersuites from RFC 4132.
10868 To minimize changes between patchlevels in the OpenSSL 0.9.8
10869 series, Camellia remains excluded from compilation unless OpenSSL
10870 is configured with 'enable-camellia'.
10874 * Disable the padding bug check when compression is in use. The padding
10875 bug check assumes the first packet is of even length, this is not
10876 necessarily true if compression is enabled and can result in false
10877 positives causing handshake failure. The actual bug test is ancient
10878 code so it is hoped that implementations will either have fixed it by
10879 now or any which still have the bug do not support compression.
10883 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10885 * When applying a cipher rule check to see if string match is an explicit
10886 cipher suite and only match that one cipher suite if it is.
10890 * Link in manifests for VC++ if needed.
10892 *Austin Ziegler <halostatue@gmail.com>*
10894 * Update support for ECC-based TLS ciphersuites according to
10895 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10896 TLS extensions, which are supported starting with the 0.9.9
10897 branch, not in the OpenSSL 0.9.8 branch).
10901 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10902 opaque EVP_CIPHER_CTX handling.
10906 * Fixes and enhancements to zlib compression code. We now only use
10907 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10908 to conform with the standards mentioned here:
10909 <http://www.zlib.net/DLL_FAQ.txt>
10910 Static zlib linking now works on Windows and the new --with-zlib-include
10911 --with-zlib-lib options to Configure can be used to supply the location
10912 of the headers and library. Gracefully handle case where zlib library
10917 * Several fixes and enhancements to the OID generation code. The old code
10918 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10919 handle numbers larger than ULONG_MAX, truncated printing and had a
10920 non standard OBJ_obj2txt() behaviour.
10924 * Add support for building of engines under engine/ as shared libraries
10925 under VC++ build system.
10929 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10930 Hopefully, we will not see any false combination of paths any more.
10934 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10936 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10937 (part of SSL_OP_ALL). This option used to disable the
10938 countermeasure against man-in-the-middle protocol-version
10939 rollback in the SSL 2.0 server implementation, which is a bad
10940 idea. ([CVE-2005-2969])
10942 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10943 for Information Security, National Institute of Advanced Industrial
10944 Science and Technology [AIST], Japan)*
10946 * Add two function to clear and return the verify parameter flags.
10950 * Keep cipherlists sorted in the source instead of sorting them at
10951 runtime, thus removing the need for a lock.
10955 * Avoid some small subgroup attacks in Diffie-Hellman.
10957 *Nick Mathewson and Ben Laurie*
10959 * Add functions for well-known primes.
10963 * Extended Windows CE support.
10965 *Satoshi Nakamura and Andy Polyakov*
10967 * Initialize SSL_METHOD structures at compile time instead of during
10968 runtime, thus removing the need for a lock.
10972 * Make PKCS7_decrypt() work even if no certificate is supplied by
10973 attempting to decrypt each encrypted key in turn. Add support to
10978 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10980 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10983 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10987 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10988 key into the same file any more.
10992 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10996 * Add -utf8 command line and config file option to 'ca'.
10998 *Stefan <stf@udoma.org*
11000 * Removed the macro des_crypt(), as it seems to conflict with some
11001 libraries. Use DES_crypt().
11005 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11006 involves renaming the source and generated shared-libs for
11007 both. The engines will accept the corrected or legacy ids
11008 ('ncipher' and '4758_cca' respectively) when binding. NB,
11009 this only applies when building 'shared'.
11011 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11013 * Add attribute functions to EVP_PKEY structure. Modify
11014 PKCS12_create() to recognize a CSP name attribute and
11015 use it. Make -CSP option work again in pkcs12 utility.
11019 * Add new functionality to the bn blinding code:
11020 - automatic re-creation of the BN_BLINDING parameters after
11021 a fixed number of uses (currently 32)
11022 - add new function for parameter creation
11023 - introduce flags to control the update behaviour of the
11024 BN_BLINDING parameters
11025 - hide BN_BLINDING structure
11026 Add a second BN_BLINDING slot to the RSA structure to improve
11027 performance when a single RSA object is shared among several
11032 * Add support for DTLS.
11034 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11036 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11037 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11041 * Remove buggy and incomplete DH cert support from
11042 ssl/ssl_rsa.c and ssl/s3_both.c
11046 * Use SHA-1 instead of MD5 as the default digest algorithm for
11047 the `apps/openssl` commands.
11051 * Compile clean with "-Wall -Wmissing-prototypes
11052 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11053 DEBUG_SAFESTACK must also be set.
11057 * Change ./Configure so that certain algorithms can be disabled by default.
11058 The new counterpiece to "no-xxx" is "enable-xxx".
11060 The patented RC5 and MDC2 algorithms will now be disabled unless
11061 "enable-rc5" and "enable-mdc2", respectively, are specified.
11063 (IDEA remains enabled despite being patented. This is because IDEA
11064 is frequently required for interoperability, and there is no license
11065 fee for non-commercial use. As before, "no-idea" can be used to
11066 avoid this algorithm.)
11070 * Add processing of proxy certificates (see RFC 3820). This work was
11071 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11072 EGEE (Enabling Grids for E-science in Europe).
11076 * RC4 performance overhaul on modern architectures/implementations, such
11077 as Intel P4, IA-64 and AMD64.
11081 * New utility extract-section.pl. This can be used specify an alternative
11082 section number in a pod file instead of having to treat each file as
11083 a separate case in Makefile. This can be done by adding two lines to the
11086 =for comment openssl_section:XXX
11088 The blank line is mandatory.
11092 * New arguments -certform, -keyform and -pass for s_client and s_server
11093 to allow alternative format key and certificate files and passphrase
11098 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11099 update associated structures and add various utility functions.
11101 Add new policy related verify parameters, include policy checking in
11102 standard verify code. Enhance 'smime' application with extra parameters
11103 to support policy checking and print out.
11107 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11108 Nehemiah processors. These extensions support AES encryption in hardware
11109 as well as RNG (though RNG support is currently disabled).
11111 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11113 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11117 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11119 *Andy Polyakov and a number of other people*
11121 * Improved PowerPC platform support. Most notably BIGNUM assembler
11122 implementation contributed by IBM.
11124 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11126 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11127 exponent rather than 'unsigned long'. There is a corresponding change to
11128 the new 'rsa_keygen' element of the RSA_METHOD structure.
11130 *Jelte Jansen, Geoff Thorpe*
11132 * Functionality for creating the initial serial number file is now
11133 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11135 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11136 number file to 1, which is bound to cause problems. To avoid
11137 the problems while respecting compatibility between different 0.9.7
11138 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11139 CA.pl for serial number initialization. With the new release 0.9.8,
11140 we can fix the problem directly in the 'ca' utility.)
11144 * Reduced header interdependencies by declaring more opaque objects in
11145 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11146 give fewer recursive includes, which could break lazy source code - so
11147 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11148 developers should define this symbol when building and using openssl to
11149 ensure they track the recommended behaviour, interfaces, [etc], but
11150 backwards-compatible behaviour prevails when this isn't defined.
11154 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11158 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11159 This will generate a random key of the appropriate length based on the
11160 cipher context. The EVP_CIPHER can provide its own random key generation
11161 routine to support keys of a specific form. This is used in the des and
11162 3des routines to generate a key of the correct parity. Update S/MIME
11163 code to use new functions and hence generate correct parity DES keys.
11164 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11165 valid (weak or incorrect parity).
11169 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11170 as looking them up. This is useful when the verified structure may contain
11171 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11172 present unless the new PKCS7_NO_CRL flag is asserted.
11176 * Extend ASN1 oid configuration module. It now additionally accepts the
11179 shortName = some long name, 1.2.3.4
11183 * Reimplemented the BN_CTX implementation. There is now no more static
11184 limitation on the number of variables it can handle nor the depth of the
11185 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11186 information can now expand as required, and rather than having a single
11187 static array of bignums, BN_CTX now uses a linked-list of such arrays
11188 allowing it to expand on demand whilst maintaining the usefulness of
11189 BN_CTX's "bundling".
11193 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11194 to allow all RSA operations to function using a single BN_CTX.
11198 * Preliminary support for certificate policy evaluation and checking. This
11199 is initially intended to pass the tests outlined in "Conformance Testing
11200 of Relying Party Client Certificate Path Processing Logic" v1.07.
11204 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11205 remained unused and not that useful. A variety of other little bignum
11206 tweaks and fixes have also been made continuing on from the audit (see
11211 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11212 associated ASN1, EVP and SSL functions and old ASN1 macros.
11216 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11217 and this should never fail. So the return value from the use of
11218 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11219 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11223 * BN_CTX_get() should return zero-valued bignums, providing the same
11224 initialised value as BN_new().
11226 *Geoff Thorpe, suggested by Ulf Möller*
11228 * Support for inhibitAnyPolicy certificate extension.
11232 * An audit of the BIGNUM code is underway, for which debugging code is
11233 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11234 is considered valid when processing BIGNUMs, and causes execution to
11235 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11236 further steps are taken to deliberately pollute unused data in BIGNUM
11237 structures to try and expose faulty code further on. For now, openssl will
11238 (in its default mode of operation) continue to tolerate the inconsistent
11239 forms that it has tolerated in the past, but authors and packagers should
11240 consider trying openssl and their own applications when compiled with
11241 these debugging symbols defined. It will help highlight potential bugs in
11242 their own code, and will improve the test coverage for OpenSSL itself. At
11243 some point, these tighter rules will become openssl's default to improve
11244 maintainability, though the assert()s and other overheads will remain only
11245 in debugging configurations. See bn.h for more details.
11247 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11249 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11250 that can only be obtained through BN_CTX_new() (which implicitly
11251 initialises it). The presence of this function only made it possible
11252 to overwrite an existing structure (and cause memory leaks).
11256 * Because of the callback-based approach for implementing LHASH as a
11257 template type, lh_insert() adds opaque objects to hash-tables and
11258 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11259 to clean up those corresponding objects before destroying the hash table
11260 (and losing the object pointers). So some over-zealous constifications in
11261 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11262 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11263 prototyped to have "const" restrictions on the object pointers they are
11264 given (and so aren't required to cast them away any more).
11268 * The tmdiff.h API was so ugly and minimal that our own timing utility
11269 (speed) prefers to use its own implementation. The two implementations
11270 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11271 its object type properly exposed (MS_TM) instead of casting to/from
11272 `char *`. This may still change yet if someone realises MS_TM and
11274 aren't necessarily the greatest nomenclatures - but this is what was used
11275 internally to the implementation so I've used that for now.
11279 * Ensure that deprecated functions do not get compiled when
11280 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11281 the self-tests were still using deprecated key-generation functions so
11282 these have been updated also.
11286 * Reorganise PKCS#7 code to separate the digest location functionality
11287 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11288 New function PKCS7_set_digest() to set the digest type for PKCS#7
11289 digestedData type. Add additional code to correctly generate the
11290 digestedData type and add support for this type in PKCS7 initialization
11295 * New function PKCS7_set0_type_other() this initializes a PKCS7
11296 structure of type "other".
11300 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11301 sure the loop does correctly stop and breaking ("division by zero")
11302 modulus operations are not performed. The (pre-generated) prime
11303 table crypto/bn/bn_prime.h was already correct, but it could not be
11304 re-generated on some platforms because of the "division by zero"
11305 situation in the script.
11307 *Ralf S. Engelschall*
11309 * Update support for ECC-based TLS ciphersuites according to
11310 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11311 SHA-1 now is only used for "small" curves (where the
11312 representation of a field element takes up to 24 bytes); for
11313 larger curves, the field element resulting from ECDH is directly
11314 used as premaster secret.
11316 *Douglas Stebila (Sun Microsystems Laboratories)*
11318 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11319 curve secp160r1 to the tests.
11321 *Douglas Stebila (Sun Microsystems Laboratories)*
11323 * Add the possibility to load symbols globally with DSO.
11325 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11327 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11328 control of the error stack.
11332 * Add support for STORE in ENGINE.
11336 * Add the STORE type. The intention is to provide a common interface
11337 to certificate and key stores, be they simple file-based stores, or
11338 HSM-type store, or LDAP stores, or...
11339 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11343 * Add a generic structure called OPENSSL_ITEM. This can be used to
11344 pass a list of arguments to any function as well as provide a way
11345 for a function to pass data back to the caller.
11349 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11350 works like BUF_strdup() but can be used to duplicate a portion of
11351 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11356 * Add the function sk_find_ex() which works like sk_find(), but will
11357 return an index to an element even if an exact match couldn't be
11358 found. The index is guaranteed to point at the element where the
11359 searched-for key would be inserted to preserve sorting order.
11363 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11364 takes an extra flags argument for optional functionality. Currently,
11365 the following flags are defined:
11367 OBJ_BSEARCH_VALUE_ON_NOMATCH
11368 This one gets OBJ_bsearch_ex() to return a pointer to the first
11369 element where the comparing function returns a negative or zero
11372 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11373 This one gets OBJ_bsearch_ex() to return a pointer to the first
11374 element where the comparing function returns zero. This is useful
11375 if there are more than one element where the comparing function
11380 * Make it possible to create self-signed certificates with 'openssl ca'
11381 in such a way that the self-signed certificate becomes part of the
11382 CA database and uses the same mechanisms for serial number generation
11383 as all other certificate signing. The new flag '-selfsign' enables
11384 this functionality. Adapt CA.sh and CA.pl.in.
11388 * Add functionality to check the public key of a certificate request
11389 against a given private. This is useful to check that a certificate
11390 request can be signed by that key (self-signing).
11394 * Make it possible to have multiple active certificates with the same
11395 subject in the CA index file. This is done only if the keyword
11396 'unique_subject' is set to 'no' in the main CA section (default
11397 if 'CA_default') of the configuration file. The value is saved
11398 with the database itself in a separate index attribute file,
11399 named like the index file with '.attr' appended to the name.
11403 * Generate multi-valued AVAs using '+' notation in config files for
11408 * Support for nameConstraints certificate extension.
11412 * Support for policyConstraints certificate extension.
11416 * Support for policyMappings certificate extension.
11420 * Make sure the default DSA_METHOD implementation only uses its
11421 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11422 and change its own handlers to be NULL so as to remove unnecessary
11423 indirection. This lets alternative implementations fallback to the
11424 default implementation more easily.
11428 * Support for directoryName in GeneralName related extensions
11433 * Make it possible to link applications using Makefile.shared.
11434 Make that possible even when linking against static libraries!
11438 * Support for single pass processing for S/MIME signing. This now
11439 means that S/MIME signing can be done from a pipe, in addition
11440 cleartext signing (multipart/signed type) is effectively streaming
11441 and the signed data does not need to be all held in memory.
11443 This is done with a new flag PKCS7_STREAM. When this flag is set
11444 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11445 is done after the data is output (and digests calculated) in
11446 SMIME_write_PKCS7().
11450 * Add full support for -rpath/-R, both in shared libraries and
11451 applications, at least on the platforms where it's known how
11456 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11457 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11458 will now compute a table of multiples of the generator that
11459 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11460 faster (notably in the case of a single point multiplication,
11461 scalar * generator).
11463 *Nils Larsch, Bodo Moeller*
11465 * IPv6 support for certificate extensions. The various extensions
11466 which use the IP:a.b.c.d can now take IPv6 addresses using the
11467 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11472 * Added an ENGINE that implements RSA by performing private key
11473 exponentiations with the GMP library. The conversions to and from
11474 GMP's mpz_t format aren't optimised nor are any montgomery forms
11475 cached, and on x86 it appears OpenSSL's own performance has caught up.
11476 However there are likely to be other architectures where GMP could
11477 provide a boost. This ENGINE is not built in by default, but it can be
11478 specified at Configure time and should be accompanied by the necessary
11479 linker additions, eg;
11480 ./config -DOPENSSL_USE_GMP -lgmp
11484 * "openssl engine" will not display ENGINE/DSO load failure errors when
11485 testing availability of engines with "-t" - the old behaviour is
11486 produced by increasing the feature's verbosity with "-tt".
11490 * ECDSA routines: under certain error conditions uninitialized BN objects
11491 could be freed. Solution: make sure initialization is performed early
11492 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11497 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11498 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11499 software implementations. For DSA and DH, parameter generation can
11500 also be overridden by providing the appropriate method callbacks.
11504 * Change the "progress" mechanism used in key-generation and
11505 primality testing to functions that take a new BN_GENCB pointer in
11506 place of callback/argument pairs. The new API functions have `_ex`
11507 postfixes and the older functions are reimplemented as wrappers for
11508 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11509 declarations of the old functions to help (graceful) attempts to
11510 migrate to the new functions. Also, the new key-generation API
11511 functions operate on a caller-supplied key-structure and return
11512 success/failure rather than returning a key or NULL - this is to
11513 help make "keygen" another member function of RSA_METHOD etc.
11515 Example for using the new callback interface:
11517 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11518 void *my_arg = ...;
11521 BN_GENCB_set(&my_cb, my_callback, my_arg);
11523 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11524 /* For the meaning of a, b in calls to my_callback(), see the
11525 * documentation of the function that calls the callback.
11526 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11527 * my_callback should return 1 if it wants BN_is_prime_ex()
11528 * to continue, or 0 to stop.
11533 * Change the ZLIB compression method to be stateful, and make it
11534 available to TLS with the number defined in
11535 draft-ietf-tls-compression-04.txt.
11539 * Add the ASN.1 structures and functions for CertificatePair, which
11540 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11542 CertificatePair ::= SEQUENCE {
11543 forward [0] Certificate OPTIONAL,
11544 reverse [1] Certificate OPTIONAL,
11545 -- at least one of the pair shall be present -- }
11547 Also implement the PEM functions to read and write certificate
11548 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11550 This needed to be defined, mostly for the sake of the LDAP
11551 attribute crossCertificatePair, but may prove useful elsewhere as
11556 * Make it possible to inhibit symlinking of shared libraries in
11557 Makefile.shared, for Cygwin's sake.
11561 * Extend the BIGNUM API by creating a function
11562 void BN_set_negative(BIGNUM *a, int neg);
11563 and a macro that behave like
11564 int BN_is_negative(const BIGNUM *a);
11566 to avoid the need to access 'a->neg' directly in applications.
11570 * Implement fast modular reduction for pseudo-Mersenne primes
11571 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11572 EC_GROUP_new_curve_GFp() will now automatically use this
11575 *Nils Larsch <nla@trustcenter.de>*
11577 * Add new lock type (CRYPTO_LOCK_BN).
11581 * Change the ENGINE framework to automatically load engines
11582 dynamically from specific directories unless they could be
11583 found to already be built in or loaded. Move all the
11584 current engines except for the cryptodev one to a new
11585 directory engines/.
11586 The engines in engines/ are built as shared libraries if
11587 the "shared" options was given to ./Configure or ./config.
11588 Otherwise, they are inserted in libcrypto.a.
11589 /usr/local/ssl/engines is the default directory for dynamic
11590 engines, but that can be overridden at configure time through
11591 the usual use of --prefix and/or --openssldir, and at run
11592 time with the environment variable OPENSSL_ENGINES.
11594 *Geoff Thorpe and Richard Levitte*
11596 * Add Makefile.shared, a helper makefile to build shared
11597 libraries. Adapt Makefile.org.
11601 * Add version info to Win32 DLLs.
11603 *Peter 'Luna' Runestig" <peter@runestig.com>*
11605 * Add new 'medium level' PKCS#12 API. Certificates and keys
11606 can be added using this API to created arbitrary PKCS#12
11607 files while avoiding the low-level API.
11609 New options to PKCS12_create(), key or cert can be NULL and
11610 will then be omitted from the output file. The encryption
11611 algorithm NIDs can be set to -1 for no encryption, the mac
11612 iteration count can be set to 0 to omit the mac.
11614 Enhance pkcs12 utility by making the -nokeys and -nocerts
11615 options work when creating a PKCS#12 file. New option -nomac
11616 to omit the mac, NONE can be set for an encryption algorithm.
11617 New code is modified to use the enhanced PKCS12_create()
11618 instead of the low-level API.
11622 * Extend ASN1 encoder to support indefinite length constructed
11623 encoding. This can output sequences tags and octet strings in
11624 this form. Modify pk7_asn1.c to support indefinite length
11625 encoding. This is experimental and needs additional code to
11626 be useful, such as an ASN1 bio and some enhanced streaming
11629 Extend template encode functionality so that tagging is passed
11630 down to the template encoder.
11634 * Let 'openssl req' fail if an argument to '-newkey' is not
11635 recognized instead of using RSA as a default.
11639 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11640 As these are not official, they are not included in "ALL";
11641 the "ECCdraft" ciphersuite group alias can be used to select them.
11643 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11645 * Add ECDH engine support.
11647 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11649 * Add ECDH in new directory crypto/ecdh/.
11651 *Douglas Stebila (Sun Microsystems Laboratories)*
11653 * Let BN_rand_range() abort with an error after 100 iterations
11654 without success (which indicates a broken PRNG).
11658 * Change BN_mod_sqrt() so that it verifies that the input value
11659 is really the square of the return value. (Previously,
11660 BN_mod_sqrt would show GIGO behaviour.)
11664 * Add named elliptic curves over binary fields from X9.62, SECG,
11665 and WAP/WTLS; add OIDs that were still missing.
11667 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11669 * Extend the EC library for elliptic curves over binary fields
11670 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11673 EC_GF2m_simple_method
11677 EC_GROUP_new_curve_GF2m
11678 EC_GROUP_set_curve_GF2m
11679 EC_GROUP_get_curve_GF2m
11680 EC_POINT_set_affine_coordinates_GF2m
11681 EC_POINT_get_affine_coordinates_GF2m
11682 EC_POINT_set_compressed_coordinates_GF2m
11684 Point compression for binary fields is disabled by default for
11685 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11688 As binary polynomials are represented as BIGNUMs, various members
11689 of the EC_GROUP and EC_POINT data structures can be shared
11690 between the implementations for prime fields and binary fields;
11691 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11692 are essentially identical to their `..._GFp` counterparts.
11693 (For simplicity, the `..._GFp` prefix has been dropped from
11694 various internal method names.)
11696 An internal 'field_div' method (similar to 'field_mul' and
11697 'field_sqr') has been added; this is used only for binary fields.
11699 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11701 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11702 through methods ('mul', 'precompute_mult').
11704 The generic implementations (now internally called 'ec_wNAF_mul'
11705 and 'ec_wNAF_precomputed_mult') remain the default if these
11706 methods are undefined.
11708 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11710 * New function EC_GROUP_get_degree, which is defined through
11711 EC_METHOD. For curves over prime fields, this returns the bit
11712 length of the modulus.
11714 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11716 * New functions EC_GROUP_dup, EC_POINT_dup.
11717 (These simply call ..._new and ..._copy).
11719 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11721 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11722 Polynomials are represented as BIGNUMs (where the sign bit is not
11723 used) in the following functions [macros]:
11726 BN_GF2m_sub [= BN_GF2m_add]
11727 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11728 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11729 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11731 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11732 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11733 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11734 BN_GF2m_cmp [= BN_ucmp]
11736 (Note that only the 'mod' functions are actually for fields GF(2^m).
11737 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11739 For some functions, an the irreducible polynomial defining a
11740 field can be given as an 'unsigned int[]' with strictly
11741 decreasing elements giving the indices of those bits that are set;
11742 i.e., p[] represents the polynomial
11743 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11745 p[0] > p[1] > ... > p[k] = 0.
11746 This applies to the following functions:
11749 BN_GF2m_mod_mul_arr
11750 BN_GF2m_mod_sqr_arr
11751 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11752 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11753 BN_GF2m_mod_exp_arr
11754 BN_GF2m_mod_sqrt_arr
11755 BN_GF2m_mod_solve_quad_arr
11759 Conversion can be performed by the following functions:
11764 bntest.c has additional tests for binary polynomial arithmetic.
11766 Two implementations for BN_GF2m_mod_div() are available.
11767 The default algorithm simply uses BN_GF2m_mod_inv() and
11768 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11769 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11770 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11772 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11774 * Add new error code 'ERR_R_DISABLED' that can be used when some
11775 functionality is disabled at compile-time.
11777 *Douglas Stebila <douglas.stebila@sun.com>*
11779 * Change default behaviour of 'openssl asn1parse' so that more
11780 information is visible when viewing, e.g., a certificate:
11782 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11783 mode the content of non-printable OCTET STRINGs is output in a
11784 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11785 avoid the appearance of a printable string.
11787 *Nils Larsch <nla@trustcenter.de>*
11789 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11791 EC_GROUP_set_asn1_flag()
11792 EC_GROUP_get_asn1_flag()
11793 EC_GROUP_set_point_conversion_form()
11794 EC_GROUP_get_point_conversion_form()
11795 These control ASN1 encoding details:
11796 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11797 has been set to OPENSSL_EC_NAMED_CURVE.
11798 - Points are encoded in uncompressed form by default; options for
11799 asn1_for are as for point2oct, namely
11800 POINT_CONVERSION_COMPRESSED
11801 POINT_CONVERSION_UNCOMPRESSED
11802 POINT_CONVERSION_HYBRID
11804 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11806 EC_GROUP_set_seed()
11807 EC_GROUP_get0_seed()
11808 EC_GROUP_get_seed_len()
11809 This is used only for ASN1 purposes (so far).
11811 *Nils Larsch <nla@trustcenter.de>*
11813 * Add 'field_type' member to EC_METHOD, which holds the NID
11814 of the appropriate field type OID. The new function
11815 EC_METHOD_get_field_type() returns this value.
11817 *Nils Larsch <nla@trustcenter.de>*
11820 EC_POINT_point2bn()
11821 EC_POINT_bn2point()
11822 EC_POINT_point2hex()
11823 EC_POINT_hex2point()
11824 providing useful interfaces to EC_POINT_point2oct() and
11825 EC_POINT_oct2point().
11827 *Nils Larsch <nla@trustcenter.de>*
11829 * Change internals of the EC library so that the functions
11830 EC_GROUP_set_generator()
11831 EC_GROUP_get_generator()
11832 EC_GROUP_get_order()
11833 EC_GROUP_get_cofactor()
11834 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11835 to methods, which would lead to unnecessary code duplication when
11836 adding different types of curves.
11838 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11840 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11841 arithmetic, and such that modified wNAFs are generated
11842 (which avoid length expansion in many cases).
11846 * Add a function EC_GROUP_check_discriminant() (defined via
11847 EC_METHOD) that verifies that the curve discriminant is non-zero.
11849 Add a function EC_GROUP_check() that makes some sanity tests
11850 on a EC_GROUP, its generator and order. This includes
11851 EC_GROUP_check_discriminant().
11853 *Nils Larsch <nla@trustcenter.de>*
11855 * Add ECDSA in new directory crypto/ecdsa/.
11857 Add applications 'openssl ecparam' and 'openssl ecdsa'
11858 (these are based on 'openssl dsaparam' and 'openssl dsa').
11860 ECDSA support is also included in various other files across the
11861 library. Most notably,
11862 - 'openssl req' now has a '-newkey ecdsa:file' option;
11863 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11864 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11865 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11866 them suitable for ECDSA where domain parameters must be
11867 extracted before the specific public key;
11868 - ECDSA engine support has been added.
11870 *Nils Larsch <nla@trustcenter.de>*
11872 * Include some named elliptic curves, and add OIDs from X9.62,
11873 SECG, and WAP/WTLS. Each curve can be obtained from the new
11875 EC_GROUP_new_by_curve_name(),
11876 and the list of available named curves can be obtained with
11877 EC_get_builtin_curves().
11878 Also add a 'curve_name' member to EC_GROUP objects, which can be
11880 EC_GROUP_set_curve_name()
11881 EC_GROUP_get_curve_name()
11883 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11885 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11886 was actually never needed) and in BN_mul(). The removal in BN_mul()
11887 required a small change in bn_mul_part_recursive() and the addition
11888 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11889 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11890 bn_sub_words() and bn_add_words() except they take arrays with
11895 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11897 * Cleanse PEM buffers before freeing them since they may contain
11900 *Benjamin Bennett <ben@psc.edu>*
11902 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11903 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11904 authentication-only ciphersuites.
11908 * Since AES128 and AES256 share a single mask bit in the logic of
11909 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11910 kludge to work properly if AES128 is available and AES256 isn't.
11914 * Expand security boundary to match 1.1.1 module.
11918 * Remove redundant features: hash file source, editing of test vectors
11919 modify fipsld to use external fips_premain.c signature.
11923 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11924 run algorithm test programs.
11928 * Make algorithm test programs more tolerant of whitespace.
11932 * Have SSL/TLS server implementation tolerate "mismatched" record
11933 protocol version while receiving ClientHello even if the
11934 ClientHello is fragmented. (The server can't insist on the
11935 particular protocol version it has chosen before the ServerHello
11936 message has informed the client about his choice.)
11940 * Load error codes if they are not already present instead of using a
11941 static variable. This allows them to be cleanly unloaded and reloaded.
11945 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11947 * Introduce limits to prevent malicious keys being able to
11948 cause a denial of service. ([CVE-2006-2940])
11950 *Steve Henson, Bodo Moeller*
11952 * Fix ASN.1 parsing of certain invalid structures that can result
11953 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11955 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11956 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11958 * Fix SSL client code which could crash if connecting to a
11959 malicious SSLv2 server. ([CVE-2006-4343])
11961 *Tavis Ormandy and Will Drewry, Google Security Team*
11963 * Change ciphersuite string processing so that an explicit
11964 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11965 will no longer include "AES128-SHA"), and any other similar
11966 ciphersuite (same bitmap) from *other* protocol versions (so that
11967 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11968 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11969 changes from 0.9.8b and 0.9.8d.
11973 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11975 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11976 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11978 * Change the Unix randomness entropy gathering to use poll() when
11979 possible instead of select(), since the latter has some
11980 undesirable limitations.
11982 *Darryl Miles via Richard Levitte and Bodo Moeller*
11984 * Disable rogue ciphersuites:
11986 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11987 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11988 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11990 The latter two were purportedly from
11991 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11994 Also deactivate the remaining ciphersuites from
11995 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11996 unofficial, and the ID has long expired.
12000 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12001 dual-core machines) and other potential thread-safety issues.
12005 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12007 * Adapt fipsld and the build system to link against the validated FIPS
12008 module in FIPS mode.
12012 * Fixes for VC++ 2005 build under Windows.
12016 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12017 from a Windows bash shell such as MSYS. It is autodetected from the
12018 "config" script when run from a VC++ environment. Modify standard VC++
12019 build to use fipscanister.o from the GNU make build.
12023 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12025 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12026 The value now differs depending on if you build for FIPS or not.
12027 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12028 safely run with a non-FIPSed libcrypto, as it may crash because of
12029 the difference induced by this change.
12033 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12035 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12036 (part of SSL_OP_ALL). This option used to disable the
12037 countermeasure against man-in-the-middle protocol-version
12038 rollback in the SSL 2.0 server implementation, which is a bad
12039 idea. ([CVE-2005-2969])
12041 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12042 for Information Security, National Institute of Advanced Industrial
12043 Science and Technology [AIST, Japan)]*
12045 * Minimal support for X9.31 signatures and PSS padding modes. This is
12046 mainly for FIPS compliance and not fully integrated at this stage.
12050 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12051 the exponentiation using a fixed-length exponent. (Otherwise,
12052 the information leaked through timing could expose the secret key
12053 after many signatures; cf. Bleichenbacher's attack on DSA with
12058 * Make a new fixed-window mod_exp implementation the default for
12059 RSA, DSA, and DH private-key operations so that the sequence of
12060 squares and multiplies and the memory access pattern are
12061 independent of the particular secret key. This will mitigate
12062 cache-timing and potential related attacks.
12064 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12065 and this is automatically used by BN_mod_exp_mont() if the new flag
12066 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12067 will use this BN flag for private exponents unless the flag
12068 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12069 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12071 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12073 * Change the client implementation for SSLv23_method() and
12074 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12075 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12076 (Previously, the SSL 2.0 backwards compatible Client Hello
12077 message format would be used even with SSL_OP_NO_SSLv2.)
12081 * Add support for smime-type MIME parameter in S/MIME messages which some
12086 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12087 a threadsafe manner. Modify rsa code to use new function and add calls
12088 to dsa and dh code (which had race conditions before).
12092 * Include the fixed error library code in the C error file definitions
12093 instead of fixing them up at runtime. This keeps the error code
12094 structures constant.
12098 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12100 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12103 * Fixes for newer kerberos headers. NB: the casts are needed because
12104 the 'length' field is signed on one version and unsigned on another
12105 with no (?) obvious way to tell the difference, without these VC++
12106 complains. Also the "definition" of FAR (blank) is no longer included
12107 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12108 some needed definitions.
12112 * Undo Cygwin change.
12116 * Added support for proxy certificates according to RFC 3820.
12117 Because they may be a security thread to unaware applications,
12118 they must be explicitly allowed in run-time. See
12119 docs/HOWTO/proxy_certificates.txt for further information.
12123 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12125 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12126 server and client random values. Previously
12127 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12128 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12130 This change has negligible security impact because:
12132 1. Server and client random values still have 24 bytes of pseudo random
12135 2. Server and client random values are sent in the clear in the initial
12138 3. The master secret is derived using the premaster secret (48 bytes in
12139 size for static RSA ciphersuites) as well as client server and random
12142 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12145 *Stephen Henson, reported by UK NISCC*
12147 * Use Windows randomness collection on Cygwin.
12151 * Fix hang in EGD/PRNGD query when communication socket is closed
12152 prematurely by EGD/PRNGD.
12154 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12156 * Prompt for pass phrases when appropriate for PKCS12 input format.
12160 * Back-port of selected performance improvements from development
12161 branch, as well as improved support for PowerPC platforms.
12165 * Add lots of checks for memory allocation failure, error codes to indicate
12166 failure and freeing up memory if a failure occurs.
12168 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12170 * Add new -passin argument to dgst.
12174 * Perform some character comparisons of different types in X509_NAME_cmp:
12175 this is needed for some certificates that re-encode DNs into UTF8Strings
12176 (in violation of RFC3280) and can't or won't issue name rollover
12181 * Make an explicit check during certificate validation to see that
12182 the CA setting in each certificate on the chain is correct. As a
12183 side effect always do the following basic checks on extensions,
12184 not just when there's an associated purpose to the check:
12186 - if there is an unhandled critical extension (unless the user
12187 has chosen to ignore this fault)
12188 - if the path length has been exceeded (if one is set at all)
12189 - that certain extensions fit the associated purpose (if one has
12194 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12196 * Avoid a race condition when CRLs are checked in a multi threaded
12197 environment. This would happen due to the reordering of the revoked
12198 entries during signature checking and serial number lookup. Now the
12199 encoding is cached and the serial number sort performed under a lock.
12200 Add new STACK function sk_is_sorted().
12204 * Add Delta CRL to the extension code.
12208 * Various fixes to s3_pkt.c so alerts are sent properly.
12210 *David Holmes <d.holmes@f5.com>*
12212 * Reduce the chances of duplicate issuer name and serial numbers (in
12213 violation of RFC3280) using the OpenSSL certificate creation utilities.
12214 This is done by creating a random 64 bit value for the initial serial
12215 number when a serial number file is created or when a self signed
12216 certificate is created using 'openssl req -x509'. The initial serial
12217 number file is created using 'openssl x509 -next_serial' in CA.pl
12218 rather than being initialized to 1.
12222 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12224 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12225 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12227 *Joe Orton, Steve Henson*
12229 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12232 *Joe Orton, Steve Henson*
12234 * Make it possible to have multiple active certificates with the same
12235 subject in the CA index file. This is done only if the keyword
12236 'unique_subject' is set to 'no' in the main CA section (default
12237 if 'CA_default') of the configuration file. The value is saved
12238 with the database itself in a separate index attribute file,
12239 named like the index file with '.attr' appended to the name.
12243 * X509 verify fixes. Disable broken certificate workarounds when
12244 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12245 keyUsage extension present. Don't accept CRLs with unhandled critical
12246 extensions: since verify currently doesn't process CRL extensions this
12247 rejects a CRL with *any* critical extensions. Add new verify error codes
12252 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12253 A clarification of RFC2560 will require the use of OCTET STRINGs and
12254 some implementations cannot handle the current raw format. Since OpenSSL
12255 copies and compares OCSP nonces as opaque blobs without any attempt at
12256 parsing them this should not create any compatibility issues.
12260 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12261 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12262 this HMAC (and other) operations are several times slower than OpenSSL
12267 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12269 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12271 * Use the correct content when signing type "other".
12275 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12277 * Fix various bugs revealed by running the NISCC test suite:
12279 Stop out of bounds reads in the ASN1 code when presented with
12280 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12282 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12284 If verify callback ignores invalid public key errors don't try to check
12285 certificate signature with the NULL public key.
12289 * New -ignore_err option in ocsp application to stop the server
12290 exiting on the first error in a request.
12294 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12295 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12300 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12301 extra data after the compression methods not only for TLS 1.0
12302 but also for SSL 3.0 (as required by the specification).
12304 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12306 * Change X509_certificate_type() to mark the key as exported/exportable
12307 when it's 512 *bits* long, not 512 bytes.
12311 * Change AES_cbc_encrypt() so it outputs exact multiple of
12312 blocks during encryption.
12316 * Various fixes to base64 BIO and non blocking I/O. On write
12317 flushes were not handled properly if the BIO retried. On read
12318 data was not being buffered properly and had various logic bugs.
12319 This also affects blocking I/O when the data being decoded is a
12324 * Various S/MIME bugfixes and compatibility changes:
12325 output correct application/pkcs7 MIME type if
12326 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12327 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12328 of files as .eml work). Correctly handle very long lines in MIME
12333 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12335 * Countermeasure against the Klima-Pokorny-Rosa extension of
12336 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12337 a protocol version number mismatch like a decryption error
12338 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12342 * Turn on RSA blinding by default in the default implementation
12343 to avoid a timing attack. Applications that don't want it can call
12344 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12345 They would be ill-advised to do so in most cases.
12347 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12349 * Change RSA blinding code so that it works when the PRNG is not
12350 seeded (in this case, the secret RSA exponent is abused as
12351 an unpredictable seed -- if it is not unpredictable, there
12352 is no point in blinding anyway). Make RSA blinding thread-safe
12353 by remembering the creator's thread ID in rsa->blinding and
12354 having all other threads use local one-time blinding factors
12355 (this requires more computation than sharing rsa->blinding, but
12356 avoids excessive locking; and if an RSA object is not shared
12357 between threads, blinding will still be very fast).
12361 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12362 ENGINE as defaults for all supported algorithms irrespective of
12363 the 'flags' parameter. 'flags' is now honoured, so applications
12364 should make sure they are passing it correctly.
12368 * Target "mingw" now allows native Windows code to be generated in
12369 the Cygwin environment as well as with the MinGW compiler.
12373 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12375 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12376 via timing by performing a MAC computation even if incorrect
12377 block cipher padding has been found. This is a countermeasure
12378 against active attacks where the attacker has to distinguish
12379 between bad padding and a MAC verification error. ([CVE-2003-0078])
12381 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12382 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12383 Martin Vuagnoux (EPFL, Ilion)*
12385 * Make the no-err option work as intended. The intention with no-err
12386 is not to have the whole error stack handling routines removed from
12387 libcrypto, it's only intended to remove all the function name and
12388 reason texts, thereby removing some of the footprint that may not
12389 be interesting if those errors aren't displayed anyway.
12391 NOTE: it's still possible for any application or module to have its
12392 own set of error texts inserted. The routines are there, just not
12393 used by default when no-err is given.
12397 * Add support for FreeBSD on IA64.
12399 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12401 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12402 Kerberos function mit_des_cbc_cksum(). Before this change,
12403 the value returned by DES_cbc_cksum() was like the one from
12404 mit_des_cbc_cksum(), except the bytes were swapped.
12406 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12408 * Allow an application to disable the automatic SSL chain building.
12409 Before this a rather primitive chain build was always performed in
12410 ssl3_output_cert_chain(): an application had no way to send the
12411 correct chain if the automatic operation produced an incorrect result.
12413 Now the chain builder is disabled if either:
12415 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12417 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12419 The reasoning behind this is that an application would not want the
12420 auto chain building to take place if extra chain certificates are
12421 present and it might also want a means of sending no additional
12422 certificates (for example the chain has two certificates and the
12427 * Add the possibility to build without the ENGINE framework.
12429 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12431 * Under Win32 gmtime() can return NULL: check return value in
12432 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12436 * DSA routines: under certain error conditions uninitialized BN objects
12437 could be freed. Solution: make sure initialization is performed early
12438 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12439 Nils Larsch <nla@trustcenter.de> via PR#459)
12443 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12444 checked on reconnect on the client side, therefore session resumption
12445 could still fail with a "ssl session id is different" error. This
12446 behaviour is masked when SSL_OP_ALL is used due to
12447 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12448 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12449 followup to PR #377.
12453 * IA-32 assembler support enhancements: unified ELF targets, support
12454 for SCO/Caldera platforms, fix for Cygwin shared build.
12458 * Add support for FreeBSD on sparc64. As a consequence, support for
12459 FreeBSD on non-x86 processors is separate from x86 processors on
12460 the config script, much like the NetBSD support.
12462 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12464 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12466 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12469 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12470 code (06) was taken as the first octet of the session ID and the last
12471 octet was ignored consequently. As a result SSLv2 client side session
12472 caching could not have worked due to the session ID mismatch between
12474 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12479 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12480 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12485 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12486 seems that in spite of existing for more than a year, many application
12487 author have done nothing to provide the necessary callbacks, which
12488 means that this particular engine will not work properly anywhere.
12489 This is a very unfortunate situation which forces us, in the name
12490 of usability, to give the hw_ncipher.c a static lock, which is part
12492 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12493 appear in 0.9.8 or later. We EXPECT application authors to have
12494 dealt properly with this when 0.9.8 is released (unless we actually
12495 make such changes in the libcrypto locking code that changes will
12496 have to be made anyway).
12500 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12501 octets have been read, EOF or an error occurs. Without this change
12502 some truncated ASN1 structures will not produce an error.
12506 * Disable Heimdal support, since it hasn't been fully implemented.
12507 Still give the possibility to force the use of Heimdal, but with
12508 warnings and a request that patches get sent to openssl-dev.
12512 * Add the VC-CE target, introduce the WINCE sysname, and add
12513 INSTALL.WCE and appropriate conditionals to make it build.
12515 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12517 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12518 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12519 edit numbers of the version.
12521 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12523 * Introduce safe string copy and catenation functions
12524 (BUF_strlcpy() and BUF_strlcat()).
12526 *Ben Laurie (CHATS) and Richard Levitte*
12528 * Avoid using fixed-size buffers for one-line DNs.
12530 *Ben Laurie (CHATS)*
12532 * Add BUF_MEM_grow_clean() to avoid information leakage when
12533 resizing buffers containing secrets, and use where appropriate.
12535 *Ben Laurie (CHATS)*
12537 * Avoid using fixed size buffers for configuration file location.
12539 *Ben Laurie (CHATS)*
12541 * Avoid filename truncation for various CA files.
12543 *Ben Laurie (CHATS)*
12545 * Use sizeof in preference to magic numbers.
12547 *Ben Laurie (CHATS)*
12549 * Avoid filename truncation in cert requests.
12551 *Ben Laurie (CHATS)*
12553 * Add assertions to check for (supposedly impossible) buffer
12556 *Ben Laurie (CHATS)*
12558 * Don't cache truncated DNS entries in the local cache (this could
12559 potentially lead to a spoofing attack).
12561 *Ben Laurie (CHATS)*
12563 * Fix various buffers to be large enough for hex/decimal
12564 representations in a platform independent manner.
12566 *Ben Laurie (CHATS)*
12568 * Add CRYPTO_realloc_clean() to avoid information leakage when
12569 resizing buffers containing secrets, and use where appropriate.
12571 *Ben Laurie (CHATS)*
12573 * Add BIO_indent() to avoid much slightly worrying code to do
12576 *Ben Laurie (CHATS)*
12578 * Convert sprintf()/BIO_puts() to BIO_printf().
12580 *Ben Laurie (CHATS)*
12582 * buffer_gets() could terminate with the buffer only half
12585 *Ben Laurie (CHATS)*
12587 * Add assertions to prevent user-supplied crypto functions from
12588 overflowing internal buffers by having large block sizes, etc.
12590 *Ben Laurie (CHATS)*
12592 * New OPENSSL_assert() macro (similar to assert(), but enabled
12595 *Ben Laurie (CHATS)*
12597 * Eliminate unused copy of key in RC4.
12599 *Ben Laurie (CHATS)*
12601 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12603 *Ben Laurie (CHATS)*
12605 * Fix off-by-one error in EGD path.
12607 *Ben Laurie (CHATS)*
12609 * If RANDFILE path is too long, ignore instead of truncating.
12611 *Ben Laurie (CHATS)*
12613 * Eliminate unused and incorrectly sized X.509 structure
12616 *Ben Laurie (CHATS)*
12618 * Eliminate unused and dangerous function knumber().
12620 *Ben Laurie (CHATS)*
12622 * Eliminate unused and dangerous structure, KSSL_ERR.
12624 *Ben Laurie (CHATS)*
12626 * Protect against overlong session ID context length in an encoded
12627 session object. Since these are local, this does not appear to be
12630 *Ben Laurie (CHATS)*
12632 * Change from security patch (see 0.9.6e below) that did not affect
12633 the 0.9.6 release series:
12635 Remote buffer overflow in SSL3 protocol - an attacker could
12636 supply an oversized master key in Kerberos-enabled versions.
12639 *Ben Laurie (CHATS)*
12641 * Change the SSL kerb5 codes to match RFC 2712.
12645 * Make -nameopt work fully for req and add -reqopt switch.
12647 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12649 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12651 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12653 * Make sure tests can be performed even if the corresponding algorithms
12654 have been removed entirely. This was also the last step to make
12655 OpenSSL compilable with DJGPP under all reasonable conditions.
12657 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12659 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12660 to allow version independent disabling of normally unselected ciphers,
12661 which may be activated as a side-effect of selecting a single cipher.
12663 (E.g., cipher list string "RSA" enables ciphersuites that are left
12664 out of "ALL" because they do not provide symmetric encryption.
12665 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12667 *Lutz Jaenicke, Bodo Moeller*
12669 * Add appropriate support for separate platform-dependent build
12670 directories. The recommended way to make a platform-dependent
12671 build directory is the following (tested on Linux), maybe with
12674 # Place yourself outside of the OpenSSL source tree. In
12675 # this example, the environment variable OPENSSL_SOURCE
12676 # is assumed to contain the absolute OpenSSL source directory.
12677 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12678 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12679 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12680 mkdir -p `dirname $F`
12681 ln -s $OPENSSL_SOURCE/$F $F
12684 To be absolutely sure not to disturb the source tree, a "make clean"
12685 is a good thing. If it isn't successful, don't worry about it,
12686 it probably means the source directory is very clean.
12690 * Make sure any ENGINE control commands make local copies of string
12691 pointers passed to them whenever necessary. Otherwise it is possible
12692 the caller may have overwritten (or deallocated) the original string
12693 data when a later ENGINE operation tries to use the stored values.
12695 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12697 * Improve diagnostics in file reading and command-line digests.
12699 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12701 * Add AES modes CFB and OFB to the object database. Correct an
12702 error in AES-CFB decryption.
12706 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12707 allows existing EVP_CIPHER_CTX structures to be reused after
12708 calling `EVP_*Final()`. This behaviour is used by encryption
12709 BIOs and some applications. This has the side effect that
12710 applications must explicitly clean up cipher contexts with
12711 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12715 * Check the values of dna and dnb in bn_mul_recursive before calling
12716 bn_mul_comba (a non zero value means the a or b arrays do not contain
12717 n2 elements) and fallback to bn_mul_normal if either is not zero.
12721 * Fix escaping of non-ASCII characters when using the -subj option
12722 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12726 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12727 form for "surname", serialNumber has no short form.
12728 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12729 therefore remove "mail" short name for "internet 7".
12730 The OID for unique identifiers in X509 certificates is
12731 x500UniqueIdentifier, not uniqueIdentifier.
12732 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12736 * Add an "init" command to the ENGINE config module and auto initialize
12737 ENGINEs. Without any "init" command the ENGINE will be initialized
12738 after all ctrl commands have been executed on it. If init=1 the
12739 ENGINE is initialized at that point (ctrls before that point are run
12740 on the uninitialized ENGINE and after on the initialized one). If
12741 init=0 then the ENGINE will not be initialized at all.
12745 * Fix the 'app_verify_callback' interface so that the user-defined
12746 argument is actually passed to the callback: In the
12747 SSL_CTX_set_cert_verify_callback() prototype, the callback
12748 declaration has been changed from
12751 int (*cb)(X509_STORE_CTX *,void *);
12752 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12753 i=s->ctx->app_verify_callback(&ctx)
12754 has been changed into
12755 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12757 To update applications using SSL_CTX_set_cert_verify_callback(),
12758 a dummy argument can be added to their callback functions.
12760 *D. K. Smetters <smetters@parc.xerox.com>*
12762 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12764 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12766 * Add and OPENSSL_LOAD_CONF define which will cause
12767 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12768 This allows older applications to transparently support certain
12769 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12770 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12771 load the config file and OPENSSL_add_all_algorithms_conf() which will
12772 always load it have also been added.
12776 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12777 Adjust NIDs and EVP layer.
12779 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12781 * Config modules support in openssl utility.
12783 Most commands now load modules from the config file,
12784 though in a few (such as version) this isn't done
12785 because it couldn't be used for anything.
12787 In the case of ca and req the config file used is
12788 the same as the utility itself: that is the -config
12789 command line option can be used to specify an
12794 * Move default behaviour from OPENSSL_config(). If appname is NULL
12795 use "openssl_conf" if filename is NULL use default openssl config file.
12799 * Add an argument to OPENSSL_config() to allow the use of an alternative
12800 config section name. Add a new flag to tolerate a missing config file
12801 and move code to CONF_modules_load_file().
12805 * Support for crypto accelerator cards from Accelerated Encryption
12806 Processing, www.aep.ie. (Use engine 'aep')
12807 The support was copied from 0.9.6c [engine] and adapted/corrected
12808 to work with the new engine framework.
12810 *AEP Inc. and Richard Levitte*
12812 * Support for SureWare crypto accelerator cards from Baltimore
12813 Technologies. (Use engine 'sureware')
12814 The support was copied from 0.9.6c [engine] and adapted
12815 to work with the new engine framework.
12819 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12820 make the newer ENGINE framework commands for the CHIL engine work.
12822 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12824 * Make it possible to produce shared libraries on ReliantUNIX.
12826 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12828 * Add the configuration target debug-linux-ppro.
12829 Make 'openssl rsa' use the general key loading routines
12830 implemented in `apps.c`, and make those routines able to
12831 handle the key format FORMAT_NETSCAPE and the variant
12834 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12836 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12838 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12840 * Add -keyform to rsautl, and document -engine.
12842 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12844 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12845 BIO_R_NO_SUCH_FILE error code rather than the generic
12846 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12850 * Add new functions
12851 ERR_peek_last_error
12852 ERR_peek_last_error_line
12853 ERR_peek_last_error_line_data.
12854 These are similar to
12856 ERR_peek_error_line
12857 ERR_peek_error_line_data,
12858 but report on the latest error recorded rather than the first one
12859 still in the error queue.
12861 *Ben Laurie, Bodo Moeller*
12863 * default_algorithms option in ENGINE config module. This allows things
12865 default_algorithms = ALL
12866 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12870 * Preliminary ENGINE config module.
12874 * New experimental application configuration code.
12878 * Change the AES code to follow the same name structure as all other
12879 symmetric ciphers, and behave the same way. Move everything to
12880 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12882 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12884 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12886 *Ben Laurie and Theo de Raadt*
12888 * Add option to output public keys in req command.
12890 *Massimiliano Pala madwolf@openca.org*
12892 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12893 (up to about 10% better than before for P-192 and P-224).
12897 * New functions/macros
12899 SSL_CTX_set_msg_callback(ctx, cb)
12900 SSL_CTX_set_msg_callback_arg(ctx, arg)
12901 SSL_set_msg_callback(ssl, cb)
12902 SSL_set_msg_callback_arg(ssl, arg)
12904 to request calling a callback function
12906 void cb(int write_p, int version, int content_type,
12907 const void *buf, size_t len, SSL *ssl, void *arg)
12909 whenever a protocol message has been completely received
12910 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12911 protocol version according to which the SSL library interprets
12912 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12913 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12914 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12915 specification (change_cipher_spec(20), alert(21), handshake(22)).
12916 'buf' and 'len' point to the actual message, 'ssl' to the
12917 SSL object, and 'arg' is the application-defined value set by
12918 SSL[_CTX]_set_msg_callback_arg().
12920 'openssl s_client' and 'openssl s_server' have new '-msg' options
12921 to enable a callback that displays all protocol messages.
12925 * Change the shared library support so shared libraries are built as
12926 soon as the corresponding static library is finished, and thereby get
12927 openssl and the test programs linked against the shared library.
12928 This still only happens when the keyword "shard" has been given to
12929 the configuration scripts.
12931 NOTE: shared library support is still an experimental thing, and
12932 backward binary compatibility is still not guaranteed.
12934 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12936 * Add support for Subject Information Access extension.
12938 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12940 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12941 additional bytes when new memory had to be allocated, not just
12942 when reusing an existing buffer.
12946 * New command line and configuration option 'utf8' for the req command.
12947 This allows field values to be specified as UTF8 strings.
12951 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12952 runs for the former and machine-readable output for the latter.
12956 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12957 of the e-mail address in the DN (i.e., it will go into a certificate
12958 extension only). The new configuration file option 'email_in_dn = no'
12959 has the same effect.
12961 *Massimiliano Pala madwolf@openca.org*
12963 * Change all functions with names starting with `des_` to be starting
12964 with `DES_` instead. Add wrappers that are compatible with libdes,
12965 but are named `_ossl_old_des_*`. Finally, add macros that map the
12966 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12967 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12968 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12971 Since we provide two compatibility mappings, the user needs to
12972 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12973 compatibility is desired. The default (i.e., when that macro
12974 isn't defined) is OpenSSL 0.9.6c compatibility.
12976 There are also macros that enable and disable the support of old
12977 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12978 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12979 are defined, the default will apply: to support the old des routines.
12981 In either case, one must include openssl/des.h to get the correct
12982 definitions. Do not try to just include openssl/des_old.h, that
12985 NOTE: This is a major break of an old API into a new one. Software
12986 authors are encouraged to switch to the `DES_` style functions. Some
12987 time in the future, des_old.h and the libdes compatibility functions
12988 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12989 default), and then completely removed.
12993 * Test for certificates which contain unsupported critical extensions.
12994 If such a certificate is found during a verify operation it is
12995 rejected by default: this behaviour can be overridden by either
12996 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12997 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12998 X509_supported_extension() has also been added which returns 1 if a
12999 particular extension is supported.
13003 * Modify the behaviour of EVP cipher functions in similar way to digests
13004 to retain compatibility with existing code.
13008 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13009 compatibility with existing code. In particular the 'ctx' parameter does
13010 not have to be to be initialized before the call to EVP_DigestInit() and
13011 it is tidied up after a call to EVP_DigestFinal(). New function
13012 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13013 EVP_MD_CTX_copy() changed to not require the destination to be
13014 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13015 requires the destination to be valid.
13017 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13018 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13022 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13023 so that complete 'Handshake' protocol structures are kept in memory
13024 instead of overwriting 'msg_type' and 'length' with 'body' data.
13028 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13030 *Massimo Santin via Richard Levitte*
13032 * Major restructuring to the underlying ENGINE code. This includes
13033 reduction of linker bloat, separation of pure "ENGINE" manipulation
13034 (initialisation, etc) from functionality dealing with implementations
13035 of specific crypto interfaces. This change also introduces integrated
13036 support for symmetric ciphers and digest implementations - so ENGINEs
13037 can now accelerate these by providing EVP_CIPHER and EVP_MD
13038 implementations of their own. This is detailed in
13039 [crypto/engine/README.md](crypto/engine/README.md)
13040 as it couldn't be adequately described here. However, there are a few
13041 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13042 were changed in the original introduction of ENGINE code have now
13043 reverted back - the hooking from this code to ENGINE is now a good
13044 deal more passive and at run-time, operations deal directly with
13045 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13046 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13047 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13048 they were not being used by the framework as there is no concept of a
13049 BIGNUM_METHOD and they could not be generalised to the new
13050 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13051 ENGINE_cpy() has been removed as it cannot be consistently defined in
13056 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13060 * Change mkdef.pl to sort symbols that get the same entry number,
13061 and make sure the automatically generated functions `ERR_load_*`
13062 become part of libeay.num as well.
13066 * New function SSL_renegotiate_pending(). This returns true once
13067 renegotiation has been requested (either SSL_renegotiate() call
13068 or HelloRequest/ClientHello received from the peer) and becomes
13069 false once a handshake has been completed.
13070 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13071 sends a HelloRequest, but does not ensure that a handshake takes
13072 place. SSL_renegotiate_pending() is useful for checking if the
13073 client has followed the request.)
13077 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13078 By default, clients may request session resumption even during
13079 renegotiation (if session ID contexts permit); with this option,
13080 session resumption is possible only in the first handshake.
13082 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13083 more bits available for options that should not be part of
13084 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13088 * Add some demos for certificate and certificate request creation.
13092 * Make maximum certificate chain size accepted from the peer application
13093 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13094 "Douglas E. Engert" <deengert@anl.gov>.
13098 * Add support for shared libraries for Unixware-7
13099 (Boyd Lynn Gerber <gerberb@zenez.com>).
13103 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13104 be done prior to destruction. Use this to unload error strings from
13105 ENGINEs that load their own error strings. NB: This adds two new API
13106 functions to "get" and "set" this destroy handler in an ENGINE.
13110 * Alter all existing ENGINE implementations (except "openssl" and
13111 "openbsd") to dynamically instantiate their own error strings. This
13112 makes them more flexible to be built both as statically-linked ENGINEs
13113 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13114 Also, add stub code to each that makes building them as self-contained
13115 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13119 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13120 implementations into applications that are completely implemented in
13121 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13122 commands that can be used to configure what shared-library to load and
13123 to control aspects of the way it is handled. Also, made an update to
13124 the [README-Engine.md](README-Engine.md) file
13125 that brings its information up-to-date and
13126 provides some information and instructions on the "dynamic" ENGINE
13127 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13131 * Make it possible to unload ranges of ERR strings with a new
13132 "ERR_unload_strings" function.
13136 * Add a copy() function to EVP_MD.
13140 * Make EVP_MD routines take a context pointer instead of just the
13141 md_data void pointer.
13145 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13146 that the digest can only process a single chunk of data
13147 (typically because it is provided by a piece of
13148 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13149 is only going to provide a single chunk of data, and hence the
13150 framework needn't accumulate the data for oneshot drivers.
13154 * As with "ERR", make it possible to replace the underlying "ex_data"
13155 functions. This change also alters the storage and management of global
13156 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13157 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13158 index counters. The API functions that use this state have been changed
13159 to take a "class_index" rather than pointers to the class's local STACK
13160 and counter, and there is now an API function to dynamically create new
13161 classes. This centralisation allows us to (a) plug a lot of the
13162 thread-safety problems that existed, and (b) makes it possible to clean
13163 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13164 such data would previously have always leaked in application code and
13165 workarounds were in place to make the memory debugging turn a blind eye
13166 to it. Application code that doesn't use this new function will still
13167 leak as before, but their memory debugging output will announce it now
13168 rather than letting it slide.
13170 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13171 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13172 has a return value to indicate success or failure.
13176 * Make it possible to replace the underlying "ERR" functions such that the
13177 global state (2 LHASH tables and 2 locks) is only used by the "default"
13178 implementation. This change also adds two functions to "get" and "set"
13179 the implementation prior to it being automatically set the first time
13180 any other ERR function takes place. Ie. an application can call "get",
13181 pass the return value to a module it has just loaded, and that module
13182 can call its own "set" function using that value. This means the
13183 module's "ERR" operations will use (and modify) the error state in the
13184 application and not in its own statically linked copy of OpenSSL code.
13188 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13189 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13190 the operation, and provides a more encapsulated way for external code
13191 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13192 to use these functions rather than manually incrementing the counts.
13194 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13198 * Add EVP test program.
13202 * Add symmetric cipher support to ENGINE. Expect the API to change!
13206 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13207 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13208 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13209 These allow a CRL to be built without having to access X509_CRL fields
13210 directly. Modify 'ca' application to use new functions.
13214 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13215 bug workarounds. Rollback attack detection is a security feature.
13216 The problem will only arise on OpenSSL servers when TLSv1 is not
13217 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13218 Software authors not wanting to support TLSv1 will have special reasons
13219 for their choice and can explicitly enable this option.
13221 *Bodo Moeller, Lutz Jaenicke*
13223 * Rationalise EVP so it can be extended: don't include a union of
13224 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13225 (similar to those existing for EVP_CIPHER_CTX).
13230 EVP_MD_CTX_init(&md); /* new function call */
13231 EVP_DigestInit(&md, EVP_sha1());
13232 EVP_DigestUpdate(&md, in, len);
13233 EVP_DigestFinal(&md, out, NULL);
13234 EVP_MD_CTX_cleanup(&md); /* new function call */
13238 * Make DES key schedule conform to the usual scheme, as well as
13239 correcting its structure. This means that calls to DES functions
13240 now have to pass a pointer to a des_key_schedule instead of a
13241 plain des_key_schedule (which was actually always a pointer
13244 des_key_schedule ks;
13246 des_set_key_checked(..., &ks);
13247 des_ncbc_encrypt(..., &ks, ...);
13249 (Note that a later change renames 'des_...' into 'DES_...'.)
13253 * Initial reduction of linker bloat: the use of some functions, such as
13254 PEM causes large amounts of unused functions to be linked in due to
13255 poor organisation. For example pem_all.c contains every PEM function
13256 which has a knock on effect of linking in large amounts of (unused)
13257 ASN1 code. Grouping together similar functions and splitting unrelated
13258 functions prevents this.
13262 * Cleanup of EVP macros.
13266 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13267 correct `_ecb suffix`.
13271 * Add initial OCSP responder support to ocsp application. The
13272 revocation information is handled using the text based index
13273 use by the ca application. The responder can either handle
13274 requests generated internally, supplied in files (for example
13275 via a CGI script) or using an internal minimal server.
13279 * Add configuration choices to get zlib compression for TLS.
13283 * Changes to Kerberos SSL for RFC 2712 compliance:
13284 1. Implemented real KerberosWrapper, instead of just using
13285 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13286 2. Implemented optional authenticator field of KerberosWrapper.
13288 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13289 and authenticator structs; see crypto/krb5/.
13291 Generalized Kerberos calls to support multiple Kerberos libraries.
13292 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13293 via Richard Levitte*
13295 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13296 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13297 values for each of the key sizes rather than having just
13298 parameters (and 'speed' generating keys each time).
13302 * Speed up EVP routines.
13305 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13306 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13307 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13308 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13310 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13311 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13312 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13315 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13317 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13321 * Added the OS2-EMX target.
13323 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13325 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13326 New functions to support `NCONF` routines in extension code.
13327 New function `CONF_set_nconf()`
13328 to allow functions which take an `NCONF` to also handle the old `LHASH`
13329 structure: this means that the old `CONF` compatible routines can be
13330 retained (in particular w.rt. extensions) without having to duplicate the
13331 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13335 * Enhance the general user interface with mechanisms for inner control
13336 and with possibilities to have yes/no kind of prompts.
13340 * Change all calls to low-level digest routines in the library and
13341 applications to use EVP. Add missing calls to HMAC_cleanup() and
13342 don't assume HMAC_CTX can be copied using memcpy().
13344 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13346 * Add the possibility to control engines through control names but with
13347 arbitrary arguments instead of just a string.
13348 Change the key loaders to take a UI_METHOD instead of a callback
13349 function pointer. NOTE: this breaks binary compatibility with earlier
13350 versions of OpenSSL [engine].
13351 Adapt the nCipher code for these new conditions and add a card insertion
13356 * Enhance the general user interface with mechanisms to better support
13357 dialog box interfaces, application-defined prompts, the possibility
13358 to use defaults (for example default passwords from somewhere else)
13359 and interrupts/cancellations.
13363 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13364 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13368 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13369 tidy up some unnecessarily weird code in 'sk_new()').
13371 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13373 * Change the key loading routines for ENGINEs to use the same kind
13374 callback (pem_password_cb) as all other routines that need this
13379 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13380 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13381 than this minimum value is recommended.
13385 * New random seeder for OpenVMS, using the system process statistics
13386 that are easily reachable.
13390 * Windows apparently can't transparently handle global
13391 variables defined in DLLs. Initialisations such as:
13393 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13395 won't compile. This is used by the any applications that need to
13396 declare their own ASN1 modules. This was fixed by adding the option
13397 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13398 needed for static libraries under Win32.
13402 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13403 setting of purpose and trust fields. New X509_STORE trust and
13404 purpose functions and tidy up setting in other SSL functions.
13408 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13409 structure. These are inherited by X509_STORE_CTX when it is
13410 initialised. This allows various defaults to be set in the
13411 X509_STORE structure (such as flags for CRL checking and custom
13412 purpose or trust settings) for functions which only use X509_STORE_CTX
13413 internally such as S/MIME.
13415 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13416 trust settings if they are not set in X509_STORE. This allows X509_STORE
13417 purposes and trust (in S/MIME for example) to override any set by default.
13419 Add command line options for CRL checking to smime, s_client and s_server
13424 * Initial CRL based revocation checking. If the CRL checking flag(s)
13425 are set then the CRL is looked up in the X509_STORE structure and
13426 its validity and signature checked, then if the certificate is found
13427 in the CRL the verify fails with a revoked error.
13429 Various new CRL related callbacks added to X509_STORE_CTX structure.
13431 Command line options added to 'verify' application to support this.
13433 This needs some additional work, such as being able to handle multiple
13434 CRLs with different times, extension based lookup (rather than just
13435 by subject name) and ultimately more complete V2 CRL extension
13440 * Add a general user interface API (crypto/ui/). This is designed
13441 to replace things like des_read_password and friends (backward
13442 compatibility functions using this new API are provided).
13443 The purpose is to remove prompting functions from the DES code
13444 section as well as provide for prompting through dialog boxes in
13445 a window system and the like.
13449 * Add "ex_data" support to ENGINE so implementations can add state at a
13450 per-structure level rather than having to store it globally.
13454 * Make it possible for ENGINE structures to be copied when retrieved by
13455 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13456 This causes the "original" ENGINE structure to act like a template,
13457 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13458 operational state can be localised to each ENGINE structure, despite the
13459 fact they all share the same "methods". New ENGINE structures returned in
13460 this case have no functional references and the return value is the single
13461 structural reference. This matches the single structural reference returned
13462 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13467 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13468 needs to match any other type at all we need to manually clear the
13473 * Changes to the "openssl engine" utility to include;
13474 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13475 about an ENGINE's available control commands.
13476 - executing control commands from command line arguments using the
13477 '-pre' and '-post' switches. '-post' is only used if '-t' is
13478 specified and the ENGINE is successfully initialised. The syntax for
13479 the individual commands are colon-separated, for example;
13480 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13484 * New dynamic control command support for ENGINEs. ENGINEs can now
13485 declare their own commands (numbers), names (strings), descriptions,
13486 and input types for run-time discovery by calling applications. A
13487 subset of these commands are implicitly classed as "executable"
13488 depending on their input type, and only these can be invoked through
13489 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13490 can be based on user input, config files, etc). The distinction is
13491 that "executable" commands cannot return anything other than a boolean
13492 result and can only support numeric or string input, whereas some
13493 discoverable commands may only be for direct use through
13494 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13495 pointers, or other custom uses. The "executable" commands are to
13496 support parameterisations of ENGINE behaviour that can be
13497 unambiguously defined by ENGINEs and used consistently across any
13498 OpenSSL-based application. Commands have been added to all the
13499 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13500 control over shared-library paths without source code alterations.
13504 * Changed all ENGINE implementations to dynamically allocate their
13505 ENGINEs rather than declaring them statically. Apart from this being
13506 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13507 this also allows the implementations to compile without using the
13508 internal engine_int.h header.
13512 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13513 'const' value. Any code that should be able to modify a RAND_METHOD
13514 should already have non-const pointers to it (ie. they should only
13515 modify their own ones).
13519 * Made a variety of little tweaks to the ENGINE code.
13520 - "atalla" and "ubsec" string definitions were moved from header files
13521 to C code. "nuron" string definitions were placed in variables
13522 rather than hard-coded - allowing parameterisation of these values
13523 later on via ctrl() commands.
13524 - Removed unused "#if 0"'d code.
13525 - Fixed engine list iteration code so it uses ENGINE_free() to release
13526 structural references.
13527 - Constified the RAND_METHOD element of ENGINE structures.
13528 - Constified various get/set functions as appropriate and added
13529 missing functions (including a catch-all ENGINE_cpy that duplicates
13530 all ENGINE values onto a new ENGINE except reference counts/state).
13531 - Removed NULL parameter checks in get/set functions. Setting a method
13532 or function to NULL is a way of cancelling out a previously set
13533 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13534 and doesn't justify the extra error symbols and code.
13535 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13536 flags from engine_int.h to engine.h.
13537 - Changed prototypes for ENGINE handler functions (init(), finish(),
13538 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13542 * Implement binary inversion algorithm for BN_mod_inverse in addition
13543 to the algorithm using long division. The binary algorithm can be
13544 used only if the modulus is odd. On 32-bit systems, it is faster
13545 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13546 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13547 up to 450 bits. In 64-bit environments, the binary algorithm
13548 appears to be advantageous for much longer moduli; here we use it
13549 for moduli up to 2048 bits.
13553 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13554 could not support the combine flag in choice fields.
13558 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13559 extensions from a certificate request to the certificate.
13563 * Allow multiple 'certopt' and 'nameopt' options to be separated
13564 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13565 file: this allows the display of the certificate about to be
13566 signed to be customised, to allow certain fields to be included
13567 or excluded and extension details. The old system didn't display
13568 multicharacter strings properly, omitted fields not in the policy
13569 and couldn't display additional details such as extensions.
13573 * Function EC_POINTs_mul for multiple scalar multiplication
13574 of an arbitrary number of elliptic curve points
13575 \sum scalars[i]*points[i],
13576 optionally including the generator defined for the EC_GROUP:
13577 scalar*generator + \sum scalars[i]*points[i].
13579 EC_POINT_mul is a simple wrapper function for the typical case
13580 that the point list has just one item (besides the optional
13585 * First EC_METHODs for curves over GF(p):
13587 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13588 operations and provides various method functions that can also
13589 operate with faster implementations of modular arithmetic.
13591 EC_GFp_mont_method() reuses most functions that are part of
13592 EC_GFp_simple_method, but uses Montgomery arithmetic.
13594 *Bodo Moeller; point addition and point doubling
13595 implementation directly derived from source code provided by
13596 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13598 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13599 crypto/ec/ec_lib.c):
13601 Curves are EC_GROUP objects (with an optional group generator)
13602 based on EC_METHODs that are built into the library.
13604 Points are EC_POINT objects based on EC_GROUP objects.
13606 Most of the framework would be able to handle curves over arbitrary
13607 finite fields, but as there are no obvious types for fields other
13608 than GF(p), some functions are limited to that for now.
13612 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13613 that the file contains a complete HTTP response.
13617 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13618 change the def and num file printf format specifier from "%-40sXXX"
13619 to "%-39s XXX". The latter will always guarantee a space after the
13620 field while the former will cause them to run together if the field
13621 is 40 of more characters long.
13625 * Constify the cipher and digest 'method' functions and structures
13626 and modify related functions to take constant EVP_MD and EVP_CIPHER
13631 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13632 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13636 * Modify `EVP_Digest*()` routines so they now return values. Although the
13637 internal software routines can never fail additional hardware versions
13642 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13644 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13645 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13648 ERR_R_NESTED_ASN1_ERROR
13650 ERR_R_MISSING_ASN1_EOS
13651 were 4 .. 9, conflicting with
13652 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13654 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13655 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13657 Add new error code 'ERR_R_INTERNAL_ERROR'.
13661 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13666 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13667 sets the subject name for a new request or supersedes the
13668 subject name in a given request. Formats that can be parsed are
13669 'CN=Some Name, OU=myOU, C=IT'
13671 'CN=Some Name/OU=myOU/C=IT'.
13673 Add options '-batch' and '-verbose' to 'openssl req'.
13675 *Massimiliano Pala <madwolf@hackmasters.net>*
13677 * Introduce the possibility to access global variables through
13678 functions on platform were that's the best way to handle exporting
13679 global variables in shared libraries. To enable this functionality,
13680 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13681 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13682 is normally done by Configure or something similar).
13684 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13685 in the source file (foo.c) like this:
13687 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13688 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13690 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13691 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13693 OPENSSL_DECLARE_GLOBAL(int,foo);
13694 #define foo OPENSSL_GLOBAL_REF(foo)
13695 OPENSSL_DECLARE_GLOBAL(double,bar);
13696 #define bar OPENSSL_GLOBAL_REF(bar)
13698 The #defines are very important, and therefore so is including the
13699 header file everywhere where the defined globals are used.
13701 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13702 of ASN.1 items, but that structure is a bit different.
13704 The largest change is in util/mkdef.pl which has been enhanced with
13705 better and easier to understand logic to choose which symbols should
13706 go into the Windows .def files as well as a number of fixes and code
13707 cleanup (among others, algorithm keywords are now sorted
13708 lexicographically to avoid constant rewrites).
13712 * In BN_div() keep a copy of the sign of 'num' before writing the
13713 result to 'rm' because if rm==num the value will be overwritten
13714 and produce the wrong result if 'num' is negative: this caused
13715 problems with BN_mod() and BN_nnmod().
13719 * Function OCSP_request_verify(). This checks the signature on an
13720 OCSP request and verifies the signer certificate. The signer
13721 certificate is just checked for a generic purpose and OCSP request
13726 * Add OCSP_check_validity() function to check the validity of OCSP
13727 responses. OCSP responses are prepared in real time and may only
13728 be a few seconds old. Simply checking that the current time lies
13729 between thisUpdate and nextUpdate max reject otherwise valid responses
13730 caused by either OCSP responder or client clock inaccuracy. Instead
13731 we allow thisUpdate and nextUpdate to fall within a certain period of
13732 the current time. The age of the response can also optionally be
13733 checked. Two new options -validity_period and -status_age added to
13738 * If signature or public key algorithm is unrecognized print out its
13739 OID rather that just UNKNOWN.
13743 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13744 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13745 ID to be generated from the issuer certificate alone which can then be
13746 passed to OCSP_id_issuer_cmp().
13750 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13751 ASN1 modules to export functions returning ASN1_ITEM pointers
13752 instead of the ASN1_ITEM structures themselves. This adds several
13753 new macros which allow the underlying ASN1 function/structure to
13754 be accessed transparently. As a result code should not use ASN1_ITEM
13755 references directly (such as &X509_it) but instead use the relevant
13756 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13757 use of the new ASN1 code on platforms where exporting structures
13758 is problematical (for example in shared libraries) but exporting
13759 functions returning pointers to structures is not.
13763 * Add support for overriding the generation of SSL/TLS session IDs.
13764 These callbacks can be registered either in an SSL_CTX or per SSL.
13765 The purpose of this is to allow applications to control, if they wish,
13766 the arbitrary values chosen for use as session IDs, particularly as it
13767 can be useful for session caching in multiple-server environments. A
13768 command-line switch for testing this (and any client code that wishes
13769 to use such a feature) has been added to "s_server".
13771 *Geoff Thorpe, Lutz Jaenicke*
13773 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13774 of the form `#if defined(...) || defined(...) || ...` and
13775 `#if !defined(...) && !defined(...) && ...`. This also avoids
13776 the growing number of special cases it was previously handling.
13780 * Make all configuration macros available for application by making
13781 sure they are available in opensslconf.h, by giving them names starting
13782 with `OPENSSL_` to avoid conflicts with other packages and by making
13783 sure e_os2.h will cover all platform-specific cases together with
13785 Additionally, it is now possible to define configuration/platform-
13786 specific names (called "system identities"). In the C code, these
13787 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13788 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13789 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13794 * New option -set_serial to 'req' and 'x509' this allows the serial
13795 number to use to be specified on the command line. Previously self
13796 signed certificates were hard coded with serial number 0 and the
13797 CA options of 'x509' had to use a serial number in a file which was
13802 * New options to 'ca' utility to support V2 CRL entry extensions.
13803 Currently CRL reason, invalidity date and hold instruction are
13804 supported. Add new CRL extensions to V3 code and some new objects.
13808 * New function EVP_CIPHER_CTX_set_padding() this is used to
13809 disable standard block padding (aka PKCS#5 padding) in the EVP
13810 API, which was previously mandatory. This means that the data is
13811 not padded in any way and so the total length much be a multiple
13812 of the block size, otherwise an error occurs.
13816 * Initial (incomplete) OCSP SSL support.
13820 * New function OCSP_parse_url(). This splits up a URL into its host,
13821 port and path components: primarily to parse OCSP URLs. New -url
13822 option to ocsp utility.
13826 * New nonce behavior. The return value of OCSP_check_nonce() now
13827 reflects the various checks performed. Applications can decide
13828 whether to tolerate certain situations such as an absent nonce
13829 in a response when one was present in a request: the ocsp application
13830 just prints out a warning. New function OCSP_add1_basic_nonce()
13831 this is to allow responders to include a nonce in a response even if
13832 the request is nonce-less.
13836 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13837 skipped when using openssl x509 multiple times on a single input file,
13838 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13842 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13843 set string type: to handle setting ASN1_TIME structures. Fix ca
13844 utility to correctly initialize revocation date of CRLs.
13848 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13849 the clients preferred ciphersuites and rather use its own preferences.
13850 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13851 Internet Explorer by ensuring unchanged hash method during stepup.
13852 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13856 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13857 to aes and add a new 'exist' option to print out symbols that don't
13862 * Additional options to ocsp utility to allow flags to be set and
13863 additional certificates supplied.
13867 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13868 OCSP client a number of certificate to only verify the response
13873 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13874 handle the new API. Currently only ECB, CBC modes supported. Add new
13877 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13878 Encryption Standard (AES) Ciphersuites for Transport Layer
13879 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13880 not enabled by default and were not part of the "ALL" ciphersuite
13881 alias because they were not yet official; they could be
13882 explicitly requested by specifying the "AESdraft" ciphersuite
13883 group alias. In the final release of OpenSSL 0.9.7, the group
13884 alias is called "AES" and is part of "ALL".)
13886 *Ben Laurie, Steve Henson, Bodo Moeller*
13888 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13889 request to response.
13893 * Functions for OCSP responders. OCSP_request_onereq_count(),
13894 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13895 extract information from a certificate request. OCSP_response_create()
13896 creates a response and optionally adds a basic response structure.
13897 OCSP_basic_add1_status() adds a complete single response to a basic
13898 response and returns the OCSP_SINGLERESP structure just added (to allow
13899 extensions to be included for example). OCSP_basic_add1_cert() adds a
13900 certificate to a basic response and OCSP_basic_sign() signs a basic
13901 response with various flags. New helper functions ASN1_TIME_check()
13902 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13903 (converts ASN1_TIME to GeneralizedTime).
13907 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13908 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13909 structure from a certificate. X509_pubkey_digest() digests the public_key
13910 contents: this is used in various key identifiers.
13914 * Make sk_sort() tolerate a NULL argument.
13916 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13918 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13919 passed by the function are trusted implicitly. If any of them signed the
13920 response then it is assumed to be valid and is not verified.
13924 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13925 to data. This was previously part of the PKCS7 ASN1 code. This
13926 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13927 *Steve Henson, reported by Kenneth R. Robinette
13928 <support@securenetterm.com>*
13930 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13931 routines: without these tracing memory leaks is very painful.
13932 Fix leaks in PKCS12 and PKCS7 routines.
13936 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13937 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13938 effectively meant GeneralizedTime would never be used. Now it
13939 is initialised to -1 but X509_time_adj() now has to check the value
13940 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13941 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13942 *Steve Henson, reported by Kenneth R. Robinette
13943 <support@securenetterm.com>*
13945 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13946 result in a zero length in the ASN1_INTEGER structure which was
13947 not consistent with the structure when d2i_ASN1_INTEGER() was used
13948 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13949 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13950 where it did not print out a minus for negative ASN1_INTEGER.
13954 * Add summary printout to ocsp utility. The various functions which
13955 convert status values to strings have been renamed to:
13956 OCSP_response_status_str(), OCSP_cert_status_str() and
13957 OCSP_crl_reason_str() and are no longer static. New options
13958 to verify nonce values and to disable verification. OCSP response
13959 printout format cleaned up.
13963 * Add additional OCSP certificate checks. These are those specified
13964 in RFC2560. This consists of two separate checks: the CA of the
13965 certificate being checked must either be the OCSP signer certificate
13966 or the issuer of the OCSP signer certificate. In the latter case the
13967 OCSP signer certificate must contain the OCSP signing extended key
13968 usage. This check is performed by attempting to match the OCSP
13969 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13970 in the OCSP_CERTID structures of the response.
13974 * Initial OCSP certificate verification added to OCSP_basic_verify()
13975 and related routines. This uses the standard OpenSSL certificate
13976 verify routines to perform initial checks (just CA validity) and
13977 to obtain the certificate chain. Then additional checks will be
13978 performed on the chain. Currently the root CA is checked to see
13979 if it is explicitly trusted for OCSP signing. This is used to set
13980 a root CA as a global signing root: that is any certificate that
13981 chains to that CA is an acceptable OCSP signing certificate.
13985 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13986 extensions from a separate configuration file.
13987 As when reading extensions from the main configuration file,
13988 the '-extensions ...' option may be used for specifying the
13991 *Massimiliano Pala <madwolf@comune.modena.it>*
13993 * New OCSP utility. Allows OCSP requests to be generated or
13994 read. The request can be sent to a responder and the output
13995 parsed, outputted or printed in text form. Not complete yet:
13996 still needs to check the OCSP response validity.
14000 * New subcommands for 'openssl ca':
14001 `openssl ca -status <serial>` prints the status of the cert with
14002 the given serial number (according to the index file).
14003 `openssl ca -updatedb` updates the expiry status of certificates
14006 *Massimiliano Pala <madwolf@comune.modena.it>*
14008 * New '-newreq-nodes' command option to CA.pl. This is like
14009 '-newreq', but calls 'openssl req' with the '-nodes' option
14010 so that the resulting key is not encrypted.
14012 *Damien Miller <djm@mindrot.org>*
14014 * New configuration for the GNU Hurd.
14016 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14018 * Initial code to implement OCSP basic response verify. This
14019 is currently incomplete. Currently just finds the signer's
14020 certificate and verifies the signature on the response.
14024 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14025 value of OPENSSLDIR. This is available via the new '-d' option
14026 to 'openssl version', and is also included in 'openssl version -a'.
14030 * Allowing defining memory allocation callbacks that will be given
14031 file name and line number information in additional arguments
14032 (a `const char*` and an int). The basic functionality remains, as
14033 well as the original possibility to just replace malloc(),
14034 realloc() and free() by functions that do not know about these
14035 additional arguments. To register and find out the current
14036 settings for extended allocation functions, the following
14037 functions are provided:
14039 CRYPTO_set_mem_ex_functions
14040 CRYPTO_set_locked_mem_ex_functions
14041 CRYPTO_get_mem_ex_functions
14042 CRYPTO_get_locked_mem_ex_functions
14044 These work the same way as CRYPTO_set_mem_functions and friends.
14045 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14046 extended allocation function is enabled.
14047 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14048 a conventional allocation function is enabled.
14050 *Richard Levitte, Bodo Moeller*
14052 * Finish off removing the remaining LHASH function pointer casts.
14053 There should no longer be any prototype-casting required when using
14054 the LHASH abstraction, and any casts that remain are "bugs". See
14055 the callback types and macros at the head of lhash.h for details
14056 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14060 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14061 If /dev/[u]random devices are not available or do not return enough
14062 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14064 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14065 /etc/entropy will be queried once each in this sequence, querying stops
14066 when enough entropy was collected without querying more sockets.
14070 * Change the Unix RAND_poll() variant to be able to poll several
14071 random devices, as specified by DEVRANDOM, until a sufficient amount
14072 of data has been collected. We spend at most 10 ms on each file
14073 (select timeout) and read in non-blocking mode. DEVRANDOM now
14074 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14075 (previously it was just the string "/dev/urandom"), so on typical
14076 platforms the 10 ms delay will never occur.
14077 Also separate out the Unix variant to its own file, rand_unix.c.
14078 For VMS, there's a currently-empty rand_vms.c.
14082 * Move OCSP client related routines to ocsp_cl.c. These
14083 provide utility functions which an application needing
14084 to issue a request to an OCSP responder and analyse the
14085 response will typically need: as opposed to those which an
14086 OCSP responder itself would need which will be added later.
14088 OCSP_request_sign() signs an OCSP request with an API similar
14089 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14090 response. OCSP_response_get1_basic() extracts basic response
14091 from response. OCSP_resp_find_status(): finds and extracts status
14092 information from an OCSP_CERTID structure (which will be created
14093 when the request structure is built). These are built from lower
14094 level functions which work on OCSP_SINGLERESP structures but
14095 won't normally be used unless the application wishes to examine
14096 extensions in the OCSP response for example.
14098 Replace nonce routines with a pair of functions.
14099 OCSP_request_add1_nonce() adds a nonce value and optionally
14100 generates a random value. OCSP_check_nonce() checks the
14101 validity of the nonce in an OCSP response.
14105 * Change function OCSP_request_add() to OCSP_request_add0_id().
14106 This doesn't copy the supplied OCSP_CERTID and avoids the
14107 need to free up the newly created id. Change return type
14108 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14109 This can then be used to add extensions to the request.
14110 Deleted OCSP_request_new(), since most of its functionality
14111 is now in OCSP_REQUEST_new() (and the case insensitive name
14112 clash) apart from the ability to set the request name which
14113 will be added elsewhere.
14117 * Update OCSP API. Remove obsolete extensions argument from
14118 various functions. Extensions are now handled using the new
14119 OCSP extension code. New simple OCSP HTTP function which
14120 can be used to send requests and parse the response.
14124 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14125 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14126 uses the special reorder version of SET OF to sort the attributes
14127 and reorder them to match the encoded order. This resolves a long
14128 standing problem: a verify on a PKCS7 structure just after signing
14129 it used to fail because the attribute order did not match the
14130 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14131 it uses the received order. This is necessary to tolerate some broken
14132 software that does not order SET OF. This is handled by encoding
14133 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14134 to produce the required SET OF.
14138 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14139 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14140 files to get correct declarations of the ASN.1 item variables.
14144 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14145 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14146 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14147 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14148 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14149 ASN1_ITEM and no wrapper functions.
14153 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14154 replace the old function pointer based I/O routines. Change most of
14155 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14159 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14160 lines, recognize more "algorithms" that can be deselected, and make
14161 it complain about algorithm deselection that isn't recognised.
14165 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14166 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14167 to use new functions. Add NO_ASN1_OLD which can be set to remove
14168 some old style ASN1 functions: this can be used to determine if old
14169 code will still work when these eventually go away.
14173 * New extension functions for OCSP structures, these follow the
14174 same conventions as certificates and CRLs.
14178 * New function X509V3_add1_i2d(). This automatically encodes and
14179 adds an extension. Its behaviour can be customised with various
14180 flags to append, replace or delete. Various wrappers added for
14181 certificates and CRLs.
14185 * Fix to avoid calling the underlying ASN1 print routine when
14186 an extension cannot be parsed. Correct a typo in the
14187 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14191 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14192 entries for variables.
14196 * Add functionality to `apps/openssl.c` for detecting locking
14197 problems: As the program is single-threaded, all we have
14198 to do is register a locking callback using an array for
14199 storing which locks are currently held by the program.
14203 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14204 SSL_get_ex_data_X509_STORE_idx(), which is used in
14205 ssl_verify_cert_chain() and thus can be called at any time
14206 during TLS/SSL handshakes so that thread-safety is essential.
14207 Unfortunately, the ex_data design is not at all suited
14208 for multi-threaded use, so it probably should be abolished.
14212 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14214 *Broadcom, tweaked and integrated by Geoff Thorpe*
14216 * Move common extension printing code to new function
14217 X509V3_print_extensions(). Reorganise OCSP print routines and
14218 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14222 * New function X509_signature_print() to remove duplication in some
14227 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14228 set (this was treated exactly the same as SET OF previously). This
14229 is used to reorder the STACK representing the structure to match the
14230 encoding. This will be used to get round a problem where a PKCS7
14231 structure which was signed could not be verified because the STACK
14232 order did not reflect the encoded order.
14236 * Reimplement the OCSP ASN1 module using the new code.
14240 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14241 for its ASN1 operations. The old style function pointers still exist
14242 for now but they will eventually go away.
14246 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14247 completely replaces the old ASN1 functionality with a table driven
14248 encoder and decoder which interprets an ASN1_ITEM structure describing
14249 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14250 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14251 has also been converted to the new form.
14255 * Change BN_mod_exp_recp so that negative moduli are tolerated
14256 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14257 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14258 for negative moduli.
14262 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14263 of not touching the result's sign bit.
14267 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14272 * Changed the LHASH code to use prototypes for callbacks, and created
14273 macros to declare and implement thin (optionally static) functions
14274 that provide type-safety and avoid function pointer casting for the
14275 type-specific callbacks.
14279 * Added Kerberos Cipher Suites to be used with TLS, as written in
14281 *Veers Staats <staatsvr@asc.hpc.mil>,
14282 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14284 * Reformat the FAQ so the different questions and answers can be divided
14285 in sections depending on the subject.
14289 * Have the zlib compression code load ZLIB.DLL dynamically under
14294 * New function BN_mod_sqrt for computing square roots modulo a prime
14295 (using the probabilistic Tonelli-Shanks algorithm unless
14296 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14297 be handled deterministically).
14299 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14301 * Make BN_mod_inverse faster by explicitly handling small quotients
14302 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14303 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14307 * New function BN_kronecker.
14311 * Fix BN_gcd so that it works on negative inputs; the result is
14312 positive unless both parameters are zero.
14313 Previously something reasonably close to an infinite loop was
14314 possible because numbers could be growing instead of shrinking
14315 in the implementation of Euclid's algorithm.
14319 * Fix BN_is_word() and BN_is_one() macros to take into account the
14320 sign of the number in question.
14322 Fix BN_is_word(a,w) to work correctly for w == 0.
14324 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14325 because its test if the absolute value of 'a' equals 'w'.
14326 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14327 it exists mostly for use in the implementations of BN_is_zero(),
14328 BN_is_one(), and BN_is_word().
14332 * New function BN_swap.
14336 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14337 the exponentiation functions are more likely to produce reasonable
14338 results on negative inputs.
14342 * Change BN_mod_mul so that the result is always non-negative.
14343 Previously, it could be negative if one of the factors was negative;
14344 I don't think anyone really wanted that behaviour.
14348 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14349 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14350 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14351 and add new functions:
14360 BN_mod_lshift1_quick
14362 BN_mod_lshift_quick
14364 These functions always generate non-negative results.
14366 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14367 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14369 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14370 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14371 be reduced modulo `m`.
14373 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14376 The following entry accidentally appeared in the CHANGES file
14377 distributed with OpenSSL 0.9.7. The modifications described in
14378 it do *not* apply to OpenSSL 0.9.7.
14380 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14381 was actually never needed) and in BN_mul(). The removal in BN_mul()
14382 required a small change in bn_mul_part_recursive() and the addition
14383 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14384 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14385 bn_sub_words() and bn_add_words() except they take arrays with
14391 * In 'openssl passwd', verify passwords read from the terminal
14392 unless the '-salt' option is used (which usually means that
14393 verification would just waste user's time since the resulting
14394 hash is going to be compared with some given password hash)
14395 or the new '-noverify' option is used.
14397 This is an incompatible change, but it does not affect
14398 non-interactive use of 'openssl passwd' (passwords on the command
14399 line, '-stdin' option, '-in ...' option) and thus should not
14400 cause any problems.
14404 * Remove all references to RSAref, since there's no more need for it.
14408 * Make DSO load along a path given through an environment variable
14409 (SHLIB_PATH) with shl_load().
14413 * Constify the ENGINE code as a result of BIGNUM constification.
14414 Also constify the RSA code and most things related to it. In a
14415 few places, most notable in the depth of the ASN.1 code, ugly
14416 casts back to non-const were required (to be solved at a later
14421 * Make it so the openssl application has all engines loaded by default.
14425 * Constify the BIGNUM routines a little more.
14429 * Add the following functions:
14431 ENGINE_load_cswift()
14433 ENGINE_load_atalla()
14434 ENGINE_load_nuron()
14435 ENGINE_load_builtin_engines()
14437 That way, an application can itself choose if external engines that
14438 are built-in in OpenSSL shall ever be used or not. The benefit is
14439 that applications won't have to be linked with libdl or other dso
14440 libraries unless it's really needed.
14442 Changed 'openssl engine' to load all engines on demand.
14443 Changed the engine header files to avoid the duplication of some
14444 declarations (they differed!).
14448 * 'openssl engine' can now list capabilities.
14452 * Better error reporting in 'openssl engine'.
14456 * Never call load_dh_param(NULL) in s_server.
14460 * Add engine application. It can currently list engines by name and
14461 identity, and test if they are actually available.
14465 * Improve RPM specification file by forcing symbolic linking and making
14466 sure the installed documentation is also owned by root.root.
14468 *Damien Miller <djm@mindrot.org>*
14470 * Give the OpenSSL applications more possibilities to make use of
14471 keys (public as well as private) handled by engines.
14475 * Add OCSP code that comes from CertCo.
14479 * Add VMS support for the Rijndael code.
14483 * Added untested support for Nuron crypto accelerator.
14487 * Add support for external cryptographic devices. This code was
14488 previously distributed separately as the "engine" branch.
14490 *Geoff Thorpe, Richard Levitte*
14492 * Rework the filename-translation in the DSO code. It is now possible to
14493 have far greater control over how a "name" is turned into a filename
14494 depending on the operating environment and any oddities about the
14495 different shared library filenames on each system.
14499 * Support threads on FreeBSD-elf in Configure.
14503 * Fix for SHA1 assembly problem with MASM: it produces
14504 warnings about corrupt line number information when assembling
14505 with debugging information. This is caused by the overlapping
14508 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14511 NCONF_get_number() has no error checking at all. As a replacement,
14512 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14513 promoted strongly. The old NCONF_get_number is kept around for
14514 binary backward compatibility.
14515 Make it possible for methods to load from something other than a BIO,
14516 by providing a function pointer that is given a name instead of a BIO.
14517 For example, this could be used to load configuration data from an
14522 * Fix for non blocking accept BIOs. Added new I/O special reason
14523 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14524 with non blocking I/O was not possible because no retry code was
14525 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14530 * Added the beginnings of Rijndael support.
14534 * Fix for bug in DirectoryString mask setting. Add support for
14535 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14536 to allow certificate printing to more controllable, additional
14537 'certopt' option to 'x509' to allow new printing options to be
14542 * Clean old EAY MD5 hack from e_os.h.
14546 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14548 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14549 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14551 *Joe Orton, Steve Henson*
14553 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14555 * Fix additional bug revealed by the NISCC test suite:
14557 Stop bug triggering large recursion when presented with
14558 certain ASN.1 tags ([CVE-2003-0851])
14562 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14564 * Fix various bugs revealed by running the NISCC test suite:
14566 Stop out of bounds reads in the ASN1 code when presented with
14567 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14569 If verify callback ignores invalid public key errors don't try to check
14570 certificate signature with the NULL public key.
14574 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14575 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14580 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14581 extra data after the compression methods not only for TLS 1.0
14582 but also for SSL 3.0 (as required by the specification).
14584 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14586 * Change X509_certificate_type() to mark the key as exported/exportable
14587 when it's 512 *bits* long, not 512 bytes.
14591 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14593 * Countermeasure against the Klima-Pokorny-Rosa extension of
14594 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14595 a protocol version number mismatch like a decryption error
14596 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14600 * Turn on RSA blinding by default in the default implementation
14601 to avoid a timing attack. Applications that don't want it can call
14602 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14603 They would be ill-advised to do so in most cases.
14605 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14607 * Change RSA blinding code so that it works when the PRNG is not
14608 seeded (in this case, the secret RSA exponent is abused as
14609 an unpredictable seed -- if it is not unpredictable, there
14610 is no point in blinding anyway). Make RSA blinding thread-safe
14611 by remembering the creator's thread ID in rsa->blinding and
14612 having all other threads use local one-time blinding factors
14613 (this requires more computation than sharing rsa->blinding, but
14614 avoids excessive locking; and if an RSA object is not shared
14615 between threads, blinding will still be very fast).
14619 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14621 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14622 via timing by performing a MAC computation even if incorrect
14623 block cipher padding has been found. This is a countermeasure
14624 against active attacks where the attacker has to distinguish
14625 between bad padding and a MAC verification error. ([CVE-2003-0078])
14627 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14628 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14629 Martin Vuagnoux (EPFL, Ilion)*
14631 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14633 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14634 memory from its contents. This is done with a counter that will
14635 place alternating values in each byte. This can be used to solve
14636 two issues: 1) the removal of calls to memset() by highly optimizing
14637 compilers, and 2) cleansing with other values than 0, since those can
14638 be read through on certain media, for example a swap space on disk.
14642 * Bugfix: client side session caching did not work with external caching,
14643 because the session->cipher setting was not restored when reloading
14644 from the external cache. This problem was masked, when
14645 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14646 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14650 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14651 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14653 *Zeev Lieber <zeev-l@yahoo.com>*
14655 * Undo an undocumented change introduced in 0.9.6e which caused
14656 repeated calls to OpenSSL_add_all_ciphers() and
14657 OpenSSL_add_all_digests() to be ignored, even after calling
14662 * Change the default configuration reader to deal with last line not
14663 being properly terminated.
14667 * Change X509_NAME_cmp() so it applies the special rules on handling
14668 DN values that are of type PrintableString, as well as RDNs of type
14669 emailAddress where the value has the type ia5String.
14671 *stefank@valicert.com via Richard Levitte*
14673 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14674 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14675 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14676 the bitwise-OR of the two for use by the majority of applications
14677 wanting this behaviour, and update the docs. The documented
14678 behaviour and actual behaviour were inconsistent and had been
14679 changing anyway, so this is more a bug-fix than a behavioural
14682 *Geoff Thorpe, diagnosed by Nadav Har'El*
14684 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14685 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14689 * Fix initialization code race conditions in
14690 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14691 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14692 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14693 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14694 ssl2_get_cipher_by_char(),
14695 ssl3_get_cipher_by_char().
14697 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14699 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14700 the cached sessions are flushed, as the remove_cb() might use ex_data
14701 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14702 (see [openssl.org #212]).
14704 *Geoff Thorpe, Lutz Jaenicke*
14706 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14707 length, instead of the encoding length to d2i_ASN1_OBJECT.
14711 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14713 * [In 0.9.6g-engine release:]
14714 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14716 *Lynn Gazis <lgazis@rainbow.com>*
14718 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14720 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14721 and get fix the header length calculation.
14722 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14723 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14725 * Use proper error handling instead of 'assertions' in buffer
14726 overflow checks added in 0.9.6e. This prevents DoS (the
14727 assertions could call abort()).
14729 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14731 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14733 * Add various sanity checks to asn1_get_length() to reject
14734 the ASN1 length bytes if they exceed sizeof(long), will appear
14735 negative or the content length exceeds the length of the
14738 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14740 * Fix cipher selection routines: ciphers without encryption had no flags
14741 for the cipher strength set and where therefore not handled correctly
14742 by the selection routines (PR #130).
14746 * Fix EVP_dsa_sha macro.
14751 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14752 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14753 that was added in OpenSSL 0.9.6d.
14755 As the countermeasure turned out to be incompatible with some
14756 broken SSL implementations, the new option is part of SSL_OP_ALL.
14757 SSL_OP_ALL is usually employed when compatibility with weird SSL
14758 implementations is desired (e.g. '-bugs' option to 's_client' and
14759 's_server'), so the new option is automatically set in many
14764 * Changes in security patch:
14766 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14767 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14768 Air Force Materiel Command, USAF, under agreement number
14771 * Add various sanity checks to asn1_get_length() to reject
14772 the ASN1 length bytes if they exceed sizeof(long), will appear
14773 negative or the content length exceeds the length of the
14774 supplied buffer. ([CVE-2002-0659])
14776 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14778 * Assertions for various potential buffer overflows, not known to
14779 happen in practice.
14781 *Ben Laurie (CHATS)*
14783 * Various temporary buffers to hold ASCII versions of integers were
14784 too small for 64 bit platforms. ([CVE-2002-0655])
14785 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14787 * Remote buffer overflow in SSL3 protocol - an attacker could
14788 supply an oversized session ID to a client. ([CVE-2002-0656])
14790 *Ben Laurie (CHATS)*
14792 * Remote buffer overflow in SSL2 protocol - an attacker could
14793 supply an oversized client master key. ([CVE-2002-0656])
14795 *Ben Laurie (CHATS)*
14797 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14799 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14800 encoded as NULL) with id-dsa-with-sha1.
14802 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14804 * Check various `X509_...()` return values in `apps/req.c`.
14806 *Nils Larsch <nla@trustcenter.de>*
14808 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14809 an end-of-file condition would erroneously be flagged, when the CRLF
14810 was just at the end of a processed block. The bug was discovered when
14811 processing data through a buffering memory BIO handing the data to a
14812 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14813 <ptsekov@syntrex.com> and Nedelcho Stanev.
14817 * Implement a countermeasure against a vulnerability recently found
14818 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14819 before application data chunks to avoid the use of known IVs
14820 with data potentially chosen by the attacker.
14824 * Fix length checks in ssl3_get_client_hello().
14828 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14829 to prevent ssl3_read_internal() from incorrectly assuming that
14830 ssl3_read_bytes() found application data while handshake
14831 processing was enabled when in fact s->s3->in_read_app_data was
14832 merely automatically cleared during the initial handshake.
14834 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14836 * Fix object definitions for Private and Enterprise: they were not
14837 recognized in their shortname (=lowercase) representation. Extend
14838 obj_dat.pl to issue an error when using undefined keywords instead
14839 of silently ignoring the problem (Svenning Sorensen
14840 <sss@sss.dnsalias.net>).
14844 * Fix DH_generate_parameters() so that it works for 'non-standard'
14845 generators, i.e. generators other than 2 and 5. (Previously, the
14846 code did not properly initialise the 'add' and 'rem' values to
14847 BN_generate_prime().)
14849 In the new general case, we do not insist that 'generator' is
14850 actually a primitive root: This requirement is rather pointless;
14851 a generator of the order-q subgroup is just as good, if not
14856 * Map new X509 verification errors to alerts. Discovered and submitted by
14857 Tom Wu <tom@arcot.com>.
14861 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14862 returning non-zero before the data has been completely received
14863 when using non-blocking I/O.
14865 *Bodo Moeller; problem pointed out by John Hughes*
14867 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14869 *Ben Laurie, Lutz Jaenicke*
14871 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14872 Yoram Zahavi <YoramZ@gilian.com>).
14876 * Add information about CygWin 1.3 and on, and preserve proper
14877 configuration for the versions before that.
14879 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14881 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14882 check whether we deal with a copy of a session and do not delete from
14883 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14884 <izhar@checkpoint.com>.
14888 * Do not store session data into the internal session cache, if it
14889 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14890 flag is set). Proposed by Aslam <aslam@funk.com>.
14894 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14899 * [In 0.9.6d-engine release:]
14900 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14902 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14904 * Add the configuration target linux-s390x.
14906 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14908 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14909 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14910 variable as an indication that a ClientHello message has been
14911 received. As the flag value will be lost between multiple
14912 invocations of ssl3_accept when using non-blocking I/O, the
14913 function may not be aware that a handshake has actually taken
14914 place, thus preventing a new session from being added to the
14917 To avoid this problem, we now set s->new_session to 2 instead of
14918 using a local variable.
14920 *Lutz Jaenicke, Bodo Moeller*
14922 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14923 if the SSL_R_LENGTH_MISMATCH error is detected.
14925 *Geoff Thorpe, Bodo Moeller*
14927 * New 'shared_ldflag' column in Configure platform table.
14931 * Fix EVP_CIPHER_mode macro.
14933 *"Dan S. Camper" <dan@bti.net>*
14935 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14936 type, we must throw them away by setting rr->length to 0.
14938 *D P Chang <dpc@qualys.com>*
14940 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14942 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14943 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14944 worked incorrectly for those cases where range = `10..._2` and
14945 `3*range` is two bits longer than range.)
14949 * Only add signing time to PKCS7 structures if it is not already
14954 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14955 OBJ_ld_ce should be OBJ_id_ce.
14956 Also some ip-pda OIDs in crypto/objects/objects.txt were
14957 incorrect (cf. RFC 3039).
14959 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14961 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14962 returns early because it has nothing to do.
14964 *Andy Schneider <andy.schneider@bjss.co.uk>*
14966 * [In 0.9.6c-engine release:]
14967 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14969 *Andy Schneider <andy.schneider@bjss.co.uk>*
14971 * [In 0.9.6c-engine release:]
14972 Add support for Cryptographic Appliance's keyserver technology.
14973 (Use engine 'keyclient')
14975 *Cryptographic Appliances and Geoff Thorpe*
14977 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14978 is called via tools/c89.sh because arguments have to be
14979 rearranged (all '-L' options must appear before the first object
14982 *Richard Shapiro <rshapiro@abinitio.com>*
14984 * [In 0.9.6c-engine release:]
14985 Add support for Broadcom crypto accelerator cards, backported
14988 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14990 * [In 0.9.6c-engine release:]
14991 Add support for SureWare crypto accelerator cards from
14992 Baltimore Technologies. (Use engine 'sureware')
14994 *Baltimore Technologies and Mark Cox*
14996 * [In 0.9.6c-engine release:]
14997 Add support for crypto accelerator cards from Accelerated
14998 Encryption Processing, www.aep.ie. (Use engine 'aep')
15000 *AEP Inc. and Mark Cox*
15002 * Add a configuration entry for gcc on UnixWare.
15004 *Gary Benson <gbenson@redhat.com>*
15006 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15007 messages are stored in a single piece (fixed-length part and
15008 variable-length part combined) and fix various bugs found on the way.
15012 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15013 instead. BIO_gethostbyname() does not know what timeouts are
15014 appropriate, so entries would stay in cache even when they have
15016 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15018 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15019 faced with a pathologically small ClientHello fragment that does
15020 not contain client_version: Instead of aborting with an error,
15021 simply choose the highest available protocol version (i.e.,
15022 TLS 1.0 unless it is disabled). In practice, ClientHello
15023 messages are never sent like this, but this change gives us
15024 strictly correct behaviour at least for TLS.
15028 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15029 never resets s->method to s->ctx->method when called from within
15030 one of the SSL handshake functions.
15032 *Bodo Moeller; problem pointed out by Niko Baric*
15034 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15035 (sent using the client's version number) if client_version is
15036 smaller than the protocol version in use. Also change
15037 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15038 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15039 the client will at least see that alert.
15043 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15048 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15049 client receives HelloRequest while in a handshake.
15051 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15053 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15054 should end in 'break', not 'goto end' which circumvents various
15055 cleanups done in state SSL_ST_OK. But session related stuff
15056 must be disabled for SSL_ST_OK in the case that we just sent a
15059 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15060 before just sending a HelloRequest.
15062 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15064 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15065 reveal whether illegal block cipher padding was found or a MAC
15066 verification error occurred. (Neither SSLerr() codes nor alerts
15067 are directly visible to potential attackers, but the information
15068 may leak via logfiles.)
15070 Similar changes are not required for the SSL 2.0 implementation
15071 because the number of padding bytes is sent in clear for SSL 2.0,
15072 and the extra bytes are just ignored. However ssl/s2_pkt.c
15073 failed to verify that the purported number of padding bytes is in
15078 * Add OpenUNIX-8 support including shared libraries
15079 (Boyd Lynn Gerber <gerberb@zenez.com>).
15083 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15084 'wristwatch attack' using huge encoding parameters (cf.
15085 James H. Manger's CRYPTO 2001 paper). Note that the
15086 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15087 encoding parameters and hence was not vulnerable.
15091 * BN_sqr() bug fix.
15093 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15095 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15096 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15097 followed by modular reduction.
15099 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15101 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15102 equivalent based on BN_pseudo_rand() instead of BN_rand().
15106 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15107 This function was broken, as the check for a new client hello message
15108 to handle SGC did not allow these large messages.
15109 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15113 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15117 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15118 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15122 * Rework the configuration and shared library support for Tru64 Unix.
15123 The configuration part makes use of modern compiler features and
15124 still retains old compiler behavior for those that run older versions
15125 of the OS. The shared library support part includes a variant that
15126 uses the RPATH feature, and is available through the special
15127 configuration target "alpha-cc-rpath", which will never be selected
15130 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15132 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15133 with the same message size as in ssl3_get_certificate_request().
15134 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15135 messages might inadvertently be reject as too long.
15137 *Petr Lampa <lampa@fee.vutbr.cz>*
15139 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15143 * Modified SSL library such that the verify_callback that has been set
15144 specifically for an SSL object with SSL_set_verify() is actually being
15145 used. Before the change, a verify_callback set with this function was
15146 ignored and the verify_callback() set in the SSL_CTX at the time of
15147 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15148 to allow the necessary settings.
15152 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15153 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15154 done automatically (in contradiction to the requirements of the C
15155 standard). This made problems when used from OpenSSH.
15159 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15160 dh->length and always used
15162 BN_rand_range(priv_key, dh->p).
15164 BN_rand_range() is not necessary for Diffie-Hellman, and this
15165 specific range makes Diffie-Hellman unnecessarily inefficient if
15166 dh->length (recommended exponent length) is much smaller than the
15167 length of dh->p. We could use BN_rand_range() if the order of
15168 the subgroup was stored in the DH structure, but we only have
15173 BN_rand(priv_key, l, ...)
15175 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15182 RSA_eay_public_encrypt
15183 RSA_eay_private_decrypt
15184 RSA_eay_private_encrypt (signing)
15185 RSA_eay_public_decrypt (signature verification)
15187 (default implementations for RSA_public_encrypt,
15188 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15189 always reject numbers >= n.
15193 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15194 to synchronize access to 'locking_thread'. This is necessary on
15195 systems where access to 'locking_thread' (an 'unsigned long'
15196 variable) is not atomic.
15200 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15201 *before* setting the 'crypto_lock_rand' flag. The previous code had
15202 a race condition if 0 is a valid thread ID.
15204 *Travis Vitek <vitek@roguewave.com>*
15206 * Add support for shared libraries under Irix.
15208 *Albert Chin-A-Young <china@thewrittenword.com>*
15210 * Add configuration option to build on Linux on both big-endian and
15211 little-endian MIPS.
15213 *Ralf Baechle <ralf@uni-koblenz.de>*
15215 * Add the possibility to create shared libraries on HP-UX.
15219 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15221 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15222 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15223 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15224 PRNG state recovery was possible based on the output of
15225 one PRNG request appropriately sized to gain knowledge on
15226 'md' followed by enough consecutive 1-byte PRNG requests
15227 to traverse all of 'state'.
15229 1. When updating 'md_local' (the current thread's copy of 'md')
15230 during PRNG output generation, hash all of the previous
15231 'md_local' value, not just the half used for PRNG output.
15233 2. Make the number of bytes from 'state' included into the hash
15234 independent from the number of PRNG bytes requested.
15236 The first measure alone would be sufficient to avoid
15237 Markku-Juhani's attack. (Actually it had never occurred
15238 to me that the half of 'md_local' used for chaining was the
15239 half from which PRNG output bytes were taken -- I had always
15240 assumed that the secret half would be used.) The second
15241 measure makes sure that additional data from 'state' is never
15242 mixed into 'md_local' in small portions; this heuristically
15243 further strengthens the PRNG.
15247 * Fix crypto/bn/asm/mips3.s.
15251 * When only the key is given to "enc", the IV is undefined. Print out
15252 an error message in this case.
15256 * Handle special case when X509_NAME is empty in X509 printing routines.
15260 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15261 positive and less than q.
15265 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15266 used: it isn't thread safe and the add_lock_callback should handle
15269 *Paul Rose <Paul.Rose@bridge.com>*
15271 * Verify that incoming data obeys the block size in
15272 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15278 *Ulf Möller, Bodo Möller*
15280 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15281 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15282 when fixing the server behaviour for backwards-compatible 'client
15283 hello' messages. (Note that the attack is impractical against
15284 SSL 3.0 and TLS 1.0 anyway because length and version checking
15285 means that the probability of guessing a valid ciphertext is
15286 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15289 Before 0.9.5, the countermeasure (hide the error by generating a
15290 random 'decryption result') did not work properly because
15291 ERR_clear_error() was missing, meaning that SSL_get_error() would
15292 detect the supposedly ignored error.
15294 Both problems are now fixed.
15298 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15299 (previously it was 1024).
15303 * Fix for compatibility mode trust settings: ignore trust settings
15304 unless some valid trust or reject settings are present.
15308 * Fix for blowfish EVP: its a variable length cipher.
15312 * Fix various bugs related to DSA S/MIME verification. Handle missing
15313 parameters in DSA public key structures and return an error in the
15314 DSA routines if parameters are absent.
15318 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15319 in the current directory if neither $RANDFILE nor $HOME was set.
15320 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15321 caused some confusion to Windows users who haven't defined $HOME.
15322 Thus RAND_file_name() is changed again: e_os.h can define a
15323 DEFAULT_HOME, which will be used if $HOME is not set.
15324 For Windows, we use "C:"; on other platforms, we still require
15325 environment variables.
15327 * Move 'if (!initialized) RAND_poll()' into regions protected by
15328 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15329 having multiple threads call RAND_poll() concurrently.
15333 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15334 combination of a flag and a thread ID variable.
15335 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15336 flag), *other* threads can enter ssleay_add_bytes without obeying
15337 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15338 that they do not hold after the first thread unsets add_do_not_lock).
15342 * Change bctest again: '-x' expressions are not available in all
15343 versions of 'test'.
15347 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15349 * Fix a couple of memory leaks in PKCS7_dataDecode()
15351 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15353 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15354 the default extension for executables, if any. Also, make the perl
15355 scripts that use symlink() to test if it really exists and use "cp"
15356 if it doesn't. All this made OpenSSL compilable and installable in
15361 * Fix for asn1_GetSequence() for indefinite length constructed data.
15362 If SEQUENCE is length is indefinite just set c->slen to the total
15363 amount of data available.
15365 *Steve Henson, reported by shige@FreeBSD.org*
15367 *This change does not apply to 0.9.7.*
15369 * Change bctest to avoid here-documents inside command substitution
15370 (workaround for FreeBSD /bin/sh bug).
15371 For compatibility with Ultrix, avoid shell functions (introduced
15372 in the bctest version that searches along $PATH).
15376 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15377 with des_encrypt() defined on some operating systems, like Solaris
15382 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15383 On the Importance of Eliminating Errors in Cryptographic
15384 Computations, J. Cryptology 14 (2001) 2, 101-119,
15385 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15389 * MIPS assembler BIGNUM division bug fix.
15393 * Disabled incorrect Alpha assembler code.
15397 * Fix PKCS#7 decode routines so they correctly update the length
15398 after reading an EOC for the EXPLICIT tag.
15402 *This change does not apply to 0.9.7.*
15404 * Fix bug in PKCS#12 key generation routines. This was triggered
15405 if a 3DES key was generated with a 0 initial byte. Include
15406 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15407 (but broken) behaviour.
15411 * Enhance bctest to search for a working bc along $PATH and print
15414 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15416 * Fix memory leaks in err.c: free err_data string if necessary;
15417 don't write to the wrong index in ERR_set_error_data.
15421 * Implement ssl23_peek (analogous to ssl23_read), which previously
15426 * Replace rdtsc with `_emit` statements for VC++ version 5.
15428 *Jeremy Cooper <jeremy@baymoo.org>*
15430 * Make it possible to reuse SSLv2 sessions.
15434 * In copy_email() check for >= 0 as a return value for
15435 X509_NAME_get_index_by_NID() since 0 is a valid index.
15437 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15439 * Avoid coredump with unsupported or invalid public keys by checking if
15440 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15441 PKCS7_verify() fails with non detached data.
15445 * Don't use getenv in library functions when run as setuid/setgid.
15446 New function OPENSSL_issetugid().
15450 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15451 due to incorrect handling of multi-threading:
15453 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15455 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15457 3. Count how many times MemCheck_off() has been called so that
15458 nested use can be treated correctly. This also avoids
15459 inband-signalling in the previous code (which relied on the
15460 assumption that thread ID 0 is impossible).
15464 * Add "-rand" option also to s_client and s_server.
15468 * Fix CPU detection on Irix 6.x.
15469 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15470 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15472 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15477 *This change does not apply to 0.9.7.*
15479 * Use the cached encoding of an X509_NAME structure rather than
15480 copying it. This is apparently the reason for the libsafe "errors"
15481 but the code is actually correct.
15485 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15486 Bleichenbacher's DSA attack.
15487 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15488 to be set and top=0 forces the highest bit to be set; top=-1 is new
15489 and leaves the highest bit random.
15491 *Ulf Moeller, Bodo Moeller*
15493 * In the `NCONF_...`-based implementations for `CONF_...` queries
15494 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15495 a temporary CONF structure with the data component set to NULL
15496 (which gives segmentation faults in lh_retrieve).
15497 Instead, use NULL for the CONF pointer in CONF_get_string and
15498 CONF_get_number (which may use environment variables) and directly
15499 return NULL from CONF_get_section.
15503 * Fix potential buffer overrun for EBCDIC.
15507 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15508 keyUsage if basicConstraints absent for a CA.
15512 * Make SMIME_write_PKCS7() write mail header values with a format that
15513 is more generally accepted (no spaces before the semicolon), since
15514 some programs can't parse those values properly otherwise. Also make
15515 sure BIO's that break lines after each write do not create invalid
15520 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15521 macros previously used would not encode an empty SEQUENCE OF
15522 and break the signature.
15526 *This change does not apply to 0.9.7.*
15528 * Zero the premaster secret after deriving the master secret in
15533 * Add some EVP_add_digest_alias registrations (as found in
15534 OpenSSL_add_all_digests()) to SSL_library_init()
15535 aka OpenSSL_add_ssl_algorithms(). This provides improved
15536 compatibility with peers using X.509 certificates
15537 with unconventional AlgorithmIdentifier OIDs.
15541 * Fix for Irix with NO_ASM.
15543 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15545 * ./config script fixes.
15547 *Ulf Moeller, Richard Levitte*
15549 * Fix 'openssl passwd -1'.
15553 * Change PKCS12_key_gen_asc() so it can cope with non null
15554 terminated strings whose length is passed in the passlen
15555 parameter, for example from PEM callbacks. This was done
15556 by adding an extra length parameter to asc2uni().
15558 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15560 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15561 call failed, free the DSA structure.
15565 * Fix to uni2asc() to cope with zero length Unicode strings.
15566 These are present in some PKCS#12 files.
15570 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15571 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15572 when writing a 32767 byte record.
15574 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15576 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15577 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15579 (RSA objects have a reference count access to which is protected
15580 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15581 so they are meant to be shared between threads.)
15582 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15583 "Reddie, Steven" <Steven.Reddie@ca.com>*
15585 * Fix a deadlock in CRYPTO_mem_leaks().
15589 * Use better test patterns in bntest.
15593 * rand_win.c fix for Borland C.
15597 * BN_rshift bugfix for n == 0.
15601 * Add a 'bctest' script that checks for some known 'bc' bugs
15602 so that 'make test' does not abort just because 'bc' is broken.
15606 * Store verify_result within SSL_SESSION also for client side to
15607 avoid potential security hole. (Re-used sessions on the client side
15608 always resulted in verify_result==X509_V_OK, not using the original
15609 result of the server certificate verification.)
15613 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15614 SSL3_RT_APPLICATION_DATA, return 0.
15615 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15620 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15621 releases, have been re-implemented by renaming the previous
15622 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15623 and ssl3_read_internal, respectively, and adding 'peek' parameters
15624 to them. The new ssl[23]_{read,peek} functions are calls to
15625 ssl[23]_read_internal with the 'peek' flag set appropriately.
15626 A 'peek' parameter has also been added to ssl3_read_bytes, which
15627 does the actual work for ssl3_read_internal.
15631 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15632 the method-specific "init()" handler. Also clean up ex_data after
15633 calling the method-specific "finish()" handler. Previously, this was
15634 happening the other way round.
15638 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15639 The previous value, 12, was not always sufficient for BN_mod_exp().
15643 * Make sure that shared libraries get the internal name engine with
15644 the full version number and not just 0. This should mark the
15645 shared libraries as not backward compatible. Of course, this should
15646 be changed again when we can guarantee backward binary compatibility.
15650 * Fix typo in get_cert_by_subject() in by_dir.c
15652 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15654 * Rework the system to generate shared libraries:
15656 - Make note of the expected extension for the shared libraries and
15657 if there is a need for symbolic links from for example libcrypto.so.0
15658 to libcrypto.so.0.9.7. There is extended info in Configure for
15661 - Make as few rebuilds of the shared libraries as possible.
15663 - Still avoid linking the OpenSSL programs with the shared libraries.
15665 - When installing, install the shared libraries separately from the
15670 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15672 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15673 and not in SSL_clear because the latter is also used by the
15674 accept/connect functions; previously, the settings made by
15675 SSL_set_read_ahead would be lost during the handshake.
15677 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15679 * Correct util/mkdef.pl to be selective about disabled algorithms.
15680 Previously, it would create entries for disabled algorithms no
15685 * Added several new manual pages for SSL_* function.
15689 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15691 * In ssl23_get_client_hello, generate an error message when faced
15692 with an initial SSL 3.0/TLS record that is too small to contain the
15693 first two bytes of the ClientHello message, i.e. client_version.
15694 (Note that this is a pathologic case that probably has never happened
15695 in real life.) The previous approach was to use the version number
15696 from the record header as a substitute; but our protocol choice
15697 should not depend on that one because it is not authenticated
15698 by the Finished messages.
15702 * More robust randomness gathering functions for Windows.
15704 *Jeffrey Altman <jaltman@columbia.edu>*
15706 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15707 not set then we don't setup the error code for issuer check errors
15708 to avoid possibly overwriting other errors which the callback does
15709 handle. If an application does set the flag then we assume it knows
15710 what it is doing and can handle the new informational codes
15715 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15716 a general "ANY" type, as such it should be able to decode anything
15717 including tagged types. However it didn't check the class so it would
15718 wrongly interpret tagged types in the same way as their universal
15719 counterpart and unknown types were just rejected. Changed so that the
15720 tagged and unknown types are handled in the same way as a SEQUENCE:
15721 that is the encoding is stored intact. There is also a new type
15722 "V_ASN1_OTHER" which is used when the class is not universal, in this
15723 case we have no idea what the actual type is so we just lump them all
15728 * On VMS, stdout may very well lead to a file that is written to
15729 in a record-oriented fashion. That means that every write() will
15730 write a separate record, which will be read separately by the
15731 programs trying to read from it. This can be very confusing.
15733 The solution is to put a BIO filter in the way that will buffer
15734 text until a linefeed is reached, and then write everything a
15735 line at a time, so every record written will be an actual line,
15736 not chunks of lines and not (usually doesn't happen, but I've
15737 seen it once) several lines in one record. BIO_f_linebuffer() is
15740 Currently, it's a VMS-only method, because that's where it has
15741 been tested well enough.
15745 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15746 it can return incorrect results.
15747 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15748 but it was in 0.9.6-beta[12].)
15752 * Disable the check for content being present when verifying detached
15753 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15754 include zero length content when signing messages.
15758 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15759 BIO_ctrl (for BIO pairs).
15763 * Add DSO method for VMS.
15767 * Bug fix: Montgomery multiplication could produce results with the
15772 * Add RPM specification openssl.spec and modify it to build three
15773 packages. The default package contains applications, application
15774 documentation and run-time libraries. The devel package contains
15775 include files, static libraries and function documentation. The
15776 doc package contains the contents of the doc directory. The original
15777 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15781 * Add a large number of documentation files for many SSL routines.
15783 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15785 * Add a configuration entry for Sony News 4.
15787 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15789 * Don't set the two most significant bits to one when generating a
15790 random number < q in the DSA library.
15794 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15795 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15796 the underlying transport is blocking) if a handshake took place.
15797 (The default behaviour is needed by applications such as s_client
15798 and s_server that use select() to determine when to use SSL_read;
15799 but for applications that know in advance when to expect data, it
15800 just makes things more complicated.)
15804 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15809 * Add a few more EBCDIC conditionals that make `req` and `x509`
15810 work better on such systems.
15812 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15814 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15815 Update PKCS12_parse() so it copies the friendlyName and the
15816 keyid to the certificates aux info.
15820 * Fix bug in PKCS7_verify() which caused an infinite loop
15821 if there was more than one signature.
15823 *Sven Uszpelkat <su@celocom.de>*
15825 * Major change in util/mkdef.pl to include extra information
15826 about each symbol, as well as presenting variables as well
15827 as functions. This change means that there's n more need
15828 to rebuild the .num files when some algorithms are excluded.
15832 * Allow the verify time to be set by an application,
15833 rather than always using the current time.
15837 * Phase 2 verify code reorganisation. The certificate
15838 verify code now looks up an issuer certificate by a
15839 number of criteria: subject name, authority key id
15840 and key usage. It also verifies self signed certificates
15841 by the same criteria. The main comparison function is
15842 X509_check_issued() which performs these checks.
15844 Lot of changes were necessary in order to support this
15845 without completely rewriting the lookup code.
15847 Authority and subject key identifier are now cached.
15849 The LHASH 'certs' is X509_STORE has now been replaced
15850 by a STACK_OF(X509_OBJECT). This is mainly because an
15851 LHASH can't store or retrieve multiple objects with
15852 the same hash value.
15854 As a result various functions (which were all internal
15855 use only) have changed to handle the new X509_STORE
15856 structure. This will break anything that messed round
15857 with X509_STORE internally.
15859 The functions X509_STORE_add_cert() now checks for an
15860 exact match, rather than just subject name.
15862 The X509_STORE API doesn't directly support the retrieval
15863 of multiple certificates matching a given criteria, however
15864 this can be worked round by performing a lookup first
15865 (which will fill the cache with candidate certificates)
15866 and then examining the cache for matches. This is probably
15867 the best we can do without throwing out X509_LOOKUP
15868 entirely (maybe later...).
15870 The X509_VERIFY_CTX structure has been enhanced considerably.
15872 All certificate lookup operations now go via a get_issuer()
15873 callback. Although this currently uses an X509_STORE it
15874 can be replaced by custom lookups. This is a simple way
15875 to bypass the X509_STORE hackery necessary to make this
15876 work and makes it possible to use more efficient techniques
15877 in future. A very simple version which uses a simple
15878 STACK for its trusted certificate store is also provided
15879 using X509_STORE_CTX_trusted_stack().
15881 The verify_cb() and verify() callbacks now have equivalents
15882 in the X509_STORE_CTX structure.
15884 X509_STORE_CTX also has a 'flags' field which can be used
15885 to customise the verify behaviour.
15889 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15890 excludes S/MIME capabilities.
15894 * When a certificate request is read in keep a copy of the
15895 original encoding of the signed data and use it when outputting
15896 again. Signatures then use the original encoding rather than
15897 a decoded, encoded version which may cause problems if the
15898 request is improperly encoded.
15902 * For consistency with other BIO_puts implementations, call
15903 buffer_write(b, ...) directly in buffer_puts instead of calling
15906 In BIO_puts, increment b->num_write as in BIO_write.
15908 *Peter.Sylvester@EdelWeb.fr*
15910 * Fix BN_mul_word for the case where the word is 0. (We have to use
15911 BN_zero, we may not return a BIGNUM with an array consisting of
15912 words set to zero.)
15916 * Avoid calling abort() from within the library when problems are
15917 detected, except if preprocessor symbols have been defined
15918 (such as REF_CHECK, BN_DEBUG etc.).
15922 * New openssl application 'rsautl'. This utility can be
15923 used for low-level RSA operations. DER public key
15924 BIO/fp routines also added.
15928 * New Configure entry and patches for compiling on QNX 4.
15930 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15932 * A demo state-machine implementation was sponsored by
15933 Nuron (<http://www.nuron.com/>) and is now available in
15934 demos/state_machine.
15938 * New options added to the 'dgst' utility for signature
15939 generation and verification.
15943 * Unrecognized PKCS#7 content types are now handled via a
15944 catch all ASN1_TYPE structure. This allows unsupported
15945 types to be stored as a "blob" and an application can
15946 encode and decode it manually.
15950 * Fix various signed/unsigned issues to make a_strex.c
15951 compile under VC++.
15953 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15955 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15956 length if passed a buffer. ASN1_INTEGER_to_BN failed
15957 if passed a NULL BN and its argument was negative.
15959 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15961 * Modification to PKCS#7 encoding routines to output definite
15962 length encoding. Since currently the whole structures are in
15963 memory there's not real point in using indefinite length
15964 constructed encoding. However if OpenSSL is compiled with
15965 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15969 * Added BIO_vprintf() and BIO_vsnprintf().
15973 * Added more prefixes to parse for in the strings written
15974 through a logging bio, to cover all the levels that are available
15975 through syslog. The prefixes are now:
15977 PANIC, EMERG, EMR => LOG_EMERG
15978 ALERT, ALR => LOG_ALERT
15979 CRIT, CRI => LOG_CRIT
15980 ERROR, ERR => LOG_ERR
15981 WARNING, WARN, WAR => LOG_WARNING
15982 NOTICE, NOTE, NOT => LOG_NOTICE
15983 INFO, INF => LOG_INFO
15984 DEBUG, DBG => LOG_DEBUG
15986 and as before, if none of those prefixes are present at the
15987 beginning of the string, LOG_ERR is chosen.
15989 On Win32, the `LOG_*` levels are mapped according to this:
15991 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15992 LOG_WARNING => EVENTLOG_WARNING_TYPE
15993 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15997 * Made it possible to reconfigure with just the configuration
15998 argument "reconf" or "reconfigure". The command line arguments
15999 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16000 and are retrieved from there when reconfiguring.
16006 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16008 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16012 * The obj_dat.pl script was messing up the sorting of object
16013 names. The reason was that it compared the quoted version
16014 of strings as a result "OCSP" > "OCSP Signing" because
16015 " > SPACE. Changed script to store unquoted versions of
16016 names and add quotes on output. It was also omitting some
16017 names from the lookup table if they were given a default
16018 value (that is if SN is missing it is given the same
16019 value as LN and vice versa), these are now added on the
16020 grounds that if an object has a name we should be able to
16021 look it up. Finally added warning output when duplicate
16022 short or long names are found.
16026 * Changes needed for Tandem NSK.
16028 *Scott Uroff <scott@xypro.com>*
16030 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16031 RSA_padding_check_SSLv23(), special padding was never detected
16032 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16033 version rollback attacks was not effective.
16035 In s23_clnt.c, don't use special rollback-attack detection padding
16036 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16037 client; similarly, in s23_srvr.c, don't do the rollback check if
16038 SSL 2.0 is the only protocol enabled in the server.
16042 * Make it possible to get hexdumps of unprintable data with 'openssl
16043 asn1parse'. By implication, the functions ASN1_parse_dump() and
16044 BIO_dump_indent() are added.
16048 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16049 these print out strings and name structures based on various
16050 flags including RFC2253 support and proper handling of
16051 multibyte characters. Added options to the 'x509' utility
16052 to allow the various flags to be set.
16056 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16057 Also change the functions X509_cmp_current_time() and
16058 X509_gmtime_adj() work with an ASN1_TIME structure,
16059 this will enable certificates using GeneralizedTime in validity
16060 dates to be checked.
16064 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16065 negative public key encodings) on by default,
16066 NO_NEG_PUBKEY_BUG can be set to disable it.
16070 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16071 content octets. An i2c_ASN1_OBJECT is unnecessary because
16072 the encoding can be trivially obtained from the structure.
16076 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16077 not read locks (`CRYPTO_r_[un]lock`).
16081 * A first attempt at creating official support for shared
16082 libraries through configuration. I've kept it so the
16083 default is static libraries only, and the OpenSSL programs
16084 are always statically linked for now, but there are
16085 preparations for dynamic linking in place.
16086 This has been tested on Linux and Tru64.
16090 * Randomness polling function for Win9x, as described in:
16091 Peter Gutmann, Software Generation of Practically Strong
16096 * Fix so PRNG is seeded in req if using an already existing
16101 * New options to smime application. -inform and -outform
16102 allow alternative formats for the S/MIME message including
16103 PEM and DER. The -content option allows the content to be
16104 specified separately. This should allow things like Netscape
16105 form signing output easier to verify.
16109 * Fix the ASN1 encoding of tags using the 'long form'.
16113 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16114 STRING types. These convert content octets to and from the
16115 underlying type. The actual tag and length octets are
16116 already assumed to have been read in and checked. These
16117 are needed because all other string types have virtually
16118 identical handling apart from the tag. By having versions
16119 of the ASN1 functions that just operate on content octets
16120 IMPLICIT tagging can be handled properly. It also allows
16121 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16122 and ASN1_INTEGER are identical apart from the tag.
16126 * Change the handling of OID objects as follows:
16128 - New object identifiers are inserted in objects.txt, following
16129 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16130 - objects.pl is used to process obj_mac.num and create a new
16132 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16135 This is currently kind of a hack, and the perl code in objects.pl
16136 isn't very elegant, but it works as I intended. The simplest way
16137 to check that it worked correctly is to look in obj_dat.h and
16138 check the array nid_objs and make sure the objects haven't moved
16139 around (this is important!). Additions are OK, as well as
16140 consistent name changes.
16144 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16148 * Addition of the command line parameter '-rand file' to 'openssl req'.
16149 The given file adds to whatever has already been seeded into the
16150 random pool through the RANDFILE configuration file option or
16151 environment variable, or the default random state file.
16155 * mkstack.pl now sorts each macro group into lexical order.
16156 Previously the output order depended on the order the files
16157 appeared in the directory, resulting in needless rewriting
16162 * Patches to make OpenSSL compile under Win32 again. Mostly
16163 work arounds for the VC++ problem that it treats func() as
16164 func(void). Also stripped out the parts of mkdef.pl that
16165 added extra typesafe functions: these no longer exist.
16169 * Reorganisation of the stack code. The macros are now all
16170 collected in safestack.h . Each macro is defined in terms of
16171 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16172 DEBUG_SAFESTACK is now handled in terms of function casts,
16173 this has the advantage of retaining type safety without the
16174 use of additional functions. If DEBUG_SAFESTACK is not defined
16175 then the non typesafe macros are used instead. Also modified the
16176 mkstack.pl script to handle the new form. Needs testing to see
16177 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16178 the default if no major problems. Similar behaviour for ASN1_SET_OF
16179 and PKCS12_STACK_OF.
16183 * When some versions of IIS use the 'NET' form of private key the
16184 key derivation algorithm is different. Normally MD5(password) is
16185 used as a 128 bit RC4 key. In the modified case
16186 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16187 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16188 as the old Netscape_RSA functions except they have an additional
16189 'sgckey' parameter which uses the modified algorithm. Also added
16190 an -sgckey command line option to the rsa utility. Thanks to
16191 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16192 algorithm to openssl-dev.
16196 * The evp_local.h macros were using 'c.##kname' which resulted in
16197 invalid expansion on some systems (SCO 5.0.5 for example).
16198 Corrected to 'c.kname'.
16200 *Phillip Porch <root@theporch.com>*
16202 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16203 a STACK of email addresses from a certificate or request, these look
16204 in the subject name and the subject alternative name extensions and
16205 omit any duplicate addresses.
16209 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16210 This makes DSA verification about 2 % faster.
16214 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16215 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16216 plus overhead for 1024 bit moduli).
16217 This makes exponentiations about 0.5 % faster for 1024 bit
16218 exponents (as measured by "openssl speed rsa2048").
16222 * Rename memory handling macros to avoid conflicts with other
16224 Malloc => OPENSSL_malloc
16225 Malloc_locked => OPENSSL_malloc_locked
16226 Realloc => OPENSSL_realloc
16227 Free => OPENSSL_free
16231 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16232 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16236 * CygWin32 support.
16238 *John Jarvie <jjarvie@newsguy.com>*
16240 * The type-safe stack code has been rejigged. It is now only compiled
16241 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16242 by default all type-specific stack functions are "#define"d back to
16243 standard stack functions. This results in more streamlined output
16244 but retains the type-safety checking possibilities of the original
16249 * The STACK code has been cleaned up, and certain type declarations
16250 that didn't make a lot of sense have been brought in line. This has
16251 also involved a cleanup of sorts in safestack.h to more correctly
16252 map type-safe stack functions onto their plain stack counterparts.
16253 This work has also resulted in a variety of "const"ifications of
16254 lots of the code, especially `_cmp` operations which should normally
16255 be prototyped with "const" parameters anyway.
16259 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16260 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16261 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16262 where all of 'md' is used each time the PRNG is used, but 'state'
16263 is used only indexed by a cyclic counter. As entropy may not be
16264 well distributed from the beginning, 'md' is important as a
16265 chaining variable. However, the output function chains only half
16266 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16267 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16268 in all of 'state' being rewritten, with the new values depending
16269 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16273 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16274 the handshake is continued after ssl_verify_cert_chain();
16275 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16276 can lead to 'unexplainable' connection aborts later.
16278 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16280 * Major EVP API cipher revision.
16281 Add hooks for extra EVP features. This allows various cipher
16282 parameters to be set in the EVP interface. Support added for variable
16283 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16284 setting of RC2 and RC5 parameters.
16286 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16289 Remove lots of duplicated code from the EVP library. For example *every*
16290 cipher init() function handles the 'iv' in the same way according to the
16291 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16292 for CFB and OFB modes they zero ctx->num.
16294 New functionality allows removal of S/MIME code RC2 hack.
16296 Most of the routines have the same form and so can be declared in terms
16299 By shifting this to the top level EVP_CipherInit() it can be removed from
16300 all individual ciphers. If the cipher wants to handle IVs or keys
16301 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16304 Change lots of functions like EVP_EncryptUpdate() to now return a
16305 value: although software versions of the algorithms cannot fail
16306 any installed hardware versions can.
16310 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16311 this option is set, tolerate broken clients that send the negotiated
16312 protocol version number instead of the requested protocol version
16317 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16318 i.e. non-zero for export ciphersuites, zero otherwise.
16319 Previous versions had this flag inverted, inconsistent with
16320 rsa_tmp_cb (..._TMP_RSA_CB).
16322 *Bodo Moeller; problem reported by Amit Chopra*
16324 * Add missing DSA library text string. Work around for some IIS
16325 key files with invalid SEQUENCE encoding.
16329 * Add a document (doc/standards.txt) that list all kinds of standards
16330 and so on that are implemented in OpenSSL.
16334 * Enhance c_rehash script. Old version would mishandle certificates
16335 with the same subject name hash and wouldn't handle CRLs at all.
16336 Added -fingerprint option to crl utility, to support new c_rehash
16341 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16345 * Fix for SSL server purpose checking. Server checking was
16346 rejecting certificates which had extended key usage present
16347 but no ssl client purpose.
16349 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16351 * Make PKCS#12 code work with no password. The PKCS#12 spec
16352 is a little unclear about how a blank password is handled.
16353 Since the password in encoded as a BMPString with terminating
16354 double NULL a zero length password would end up as just the
16355 double NULL. However no password at all is different and is
16356 handled differently in the PKCS#12 key generation code. NS
16357 treats a blank password as zero length. MSIE treats it as no
16358 password on export: but it will try both on import. We now do
16359 the same: PKCS12_parse() tries zero length and no password if
16360 the password is set to "" or NULL (NULL is now a valid password:
16361 it wasn't before) as does the pkcs12 application.
16365 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16366 perror when PEM_read_bio_X509_REQ fails, the error message must
16367 be obtained from the error queue.
16371 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16372 it in ERR_remove_state if appropriate, and change ERR_get_state
16373 accordingly to avoid race conditions (this is necessary because
16374 thread_hash is no longer constant once set).
16378 * Bugfix for linux-elf makefile.one.
16382 * RSA_get_default_method() will now cause a default
16383 RSA_METHOD to be chosen if one doesn't exist already.
16384 Previously this was only set during a call to RSA_new()
16385 or RSA_new_method(NULL) meaning it was possible for
16386 RSA_get_default_method() to return NULL.
16390 * Added native name translation to the existing DSO code
16391 that will convert (if the flag to do so is set) filenames
16392 that are sufficiently small and have no path information
16393 into a canonical native form. Eg. "blah" converted to
16394 "libblah.so" or "blah.dll" etc.
16398 * New function ERR_error_string_n(e, buf, len) which is like
16399 ERR_error_string(e, buf), but writes at most 'len' bytes
16400 including the 0 terminator. For ERR_error_string_n, 'buf'
16403 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16405 * CONF library reworked to become more general. A new CONF
16406 configuration file reader "class" is implemented as well as a
16407 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16408 old `CONF_*` functions are still there, but are reimplemented to
16409 work in terms of the new functions. Also, a set of functions
16410 to handle the internal storage of the configuration data is
16411 provided to make it easier to write new configuration file
16412 reader "classes" (I can definitely see something reading a
16413 configuration file in XML format, for example), called `_CONF_*`,
16414 or "the configuration storage API"...
16416 The new configuration file reading functions are:
16418 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16419 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16421 NCONF_default, NCONF_WIN32
16423 NCONF_dump_fp, NCONF_dump_bio
16425 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16426 NCONF_new creates a new CONF object. This works in the same way
16427 as other interfaces in OpenSSL, like the BIO interface.
16428 `NCONF_dump_*` dump the internal storage of the configuration file,
16429 which is useful for debugging. All other functions take the same
16430 arguments as the old `CONF_*` functions with the exception of the
16431 first that must be a `CONF *` instead of a `LHASH *`.
16433 To make it easier to use the new classes with the old `CONF_*` functions,
16434 the function CONF_set_default_method is provided.
16438 * Add '-tls1' option to 'openssl ciphers', which was already
16439 mentioned in the documentation but had not been implemented.
16440 (This option is not yet really useful because even the additional
16441 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16445 * Initial DSO code added into libcrypto for letting OpenSSL (and
16446 OpenSSL-based applications) load shared libraries and bind to
16447 them in a portable way.
16449 *Geoff Thorpe, with contributions from Richard Levitte*
16451 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16453 * Make sure _lrotl and _lrotr are only used with MSVC.
16455 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16456 (the default implementation of RAND_status).
16458 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16459 to '-clrext' (= clear extensions), as intended and documented.
16460 *Bodo Moeller; inconsistency pointed out by Michael Attili
16461 <attili@amaxo.com>*
16463 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16464 was larger than the MD block size.
16466 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16468 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16469 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16470 using the passed key: if the passed key was a private key the result
16471 of X509_print(), for example, would be to print out all the private key
16476 * des_quad_cksum() byte order bug fix.
16477 *Ulf Möller, using the problem description in krb4-0.9.7, where
16478 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16480 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16483 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16485 * For easily testing in shell scripts whether some command
16486 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16487 returns with exit code 0 iff no command of the given name is available.
16488 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16489 the output goes to stdout and nothing is printed to stderr.
16490 Additional arguments are always ignored.
16492 Since for each cipher there is a command of the same name,
16493 the 'no-cipher' compilation switches can be tested this way.
16495 ('openssl no-XXX' is not able to detect pseudo-commands such
16496 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16500 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16504 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16505 is set; it will be thrown away anyway because each handshake creates
16507 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16508 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16509 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16510 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16514 * New s_client option -ign_eof: EOF at stdin is ignored, and
16515 'Q' and 'R' lose their special meanings (quit/renegotiate).
16516 This is part of what -quiet does; unlike -quiet, -ign_eof
16517 does not suppress any output.
16521 * Add compatibility options to the purpose and trust code. The
16522 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16523 accepts a certificate or CA, this was the previous behaviour,
16524 with all the associated security issues.
16526 X509_TRUST_COMPAT is the old trust behaviour: only and
16527 automatically trust self signed roots in certificate store. A
16528 new trust setting X509_TRUST_DEFAULT is used to specify that
16529 a purpose has no associated trust setting and it should instead
16530 use the value in the default purpose.
16534 * Fix the PKCS#8 DSA private key code so it decodes keys again
16535 and fix a memory leak.
16539 * In util/mkerr.pl (which implements 'make errors'), preserve
16540 reason strings from the previous version of the .c file, as
16541 the default to have only downcase letters (and digits) in
16542 automatically generated reasons codes is not always appropriate.
16546 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16547 using strerror. Previously, ERR_reason_error_string() returned
16548 library names as reason strings for SYSerr; but SYSerr is a special
16549 case where small numbers are errno values, not library numbers.
16553 * Add '-dsaparam' option to 'openssl dhparam' application. This
16554 converts DSA parameters into DH parameters. (When creating parameters,
16555 DSA_generate_parameters is used.)
16559 * Include 'length' (recommended exponent length) in C code generated
16560 by 'openssl dhparam -C'.
16564 * The second argument to set_label in perlasm was already being used
16565 so couldn't be used as a "file scope" flag. Moved to third argument
16570 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16571 instead of RAND_bytes for encryption IVs and salts.
16575 * Include RAND_status() into RAND_METHOD instead of implementing
16576 it only for md_rand.c Otherwise replacing the PRNG by calling
16577 RAND_set_rand_method would be impossible.
16581 * Don't let DSA_generate_key() enter an infinite loop if the random
16582 number generation fails.
16586 * New 'rand' application for creating pseudo-random output.
16590 * Added configuration support for Linux/IA64
16592 *Rolf Haberrecker <rolf@suse.de>*
16594 * Assembler module support for Mingw32.
16598 * Shared library support for HPUX (in shlib/).
16600 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16602 * Shared library support for Solaris gcc.
16604 *Lutz Behnke <behnke@trustcenter.de>*
16606 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16608 * PKCS7_encrypt() was adding text MIME headers twice because they
16609 were added manually and by SMIME_crlf_copy().
16613 * In bntest.c don't call BN_rand with zero bits argument.
16615 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16617 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16618 case was implemented. This caused BN_div_recp() to fail occasionally.
16622 * Add an optional second argument to the set_label() in the perl
16623 assembly language builder. If this argument exists and is set
16624 to 1 it signals that the assembler should use a symbol whose
16625 scope is the entire file, not just the current function. This
16626 is needed with MASM which uses the format label:: for this scope.
16628 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16630 * Change the ASN1 types so they are typedefs by default. Before
16631 almost all types were #define'd to ASN1_STRING which was causing
16632 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16637 * Change names of new functions to the new get1/get0 naming
16638 convention: After 'get1', the caller owns a reference count
16639 and has to call `..._free`; 'get0' returns a pointer to some
16640 data structure without incrementing reference counters.
16641 (Some of the existing 'get' functions increment a reference
16642 counter, some don't.)
16643 Similarly, 'set1' and 'add1' functions increase reference
16644 counters or duplicate objects.
16648 * Allow for the possibility of temp RSA key generation failure:
16649 the code used to assume it always worked and crashed on failure.
16653 * Fix potential buffer overrun problem in BIO_printf().
16654 *Ulf Möller, using public domain code by Patrick Powell; problem
16655 pointed out by David Sacerdote <das33@cornell.edu>*
16657 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16658 RAND_egd() and RAND_status(). In the command line application,
16659 the EGD socket can be specified like a seed file using RANDFILE
16664 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16665 Some CAs (e.g. Verisign) distribute certificates in this form.
16669 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16670 list to exclude them. This means that no special compilation option
16671 is needed to use anonymous DH: it just needs to be included in the
16676 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16677 EVP_MD_type. The old functionality is available in a new macro called
16678 EVP_MD_md(). Change code that uses it and update docs.
16682 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16683 where the `void *` argument is replaced by a function pointer argument.
16684 Previously `void *` was abused to point to functions, which works on
16685 many platforms, but is not correct. As these functions are usually
16686 called by macros defined in OpenSSL header files, most source code
16687 should work without changes.
16691 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16692 sections with information on -D... compiler switches used for
16693 compiling the library so that applications can see them. To enable
16694 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16695 must be defined. E.g.,
16696 #define OPENSSL_ALGORITHM_DEFINES
16697 #include <openssl/opensslconf.h>
16698 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16700 *Richard Levitte, Ulf and Bodo Möller*
16702 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16707 * Change the 'other' type in certificate aux info to a STACK_OF
16708 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16709 the required ASN1 format: arbitrary types determined by an OID.
16713 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16714 argument to 'req'. This is not because the function is newer or
16715 better than others it just uses the work 'NEW' in the certificate
16716 request header lines. Some software needs this.
16720 * Reorganise password command line arguments: now passwords can be
16721 obtained from various sources. Delete the PEM_cb function and make
16722 it the default behaviour: i.e. if the callback is NULL and the
16723 usrdata argument is not NULL interpret it as a null terminated pass
16724 phrase. If usrdata and the callback are NULL then the pass phrase
16725 is prompted for as usual.
16729 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16730 the support is automatically enabled. The resulting binaries will
16731 autodetect the card and use it if present.
16733 *Ben Laurie and Compaq Inc.*
16735 * Work around for Netscape hang bug. This sends certificate request
16736 and server done in one record. Since this is perfectly legal in the
16737 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16738 the bugs/SSLv3 entry for more info.
16742 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16746 * Add -rand argument to smime and pkcs12 applications and read/write
16751 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16755 * Add command line password options to the remaining applications.
16759 * Bug fix for BN_div_recp() for numerators with an even number of
16764 * More tests in bntest.c, and changed test_bn output.
16768 * ./config recognizes MacOS X now.
16772 * Bug fix for BN_div() when the first words of num and divisor are
16773 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16777 * Add support for various broken PKCS#8 formats, and command line
16778 options to produce them.
16782 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16783 get temporary BIGNUMs from a BN_CTX.
16787 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16792 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16793 include a #define from the old name to the new. The original intent
16794 was that statically linked binaries could for example just call
16795 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16796 link with digests. This never worked because SSLeay_add_all_digests()
16797 and SSLeay_add_all_ciphers() were in the same source file so calling
16798 one would link with the other. They are now in separate source files.
16802 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16806 * Use a less unusual form of the Miller-Rabin primality test (it used
16807 a binary algorithm for exponentiation integrated into the Miller-Rabin
16808 loop, our standard modexp algorithms are faster).
16812 * Support for the EBCDIC character set completed.
16814 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16816 * Source code cleanups: use const where appropriate, eliminate casts,
16817 use `void *` instead of `char *` in lhash.
16821 * Bugfix: ssl3_send_server_key_exchange was not restartable
16822 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16823 this the server could overwrite ephemeral keys that the client
16828 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16829 using 50 iterations of the Rabin-Miller test.
16831 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16832 iterations of the Rabin-Miller test as required by the appendix
16833 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16834 As BN_is_prime_fasttest includes trial division, DSA parameter
16835 generation becomes much faster.
16837 This implies a change for the callback functions in DSA_is_prime
16838 and DSA_generate_parameters: The callback function is called once
16839 for each positive witness in the Rabin-Miller test, not just
16840 occasionally in the inner loop; and the parameters to the
16841 callback function now provide an iteration count for the outer
16842 loop rather than for the current invocation of the inner loop.
16843 DSA_generate_parameters additionally can call the callback
16844 function with an 'iteration count' of -1, meaning that a
16845 candidate has passed the trial division test (when q is generated
16846 from an application-provided seed, trial division is skipped).
16850 * New function BN_is_prime_fasttest that optionally does trial
16851 division before starting the Rabin-Miller test and has
16852 an additional BN_CTX * argument (whereas BN_is_prime always
16853 has to allocate at least one BN_CTX).
16854 'callback(1, -1, cb_arg)' is called when a number has passed the
16855 trial division stage.
16859 * Fix for bug in CRL encoding. The validity dates weren't being handled
16864 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16868 * New function BN_pseudo_rand().
16872 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16873 bignum version of BN_from_montgomery() with the working code from
16874 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16879 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16880 made it impossible to use the same SSL_SESSION data structure in
16881 SSL2 clients in multiple threads.
16885 * The return value of RAND_load_file() no longer counts bytes obtained
16886 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16887 to seed the PRNG (previously an explicit byte count was required).
16889 *Ulf Möller, Bodo Möller*
16891 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16892 used `char *` instead of `void *` and had casts all over the place.
16896 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16900 * Retain source code compatibility for BN_prime_checks macro:
16901 BN_is_prime(..., BN_prime_checks, ...) now uses
16902 BN_prime_checks_for_size to determine the appropriate number of
16903 Rabin-Miller iterations.
16907 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16908 DH_CHECK_P_NOT_SAFE_PRIME.
16909 (Check if this is true? OpenPGP calls them "strong".)
16913 * Merge the functionality of "dh" and "gendh" programs into a new program
16914 "dhparam". The old programs are retained for now but will handle DH keys
16915 (instead of parameters) in future.
16919 * Make the ciphers, s_server and s_client programs check the return values
16920 when a new cipher list is set.
16924 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16925 ciphers. Before when the 56bit ciphers were enabled the sorting was
16928 The syntax for the cipher sorting has been extended to support sorting by
16929 cipher-strength (using the strength_bits hard coded in the tables).
16930 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16932 Fix a bug in the cipher-command parser: when supplying a cipher command
16933 string with an "undefined" symbol (neither command nor alphanumeric
16934 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16935 an error is flagged.
16937 Due to the strength-sorting extension, the code of the
16938 ssl_create_cipher_list() function was completely rearranged. I hope that
16939 the readability was also increased :-)
16941 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16943 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16944 for the first serial number and places 2 in the serial number file. This
16945 avoids problems when the root CA is created with serial number zero and
16946 the first user certificate has the same issuer name and serial number
16951 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16952 the new code. Add documentation for this stuff.
16956 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16957 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16958 structures and behave in an analogous way to the X509v3 functions:
16959 they shouldn't be called directly but wrapper functions should be used
16962 So we also now have some wrapper functions that call the X509at functions
16963 when passed certificate requests. (TO DO: similar things can be done with
16964 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16965 things. Some of these need some d2i or i2d and print functionality
16966 because they handle more complex structures.)
16970 * Add missing #ifndefs that caused missing symbols when building libssl
16971 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16972 NO_RSA in `ssl/s2*.c`.
16974 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16976 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16977 has a return value which indicates the quality of the random data
16978 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16979 error queue. New function RAND_pseudo_bytes() generates output that is
16980 guaranteed to be unique but not unpredictable. RAND_add is like
16981 RAND_seed, but takes an extra argument for an entropy estimate
16982 (RAND_seed always assumes full entropy).
16986 * Do more iterations of Rabin-Miller probable prime test (specifically,
16987 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16988 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16989 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16990 false-positive rate of at most 2^-80 for random input.
16994 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16998 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16999 in the 0.9.5 release), this returns the chain
17000 from an X509_CTX structure with a dup of the stack and all
17001 the X509 reference counts upped: so the stack will exist
17002 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17005 Also make SSL_SESSION_print() print out the verify return
17010 * Add manpage for the pkcs12 command. Also change the default
17011 behaviour so MAC iteration counts are used unless the new
17012 -nomaciter option is used. This improves file security and
17013 only older versions of MSIE (4.0 for example) need it.
17017 * Honor the no-xxx Configure options when creating .DEF files.
17021 * Add PKCS#10 attributes to field table: challengePassword,
17022 unstructuredName and unstructuredAddress. These are taken from
17023 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17024 international characters are used.
17026 More changes to X509_ATTRIBUTE code: allow the setting of types
17027 based on strings. Remove the 'loc' parameter when adding
17028 attributes because these will be a SET OF encoding which is sorted
17033 * Initial changes to the 'req' utility to allow request generation
17034 automation. This will allow an application to just generate a template
17035 file containing all the field values and have req construct the
17038 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17039 used all over the place including certificate requests and PKCS#7
17040 structures. They are currently handled manually where necessary with
17041 some primitive wrappers for PKCS#7. The new functions behave in a
17042 manner analogous to the X509 extension functions: they allow
17043 attributes to be looked up by NID and added.
17045 Later something similar to the X509V3 code would be desirable to
17046 automatically handle the encoding, decoding and printing of the
17047 more complex types. The string types like challengePassword can
17048 be handled by the string table functions.
17050 Also modified the multi byte string table handling. Now there is
17051 a 'global mask' which masks out certain types. The table itself
17052 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17053 is useful when for example there is only one permissible type
17054 (as in countryName) and using the mask might result in no valid
17059 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17060 SSL_get_peer_finished to allow applications to obtain the latest
17061 Finished messages sent to the peer or expected from the peer,
17062 respectively. (SSL_get_peer_finished is usually the Finished message
17063 actually received from the peer, otherwise the protocol will be aborted.)
17065 As the Finished message are message digests of the complete handshake
17066 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17067 be used for external authentication procedures when the authentication
17068 provided by SSL/TLS is not desired or is not enough.
17072 * Enhanced support for Alpha Linux is added. Now ./config checks if
17073 the host supports BWX extension and if Compaq C is present on the
17074 $PATH. Just exploiting of the BWX extension results in 20-30%
17075 performance kick for some algorithms, e.g. DES and RC4 to mention
17076 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17081 * Add support for MS "fast SGC". This is arguably a violation of the
17082 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17083 weak crypto and after checking the certificate is SGC a second one
17084 with strong crypto. MS SGC stops the first handshake after receiving
17085 the server certificate message and sends a second client hello. Since
17086 a server will typically do all the time consuming operations before
17087 expecting any further messages from the client (server key exchange
17088 is the most expensive) there is little difference between the two.
17090 To get OpenSSL to support MS SGC we have to permit a second client
17091 hello message after we have sent server done. In addition we have to
17092 reset the MAC if we do get this second client hello.
17096 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17097 if a DER encoded private key is RSA or DSA traditional format. Changed
17098 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17099 format DER encoded private key. Newer code should use PKCS#8 format which
17100 has the key type encoded in the ASN1 structure. Added DER private key
17101 support to pkcs8 application.
17105 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17106 ciphersuites has been selected (as required by the SSL 3/TLS 1
17107 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17108 is set, we interpret this as a request to violate the specification
17109 (the worst that can happen is a handshake failure, and 'correct'
17110 behaviour would result in a handshake failure anyway).
17114 * In SSL_CTX_add_session, take into account that there might be multiple
17115 SSL_SESSION structures with the same session ID (e.g. when two threads
17116 concurrently obtain them from an external cache).
17117 The internal cache can handle only one SSL_SESSION with a given ID,
17118 so if there's a conflict, we now throw out the old one to achieve
17123 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17124 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17125 some routines that use cipher OIDs: some ciphers do not have OIDs
17126 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17131 * Simplify the trust setting structure and code. Now we just have
17132 two sequences of OIDs for trusted and rejected settings. These will
17133 typically have values the same as the extended key usage extension
17134 and any application specific purposes.
17136 The trust checking code now has a default behaviour: it will just
17137 check for an object with the same NID as the passed id. Functions can
17138 be provided to override either the default behaviour or the behaviour
17139 for a given id. SSL client, server and email already have functions
17140 in place for compatibility: they check the NID and also return "trusted"
17141 if the certificate is self signed.
17145 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17146 traditional format into an EVP_PKEY structure.
17150 * Add a password callback function PEM_cb() which either prompts for
17151 a password if usr_data is NULL or otherwise assumes it is a null
17152 terminated password. Allow passwords to be passed on command line
17153 environment or config files in a few more utilities.
17157 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17158 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17159 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17160 Update documentation.
17164 * Support for ASN1 "NULL" type. This could be handled before by using
17165 ASN1_TYPE but there wasn't any function that would try to read a NULL
17166 and produce an error if it couldn't. For compatibility we also have
17167 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17168 don't allocate anything because they don't need to.
17172 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17175 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17177 * Rebuild of the memory allocation routines used by OpenSSL code and
17178 possibly others as well. The purpose is to make an interface that
17179 provide hooks so anyone can build a separate set of allocation and
17180 deallocation routines to be used by OpenSSL, for example memory
17181 pool implementations, or something else, which was previously hard
17182 since Malloc(), Realloc() and Free() were defined as macros having
17183 the values malloc, realloc and free, respectively (except for Win32
17184 compilations). The same is provided for memory debugging code.
17185 OpenSSL already comes with functionality to find memory leaks, but
17186 this gives people a chance to debug other memory problems.
17188 With these changes, a new set of functions and macros have appeared:
17190 CRYPTO_set_mem_debug_functions() [F]
17191 CRYPTO_get_mem_debug_functions() [F]
17192 CRYPTO_dbg_set_options() [F]
17193 CRYPTO_dbg_get_options() [F]
17194 CRYPTO_malloc_debug_init() [M]
17196 The memory debug functions are NULL by default, unless the library
17197 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17198 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17199 gives the standard debugging functions that come with OpenSSL) or
17200 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17201 provided by the library user) must be used. When the standard
17202 debugging functions are used, CRYPTO_dbg_set_options can be used to
17203 request additional information:
17204 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17205 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17207 Also, things like CRYPTO_set_mem_functions will always give the
17208 expected result (the new set of functions is used for allocation
17209 and deallocation) at all times, regardless of platform and compiler
17212 To finish it up, some functions that were never use in any other
17213 way than through macros have a new API and new semantic:
17215 CRYPTO_dbg_malloc()
17216 CRYPTO_dbg_realloc()
17219 All macros of value have retained their old syntax.
17221 *Richard Levitte and Bodo Moeller*
17223 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17224 ordering of SMIMECapabilities wasn't in "strength order" and there
17225 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17230 * Some ASN1 types with illegal zero length encoding (INTEGER,
17231 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17233 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17235 * Merge in my S/MIME library for OpenSSL. This provides a simple
17236 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17237 functionality to handle multipart/signed properly) and a utility
17238 called 'smime' to call all this stuff. This is based on code I
17239 originally wrote for Celo who have kindly allowed it to be
17240 included in OpenSSL.
17244 * Add variants des_set_key_checked and des_set_key_unchecked of
17245 des_set_key (aka des_key_sched). Global variable des_check_key
17246 decides which of these is called by des_set_key; this way
17247 des_check_key behaves as it always did, but applications and
17248 the library itself, which was buggy for des_check_key == 1,
17249 have a cleaner way to pick the version they need.
17253 * New function PKCS12_newpass() which changes the password of a
17258 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17259 dynamic mix. In both cases the ids can be used as an index into the
17260 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17261 functions so they accept a list of the field values and the
17262 application doesn't need to directly manipulate the X509_TRUST
17267 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17272 * Modify the way the V3 extension code looks up extensions. This now
17273 works in a similar way to the object code: we have some "standard"
17274 extensions in a static table which is searched with OBJ_bsearch()
17275 and the application can add dynamic ones if needed. The file
17276 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17277 updated whenever a new extension is added to the core code and kept
17278 in ext_nid order. There is a simple program 'tabtest.c' which checks
17279 this. New extensions are not added too often so this file can readily
17280 be maintained manually.
17282 There are two big advantages in doing things this way. The extensions
17283 can be looked up immediately and no longer need to be "added" using
17284 X509V3_add_standard_extensions(): this function now does nothing.
17285 Side note: I get *lots* of email saying the extension code doesn't
17286 work because people forget to call this function.
17287 Also no dynamic allocation is done unless new extensions are added:
17288 so if we don't add custom extensions there is no need to call
17289 X509V3_EXT_cleanup().
17293 * Modify enc utility's salting as follows: make salting the default. Add a
17294 magic header, so unsalted files fail gracefully instead of just decrypting
17295 to garbage. This is because not salting is a big security hole, so people
17296 should be discouraged from doing it.
17300 * Fixes and enhancements to the 'x509' utility. It allowed a message
17301 digest to be passed on the command line but it only used this
17302 parameter when signing a certificate. Modified so all relevant
17303 operations are affected by the digest parameter including the
17304 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17305 DSA key was used because it didn't fix the digest.
17309 * Initial certificate chain verify code. Currently tests the untrusted
17310 certificates for consistency with the verify purpose (which is set
17311 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17313 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17314 this is because it will reject chains with invalid extensions whereas
17315 every previous version of OpenSSL and SSLeay made no checks at all.
17317 Trust code: checks the root CA for the relevant trust settings. Trust
17318 settings have an initial value consistent with the verify purpose: e.g.
17319 if the verify purpose is for SSL client use it expects the CA to be
17320 trusted for SSL client use. However the default value can be changed to
17321 permit custom trust settings: one example of this would be to only trust
17322 certificates from a specific "secure" set of CAs.
17324 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17325 which should be used for version portability: especially since the
17326 verify structure is likely to change more often now.
17328 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17329 to set them. If not set then assume SSL clients will verify SSL servers
17332 Two new options to the verify program: -untrusted allows a set of
17333 untrusted certificates to be passed in and -purpose which sets the
17334 intended purpose of the certificate. If a purpose is set then the
17335 new chain verify code is used to check extension consistency.
17339 * Support for the authority information access extension.
17343 * Modify RSA and DSA PEM read routines to transparently handle
17344 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17345 public keys in a format compatible with certificate
17346 SubjectPublicKeyInfo structures. Unfortunately there were already
17347 functions called *_PublicKey_* which used various odd formats so
17348 these are retained for compatibility: however the DSA variants were
17349 never in a public release so they have been deleted. Changed dsa/rsa
17350 utilities to handle the new format: note no releases ever handled public
17351 keys so we should be OK.
17353 The primary motivation for this change is to avoid the same fiasco
17354 that dogs private keys: there are several incompatible private key
17355 formats some of which are standard and some OpenSSL specific and
17356 require various evil hacks to allow partial transparent handling and
17357 even then it doesn't work with DER formats. Given the option anything
17358 other than PKCS#8 should be dumped: but the other formats have to
17359 stay in the name of compatibility.
17361 With public keys and the benefit of hindsight one standard format
17362 is used which works with EVP_PKEY, RSA or DSA structures: though
17363 it clearly returns an error if you try to read the wrong kind of key.
17365 Added a -pubkey option to the 'x509' utility to output the public key.
17366 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17367 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17368 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17369 that do the same as the `EVP_PKEY_assign_*()` except they up the
17370 reference count of the added key (they don't "swallow" the
17375 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17376 CRLs would fail if the file contained no certificates or no CRLs:
17377 added a new function to read in both types and return the number
17378 read: this means that if none are read it will be an error. The
17379 DER versions of the certificate and CRL reader would always fail
17380 because it isn't possible to mix certificates and CRLs in DER format
17381 without choking one or the other routine. Changed this to just read
17382 a certificate: this is the best we can do. Also modified the code
17383 in `apps/verify.c` to take notice of return codes: it was previously
17384 attempting to read in certificates from NULL pointers and ignoring
17385 any errors: this is one reason why the cert and CRL reader seemed
17386 to work. It doesn't check return codes from the default certificate
17387 routines: these may well fail if the certificates aren't installed.
17391 * Code to support otherName option in GeneralName.
17395 * First update to verify code. Change the verify utility
17396 so it warns if it is passed a self signed certificate:
17397 for consistency with the normal behaviour. X509_verify
17398 has been modified to it will now verify a self signed
17399 certificate if *exactly* the same certificate appears
17400 in the store: it was previously impossible to trust a
17401 single self signed certificate. This means that:
17402 openssl verify ss.pem
17403 now gives a warning about a self signed certificate but
17404 openssl verify -CAfile ss.pem ss.pem
17409 * For servers, store verify_result in SSL_SESSION data structure
17410 (and add it to external session representation).
17411 This is needed when client certificate verifications fails,
17412 but an application-provided verification callback (set by
17413 SSL_CTX_set_cert_verify_callback) allows accepting the session
17414 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17415 but returns 1): When the session is reused, we have to set
17416 ssl->verify_result to the appropriate error code to avoid
17419 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17421 * Fix a bug in the new PKCS#7 code: it didn't consider the
17422 case in PKCS7_dataInit() where the signed PKCS7 structure
17423 didn't contain any existing data because it was being created.
17425 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17427 * Add a salt to the key derivation routines in enc.c. This
17428 forms the first 8 bytes of the encrypted file. Also add a
17429 -S option to allow a salt to be input on the command line.
17433 * New function X509_cmp(). Oddly enough there wasn't a function
17434 to compare two certificates. We do this by working out the SHA1
17435 hash and comparing that. X509_cmp() will be needed by the trust
17440 * SSL_get1_session() is like SSL_get_session(), but increments
17441 the reference count in the SSL_SESSION returned.
17443 *Geoff Thorpe <geoff@eu.c2.net>*
17445 * Fix for 'req': it was adding a null to request attributes.
17446 Also change the X509_LOOKUP and X509_INFO code to handle
17447 certificate auxiliary information.
17451 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17456 * Add the possibility to add extra information to the memory leak
17457 detecting output, to form tracebacks, showing from where each
17458 allocation was originated: CRYPTO_push_info("constant string") adds
17459 the string plus current file name and line number to a per-thread
17460 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17461 is like calling CYRPTO_pop_info() until the stack is empty.
17462 Also updated memory leak detection code to be multi-thread-safe.
17466 * Add options -text and -noout to pkcs7 utility and delete the
17467 encryption options which never did anything. Update docs.
17471 * Add options to some of the utilities to allow the pass phrase
17472 to be included on either the command line (not recommended on
17473 OSes like Unix) or read from the environment. Update the
17474 manpages and fix a few bugs.
17478 * Add a few manpages for some of the openssl commands.
17482 * Fix the -revoke option in ca. It was freeing up memory twice,
17483 leaking and not finding already revoked certificates.
17487 * Extensive changes to support certificate auxiliary information.
17488 This involves the use of X509_CERT_AUX structure and X509_AUX
17489 functions. An X509_AUX function such as PEM_read_X509_AUX()
17490 can still read in a certificate file in the usual way but it
17491 will also read in any additional "auxiliary information". By
17492 doing things this way a fair degree of compatibility can be
17493 retained: existing certificates can have this information added
17494 using the new 'x509' options.
17496 Current auxiliary information includes an "alias" and some trust
17497 settings. The trust settings will ultimately be used in enhanced
17498 certificate chain verification routines: currently a certificate
17499 can only be trusted if it is self signed and then it is trusted
17504 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17505 The problem was that one of the replacement routines had not been working
17506 since SSLeay releases. For now the offending routine has been replaced
17507 with non-optimised assembler. Even so, this now gives around 95%
17508 performance improvement for 1024 bit RSA signs.
17512 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17513 handling. Most clients have the effective key size in bits equal to
17514 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17515 A few however don't do this and instead use the size of the decrypted key
17516 to determine the RC2 key length and the AlgorithmIdentifier to determine
17517 the effective key length. In this case the effective key length can still
17518 be 40 bits but the key length can be 168 bits for example. This is fixed
17519 by manually forcing an RC2 key into the EVP_PKEY structure because the
17520 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17521 the key length and effective key length are equal.
17525 * Add a bunch of functions that should simplify the creation of
17526 X509_NAME structures. Now you should be able to do:
17527 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17528 and have it automatically work out the correct field type and fill in
17529 the structures. The more adventurous can try:
17530 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17531 and it will (hopefully) work out the correct multibyte encoding.
17535 * Change the 'req' utility to use the new field handling and multibyte
17536 copy routines. Before the DN field creation was handled in an ad hoc
17537 way in req, ca, and x509 which was rather broken and didn't support
17538 BMPStrings or UTF8Strings. Since some software doesn't implement
17539 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17540 using the dirstring_type option. See the new comment in the default
17541 openssl.cnf for more info.
17545 * Make crypto/rand/md_rand.c more robust:
17546 - Assure unique random numbers after fork().
17547 - Make sure that concurrent threads access the global counter and
17548 md serializably so that we never lose entropy in them
17549 or use exactly the same state in multiple threads.
17550 Access to the large state is not always serializable because
17551 the additional locking could be a performance killer, and
17552 md should be large enough anyway.
17556 * New file `apps/app_rand.c` with commonly needed functionality
17557 for handling the random seed file.
17559 Use the random seed file in some applications that previously did not:
17561 dsaparam -genkey (which also ignored its '-rand' option),
17564 x509 (when signing).
17565 Except on systems with /dev/urandom, it is crucial to have a random
17566 seed file at least for key creation, DSA signing, and for DH exchanges;
17567 for RSA signatures we could do without one.
17569 gendh and gendsa (unlike genrsa) used to read only the first byte
17570 of each file listed in the '-rand' option. The function as previously
17571 found in genrsa is now in app_rand.c and is used by all programs
17572 that support '-rand'.
17576 * In RAND_write_file, use mode 0600 for creating files;
17577 don't just chmod when it may be too late.
17581 * Report an error from X509_STORE_load_locations
17582 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17586 * New function ASN1_mbstring_copy() this copies a string in either
17587 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17588 into an ASN1_STRING type. A mask of permissible types is passed
17589 and it chooses the "minimal" type to use or an error if not type
17594 * Add function equivalents to the various macros in asn1.h. The old
17595 macros are retained with an `M_` prefix. Code inside the library can
17596 use the `M_` macros. External code (including the openssl utility)
17597 should *NOT* in order to be "shared library friendly".
17601 * Add various functions that can check a certificate's extensions
17602 to see if it usable for various purposes such as SSL client,
17603 server or S/MIME and CAs of these types. This is currently
17604 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17605 verification. Also added a -purpose flag to x509 utility to
17606 print out all the purposes.
17610 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17615 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17616 for, obtain and decode and extension and obtain its critical flag.
17617 This allows all the necessary extension code to be handled in a
17618 single function call.
17622 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17623 platforms. See crypto/rc4/rc4_enc.c for further details.
17627 * New -noout option to asn1parse. This causes no output to be produced
17628 its main use is when combined with -strparse and -out to extract data
17629 from a file (which may not be in ASN.1 format).
17633 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17634 when producing the local key id.
17636 *Richard Levitte <levitte@stacken.kth.se>*
17638 * New option -dhparam in s_server. This allows a DH parameter file to be
17639 stated explicitly. If it is not stated then it tries the first server
17640 certificate file. The previous behaviour hard coded the filename
17645 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17646 a public key to be input or output. For example:
17647 openssl rsa -in key.pem -pubout -out pubkey.pem
17648 Also added necessary DSA public key functions to handle this.
17652 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17653 in the message. This was handled by allowing
17654 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17656 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17658 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17659 to the end of the strings whereas this didn't. This would cause problems
17660 if strings read with d2i_ASN1_bytes() were later modified.
17662 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17664 * Fix for base64 decode bug. When a base64 bio reads only one line of
17665 data and it contains EOF it will end up returning an error. This is
17666 caused by input 46 bytes long. The cause is due to the way base64
17667 BIOs find the start of base64 encoded data. They do this by trying a
17668 trial decode on each line until they find one that works. When they
17669 do a flag is set and it starts again knowing it can pass all the
17670 data directly through the decoder. Unfortunately it doesn't reset
17671 the context it uses. This means that if EOF is reached an attempt
17672 is made to pass two EOFs through the context and this causes the
17673 resulting error. This can also cause other problems as well. As is
17674 usual with these problems it takes *ages* to find and the fix is
17675 trivial: move one line.
17677 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17679 * Ugly workaround to get s_client and s_server working under Windows. The
17680 old code wouldn't work because it needed to select() on sockets and the
17681 tty (for keypresses and to see if data could be written). Win32 only
17682 supports select() on sockets so we select() with a 1s timeout on the
17683 sockets and then see if any characters are waiting to be read, if none
17684 are present then we retry, we also assume we can always write data to
17685 the tty. This isn't nice because the code then blocks until we've
17686 received a complete line of data and it is effectively polling the
17687 keyboard at 1s intervals: however it's quite a bit better than not
17688 working at all :-) A dedicated Windows application might handle this
17689 with an event loop for example.
17693 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17694 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17695 will be called when RSA_sign() and RSA_verify() are used. This is useful
17696 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17697 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17698 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17699 This necessitated the support of an extra signature type NID_md5_sha1
17700 for SSL signatures and modifications to the SSL library to use it instead
17701 of calling RSA_public_decrypt() and RSA_private_encrypt().
17705 * Add new -verify -CAfile and -CApath options to the crl program, these
17706 will lookup a CRL issuers certificate and verify the signature in a
17707 similar way to the verify program. Tidy up the crl program so it
17708 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17709 less strict. It will now permit CRL extensions even if it is not
17710 a V2 CRL: this will allow it to tolerate some broken CRLs.
17714 * Initialize all non-automatic variables each time one of the openssl
17715 sub-programs is started (this is necessary as they may be started
17716 multiple times from the "OpenSSL>" prompt).
17718 *Lennart Bang, Bodo Moeller*
17720 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17721 removing all other RSA functionality (this is what NO_RSA does). This
17722 is so (for example) those in the US can disable those operations covered
17723 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17728 * Non-copying interface to BIO pairs.
17729 (still largely untested)
17733 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17734 ASCII string. This was handled independently in various places before.
17738 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17739 UTF8 strings a character at a time.
17743 * Use client_version from client hello to select the protocol
17744 (s23_srvr.c) and for RSA client key exchange verification
17745 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17749 * Add various utility functions to handle SPKACs, these were previously
17750 handled by poking round in the structure internals. Added new function
17751 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17752 print, verify and generate SPKACs. Based on an original idea from
17753 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17757 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17761 * Allow the config file extension section to be overwritten on the
17762 command line. Based on an original idea from Massimiliano Pala
17763 <madwolf@comune.modena.it>. The new option is called -extensions
17764 and can be applied to ca, req and x509. Also -reqexts to override
17765 the request extensions in req and -crlexts to override the crl extensions
17770 * Add new feature to the SPKAC handling in ca. Now you can include
17771 the same field multiple times by preceding it by "XXXX." for example:
17774 this is the same syntax as used in the req config file.
17778 * Allow certificate extensions to be added to certificate requests. These
17779 are specified in a 'req_extensions' option of the req section of the
17780 config file. They can be printed out with the -text option to req but
17781 are otherwise ignored at present.
17785 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17786 data read consists of only the final block it would not decrypted because
17787 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17788 A misplaced 'break' also meant the decrypted final block might not be
17789 copied until the next read.
17793 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17794 a few extra parameters to the DH structure: these will be useful if
17795 for example we want the value of 'q' or implement X9.42 DH.
17799 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17800 provides hooks that allow the default DSA functions or functions on a
17801 "per key" basis to be replaced. This allows hardware acceleration and
17802 hardware key storage to be handled without major modification to the
17803 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17804 associated functions.
17808 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17809 as "read only": it can't be written to and the buffer it points to will
17810 not be freed. Reading from a read only BIO is much more efficient than
17811 a normal memory BIO. This was added because there are several times when
17812 an area of memory needs to be read from a BIO. The previous method was
17813 to create a memory BIO and write the data to it, this results in two
17814 copies of the data and an O(n^2) reading algorithm. There is a new
17815 function BIO_new_mem_buf() which creates a read only memory BIO from
17816 an area of memory. Also modified the PKCS#7 routines to use read only
17821 * Bugfix: ssl23_get_client_hello did not work properly when called in
17822 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17823 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17824 but a retry condition occurred while trying to read the rest.
17828 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17829 NID_pkcs7_encrypted by default: this was wrong since this should almost
17830 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17831 the encrypted data type: this is a more sensible place to put it and it
17832 allows the PKCS#12 code to be tidied up that duplicated this
17837 * Changed obj_dat.pl script so it takes its input and output files on
17838 the command line. This should avoid shell escape redirection problems
17843 * Initial support for certificate extension requests, these are included
17844 in things like Xenroll certificate requests. Included functions to allow
17845 extensions to be obtained and added.
17849 * -crlf option to s_client and s_server for sending newlines as
17850 CRLF (as required by many protocols).
17854 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17856 * Install libRSAglue.a when OpenSSL is built with RSAref.
17858 *Ralf S. Engelschall*
17860 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17862 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17864 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17869 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17870 DH parameters/keys (q is lost during that conversion, but the resulting
17871 DH parameters contain its length).
17873 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17874 much faster than DH_generate_parameters (which creates parameters
17875 where `p = 2*q + 1`), and also the smaller q makes DH computations
17876 much more efficient (160-bit exponentiation instead of 1024-bit
17877 exponentiation); so this provides a convenient way to support DHE
17878 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17879 utter importance to use
17880 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17882 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17883 when such DH parameters are used, because otherwise small subgroup
17884 attacks may become possible!
17888 * Avoid memory leak in i2d_DHparams.
17892 * Allow the -k option to be used more than once in the enc program:
17893 this allows the same encrypted message to be read by multiple recipients.
17897 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17898 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17899 it will always use the numerical form of the OID, even if it has a short
17904 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17905 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17906 otherwise bn_mod_exp was called. In the case of hardware keys for example
17907 no private key components need be present and it might store extra data
17908 in the RSA structure, which cannot be accessed from bn_mod_exp.
17909 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17910 private key operations.
17914 * Added support for SPARC Linux.
17918 * pem_password_cb function type incompatibly changed from
17919 typedef int pem_password_cb(char *buf, int size, int rwflag);
17921 ....(char *buf, int size, int rwflag, void *userdata);
17922 so that applications can pass data to their callbacks:
17923 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17924 additional void * argument, which is just handed through whenever
17925 the password callback is called.
17927 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17929 New function SSL_CTX_set_default_passwd_cb_userdata.
17931 Compatibility note: As many C implementations push function arguments
17932 onto the stack in reverse order, the new library version is likely to
17933 interoperate with programs that have been compiled with the old
17934 pem_password_cb definition (PEM_whatever takes some data that
17935 happens to be on the stack as its last argument, and the callback
17936 just ignores this garbage); but there is no guarantee whatsoever that
17939 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17940 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17941 problems not only on Windows, but also on some Unix platforms.
17942 To avoid problematic command lines, these definitions are now in an
17943 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17944 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17948 * MIPS III/IV assembler module is reimplemented.
17952 * More DES library cleanups: remove references to srand/rand and
17953 delete an unused file.
17957 * Add support for the free Netwide assembler (NASM) under Win32,
17958 since not many people have MASM (ml) and it can be hard to obtain.
17959 This is currently experimental but it seems to work OK and pass all
17960 the tests. Check out INSTALL.W32 for info.
17964 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17965 without temporary keys kept an extra copy of the server key,
17966 and connections with temporary keys did not free everything in case
17971 * New function RSA_check_key and new openssl rsa option -check
17972 for verifying the consistency of RSA keys.
17974 *Ulf Moeller, Bodo Moeller*
17976 * Various changes to make Win32 compile work:
17977 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17978 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17979 comparison" warnings.
17980 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17984 * Add a debugging option to PKCS#5 v2 key generation function: when
17985 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17986 derived keys are printed to stderr.
17990 * Copy the flags in ASN1_STRING_dup().
17992 *Roman E. Pavlov <pre@mo.msk.ru>*
17994 * The x509 application mishandled signing requests containing DSA
17995 keys when the signing key was also DSA and the parameters didn't match.
17997 It was supposed to omit the parameters when they matched the signing key:
17998 the verifying software was then supposed to automatically use the CA's
17999 parameters if they were absent from the end user certificate.
18001 Omitting parameters is no longer recommended. The test was also
18002 the wrong way round! This was probably due to unusual behaviour in
18003 EVP_cmp_parameters() which returns 1 if the parameters match.
18004 This meant that parameters were omitted when they *didn't* match and
18005 the certificate was useless. Certificates signed with 'ca' didn't have
18008 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18010 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18011 The interface is as follows:
18012 Applications can use
18013 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18014 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18015 "off" is now the default.
18016 The library internally uses
18017 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18018 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18019 to disable memory-checking temporarily.
18021 Some inconsistent states that previously were possible (and were
18022 even the default) are now avoided.
18024 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18025 with each memory chunk allocated; this is occasionally more helpful
18026 than just having a counter.
18028 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18030 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18035 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18036 which largely parallels "options", but is for changing API behaviour,
18037 whereas "options" are about protocol behaviour.
18038 Initial "mode" flags are:
18040 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18041 a single record has been written.
18042 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18043 retries use the same buffer location.
18044 (But all of the contents must be
18049 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18052 * Fix problems with no-hmac etc.
18054 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18056 * New functions RSA_get_default_method(), RSA_set_method() and
18057 RSA_get_method(). These allows replacement of RSA_METHODs without having
18058 to mess around with the internals of an RSA structure.
18062 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18063 Also really enable memory leak checks in openssl.c and in some
18066 *Chad C. Mulligan, Bodo Moeller*
18068 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18069 up the length of negative integers. This has now been simplified to just
18070 store the length when it is first determined and use it later, rather
18071 than trying to keep track of where data is copied and updating it to
18073 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18075 * Add a new function PKCS7_signatureVerify. This allows the verification
18076 of a PKCS#7 signature but with the signing certificate passed to the
18077 function itself. This contrasts with PKCS7_dataVerify which assumes the
18078 certificate is present in the PKCS#7 structure. This isn't always the
18079 case: certificates can be omitted from a PKCS#7 structure and be
18080 distributed by "out of band" means (such as a certificate database).
18084 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18085 function prototypes in pem.h, also change util/mkdef.pl to add the
18086 necessary function names.
18090 * mk1mf.pl (used by Windows builds) did not properly read the
18091 options set by Configure in the top level Makefile, and Configure
18092 was not even able to write more than one option correctly.
18093 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18097 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18098 file to be loaded from a BIO or FILE pointer. The BIO version will
18099 for example allow memory BIOs to contain config info.
18103 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18104 Whoever hopes to achieve shared-library compatibility across versions
18105 must use this, not the compile-time macro.
18106 (Exercise 0.9.4: Which is the minimum library version required by
18108 Note: All this applies only to multi-threaded programs, others don't
18113 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18114 through a BIO pair triggered the default case, i.e.
18115 SSLerr(...,SSL_R_UNKNOWN_STATE).
18119 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18120 can use the SSL library even if none of the specific BIOs is
18125 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18126 for the encoded length.
18128 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18130 * Add initial documentation of the X509V3 functions.
18134 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18135 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18136 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18137 secure PKCS#8 private key format with a high iteration count.
18141 * Fix determination of Perl interpreter: A perl or perl5
18142 *directory* in $PATH was also accepted as the interpreter.
18144 *Ralf S. Engelschall*
18146 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18147 wrong with it but it was very old and did things like calling
18148 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18149 unusual formatting.
18153 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18154 to use the new extension code.
18158 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18159 with macros. This should make it easier to change their form, add extra
18160 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18165 * Add to configuration table a new entry that can specify an alternative
18166 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18167 according to Mark Crispin <MRC@Panda.COM>.
18171 * DES CBC did not update the IV. Weird.
18175 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18176 Changing the behaviour of the former might break existing programs --
18177 where IV updating is needed, des_ncbc_encrypt can be used.
18180 * When bntest is run from "make test" it drives bc to check its
18181 calculations, as well as internally checking them. If an internal check
18182 fails, it needs to cause bc to give a non-zero result or make test carries
18183 on without noticing the failure. Fixed.
18187 * DES library cleanups.
18191 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18192 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18193 ciphers. NOTE: although the key derivation function has been verified
18194 against some published test vectors it has not been extensively tested
18195 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18200 * Instead of "mkdir -p", which is not fully portable, use new
18201 Perl script "util/mkdir-p.pl".
18205 * Rewrite the way password based encryption (PBE) is handled. It used to
18206 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18207 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18208 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18209 the 'parameter' field of the AlgorithmIdentifier is passed to the
18210 underlying key generation function so it must do its own ASN1 parsing.
18211 This has also changed the EVP_PBE_CipherInit() function which now has a
18212 'parameter' argument instead of literal salt and iteration count values
18213 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18217 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18218 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18219 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18220 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18221 value was just used as a "magic string" and not used directly its
18222 value doesn't matter.
18226 * Introduce some semblance of const correctness to BN. Shame C doesn't
18231 * "linux-sparc64" configuration (ultrapenguin).
18233 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18234 "linux-sparc" configuration.
18236 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18238 * config now generates no-xxx options for missing ciphers.
18242 * Support the EBCDIC character set (work in progress).
18243 File ebcdic.c not yet included because it has a different license.
18245 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18247 * Support BS2000/OSD-POSIX.
18249 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18251 * Make callbacks for key generation use `void *` instead of `char *`.
18255 * Make S/MIME samples compile (not yet tested).
18259 * Additional typesafe stacks.
18263 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18267 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18269 * New configuration variant "sco5-gcc".
18271 * Updated some demos.
18273 *Sean O Riordain, Wade Scholine*
18275 * Add missing BIO_free at exit of pkcs12 application.
18279 * Fix memory leak in conf.c.
18283 * Updates for Win32 to assembler version of MD5.
18287 * Set #! path to perl in `apps/der_chop` to where we found it
18288 instead of using a fixed path.
18292 * SHA library changes for irix64-mips4-cc.
18296 * Improvements for VMS support.
18300 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18302 * Bignum library bug fix. IRIX 6 passes "make test" now!
18303 This also avoids the problems with SC4.2 and unpatched SC5.
18305 *Andy Polyakov <appro@fy.chalmers.se>*
18307 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18308 These are required because of the typesafe stack would otherwise break
18309 existing code. If old code used a structure member which used to be STACK
18310 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18311 sk_num or sk_value it would produce an error because the num, data members
18312 are not present in STACK_OF. Now it just produces a warning. sk_set
18313 replaces the old method of assigning a value to sk_value
18314 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18315 that does this will no longer work (and should use sk_set instead) but
18316 this could be regarded as a "questionable" behaviour anyway.
18320 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18321 correctly handle encrypted S/MIME data.
18325 * Change type of various DES function arguments from des_cblock
18326 (which means, in function argument declarations, pointer to char)
18327 to des_cblock * (meaning pointer to array with 8 char elements),
18328 which allows the compiler to do more typechecking; it was like
18329 that back in SSLeay, but with lots of ugly casts.
18331 Introduce new type const_des_cblock.
18335 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18336 problems: find RecipientInfo structure that matches recipient certificate
18337 and initialise the ASN1 structures properly based on passed cipher.
18341 * Belatedly make the BN tests actually check the results.
18345 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18346 to and from BNs: it was completely broken. New compilation option
18347 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18348 key elements as negative integers.
18352 * Reorganize and speed up MD5.
18354 *Andy Polyakov <appro@fy.chalmers.se>*
18358 *Richard Levitte <richard@levitte.org>*
18360 * New option -out to asn1parse to allow the parsed structure to be
18361 output to a file. This is most useful when combined with the -strparse
18362 option to examine the output of things like OCTET STRINGS.
18366 * Make SSL library a little more fool-proof by not requiring any longer
18367 that `SSL_set_{accept,connect}_state` be called before
18368 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18369 in many applications because usually everything *appeared* to work as
18370 intended anyway -- now it really works as intended).
18374 * Move openssl.cnf out of lib/.
18378 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18379 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18380 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18382 *Ralf S. Engelschall*
18384 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18385 handle PKCS#7 enveloped data properly.
18387 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18389 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18390 copying pointers. The cert_st handling is changed by this in
18391 various ways (and thus what used to be known as ctx->default_cert
18392 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18393 any longer when s->cert does not give us what we need).
18394 ssl_cert_instantiate becomes obsolete by this change.
18395 As soon as we've got the new code right (possibly it already is?),
18396 we have solved a couple of bugs of the earlier code where s->cert
18397 was used as if it could not have been shared with other SSL structures.
18399 Note that using the SSL API in certain dirty ways now will result
18400 in different behaviour than observed with earlier library versions:
18401 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18402 does not influence s as it used to.
18404 In order to clean up things more thoroughly, inside SSL_SESSION
18405 we don't use CERT any longer, but a new structure SESS_CERT
18406 that holds per-session data (if available); currently, this is
18407 the peer's certificate chain and, for clients, the server's certificate
18408 and temporary key. CERT holds only those values that can have
18409 meaningful defaults in an SSL_CTX.
18413 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18414 from the internal representation. Various PKCS#7 fixes: remove some
18415 evil casts and set the enc_dig_alg field properly based on the signing
18420 * Allow PKCS#12 password to be set from the command line or the
18421 environment. Let 'ca' get its config file name from the environment
18422 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18427 * Allow certificate policies extension to use an IA5STRING for the
18428 organization field. This is contrary to the PKIX definition but
18429 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18434 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18435 without disallowing inline assembler and the like for non-pedantic builds.
18439 * Support Borland C++ builder.
18441 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18447 * SHA-1 cleanups and performance enhancements.
18449 *Andy Polyakov <appro@fy.chalmers.se>*
18451 * Sparc v8plus assembler for the bignum library.
18453 *Andy Polyakov <appro@fy.chalmers.se>*
18455 * Accept any -xxx and +xxx compiler options in Configure.
18459 * Update HPUX configuration.
18463 * Add missing `sk_<type>_unshift()` function to safestack.h
18465 *Ralf S. Engelschall*
18467 * New function SSL_CTX_use_certificate_chain_file that sets the
18468 "extra_cert"s in addition to the certificate. (This makes sense
18469 only for "PEM" format files, as chains as a whole are not
18474 * Support verify_depth from the SSL API.
18475 x509_vfy.c had what can be considered an off-by-one-error:
18476 Its depth (which was not part of the external interface)
18477 was actually counting the number of certificates in a chain;
18478 now it really counts the depth.
18482 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18483 instead of X509err, which often resulted in confusing error
18484 messages since the error codes are not globally unique
18485 (e.g. an alleged error in ssl3_accept when a certificate
18486 didn't match the private key).
18488 * New function SSL_CTX_set_session_id_context that allows to set a default
18489 value (so that you don't need SSL_set_session_id_context for each
18490 connection using the SSL_CTX).
18494 * OAEP decoding bug fix.
18498 * Support INSTALL_PREFIX for package builders, as proposed by
18503 * New Configure options "threads" and "no-threads". For systems
18504 where the proper compiler options are known (currently Solaris
18505 and Linux), "threads" is the default.
18509 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18513 * Install various scripts to $(OPENSSLDIR)/misc, not to
18514 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18515 such as /usr/local/bin.
18519 * "make linux-shared" to build shared libraries.
18521 *Niels Poppe <niels@netbox.org>*
18523 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18527 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18528 extension adding in x509 utility.
18532 * Remove NOPROTO sections and error code comments.
18536 * Partial rewrite of the DEF file generator to now parse the ANSI
18541 * New Configure options --prefix=DIR and --openssldir=DIR.
18545 * Complete rewrite of the error code script(s). It is all now handled
18546 by one script at the top level which handles error code gathering,
18547 header rewriting and C source file generation. It should be much better
18548 than the old method: it now uses a modified version of Ulf's parser to
18549 read the ANSI prototypes in all header files (thus the old K&R definitions
18550 aren't needed for error creation any more) and do a better job of
18551 translating function codes into names. The old 'ASN1 error code embedded
18552 in a comment' is no longer necessary and it doesn't use .err files which
18553 have now been deleted. Also the error code call doesn't have to appear all
18554 on one line (which resulted in some large lines...).
18558 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18562 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18563 0 (which usually indicates a closed connection), but continue reading.
18567 * Fix some race conditions.
18571 * Add support for CRL distribution points extension. Add Certificate
18572 Policies and CRL distribution points documentation.
18576 * Move the autogenerated header file parts to crypto/opensslconf.h.
18580 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18581 8 of keying material. Merlin has also confirmed interop with this fix
18582 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18584 *Merlin Hughes <merlin@baltimore.ie>*
18586 * Fix lots of warnings.
18588 *Richard Levitte <levitte@stacken.kth.se>*
18590 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18591 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18593 *Richard Levitte <levitte@stacken.kth.se>*
18595 * Fix problems with sizeof(long) == 8.
18597 *Andy Polyakov <appro@fy.chalmers.se>*
18599 * Change functions to ANSI C.
18603 * Fix typos in error codes.
18605 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18607 * Remove defunct assembler files from Configure.
18611 * SPARC v8 assembler BIGNUM implementation.
18613 *Andy Polyakov <appro@fy.chalmers.se>*
18615 * Support for Certificate Policies extension: both print and set.
18616 Various additions to support the r2i method this uses.
18620 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18621 return a const string when you are expecting an allocated buffer.
18625 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18626 types DirectoryString and DisplayText.
18630 * Add code to allow r2i extensions to access the configuration database,
18631 add an LHASH database driver and add several ctx helper functions.
18635 * Fix an evil bug in bn_expand2() which caused various BN functions to
18636 fail when they extended the size of a BIGNUM.
18640 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18641 support typesafe stack.
18645 * Fix typo in SSL_[gs]et_options().
18647 *Nils Frostberg <nils@medcom.se>*
18649 * Delete various functions and files that belonged to the (now obsolete)
18650 old X509V3 handling code.
18654 * New Configure option "rsaref".
18658 * Don't auto-generate pem.h.
18662 * Introduce type-safe ASN.1 SETs.
18666 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18668 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18670 * Introduce type-safe STACKs. This will almost certainly break lots of code
18671 that links with OpenSSL (well at least cause lots of warnings), but fear
18672 not: the conversion is trivial, and it eliminates loads of evil casts. A
18673 few STACKed things have been converted already. Feel free to convert more.
18674 In the fullness of time, I'll do away with the STACK type altogether.
18678 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18679 specified in `<certfile>` by updating the entry in the index.txt file.
18680 This way one no longer has to edit the index.txt file manually for
18681 revoking a certificate. The -revoke option does the gory details now.
18683 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18685 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18686 `-text` option at all and this way the `-noout -text` combination was
18687 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18689 *Ralf S. Engelschall*
18691 * Make sure a corresponding plain text error message exists for the
18692 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18693 verify callback function determined that a certificate was revoked.
18695 *Ralf S. Engelschall*
18697 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18698 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18699 all available ciphers including rc5, which was forgotten until now.
18700 In order to let the testing shell script know which algorithms
18701 are available, a new (up to now undocumented) command
18702 `openssl list-cipher-commands` is used.
18706 * Bugfix: s_client occasionally would sleep in select() when
18707 it should have checked SSL_pending() first.
18711 * New functions DSA_do_sign and DSA_do_verify to provide access to
18712 the raw DSA values prior to ASN.1 encoding.
18716 * Tweaks to Configure
18718 *Niels Poppe <niels@netbox.org>*
18720 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18725 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18729 * New config option to avoid instructions that are illegal on the 80386.
18730 The default code is faster, but requires at least a 486.
18734 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18735 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18736 same as SSL2_VERSION anyway.
18740 * New "-showcerts" option for s_client.
18744 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18745 application. Various cleanups and fixes.
18749 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18750 modify error routines to work internally. Add error codes and PBE init
18751 to library startup routines.
18755 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18756 packing functions to asn1 and evp. Changed function names and error
18757 codes along the way.
18761 * PKCS12 integration: and so it begins... First of several patches to
18762 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18763 objects to objects.h
18767 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18768 and display support for Thawte strong extranet extension.
18772 * Add LinuxPPC support.
18774 *Jeff Dubrule <igor@pobox.org>*
18776 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18777 bn_div_words in alpha.s.
18779 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18781 * Make sure the RSA OAEP test is skipped under -DRSAref because
18782 OAEP isn't supported when OpenSSL is built with RSAref.
18784 *Ulf Moeller <ulf@fitug.de>*
18786 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18787 so they no longer are missing under -DNOPROTO.
18789 *Soren S. Jorvang <soren@t.dk>*
18791 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18793 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18794 doesn't work when the session is reused. Coming soon!
18798 * Fix a security hole, that allows sessions to be reused in the wrong
18799 context thus bypassing client cert protection! All software that uses
18800 client certs and session caches in multiple contexts NEEDS PATCHING to
18801 allow session reuse! A fuller solution is in the works.
18803 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18805 * Some more source tree cleanups (removed obsolete files
18806 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18807 permission on "config" script to be executable) and a fix for the INSTALL
18810 *Ulf Moeller <ulf@fitug.de>*
18812 * Remove some legacy and erroneous uses of malloc, free instead of
18815 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18817 * Make rsa_oaep_test return non-zero on error.
18819 *Ulf Moeller <ulf@fitug.de>*
18821 * Add support for native Solaris shared libraries. Configure
18822 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18823 if someone would make that last step automatic.
18825 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18827 * ctx_size was not built with the right compiler during "make links". Fixed.
18831 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18832 except NULL ciphers". This means the default cipher list will no longer
18833 enable NULL ciphers. They need to be specifically enabled e.g. with
18834 the string "DEFAULT:eNULL".
18838 * Fix to RSA private encryption routines: if p < q then it would
18839 occasionally produce an invalid result. This will only happen with
18840 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18844 * Be less restrictive and allow also `perl util/perlpath.pl
18845 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18846 because this way one can also use an interpreter named `perl5` (which is
18847 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18848 installed as `perl`).
18850 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18852 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18854 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18856 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18857 advapi32.lib to Win32 build and change the pem test comparison
18858 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18859 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18860 and crypto/des/ede_cbcm_enc.c.
18864 * DES quad checksum was broken on big-endian architectures. Fixed.
18868 * Comment out two functions in bio.h that aren't implemented. Fix up the
18869 Win32 test batch file so it (might) work again. The Win32 test batch file
18870 is horrible: I feel ill....
18874 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18875 in e_os.h. Audit of header files to check ANSI and non ANSI
18876 sections: 10 functions were absent from non ANSI section and not exported
18877 from Windows DLLs. Fixed up libeay.num for new functions.
18881 * Make `openssl version` output lines consistent.
18883 *Ralf S. Engelschall*
18885 * Fix Win32 symbol export lists for BIO functions: Added
18886 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18887 to ms/libeay{16,32}.def.
18889 *Ralf S. Engelschall*
18891 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18892 fine under Unix and passes some trivial tests I've now added. But the
18893 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18894 added to make sure no one expects that this stuff really works in the
18895 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18896 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18899 *Ralf S. Engelschall*
18901 * Fix the generation of two part addresses in perl.
18903 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18905 * Add config entry for Linux on MIPS.
18907 *John Tobey <jtobey@channel1.com>*
18909 * Make links whenever Configure is run, unless we are on Windoze.
18913 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18914 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18919 * Add a useful kludge to allow package maintainers to specify compiler and
18920 other platforms details on the command line without having to patch the
18921 Configure script every time: One now can use
18922 `perl Configure <id>:<details>`,
18923 i.e. platform ids are allowed to have details appended
18924 to them (separated by colons). This is treated as there would be a static
18925 pre-configured entry in Configure's %table under key `<id>` with value
18926 `<details>` and `perl Configure <id>` is called. So, when you want to
18927 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18928 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18929 now, which overrides the FreeBSD-elf entry on-the-fly.
18931 *Ralf S. Engelschall*
18933 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18937 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18938 on the `perl Configure ...` command line. This way one can compile
18939 OpenSSL libraries with Position Independent Code (PIC) which is needed
18940 for linking it into DSOs.
18942 *Ralf S. Engelschall*
18944 * Remarkably, export ciphers were totally broken and no-one had noticed!
18949 * Cleaned up the LICENSE document: The official contact for any license
18950 questions now is the OpenSSL core team under openssl-core@openssl.org.
18951 And add a paragraph about the dual-license situation to make sure people
18952 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18953 to the OpenSSL toolkit.
18955 *Ralf S. Engelschall*
18957 * General source tree makefile cleanups: Made `making xxx in yyy...`
18958 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18959 Additionally cleaned up the `make links` target: Remove unnecessary
18960 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18961 to speed processing and no longer clutter the display with confusing
18962 stuff. Instead only the actually done links are displayed.
18964 *Ralf S. Engelschall*
18966 * Permit null encryption ciphersuites, used for authentication only. It used
18967 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18968 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18973 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18974 signed attributes when verifying signatures (this would break them),
18975 the detached data encoding was wrong and public keys obtained using
18976 X509_get_pubkey() weren't freed.
18980 * Add text documentation for the BUFFER functions. Also added a work around
18981 to a Win95 console bug. This was triggered by the password read stuff: the
18982 last character typed gets carried over to the next fread(). If you were
18983 generating a new cert request using 'req' for example then the last
18984 character of the passphrase would be CR which would then enter the first
18989 * Added the new 'Includes OpenSSL Cryptography Software' button as
18990 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18991 button and can be used by applications based on OpenSSL to show the
18992 relationship to the OpenSSL project.
18994 *Ralf S. Engelschall*
18996 * Remove confusing variables in function signatures in files
18997 ssl/ssl_lib.c and ssl/ssl.h.
18999 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19001 * Don't install bss_file.c under PREFIX/include/
19003 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19005 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19006 functions that return function pointers and has support for NT specific
19007 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19008 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19009 unsigned to signed types: this was killing the Win32 compile.
19013 * Add new certificate file to stack functions,
19014 SSL_add_dir_cert_subjects_to_stack() and
19015 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19016 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19017 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19018 This means that Apache-SSL and similar packages don't have to mess around
19019 to add as many CAs as they want to the preferred list.
19023 * Experiment with doxygen documentation. Currently only partially applied to
19025 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19026 openssl.doxy as the configuration file.
19030 * Get rid of remaining C++-style comments which strict C compilers hate.
19032 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19034 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19035 compiled in by default: it has problems with large keys.
19039 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19040 DH private keys and/or callback functions which directly correspond to
19041 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19042 is needed for applications which have to configure certificates on a
19043 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19045 For the RSA certificate situation is makes no difference, but
19046 for the DSA certificate situation this fixes the "no shared cipher"
19047 problem where the OpenSSL cipher selection procedure failed because the
19048 temporary keys were not overtaken from the context and the API provided
19049 no way to reconfigure them.
19050 The new functions now let applications reconfigure the stuff and they
19051 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19052 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19053 non-public-API function ssl_cert_instantiate() is used as a helper
19054 function and also to reduce code redundancy inside ssl_rsa.c.
19056 *Ralf S. Engelschall*
19058 * Move s_server -dcert and -dkey options out of the undocumented feature
19059 area because they are useful for the DSA situation and should be
19060 recognized by the users.
19062 *Ralf S. Engelschall*
19064 * Fix the cipher decision scheme for export ciphers: the export bits are
19065 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19066 SSL_EXP_MASK. So, the original variable has to be used instead of the
19067 already masked variable.
19069 *Richard Levitte <levitte@stacken.kth.se>*
19071 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19073 *Richard Levitte <levitte@stacken.kth.se>*
19075 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19076 from `int` to `unsigned int` because it is a length and initialized by
19077 EVP_DigestFinal() which expects an `unsigned int *`.
19079 *Richard Levitte <levitte@stacken.kth.se>*
19081 * Don't hard-code path to Perl interpreter on shebang line of Configure
19082 script. Instead use the usual Shell->Perl transition trick.
19084 *Ralf S. Engelschall*
19086 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19087 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19088 -noout -modulus` as it's already the case for `openssl rsa -noout
19089 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19090 currently the public key is printed (a decision which was already done by
19091 `openssl dsa -modulus` in the past) which serves a similar purpose.
19092 Additionally the NO_RSA no longer completely removes the whole -modulus
19093 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19096 *Ralf S. Engelschall*
19098 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19099 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19101 *Arne Ansper <arne@ats.cyber.ee>*
19103 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19104 to be added. Now both 'req' and 'ca' can use new objects defined in the
19109 * Add cool BIO that does syslog (or event log on NT).
19111 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19113 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19114 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19115 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19116 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19120 * Add preliminary config info for new extension code.
19124 * Make RSA_NO_PADDING really use no padding.
19126 *Ulf Moeller <ulf@fitug.de>*
19128 * Generate errors when private/public key check is done.
19132 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19133 for some CRL extensions and new objects added.
19137 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19138 key usage extension and fuller support for authority key id.
19142 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19143 padding method for RSA, which is recommended for new applications in PKCS
19144 #1 v2.0 (RFC 2437, October 1998).
19145 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19146 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19147 against Bleichbacher's attack on RSA.
19148 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19151 * Updates to the new SSL compression code
19153 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19155 * Fix so that the version number in the master secret, when passed
19156 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19157 (because the server will not accept higher), that the version number
19158 is 0x03,0x01, not 0x03,0x00
19160 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19162 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19163 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19164 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19168 * Support for RAW extensions where an arbitrary extension can be
19169 created by including its DER encoding. See `apps/openssl.cnf` for
19174 * Make sure latest Perl versions don't interpret some generated C array
19175 code as Perl array code in the crypto/err/err_genc.pl script.
19177 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19179 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19180 not many people have the assembler. Various Win32 compilation fixes and
19181 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19182 build instructions.
19186 * Modify configure script 'Configure' to automatically create crypto/date.h
19187 file under Win32 and also build pem.h from pem.org. New script
19188 util/mkfiles.pl to create the MINFO file on environments that can't do a
19189 'make files': perl util/mkfiles.pl >MINFO should work.
19193 * Major rework of DES function declarations, in the pursuit of correctness
19194 and purity. As a result, many evil casts evaporated, and some weirdness,
19195 too. You may find this causes warnings in your code. Zapping your evil
19196 casts will probably fix them. Mostly.
19200 * Fix for a typo in asn1.h. Bug fix to object creation script
19201 obj_dat.pl. It considered a zero in an object definition to mean
19202 "end of object": none of the objects in objects.h have any zeros
19203 so it wasn't spotted.
19205 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19207 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19208 Masking (CBCM). In the absence of test vectors, the best I have been able
19209 to do is check that the decrypt undoes the encrypt, so far. Send me test
19210 vectors if you have them.
19214 * Correct calculation of key length for export ciphers (too much space was
19215 allocated for null ciphers). This has not been tested!
19219 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19220 message is now correct (it understands "crypto" and "ssl" on its
19221 command line). There is also now an "update" option. This will update
19222 the util/ssleay.num and util/libeay.num files with any new functions.
19224 perl util/mkdef.pl crypto ssl update
19225 it will update them.
19229 * Overhauled the Perl interface:
19230 - ported BN stuff to OpenSSL's different BN library
19231 - made the perl/ source tree CVS-aware
19232 - renamed the package from SSLeay to OpenSSL (the files still contain
19233 their history because I've copied them in the repository)
19234 - removed obsolete files (the test scripts will be replaced
19235 by better Test::Harness variants in the future)
19237 *Ralf S. Engelschall*
19239 * First cut for a very conservative source tree cleanup:
19240 1. merge various obsolete readme texts into doc/ssleay.txt
19241 where we collect the old documents and readme texts.
19242 2. remove the first part of files where I'm already sure that we no
19243 longer need them because of three reasons: either they are just temporary
19244 files which were left by Eric or they are preserved original files where
19245 I've verified that the diff is also available in the CVS via "cvs diff
19246 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19247 the crypto/md/ stuff).
19249 *Ralf S. Engelschall*
19251 * More extension code. Incomplete support for subject and issuer alt
19252 name, issuer and authority key id. Change the i2v function parameters
19253 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19254 what that's for :-) Fix to ASN1 macro which messed up
19255 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19259 * Preliminary support for ENUMERATED type. This is largely copied from the
19264 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19266 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19268 * Make sure `make rehash` target really finds the `openssl` program.
19270 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19272 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19273 like to hear about it if this slows down other processors.
19277 * Add CygWin32 platform information to Configure script.
19279 *Alan Batie <batie@aahz.jf.intel.com>*
19281 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19283 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19285 * New program nseq to manipulate netscape certificate sequences
19289 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19294 * Fixes to BN code. Previously the default was to define BN_RECURSION
19295 but the BN code had some problems that would cause failures when
19296 doing certificate verification and some other functions.
19298 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19300 * Add ASN1 and PEM code to support netscape certificate sequences.
19304 * Add ASN1 and PEM code to support netscape certificate sequences.
19308 * Add several PKIX and private extended key usage OIDs.
19312 * Modify the 'ca' program to handle the new extension code. Modify
19313 openssl.cnf for new extension format, add comments.
19317 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19318 and add a sample to openssl.cnf so req -x509 now adds appropriate
19323 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19324 error code, add initial support to X509_print() and x509 application.
19328 * Takes a deep breath and start adding X509 V3 extension support code. Add
19329 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19330 stuff is currently isolated and isn't even compiled yet.
19334 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19335 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19336 Removed the versions check from X509 routines when loading extensions:
19337 this allows certain broken certificates that don't set the version
19338 properly to be processed.
19342 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19343 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19344 can still be regenerated with "make depend".
19348 * Spelling mistake in C version of CAST-128.
19350 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19352 * Changes to the error generation code. The perl script err-code.pl
19353 now reads in the old error codes and retains the old numbers, only
19354 adding new ones if necessary. It also only changes the .err files if new
19355 codes are added. The makefiles have been modified to only insert errors
19356 when needed (to avoid needlessly modifying header files). This is done
19357 by only inserting errors if the .err file is newer than the auto generated
19358 C file. To rebuild all the error codes from scratch (the old behaviour)
19359 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19360 or delete all the .err files.
19364 * CAST-128 was incorrectly implemented for short keys. The C version has
19365 been fixed, but is untested. The assembler versions are also fixed, but
19366 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19367 to regenerate it if needed.
19368 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19369 Hagino <itojun@kame.net>*
19371 * File was opened incorrectly in randfile.c.
19373 *Ulf Möller <ulf@fitug.de>*
19375 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19376 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19377 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19378 al: it's just almost always a UTCTime. Note this patch adds new error
19379 codes so do a "make errors" if there are problems.
19383 * Correct Linux 1 recognition in config.
19385 *Ulf Möller <ulf@fitug.de>*
19387 * Remove pointless MD5 hash when using DSA keys in ca.
19389 *Anonymous <nobody@replay.com>*
19391 * Generate an error if given an empty string as a cert directory. Also
19392 generate an error if handed NULL (previously returned 0 to indicate an
19393 error, but didn't set one).
19395 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19397 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19401 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19402 parameters. This was causing a warning which killed off the Win32 compile.
19406 * Remove C++ style comments from crypto/bn/bn_local.h.
19408 *Neil Costigan <neil.costigan@celocom.com>*
19410 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19411 based on a text string, looking up short and long names and finally
19412 "dot" format. The "dot" format stuff didn't work. Added new function
19413 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19414 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19415 OID is not part of the table.
19419 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19420 X509_LOOKUP_by_alias().
19424 * Sort openssl functions by name.
19428 * Get the `gendsa` command working and add it to the `list` command. Remove
19429 encryption from sample DSA keys (in case anyone is interested the password
19434 * Make *all* `*_free` functions accept a NULL pointer.
19436 *Frans Heymans <fheymans@isaserver.be>*
19438 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19441 *Anonymous <nobody@replay.com>*
19443 * s_server should send the CAfile as acceptable CAs, not its own cert.
19445 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19447 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19449 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19451 * Temp key "for export" tests were wrong in s3_srvr.c.
19453 *Anonymous <nobody@replay.com>*
19455 * Add prototype for temp key callback functions
19456 SSL_CTX_set_tmp_{rsa,dh}_callback().
19460 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19461 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19465 * X509_name_add_entry() freed the wrong thing after an error.
19467 *Arne Ansper <arne@ats.cyber.ee>*
19469 * rsa_eay.c would attempt to free a NULL context.
19471 *Arne Ansper <arne@ats.cyber.ee>*
19473 * BIO_s_socket() had a broken should_retry() on Windoze.
19475 *Arne Ansper <arne@ats.cyber.ee>*
19477 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19479 *Arne Ansper <arne@ats.cyber.ee>*
19481 * Make sure the already existing X509_STORE->depth variable is initialized
19482 in X509_STORE_new(), but document the fact that this variable is still
19483 unused in the certificate verification process.
19485 *Ralf S. Engelschall*
19487 * Fix the various library and `apps/` files to free up pkeys obtained from
19488 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19492 * Fix reference counting in X509_PUBKEY_get(). This makes
19493 demos/maurice/example2.c work, amongst others, probably.
19495 *Steve Henson and Ben Laurie*
19497 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19498 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19499 are no longer created. This way we have a single and consistent command
19500 line interface `openssl <command>`, similar to `cvs <command>`.
19502 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19504 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19505 BIT STRING wrapper always have zero unused bits.
19509 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19513 * Make the top-level INSTALL documentation easier to understand.
19517 * Makefiles updated to exit if an error occurs in a sub-directory
19518 make (including if user presses ^C) [Paul Sutton]
19520 * Make Montgomery context stuff explicit in RSA data structure.
19524 * Fix build order of pem and err to allow for generated pem.h.
19528 * Fix renumbering bug in X509_NAME_delete_entry().
19532 * Enhanced the err-ins.pl script so it makes the error library number
19533 global and can add a library name. This is needed for external ASN1 and
19534 other error libraries.
19538 * Fixed sk_insert which never worked properly.
19542 * Fix ASN1 macros so they can handle indefinite length constructed
19543 EXPLICIT tags. Some non standard certificates use these: they can now
19548 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19549 into a single doc/ssleay.txt bundle. This way the information is still
19550 preserved but no longer messes up this directory. Now it's new room for
19551 the new set of documentation files.
19553 *Ralf S. Engelschall*
19555 * SETs were incorrectly DER encoded. This was a major pain, because they
19556 shared code with SEQUENCEs, which aren't coded the same. This means that
19557 almost everything to do with SETs or SEQUENCEs has either changed name or
19558 number of arguments.
19560 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19562 * Fix test data to work with the above.
19566 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19567 was already fixed by Eric for 0.9.1 it seems.
19569 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19571 * Autodetect FreeBSD3.
19575 * Fix various bugs in Configure. This affects the following platforms:
19579 unixware-2.0-pentium
19584 * Eliminate generated files from CVS. Reorder tests to regenerate files
19585 before they are needed.
19589 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19593 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19595 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19596 changed SSLeay to OpenSSL in version strings.
19598 *Ralf S. Engelschall*
19600 * Some fixups to the top-level documents.
19604 * Fixed the nasty bug where rsaref.h was not found under compile-time
19605 because the symlink to include/ was missing.
19607 *Ralf S. Engelschall*
19609 * Incorporated the popular no-RSA/DSA-only patches
19610 which allow to compile an RSA-free SSLeay.
19612 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19614 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19615 when "ssleay" is still not found.
19617 *Ralf S. Engelschall*
19619 * Added more platforms to Configure: Cray T3E, HPUX 11,
19621 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19623 * Updated the README file.
19625 *Ralf S. Engelschall*
19627 * Added various .cvsignore files in the CVS repository subdirs
19628 to make a "cvs update" really silent.
19630 *Ralf S. Engelschall*
19632 * Recompiled the error-definition header files and added
19633 missing symbols to the Win32 linker tables.
19635 *Ralf S. Engelschall*
19637 * Cleaned up the top-level documents;
19638 o new files: CHANGES and LICENSE
19639 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19640 o merged COPYRIGHT into LICENSE
19641 o removed obsolete TODO file
19642 o renamed MICROSOFT to INSTALL.W32
19644 *Ralf S. Engelschall*
19646 * Removed dummy files from the 0.9.1b source tree:
19647 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19648 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19649 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19650 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19651 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19653 *Ralf S. Engelschall*
19655 * Added various platform portability fixes.
19659 * The Genesis of the OpenSSL rpject:
19660 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19661 Young and Tim J. Hudson created while they were working for C2Net until
19664 *The OpenSSL Project*
19666 ### Changes between 0.9.0b and 0.9.1b [not released]
19668 * Updated a few CA certificates under certs/
19672 * Changed some BIGNUM api stuff.
19676 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19677 DGUX x86, Linux Alpha, etc.
19681 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19682 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19687 * Add -strparse option to asn1pars program which parses nested
19690 *Dr Stephen Henson <shenson@bigfoot.com>*
19692 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19696 * DSA fix for "ca" program.
19700 * Added "-genkey" option to "dsaparam" program.
19704 * Added RIPE MD160 (rmd160) message digest.
19708 * Added -a (all) option to "ssleay version" command.
19712 * Added PLATFORM define which is the id given to Configure.
19716 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19720 * Extended the ASN.1 parser routines.
19724 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19728 * Added a BN_CTX to the BN library.
19732 * Fixed the weak key values in DES library
19736 * Changed API in EVP library for cipher aliases.
19740 * Added support for RC2/64bit cipher.
19744 * Converted the lhash library to the crypto/mem.c functions.
19748 * Added more recognized ASN.1 object ids.
19752 * Added more RSA padding checks for SSL/TLS.
19756 * Added BIO proxy/filter functionality.
19760 * Added extra_certs to SSL_CTX which can be used
19761 send extra CA certificates to the client in the CA cert chain sending
19762 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19766 * Now Fortezza is denied in the authentication phase because
19767 this is key exchange mechanism is not supported by SSLeay at all.
19771 * Additional PKCS1 checks.
19775 * Support the string "TLSv1" for all TLS v1 ciphers.
19779 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19780 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19784 * Fixed a few memory leaks.
19788 * Fixed various code and comment typos.
19792 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19793 bytes sent in the client random.
19795 *Edward Bishop <ebishop@spyglass.com>*
19799 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19800 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19801 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19802 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19803 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19804 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19805 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19806 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19807 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19808 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19809 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19810 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19811 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19812 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19813 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19814 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19815 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19816 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19817 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19818 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19819 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19820 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19821 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19822 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19823 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19824 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19825 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19826 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19827 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19828 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19829 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19830 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19831 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19832 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19833 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19834 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19835 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19836 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19837 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19838 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19839 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19840 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19841 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19842 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19843 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19844 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19845 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19846 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19847 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19848 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19849 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19850 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19851 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19852 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19853 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19854 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19855 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19856 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19857 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19858 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19859 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19860 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19861 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19862 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19863 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19864 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19865 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19866 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19867 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19868 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19869 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19870 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19871 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19872 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19873 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19874 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19875 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19876 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19877 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19878 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19879 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19880 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19881 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19882 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19883 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19884 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19885 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19886 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19887 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19888 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19889 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19890 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19891 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19892 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19893 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19894 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19895 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19896 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19897 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19898 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19899 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19900 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19901 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19902 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19903 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19904 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19905 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19906 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19907 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19908 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19909 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19910 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19911 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19912 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19913 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19914 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19915 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19916 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19917 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19918 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19919 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19920 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19921 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19922 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19923 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19924 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19925 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19926 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19927 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19928 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19929 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19930 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19931 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19932 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19933 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19934 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19935 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19936 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19937 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19938 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19939 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19940 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19941 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19942 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19943 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19944 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19945 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19946 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19947 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19948 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19949 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19950 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19951 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19952 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19953 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19954 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19955 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19956 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19957 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19958 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19959 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19960 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19961 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19962 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19963 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19964 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19965 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19966 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19967 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19968 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19969 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19970 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19971 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19972 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19973 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19974 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19975 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19976 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19977 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19978 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655