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 3.1.2 and 3.1.3 [xx XXX xxxx]
29 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
31 * Fix excessive time spent checking DH q parameter value.
33 The function DH_check() performs various checks on DH parameters. After
34 fixing CVE-2023-3446 it was discovered that a large q parameter value can
35 also trigger an overly long computation during some of these checks.
36 A correct q value, if present, cannot be larger than the modulus p
37 parameter, thus it is unnecessary to perform these checks if q is larger
40 If DH_check() is called with such q parameter value,
41 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
42 intensive checks are skipped.
48 * Fix DH_check() excessive time with over sized modulus.
50 The function DH_check() performs various checks on DH parameters. One of
51 those checks confirms that the modulus ("p" parameter) is not too large.
52 Trying to use a very large modulus is slow and OpenSSL will not normally use
53 a modulus which is over 10,000 bits in length.
55 However the DH_check() function checks numerous aspects of the key or
56 parameters that have been supplied. Some of those checks use the supplied
57 modulus value even if it has already been found to be too large.
59 A new limit has been added to DH_check of 32,768 bits. Supplying a
60 key/parameters with a modulus over this size will simply cause DH_check() to
67 * Do not ignore empty associated data entries with AES-SIV.
69 The AES-SIV algorithm allows for authentication of multiple associated
70 data entries along with the encryption. To authenticate empty data the
71 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
72 with NULL pointer as the output buffer and 0 as the input buffer length.
73 The AES-SIV implementation in OpenSSL just returns success for such call
74 instead of performing the associated data authentication operation.
75 The empty data thus will not be authenticated. ([CVE-2023-2975])
77 Thanks to Juerg Wullschleger (Google) for discovering the issue.
79 The fix changes the authentication tag value and the ciphertext for
80 applications that use empty associated data entries with AES-SIV.
81 To decrypt data encrypted with previous versions of OpenSSL the application
82 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
87 * When building with the `enable-fips` option and using the resulting
88 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
89 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
90 not operate with truncated digests (FIPS 140-3 IG G.R).
94 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
96 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
97 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
99 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
100 numeric text form. For gigantic sub-identifiers, this would take a very
101 long time, the time complexity being O(n^2) where n is the size of that
102 sub-identifier. ([CVE-2023-2650])
104 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
105 IDENTIFIER to canonical numeric text form if the size of that OBJECT
106 IDENTIFIER is 586 bytes or less, and fail otherwise.
108 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
109 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
110 most 128 sub-identifiers, and that the maximum value that each sub-
111 identifier may have is 2^32-1 (4294967295 decimal).
113 For each byte of every sub-identifier, only the 7 lower bits are part of
114 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
115 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
120 * Multiple algorithm implementation fixes for ARM BE platforms.
124 * Added a -pedantic option to fipsinstall that adjusts the various
125 settings to ensure strict FIPS compliance rather than backwards
130 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
131 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
132 trigger a crash of an application using AES-XTS decryption if the memory
133 just after the buffer being decrypted is not mapped.
134 Thanks to Anton Romanov (Amazon) for discovering the issue.
139 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
140 The previous fix for this timing side channel turned out to cause
141 a severe 2-3x performance regression in the typical use case
142 compared to 3.0.7. The new fix uses existing constant time
143 code paths, and restores the previous performance level while
144 fully eliminating all existing timing side channels.
145 The fix was developed by Bernd Edlinger with testing support
150 * Add FIPS provider configuration option to disallow the use of
151 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
152 The option '-no_drbg_truncated_digests' can optionally be
153 supplied to 'openssl fipsinstall'.
157 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
158 that it does not enable policy checking. Thanks to David Benjamin for
159 discovering this issue.
164 * Fixed an issue where invalid certificate policies in leaf certificates are
165 silently ignored by OpenSSL and other certificate policy checks are skipped
166 for that certificate. A malicious CA could use this to deliberately assert
167 invalid certificate policies in order to circumvent policy checking on the
168 certificate altogether.
173 * Limited the number of nodes created in a policy tree to mitigate
174 against CVE-2023-0464. The default limit is set to 1000 nodes, which
175 should be sufficient for most installations. If required, the limit
176 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
177 time define to a desired maximum number of nodes or zero to allow
183 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
185 * Add FIPS provider configuration option to enforce the
186 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
187 The option '-ems-check' can optionally be supplied to
188 'openssl fipsinstall'.
192 * The FIPS provider includes a few non-approved algorithms for
193 backward compatibility purposes and the "fips=yes" property query
194 must be used for all algorithm fetches to ensure FIPS compliance.
196 The algorithms that are included but not approved are Triple DES ECB,
197 Triple DES CBC and EdDSA.
201 * Added support for KMAC in KBKDF.
205 * RNDR and RNDRRS support in provider functions to provide
206 random number generation for Arm CPUs (aarch64).
210 * s_client and s_server apps now explicitly say when the TLS version
211 does not include the renegotiation mechanism. This avoids confusion
212 between that scenario versus when the TLS version includes secure
213 renegotiation but the peer lacks support for it.
217 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
219 *Tomasz Kantecki, Andrey Matyukov*
221 * The various OBJ_* functions have been made thread safe.
225 * Parallel dual-prime 1536/2048-bit modular exponentiation for
226 AVX512_IFMA capable processors.
228 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
230 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
231 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
232 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
233 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
234 `OPENSSL_NO_DEPRECATED_3_1`.
236 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
237 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
238 definitions for these functions regardless of whether
239 `OPENSSL_NO_DEPRECATED_3_1` is defined.
241 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
242 functions regardless of whether they are using them. It is recommended that
243 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
247 * When generating safe-prime DH parameters set the recommended private key
248 length equivalent to minimum key lengths as in RFC 7919.
252 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
253 maximum size that is smaller or equal to the digest length to comply with
254 FIPS 186-4 section 5. This is implemented by a new option
255 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
256 `rsa_pss_saltlen` parameter, which is now the default. Signature
257 verification is not affected by this change and continues to work as before.
264 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
265 listed here are only a brief description.
266 The migration guide contains more detailed information related to new features,
267 breaking changes, and mappings for the large list of deprecated functions.
269 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
271 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
273 * Fixed NULL dereference during PKCS7 data verification.
275 A NULL pointer can be dereferenced when signatures are being
276 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
277 algorithm used for the signature is known to the OpenSSL library but
278 the implementation of the hash algorithm is not available the digest
279 initialization will fail. There is a missing check for the return
280 value from the initialization function which later leads to invalid
281 usage of the digest API most likely leading to a crash.
284 PKCS7 data is processed by the SMIME library calls and also by the
285 time stamp (TS) library calls. The TLS implementation in OpenSSL does
286 not call these functions however third party applications would be
287 affected if they call these functions to verify signatures on untrusted
292 * Fixed X.400 address type confusion in X.509 GeneralName.
294 There is a type confusion vulnerability relating to X.400 address processing
295 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
296 but the public structure definition for GENERAL_NAME incorrectly specified
297 the type of the x400Address field as ASN1_TYPE. This field is subsequently
298 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
301 When CRL checking is enabled (i.e. the application sets the
302 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
303 pass arbitrary pointers to a memcmp call, enabling them to read memory
304 contents or enact a denial of service.
309 * Fixed NULL dereference validating DSA public key.
311 An invalid pointer dereference on read can be triggered when an
312 application tries to check a malformed DSA public key by the
313 EVP_PKEY_public_check() function. This will most likely lead
314 to an application crash. This function can be called on public
315 keys supplied from untrusted sources which could allow an attacker
316 to cause a denial of service attack.
318 The TLS implementation in OpenSSL does not call this function
319 but applications might call the function if there are additional
320 security requirements imposed by standards such as FIPS 140-3.
323 *Shane Lontis, Tomáš Mráz*
325 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
327 An invalid pointer dereference on read can be triggered when an
328 application tries to load malformed PKCS7 data with the
329 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
331 The result of the dereference is an application crash which could
332 lead to a denial of service attack. The TLS implementation in OpenSSL
333 does not call this function however third party applications might
334 call these functions on untrusted data.
339 * Fixed Use-after-free following BIO_new_NDEF.
341 The public API function BIO_new_NDEF is a helper function used for
342 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
343 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
344 be called directly by end user applications.
346 The function receives a BIO from the caller, prepends a new BIO_f_asn1
347 filter BIO onto the front of it to form a BIO chain, and then returns
348 the new head of the BIO chain to the caller. Under certain conditions,
349 for example if a CMS recipient public key is invalid, the new filter BIO
350 is freed and the function returns a NULL result indicating a failure.
351 However, in this case, the BIO chain is not properly cleaned up and the
352 BIO passed by the caller still retains internal pointers to the previously
353 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
354 then a use-after-free will occur. This will most likely result in a crash.
357 *Viktor Dukhovni, Matt Caswell*
359 * Fixed Double free after calling PEM_read_bio_ex.
361 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
362 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
363 data. If the function succeeds then the "name_out", "header" and "data"
364 arguments are populated with pointers to buffers containing the relevant
365 decoded data. The caller is responsible for freeing those buffers. It is
366 possible to construct a PEM file that results in 0 bytes of payload data.
367 In this case PEM_read_bio_ex() will return a failure code but will populate
368 the header argument with a pointer to a buffer that has already been freed.
369 If the caller also frees this buffer then a double free will occur. This
370 will most likely lead to a crash.
372 The functions PEM_read_bio() and PEM_read() are simple wrappers around
373 PEM_read_bio_ex() and therefore these functions are also directly affected.
375 These functions are also called indirectly by a number of other OpenSSL
376 functions including PEM_X509_INFO_read_bio_ex() and
377 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
378 internal uses of these functions are not vulnerable because the caller does
379 not free the header argument if PEM_read_bio_ex() returns a failure code.
382 *Kurt Roeckx, Matt Caswell*
384 * Fixed Timing Oracle in RSA Decryption.
386 A timing based side channel exists in the OpenSSL RSA Decryption
387 implementation which could be sufficient to recover a plaintext across
388 a network in a Bleichenbacher style attack. To achieve a successful
389 decryption an attacker would have to be able to send a very large number
390 of trial messages for decryption. The vulnerability affects all RSA padding
391 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
394 *Dmitry Belyavsky, Hubert Kario*
396 * Fixed X.509 Name Constraints Read Buffer Overflow.
398 A read buffer overrun can be triggered in X.509 certificate verification,
399 specifically in name constraint checking. The read buffer overrun might
400 result in a crash which could lead to a denial of service attack.
401 In a TLS client, this can be triggered by connecting to a malicious
402 server. In a TLS server, this can be triggered if the server requests
403 client authentication and a malicious client connects.
408 * Fixed X.509 Policy Constraints Double Locking security issue.
410 If an X.509 certificate contains a malformed policy constraint and
411 policy processing is enabled, then a write lock will be taken twice
412 recursively. On some operating systems (most widely: Windows) this
413 results in a denial of service when the affected process hangs. Policy
414 processing being enabled on a publicly facing server is not considered
415 to be a common setup.
420 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
421 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
422 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
423 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
424 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
425 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
426 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
427 for legacy EC and SM2 keys is also changed similarly to honor the
428 equivalent conversion format flag as specified in the underlying
429 `EC_KEY` object being exported to a provider, when this function is
430 called through `EVP_PKEY_export()`.
434 * Fixed a type confusion vulnerability relating to X.400 address processing
435 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
436 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
437 vulnerability may allow an attacker who can provide a certificate chain and
438 CRL (neither of which need have a valid signature) to pass arbitrary pointers
439 to a `memcmp` call, creating a possible read primitive, subject to some
440 constraints. Refer to the advisory for more information. Thanks to David
441 Benjamin for discovering this issue. ([CVE-2023-0286])
443 This issue has been fixed by changing the public header file definition of
444 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
445 possible for any existing application to successfully use the existing
446 definition; however, if any application references the `x400Address` field
447 (e.g. in dead code), note that the type of this field has changed. There is
452 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
454 * Fixed two buffer overflows in punycode decoding functions.
456 A buffer overrun can be triggered in X.509 certificate verification,
457 specifically in name constraint checking. Note that this occurs after
458 certificate chain signature verification and requires either a CA to
459 have signed the malicious certificate or for the application to continue
460 certificate verification despite failure to construct a path to a trusted
463 In a TLS client, this can be triggered by connecting to a malicious
464 server. In a TLS server, this can be triggered if the server requests
465 client authentication and a malicious client connects.
467 An attacker can craft a malicious email address to overflow
468 an arbitrary number of bytes containing the `.` character (decimal 46)
469 on the stack. This buffer overflow could result in a crash (causing a
473 An attacker can craft a malicious email address to overflow four
474 attacker-controlled bytes on the stack. This buffer overflow could
475 result in a crash (causing a denial of service) or potentially remote code
476 execution depending on stack layout for any given platform/compiler.
481 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
482 parameters in OpenSSL code.
483 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
484 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
485 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
486 Using these invalid names may cause algorithms to use slower methods
487 that ignore the CRT parameters.
491 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
496 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
497 data to be signed before signing the certificate.
501 * Added RIPEMD160 to the default provider.
505 * Ensured that the key share group sent or accepted for the key exchange
506 is allowed for the protocol version.
510 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
512 * OpenSSL supports creating a custom cipher via the legacy
513 EVP_CIPHER_meth_new() function and associated function calls. This function
514 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
515 to use the new provider mechanism in order to implement custom ciphers.
517 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
518 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
519 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
520 and decryption initialisation functions). Instead of using the custom cipher
521 directly it incorrectly tries to fetch an equivalent cipher from the
522 available providers. An equivalent cipher is found based on the NID passed to
523 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
524 given cipher. However it is possible for an application to incorrectly pass
525 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
526 is used in this way the OpenSSL encryption/decryption initialisation function
527 will match the NULL cipher as being equivalent and will fetch this from the
528 available providers. This will succeed if the default provider has been
529 loaded (or if a third party provider has been loaded that offers this
530 cipher). Using the NULL cipher means that the plaintext is emitted as the
533 Applications are only affected by this issue if they call
534 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
535 encryption/decryption initialisation function. Applications that only use
536 SSL/TLS are not impacted by this issue.
541 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
546 * Fixed the linux-mips64 Configure target which was missing the
547 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
552 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
557 * Correctly handle a retransmitted ClientHello in DTLS
561 * Fixed detection of ktls support in cross-compile environment on Linux
565 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
570 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
571 report correct results in some cases
575 * Fix UWP builds by defining VirtualLock
579 * For known safe primes use the minimum key length according to RFC 7919.
580 Longer private key sizes unnecessarily raise the cycles needed to compute the
581 shared secret without any increase of the real security. This fixes a
582 regression from 1.1.1 where these shorter keys were generated for the known
587 * Added the loongarch64 target
591 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
592 only passed to the FIPS provider and not to the default or legacy provider.
596 * Fixed reported performance degradation on aarch64. Restored the
597 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
598 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
599 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
600 The new algorithm is still used for 32 bit targets.
604 * Added a missing header for memcmp that caused compilation failure on some
609 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
611 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
612 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
613 This issue makes the RSA implementation with 2048 bit private keys
614 incorrect on such machines and memory corruption will happen during
615 the computation. As a consequence of the memory corruption an attacker
616 may be able to trigger a remote code execution on the machine performing
619 SSL/TLS servers or other servers using 2048 bit RSA private keys running
620 on machines supporting AVX512IFMA instructions of the X86_64 architecture
621 are affected by this issue.
626 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
627 implementation would not encrypt the entirety of the data under some
628 circumstances. This could reveal sixteen bytes of data that was
629 preexisting in the memory that wasn't written. In the special case of
630 "in place" encryption, sixteen bytes of the plaintext would be revealed.
632 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
633 they are both unaffected.
636 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
638 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
640 * In addition to the c_rehash shell command injection identified in
641 CVE-2022-1292, further bugs where the c_rehash script does not
642 properly sanitise shell metacharacters to prevent command injection have been
645 When the CVE-2022-1292 was fixed it was not discovered that there
646 are other places in the script where the file names of certificates
647 being hashed were possibly passed to a command executed through the shell.
649 This script is distributed by some operating systems in a manner where
650 it is automatically executed. On such operating systems, an attacker
651 could execute arbitrary commands with the privileges of the script.
653 Use of the c_rehash script is considered obsolete and should be replaced
654 by the OpenSSL rehash command line tool.
657 *Daniel Fiala, Tomáš Mráz*
659 * Case insensitive string comparison no longer uses locales. It has instead
660 been directly implemented.
664 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
666 * Case insensitive string comparison is reimplemented via new locale-agnostic
667 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
668 comparison. The previous implementation had problems when the Turkish locale
673 * Fixed a bug in the c_rehash script which was not properly sanitising shell
674 metacharacters to prevent command injection. This script is distributed by
675 some operating systems in a manner where it is automatically executed. On
676 such operating systems, an attacker could execute arbitrary commands with the
677 privileges of the script.
679 Use of the c_rehash script is considered obsolete and should be replaced
680 by the OpenSSL rehash command line tool.
685 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
686 certificate on an OCSP response. The bug caused the function in the case
687 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
688 response (meaning a successful verification) even in the case where the
689 response signing certificate fails to verify.
691 It is anticipated that most users of `OCSP_basic_verify` will not use the
692 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
693 a negative value (indicating a fatal error) in the case of a certificate
694 verification failure. The normal expected return value in this case would be
697 This issue also impacts the command line OpenSSL "ocsp" application. When
698 verifying an ocsp response with the "-no_cert_checks" option the command line
699 application will report that the verification is successful even though it
700 has in fact failed. In this case the incorrect successful response will also
701 be accompanied by error messages showing the failure and contradicting the
702 apparently successful result.
707 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
708 AAD data as the MAC key. This made the MAC key trivially predictable.
710 An attacker could exploit this issue by performing a man-in-the-middle attack
711 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
712 that the modified data would still pass the MAC integrity check.
714 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
715 endpoint will always be rejected by the recipient and the connection will
716 fail at that point. Many application protocols require data to be sent from
717 the client to the server first. Therefore, in such a case, only an OpenSSL
718 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
720 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
721 sent in both directions. In this case both clients and servers could be
722 affected, regardless of the application protocol.
724 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
725 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
726 the handshake when using this ciphersuite.
728 The confidentiality of data is not impacted by this issue, i.e. an attacker
729 cannot decrypt data that has been encrypted using this ciphersuite - they can
732 In order for this attack to work both endpoints must legitimately negotiate
733 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
734 OpenSSL 3.0, and is not available within the default provider or the default
735 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
736 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
737 following must have occurred:
739 1) OpenSSL must have been compiled with the (non-default) compile time option
740 enable-weak-ssl-ciphers
742 2) OpenSSL must have had the legacy provider explicitly loaded (either
743 through application code or via configuration)
745 3) The ciphersuite must have been explicitly added to the ciphersuite list
747 4) The libssl security level must have been set to 0 (default is 1)
749 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
751 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
752 others that both endpoints have in common
757 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
758 occupied by the removed hash table entries.
760 This function is used when decoding certificates or keys. If a long lived
761 process periodically decodes certificates or keys its memory usage will
762 expand without bounds and the process might be terminated by the operating
763 system causing a denial of service. Also traversing the empty hash table
764 entries will take increasingly more time.
766 Typically such long lived processes might be TLS clients or TLS servers
767 configured to accept client certificate authentication.
770 *Hugo Landau, Aliaksei Levin*
772 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
773 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
774 statistics are no longer supported. For compatibility, these statistics are
775 still listed in the output but are now always reported as zero.
779 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
781 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
782 for non-prime moduli.
784 Internally this function is used when parsing certificates that contain
785 elliptic curve public keys in compressed form or explicit elliptic curve
786 parameters with a base point encoded in compressed form.
788 It is possible to trigger the infinite loop by crafting a certificate that
789 has invalid explicit curve parameters.
791 Since certificate parsing happens prior to verification of the certificate
792 signature, any process that parses an externally supplied certificate may thus
793 be subject to a denial of service attack. The infinite loop can also be
794 reached when parsing crafted private keys as they can contain explicit
795 elliptic curve parameters.
797 Thus vulnerable situations include:
799 - TLS clients consuming server certificates
800 - TLS servers consuming client certificates
801 - Hosting providers taking certificates or private keys from customers
802 - Certificate authorities parsing certification requests from subscribers
803 - Anything else which parses ASN.1 elliptic curve parameters
805 Also any other applications that use the BN_mod_sqrt() where the attacker
806 can control the parameter values are vulnerable to this DoS issue.
811 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
812 to the list of ciphersuites providing Perfect Forward Secrecy as
813 required by SECLEVEL >= 3.
815 *Dmitry Belyavskiy, Nicola Tuveri*
817 * Made the AES constant time code for no-asm configurations
818 optional due to the resulting 95% performance degradation.
819 The AES constant time code can be enabled, for no assembly
820 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
824 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
829 * The negative return value handling of the certificate verification callback
830 was reverted. The replacement is to set the verification retry state with
831 the SSL_set_retry_verify() function.
835 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
837 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
838 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
839 verify a certificate supplied by a server. That function may return a
840 negative return value to indicate an internal error (for example out of
841 memory). Such a negative return value is mishandled by OpenSSL and will cause
842 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
843 success and a subsequent call to SSL_get_error() to return the value
844 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
845 returned by OpenSSL if the application has previously called
846 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
847 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
848 totally unexpected and applications may not behave correctly as a result. The
849 exact behaviour will depend on the application but it could result in
850 crashes, infinite loops or other similar incorrect responses.
852 This issue is made more serious in combination with a separate bug in OpenSSL
853 3.0 that will cause X509_verify_cert() to indicate an internal error when
854 processing a certificate chain. This will occur where a certificate does not
855 include the Subject Alternative Name extension but where a Certificate
856 Authority has enforced name constraints. This issue can occur even with valid
862 * Corrected a few file name and file reference bugs in the build,
863 installation and setup scripts, which lead to installation verification
864 failures. Slightly enhanced the installation verification script.
868 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
873 * Fixed PVK encoder to properly query for the passphrase.
877 * Multiple fixes in the OSSL_HTTP API functions.
881 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
882 OSSL_PARAM_INTEGER data type and return error on negative numbers
883 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
884 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
888 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
892 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
896 * Multiple threading fixes.
900 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
904 * Allow fetching an operation from the provider that owns an unexportable key
905 as a fallback if that is still allowed by the property query.
909 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
911 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
916 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
917 S390X capability vector to zero. This simplifies testing of different code
918 paths on S390X architecture.
922 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
923 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
924 SP 800-38D". The communication will fail at this point.
928 * The EC_GROUP_clear_free() function is deprecated as there is nothing
929 confidential in EC_GROUP data.
933 * The byte order mark (BOM) character is ignored if encountered at the
934 beginning of a PEM-formatted file.
938 * Added CMS support for the Russian GOST algorithms.
942 * Due to move of the implementation of cryptographic operations
943 to the providers, validation of various operation parameters can
944 be postponed until the actual operation is executed where previously
945 it happened immediately when an operation parameter was set.
947 For example when setting an unsupported curve with
948 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
949 fail but later keygen operations with the EVP_PKEY_CTX will fail.
951 *OpenSSL team members and many third party contributors*
953 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
954 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
955 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
956 instead to retrieve these algorithms from a provider.
960 * On build targets where the multilib postfix is set in the build
961 configuration the libdir directory was changing based on whether
962 the lib directory with the multilib postfix exists on the system
963 or not. This unpredictable behavior was removed and eventual
964 multilib postfix is now always added to the default libdir. Use
965 `--libdir=lib` to override the libdir if adding the postfix is
970 * The triple DES key wrap functionality now conforms to RFC 3217 but is
971 no longer interoperable with OpenSSL 1.1.1.
975 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
976 function codes, this function can only cause problems for calling
981 * Add a configurable flag to output date formats as ISO 8601. Does not
982 change the default date format.
986 * Version of MSVC earlier than 1300 could get link warnings, which could
987 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
988 Support for this flag has been removed.
992 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
993 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
994 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
995 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
996 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1000 * The signatures of the functions to get and set options on SSL and
1001 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1002 Some source code changes may be required.
1006 * The public definitions of conf_method_st and conf_st have been
1007 deprecated. They will be made opaque in a future release.
1009 *Rich Salz and Tomáš Mráz*
1011 * Client-initiated renegotiation is disabled by default. To allow it, use
1012 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1013 flag, or the "ClientRenegotiation" config parameter as appropriate.
1017 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1018 or modify relative pathname inclusion.
1022 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1023 validated. Please consult the README-FIPS and
1024 README-PROVIDERS files, as well as the migration guide.
1026 *OpenSSL team members and many third party contributors*
1028 * For the key types DH and DHX the allowed settable parameters are now different.
1032 * The openssl commands that read keys, certificates, and CRLs now
1033 automatically detect the PEM or DER format of the input files.
1035 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1037 * Added enhanced PKCS#12 APIs which accept a library context.
1041 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1045 * Added support for Kernel TLS (KTLS).
1047 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1049 * Support for RFC 5746 secure renegotiation is now required by default for
1050 SSL or TLS connections to succeed.
1054 * The signature of the `copy` functional parameter of the
1055 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1056 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1057 the signature of the `pub_decode` functional parameter of the
1058 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1059 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1063 * The error return values from some control calls (ctrl) have changed.
1067 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1071 * Many functions in the EVP_ namespace that are getters of values from
1072 implementations or contexts were renamed to include get or get0 in their
1073 names. Old names are provided as macro aliases for compatibility and
1078 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1079 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1080 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1085 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1088 * The output from the command line applications may have minor
1093 * The output from numerous "printing" may have minor changes.
1097 * Windows thread synchronization uses read/write primitives (SRWLock) when
1098 supported by the OS, otherwise CriticalSection continues to be used.
1102 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1103 work on read only BIO source/sinks that do not support these functions.
1104 This allows piping or redirection of a file BIO using stdin to be buffered
1105 into memory. This is used internally in OSSL_DECODER_from_bio().
1109 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1110 this function would return one of the values OSSL_STORE_INFO_NAME,
1111 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1112 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1113 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1114 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1115 using this function should be amended to handle the changed return value.
1119 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1120 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1121 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1122 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1123 contains more than one certificate identifier: This means that all
1124 certificates referenced there MUST be part of the validation chain.
1128 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1129 RC5, DESX and DES have been moved to the legacy provider.
1133 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1134 RIPEMD-160 have been moved to the legacy provider.
1138 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1143 * The deprecated functions EVP_PKEY_get0_RSA(),
1144 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1145 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1146 well as the similarly named "get1" functions behave differently in
1151 * A number of functions handling low-level keys or engines were deprecated
1152 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1153 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1154 EVP_PKEY_get0_siphash().
1158 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1159 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1160 will need to load the legacy crypto provider. This includes these PBE
1161 algorithms which use this KDF:
1162 - NID_pbeWithMD2AndDES_CBC
1163 - NID_pbeWithMD5AndDES_CBC
1164 - NID_pbeWithSHA1AndRC2_CBC
1165 - NID_pbeWithMD2AndRC2_CBC
1166 - NID_pbeWithMD5AndRC2_CBC
1167 - NID_pbeWithSHA1AndDES_CBC
1171 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1172 BIO_debug_callback() functions.
1176 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1177 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1181 * The RAND_METHOD APIs have been deprecated.
1185 * The SRP APIs have been deprecated.
1189 * Add a compile time option to prevent the caching of provider fetched
1190 algorithms. This is enabled by including the no-cached-fetch option
1191 at configuration time.
1195 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1196 count of PKCS12_DEFAULT_ITER.
1198 *Tomáš Mráz and Sahana Prasad*
1200 * The openssl speed command does not use low-level API calls anymore.
1204 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1207 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1209 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1213 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1214 providers may supply their own group implementations (using either the "key
1215 exchange" or the "key encapsulation" methods) which will automatically be
1216 detected and used by libssl.
1218 *Matt Caswell, Nicola Tuveri*
1220 * The undocumented function X509_certificate_type() has been deprecated;
1224 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1228 * Removed RSA padding mode for SSLv23 (which was only used for
1229 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1230 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1235 * Deprecated the obsolete X9.31 RSA key generation related functions.
1237 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1238 is not allowed to return a value > 1, this is no more taken as failure.
1240 *Viktor Dukhovni and David von Oheimb*
1242 * Deprecated the obsolete X9.31 RSA key generation related functions
1243 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1244 BN_X931_generate_prime_ex().
1248 * The default key generation method for the regular 2-prime RSA keys was
1249 changed to the FIPS 186-4 B.3.6 method.
1253 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1257 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1261 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1262 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1264 *Rich Salz, Richard Levitte, and David von Oheimb*
1266 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1270 * Deprecated `OCSP_parse_url()`.
1274 * Validation of SM2 keys has been separated from the validation of regular EC
1279 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1280 switches: a validation failure triggers an early exit, returning a failure
1281 exit status to the parent process.
1285 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1286 to ignore unknown ciphers.
1290 * The `-cipher-commands` and `-digest-commands` options
1291 of the command line utility `list` have been deprecated.
1292 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1296 * Added convenience functions for generating asymmetric key pairs:
1297 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1298 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1302 * All of the low-level EC_KEY functions have been deprecated.
1304 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1306 * Deprecated all the libcrypto and libssl error string loading
1311 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1312 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1317 * The `-crypt` option to the `passwd` command line tool has been removed.
1321 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1326 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1330 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1331 EVP_PKEY_get1_tls_encodedpoint().
1335 * The security callback, which can be customised by application code, supports
1336 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1337 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1341 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1342 interface. Their functionality remains unchanged.
1346 * Added new option for 'openssl list', '-providers', which will display the
1347 list of loaded providers, their names, version and status. It optionally
1348 displays their gettable parameters.
1352 * Removed EVP_PKEY_set_alias_type().
1356 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1357 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1361 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1362 parameters are still checked for type safety at compile time via helper
1367 * Remove the RAND_DRBG API
1369 *Paul Dale and Matthias St. Pierre*
1371 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1372 as well as actual hostnames.
1376 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1377 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1378 conversely, silently ignore DTLS protocol version bounds when configuring
1379 TLS-based contexts. The commands can be repeated to set bounds of both
1380 types. The same applies with the corresponding "min_protocol" and
1381 "max_protocol" command-line switches, in case some application uses both TLS
1384 SSL_CTX instances that are created for a fixed protocol version (e.g.
1385 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1386 attempts to apply bounds to these protocol versions would result in an
1387 error. Now only the "version-flexible" SSL_CTX instances are subject to
1388 limits in configuration files in command-line options.
1392 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1397 * Reworked the recorded ERR codes to make better space for system errors.
1398 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1399 given code is a system error (true) or an OpenSSL error (false).
1403 * Reworked the test perl framework to better allow parallel testing.
1405 *Nicola Tuveri and David von Oheimb*
1407 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1408 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1412 * 'Configure' has been changed to figure out the configuration target if
1413 none is given on the command line. Consequently, the 'config' script is
1414 now only a mere wrapper. All documentation is changed to only mention
1417 *Rich Salz and Richard Levitte*
1419 * Added a library context `OSSL_LIB_CTX` that applications as well as
1420 other libraries can use to form a separate context within which
1421 libcrypto operations are performed.
1425 * Added various `_ex` functions to the OpenSSL API that support using
1426 a non-default `OSSL_LIB_CTX`.
1430 * Handshake now fails if Extended Master Secret extension is dropped
1435 * Dropped interactive mode from the `openssl` program.
1439 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1441 *David von Oheimb and Shane Lontis*
1443 * Deprecated `EC_METHOD_get_field_type()`.
1447 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1448 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1449 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1453 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1457 * Add CAdES-BES signature verification support, mostly derived
1458 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1460 *Filipe Raimundo da Silva*
1462 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1466 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1467 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1471 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1475 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1476 EC_KEY_precompute_mult().
1480 * Deprecated EC_POINTs_mul().
1484 * Removed FIPS_mode() and FIPS_mode_set().
1488 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1492 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1493 EC_POINT_get_Jprojective_coordinates_GFp().
1497 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1498 arrays to be more easily constructed via a series of utility functions.
1499 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1500 the various push functions and finally convert to a passable OSSL_PARAM
1501 array using OSSL_PARAM_BLD_to_param().
1505 * The security strength of SHA1 and MD5 based signatures in TLS has been
1510 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1511 contain a provider side internal key.
1515 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1519 * Project text documents not yet having a proper file name extension
1520 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1521 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1525 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1526 have been converted to Markdown with the goal to produce documents
1527 which not only look pretty when viewed online in the browser, but
1528 remain well readable inside a plain text editor.
1530 To achieve this goal, a 'minimalistic' Markdown style has been applied
1531 which avoids formatting elements that interfere too much with the
1532 reading flow in the text file. For example, it
1534 * avoids [ATX headings][] and uses [setext headings][] instead
1535 (which works for `<h1>` and `<h2>` headings only).
1536 * avoids [inline links][] and uses [reference links][] instead.
1537 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1539 [ATX headings]: https://github.github.com/gfm/#atx-headings
1540 [setext headings]: https://github.github.com/gfm/#setext-headings
1541 [inline links]: https://github.github.com/gfm/#inline-link
1542 [reference links]: https://github.github.com/gfm/#reference-link
1543 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1544 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1546 *Matthias St. Pierre*
1548 * The test suite is changed to preserve results of each test recipe.
1549 A new directory test-runs/ with subdirectories named like the
1550 test recipes are created in the build tree for this purpose.
1554 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1555 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1556 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1558 *David von Oheimb, Martin Peylo*
1560 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1561 It supports arbitrary request and response content types, GET redirection,
1562 TLS, connections via HTTP(S) proxies, connections and exchange via
1563 user-defined BIOs (allowing implicit connections), persistent connections,
1564 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1565 The legacy OCSP-focused (and only partly documented) API
1566 is retained for backward compatibility, while most of it is deprecated.
1570 * Added `util/check-format.pl`, a tool for checking adherence to the
1571 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1572 The checks performed are incomplete and yield some false positives.
1573 Still the tool should be useful for detecting most typical glitches.
1577 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1578 If domain name resolution yields multiple IP addresses all of them are tried
1579 after `connect()` failures.
1583 * All of the low-level RSA functions have been deprecated.
1587 * X509 certificates signed using SHA1 are no longer allowed at security
1592 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1593 modified to use PKEY APIs. These commands are now in maintenance mode
1594 and no new features will be added to them.
1598 * The command line utility rsautl has been deprecated.
1602 * The command line utilities genrsa and rsa have been modified to use PKEY
1603 APIs. They now write PKCS#8 keys by default. These commands are now in
1604 maintenance mode and no new features will be added to them.
1608 * All of the low-level DH functions have been deprecated.
1610 *Paul Dale and Matt Caswell*
1612 * All of the low-level DSA functions have been deprecated.
1616 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1617 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1621 * Deprecated low-level ECDH and ECDSA functions.
1625 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1629 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1630 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1631 a new formulation to include all the things it can be used for,
1632 as well as words of caution.
1636 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1640 * All of the low-level HMAC functions have been deprecated.
1642 *Paul Dale and David von Oheimb*
1644 * Over two thousand fixes were made to the documentation, including:
1645 - Common options (such as -rand/-writerand, TLS version control, etc)
1646 were refactored and point to newly-enhanced descriptions in openssl.pod.
1647 - Added style conformance for all options (with help from Richard Levitte),
1648 documented all reported missing options, added a CI build to check
1649 that all options are documented and that no unimplemented options
1651 - Documented some internals, such as all use of environment variables.
1652 - Addressed all internal broken L<> references.
1656 * All of the low-level CMAC functions have been deprecated.
1660 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1661 functions have been deprecated.
1663 *Paul Dale and David von Oheimb*
1665 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1666 set of functions. The documentation mentioned negative values for some
1667 errors, but this was never the case, so the mention of negative values
1670 Code that followed the documentation and thereby check with something
1671 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1675 * All of the low-level cipher functions have been deprecated.
1677 *Matt Caswell and Paul Dale*
1679 * Removed include/openssl/opensslconf.h.in and replaced it with
1680 include/openssl/configuration.h.in, which differs in not including
1681 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1682 was added to include both.
1684 This allows internal hacks where one might need to modify the set
1685 of configured macros, for example this if deprecated symbols are
1686 still supposed to be available internally:
1688 #include <openssl/configuration.h>
1690 #undef OPENSSL_NO_DEPRECATED
1691 #define OPENSSL_SUPPRESS_DEPRECATED
1693 #include <openssl/macros.h>
1695 This should not be used by applications that use the exported
1696 symbols, as that will lead to linking errors.
1700 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1701 used in exponentiation with 512-bit moduli. No EC algorithms are
1702 affected. Analysis suggests that attacks against 2-prime RSA1024,
1703 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1704 difficult to perform and are not believed likely. Attacks against DH512
1705 are considered just feasible. However, for an attack the target would
1706 have to re-use the DH512 private key, which is not recommended anyway.
1707 Also applications directly using the low-level API BN_mod_exp may be
1708 affected if they use BN_FLG_CONSTTIME.
1713 * Most memory-debug features have been deprecated, and the functionality
1714 replaced with no-ops.
1718 * Added documentation for the STACK API.
1722 * Introduced a new method type and API, OSSL_ENCODER, to represent
1723 generic encoders. These do the same sort of job that PEM writers
1724 and d2i functions do, but with support for methods supplied by
1725 providers, and the possibility for providers to support other
1730 * Introduced a new method type and API, OSSL_DECODER, to represent
1731 generic decoders. These do the same sort of job that PEM readers
1732 and i2d functions do, but with support for methods supplied by
1733 providers, and the possibility for providers to support other
1738 * Added a .pragma directive to the syntax of configuration files, to
1739 allow varying behavior in a supported and predictable manner.
1740 Currently added pragma:
1744 This allows dollar signs to be a keyword character unless it's
1745 followed by a opening brace or parenthesis. This is useful for
1746 platforms where dollar signs are commonly used in names, such as
1747 volume names and system directory names on VMS.
1751 * Added functionality to create an EVP_PKEY from user data.
1755 * Change the interpretation of the '--api' configuration option to
1756 mean that this is a desired API compatibility level with no
1757 further meaning. The previous interpretation, that this would
1758 also mean to remove all deprecated symbols up to and including
1759 the given version, no requires that 'no-deprecated' is also used
1760 in the configuration.
1762 When building applications, the desired API compatibility level
1763 can be set with the OPENSSL_API_COMPAT macro like before. For
1764 API compatibility version below 3.0, the old style numerical
1765 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1766 For version 3.0 and on, the value is expected to be the decimal
1767 value calculated from the major and minor version like this:
1769 MAJOR * 10000 + MINOR * 100
1773 -DOPENSSL_API_COMPAT=30000 For 3.0
1774 -DOPENSSL_API_COMPAT=30200 For 3.2
1776 To hide declarations that are deprecated up to and including the
1777 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1778 given when building the application as well.
1782 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1783 access to certificate and CRL stores via URIs and OSSL_STORE
1786 This adds the following functions:
1788 - X509_LOOKUP_store()
1789 - X509_STORE_load_file()
1790 - X509_STORE_load_path()
1791 - X509_STORE_load_store()
1792 - SSL_add_store_cert_subjects_to_stack()
1793 - SSL_CTX_set_default_verify_store()
1794 - SSL_CTX_load_verify_file()
1795 - SSL_CTX_load_verify_dir()
1796 - SSL_CTX_load_verify_store()
1800 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1801 The presence of this system service is determined at run-time.
1805 * Added functionality to create an EVP_PKEY context based on data
1806 for methods from providers. This takes an algorithm name and a
1807 property query string and simply stores them, with the intent
1808 that any operation that uses this context will use those strings
1809 to fetch the needed methods implicitly, thereby making the port
1810 of application written for pre-3.0 OpenSSL easier.
1814 * The undocumented function NCONF_WIN32() has been deprecated; for
1815 conversion details see the HISTORY section of doc/man5/config.pod
1819 * Introduced the new functions EVP_DigestSignInit_ex() and
1820 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1821 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1822 pages for further details.
1826 * Over two thousand fixes were made to the documentation, including:
1827 adding missing command flags, better style conformance, documentation
1830 *Rich Salz, Richard Levitte*
1832 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1833 X25519, X448, Ed25519 and Ed448.
1837 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1842 * Deprecated the public definition of `ERR_STATE` as well as the function
1843 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1848 * Added ERR functionality to give callers access to the stored function
1849 names that have replaced the older function code based functions.
1851 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1852 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1853 ERR_peek_error_all() and ERR_peek_last_error_all().
1855 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1856 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1857 ERR_func_error_string().
1861 * Extended testing to be verbose for failing tests only. The make variables
1862 VERBOSE_FAILURE or VF can be used to enable this:
1864 $ make VF=1 test # Unix
1865 $ mms /macro=(VF=1) test ! OpenVMS
1866 $ nmake VF=1 test # Windows
1870 * Added the `-copy_extensions` option to the `x509` command for use with
1871 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1872 all extensions in the request are copied to the certificate or vice versa.
1874 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1876 * Added the `-copy_extensions` option to the `req` command for use with
1877 `-x509`. When given with the `copy` or `copyall` argument,
1878 all extensions in the certification request are copied to the certificate.
1882 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1883 they generate are by default RFC 5280 compliant in the following sense:
1884 There is a subjectKeyIdentifier extension with a hash value of the public key
1885 and for not self-signed certs there is an authorityKeyIdentifier extension
1886 with a keyIdentifier field or issuer information identifying the signing key.
1887 This is done unless some configuration overrides the new default behavior,
1888 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1892 * Added several checks to `X509_verify_cert()` according to requirements in
1893 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1894 (which may be done by using the CLI option `-x509_strict`):
1895 * The basicConstraints of CA certificates must be marked critical.
1896 * CA certificates must explicitly include the keyUsage extension.
1897 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1898 * The issuer name of any certificate must not be empty.
1899 * The subject name of CA certs, certs with keyUsage crlSign,
1900 and certs without subjectAlternativeName must not be empty.
1901 * If a subjectAlternativeName extension is given it must not be empty.
1902 * The signatureAlgorithm field and the cert signature must be consistent.
1903 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1904 must not be marked critical.
1905 * The authorityKeyIdentifier must be given for X.509v3 certs
1906 unless they are self-signed.
1907 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1911 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1912 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1916 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1917 used even when parsing explicit parameters, when loading a encoded key
1918 or calling `EC_GROUP_new_from_ecpkparameters()`/
1919 `EC_GROUP_new_from_ecparameters()`.
1920 This prevents bypass of security hardening and performance gains,
1921 especially for curves with specialized EC_METHODs.
1922 By default, if a key encoded with explicit parameters is loaded and later
1923 encoded, the output is still encoded with explicit parameters, even if
1924 internally a "named" EC_GROUP is used for computation.
1928 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1929 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1930 NULL. After this change, only the cofactor parameter can be NULL. It also
1931 does some minimal sanity checks on the passed order.
1936 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1937 An attack is simple, if the first CMS_recipientInfo is valid but the
1938 second CMS_recipientInfo is chosen ciphertext. If the second
1939 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1940 encryption key will be replaced by garbage, and the message cannot be
1941 decoded, but if the RSA decryption fails, the correct encryption key is
1942 used and the recipient will not notice the attack.
1943 As a work around for this potential attack the length of the decrypted
1944 key must be equal to the cipher default key length, in case the
1945 certificate is not given and all recipientInfo are tried out.
1946 The old behaviour can be re-enabled in the CMS code by setting the
1947 CMS_DEBUG_DECRYPT flag.
1951 * Early start up entropy quality from the DEVRANDOM seed source has been
1952 improved for older Linux systems. The RAND subsystem will wait for
1953 /dev/random to be producing output before seeding from /dev/urandom.
1954 The seeded state is stored for future library initialisations using
1955 a system global shared memory segment. The shared memory identifier
1956 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1957 the desired value. The default identifier is 114.
1961 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1962 when primes for RSA keys are computed.
1963 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1964 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1965 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1966 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1967 This avoids possible fingerprinting of newly generated RSA modules.
1971 * Correct the extended master secret constant on EBCDIC systems. Without this
1972 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1973 negotiate EMS will fail. Unfortunately this also means that TLS connections
1974 between EBCDIC systems with this fix, and EBCDIC systems without this
1975 fix will fail if they negotiate EMS.
1979 * Changed the library initialisation so that the config file is now loaded
1980 by default. This was already the case for libssl. It now occurs for both
1981 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1982 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1986 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1987 where the former acts as a replacement for `ERR_put_error()`, and the
1988 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1989 `ERR_raise_data()` adds more flexibility by taking a format string and
1990 an arbitrary number of arguments following it, to be processed with
1995 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1996 to check if a named provider is loaded and available. When called, it
1997 will also activate all fallback providers if such are still present.
2001 * Enforce a minimum DH modulus size of 512 bits.
2005 * Changed DH parameters to generate the order q subgroup instead of 2q.
2006 Previously generated DH parameters are still accepted by DH_check
2007 but DH_generate_key works around that by clearing bit 0 of the
2008 private key for those. This avoids leaking bit 0 of the private key.
2012 * Significantly reduce secure memory usage by the randomness pools.
2016 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2021 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2022 algorithms. An implementation of a key exchange algorithm can be obtained
2023 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2024 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2025 the older EVP_PKEY_derive_init() function. See the man pages for the new
2026 functions for further details.
2030 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2034 * Removed the function names from error messages and deprecated the
2039 * Removed NextStep support and the macro OPENSSL_UNISTD
2043 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2044 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2045 Also removed "export var as function" capability; we do not export
2046 variables, only functions.
2050 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2051 an error and 1 indicating success. In previous versions of OpenSSL this
2052 was a void type. If a key was set longer than the maximum possible this
2057 * Support SM2 signing and verification schemes with X509 certificate.
2061 * Use SHA256 as the default digest for TS query in the `ts` app.
2065 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2069 * Default cipher lists/suites are now available via a function, the
2070 #defines are deprecated.
2074 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2075 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2076 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2080 * Join the directories crypto/x509 and crypto/x509v3
2084 * Added command 'openssl kdf' that uses the EVP_KDF API.
2088 * Added command 'openssl mac' that uses the EVP_MAC API.
2092 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2093 as default directories. Also added the command 'openssl info'
2094 for scripting purposes.
2098 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2103 * Add prediction resistance to the DRBG reseeding process.
2107 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2108 mandated by IEEE Std 1619-2018.
2112 * Added newline escaping functionality to a filename when using openssl dgst.
2113 This output format is to replicate the output format found in the `*sum`
2114 checksum programs. This aims to preserve backward compatibility.
2116 *Matt Eaton, Richard Levitte, and Paul Dale*
2118 * Removed the heartbeat message in DTLS feature, as it has very
2119 little usage and doesn't seem to fulfill a valuable purpose.
2120 The configuration option is now deprecated.
2124 * Changed the output of 'openssl {digestname} < file' to display the
2125 digest name in its output.
2129 * Added a new generic trace API which provides support for enabling
2130 instrumentation through trace output.
2132 *Richard Levitte & Matthias St. Pierre*
2134 * Added build tests for C++. These are generated files that only do one
2135 thing, to include one public OpenSSL head file each. This tests that
2136 the public header files can be usefully included in a C++ application.
2138 This test isn't enabled by default. It can be enabled with the option
2139 'enable-buildtest-c++'.
2143 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2147 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2151 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2155 * Added KMAC to EVP_MAC.
2159 * Added property based algorithm implementation selection framework to
2164 * Added SCA hardening for modular field inversion in EC_GROUP through
2165 a new dedicated field_inv() pointer in EC_METHOD.
2166 This also addresses a leakage affecting conversions from projective
2167 to affine coordinates.
2169 *Billy Bob Brumley, Nicola Tuveri*
2171 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2172 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2173 those algorithms that were already supported through the EVP_PKEY API
2174 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2175 and scrypt are now wrappers that call EVP_KDF.
2179 * Build devcrypto engine as a dynamic engine.
2181 *Eneas U de Queiroz*
2183 * Add keyed BLAKE2 to EVP_MAC.
2187 * Fix a bug in the computation of the endpoint-pair shared secret used
2188 by DTLS over SCTP. This breaks interoperability with older versions
2189 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2190 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2191 interoperability with such broken implementations. However, enabling
2192 this switch breaks interoperability with correct implementations.
2194 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2195 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2199 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2203 * Changed the license to the Apache License v2.0.
2207 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2209 - Major releases (indicated by incrementing the MAJOR release number)
2210 may introduce incompatible API/ABI changes.
2211 - Minor releases (indicated by incrementing the MINOR release number)
2212 may introduce new features but retain API/ABI compatibility.
2213 - Patch releases (indicated by incrementing the PATCH number)
2214 are intended for bug fixes and other improvements of existing
2215 features only (like improving performance or adding documentation)
2216 and retain API/ABI compatibility.
2220 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2224 * Remove the 'dist' target and add a tarball building script. The
2225 'dist' target has fallen out of use, and it shouldn't be
2226 necessary to configure just to create a source distribution.
2230 * Recreate the OS390-Unix config target. It no longer relies on a
2231 special script like it did for OpenSSL pre-1.1.0.
2235 * Instead of having the source directories listed in Configure, add
2236 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2241 * Add GMAC to EVP_MAC.
2245 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2249 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2250 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2251 to facilitate the continued use of MACs through raw private keys in
2252 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2256 * Deprecate ECDH_KDF_X9_62().
2260 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2261 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2262 are retained for backwards compatibility.
2266 * AES-XTS mode now enforces that its two keys are different to mitigate
2267 the attacked described in "Efficient Instantiations of Tweakable
2268 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2269 Details of this attack can be obtained from:
2270 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2274 * Rename the object files, i.e. give them other names than in previous
2275 versions. Their names now include the name of the final product, as
2276 well as its type mnemonic (bin, lib, shlib).
2280 * Added new option for 'openssl list', '-objects', which will display the
2281 list of built in objects, i.e. OIDs with names.
2285 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2286 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2291 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2292 improves application performance by removing data copies and providing
2293 applications with zero-copy system calls such as sendfile and splice.
2297 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2301 * `PKCS12_parse` now maintains the order of the parsed certificates
2302 when outputting them via `*ca` (rather than reversing it).
2306 * Deprecated pthread fork support methods.
2310 * Added support for FFDHE key exchange in TLS 1.3.
2314 * Added a new concept for OpenSSL plugability: providers. This
2315 functionality is designed to replace the ENGINE API and ENGINE
2316 implementations, and to be much more dynamic, allowing provider
2317 authors to introduce new algorithms among other things, as long as
2318 there's an API that supports the algorithm type.
2320 With this concept comes a new core API for interaction between
2321 libcrypto and provider implementations. Public libcrypto functions
2322 that want to use providers do so through this core API.
2324 The main documentation for this core API is found in
2325 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2326 refer to other manuals describing the API specific for supported
2327 algorithm types (also called operations).
2334 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2336 * Avoid loading of a dynamic engine twice.
2340 * Prioritise DANE TLSA issuer certs over peer certs
2344 * Fixed random API for MacOS prior to 10.12
2346 These MacOS versions don't support the CommonCrypto APIs
2350 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2352 * Fixed an SM2 Decryption Buffer Overflow.
2354 In order to decrypt SM2 encrypted data an application is expected to
2355 call the API function EVP_PKEY_decrypt(). Typically an application will
2356 call this function twice. The first time, on entry, the "out" parameter
2357 can be NULL and, on exit, the "outlen" parameter is populated with the
2358 buffer size required to hold the decrypted plaintext. The application
2359 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2360 again, but this time passing a non-NULL value for the "out" parameter.
2362 A bug in the implementation of the SM2 decryption code means that the
2363 calculation of the buffer size required to hold the plaintext returned
2364 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2365 size required by the second call. This can lead to a buffer overflow
2366 when EVP_PKEY_decrypt() is called by the application a second time with
2367 a buffer that is too small.
2369 A malicious attacker who is able present SM2 content for decryption to
2370 an application could cause attacker chosen data to overflow the buffer
2371 by up to a maximum of 62 bytes altering the contents of other data held
2372 after the buffer, possibly changing application behaviour or causing
2373 the application to crash. The location of the buffer is application
2374 dependent but is typically heap allocated.
2379 * Fixed various read buffer overruns processing ASN.1 strings
2381 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2382 structure which contains a buffer holding the string data and a field
2383 holding the buffer length. This contrasts with normal C strings which
2384 are represented as a buffer for the string data which is terminated
2385 with a NUL (0) byte.
2387 Although not a strict requirement, ASN.1 strings that are parsed using
2388 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2389 well as any string whose value has been set with the ASN1_STRING_set()
2390 function will additionally NUL terminate the byte array in the
2391 ASN1_STRING structure.
2393 However, it is possible for applications to directly construct valid
2394 ASN1_STRING structures which do not NUL terminate the byte array by
2395 directly setting the "data" and "length" fields in the ASN1_STRING
2396 array. This can also happen by using the ASN1_STRING_set0() function.
2398 Numerous OpenSSL functions that print ASN.1 data have been found to
2399 assume that the ASN1_STRING byte array will be NUL terminated, even
2400 though this is not guaranteed for strings that have been directly
2401 constructed. Where an application requests an ASN.1 structure to be
2402 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2403 been directly constructed by the application without NUL terminating
2404 the "data" field, then a read buffer overrun can occur.
2406 The same thing can also occur during name constraints processing
2407 of certificates (for example if a certificate has been directly
2408 constructed by the application instead of loading it via the OpenSSL
2409 parsing functions, and the certificate contains non NUL terminated
2410 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2411 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2413 If a malicious actor can cause an application to directly construct an
2414 ASN1_STRING and then process it through one of the affected OpenSSL
2415 functions then this issue could be hit. This might result in a crash
2416 (causing a Denial of Service attack). It could also result in the
2417 disclosure of private memory contents (such as private keys, or
2418 sensitive plaintext).
2423 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2425 * Fixed a problem with verifying a certificate chain when using the
2426 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2427 the certificates present in a certificate chain. It is not set by default.
2429 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2430 the chain that have explicitly encoded elliptic curve parameters was added
2431 as an additional strict check.
2433 An error in the implementation of this check meant that the result of a
2434 previous check to confirm that certificates in the chain are valid CA
2435 certificates was overwritten. This effectively bypasses the check
2436 that non-CA certificates must not be able to issue other certificates.
2438 If a "purpose" has been configured then there is a subsequent opportunity
2439 for checks that the certificate is a valid CA. All of the named "purpose"
2440 values implemented in libcrypto perform this check. Therefore, where
2441 a purpose is set the certificate chain will still be rejected even when the
2442 strict flag has been used. A purpose is set by default in libssl client and
2443 server certificate verification routines, but it can be overridden or
2444 removed by an application.
2446 In order to be affected, an application must explicitly set the
2447 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2448 for the certificate verification or, in the case of TLS client or server
2449 applications, override the default purpose.
2454 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2455 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2456 renegotiation ClientHello omits the signature_algorithms extension (where it
2457 was present in the initial ClientHello), but includes a
2458 signature_algorithms_cert extension then a NULL pointer dereference will
2459 result, leading to a crash and a denial of service attack.
2461 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2462 (which is the default configuration). OpenSSL TLS clients are not impacted by
2466 *Peter Kästle and Samuel Sapalski*
2468 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2470 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2471 create a unique hash value based on the issuer and serial number data
2472 contained within an X509 certificate. However, it was failing to correctly
2473 handle any errors that may occur while parsing the issuer field (which might
2474 occur if the issuer field is maliciously constructed). This may subsequently
2475 result in a NULL pointer deref and a crash leading to a potential denial of
2481 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2482 padding mode to correctly check for rollback attacks. This is considered a
2483 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2488 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2489 functions. Previously they could overflow the output length argument in some
2490 cases where the input length is close to the maximum permissible length for
2491 an integer on the platform. In such cases the return value from the function
2492 call would be 1 (indicating success), but the output length value would be
2493 negative. This could cause applications to behave incorrectly or crash.
2498 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2499 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2500 could be exploited in a side channel attack to recover the password. Since
2501 the attack is local host only this is outside of the current OpenSSL
2502 threat model and therefore no CVE is assigned.
2504 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2509 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2511 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2512 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2513 If an attacker can control both items being compared then this could lead
2514 to a possible denial of service attack. OpenSSL itself uses the
2515 GENERAL_NAME_cmp function for two purposes:
2516 1) Comparing CRL distribution point names between an available CRL and a
2517 CRL distribution point embedded in an X509 certificate
2518 2) When verifying that a timestamp response token signer matches the
2519 timestamp authority name (exposed via the API functions
2520 TS_RESP_verify_response and TS_RESP_verify_token)
2525 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2527 * Certificates with explicit curve parameters are now disallowed in
2528 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2532 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2533 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2534 conversely, silently ignore DTLS protocol version bounds when configuring
2535 TLS-based contexts. The commands can be repeated to set bounds of both
2536 types. The same applies with the corresponding "min_protocol" and
2537 "max_protocol" command-line switches, in case some application uses both TLS
2540 SSL_CTX instances that are created for a fixed protocol version (e.g.
2541 TLSv1_server_method()) also silently ignore version bounds. Previously
2542 attempts to apply bounds to these protocol versions would result in an
2543 error. Now only the "version-flexible" SSL_CTX instances are subject to
2544 limits in configuration files in command-line options.
2548 * Handshake now fails if Extended Master Secret extension is dropped
2553 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2555 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2557 * Fixed segmentation fault in SSL_check_chain()
2558 Server or client applications that call the SSL_check_chain() function
2559 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2560 dereference as a result of incorrect handling of the
2561 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2562 or unrecognised signature algorithm is received from the peer. This could
2563 be exploited by a malicious peer in a Denial of Service attack.
2568 * Added AES consttime code for no-asm configurations
2569 an optional constant time support for AES was added
2570 when building openssl for no-asm.
2571 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2572 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2573 At this time this feature is by default disabled.
2574 It will be enabled by default in 3.0.
2578 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2580 * Revert the change of EOF detection while reading in libssl to avoid
2581 regressions in applications depending on the current way of reporting
2582 the EOF. As the existing method is not fully accurate the change to
2583 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2584 branch and will be present in the 3.0 release.
2588 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2589 when primes for RSA keys are computed.
2590 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2591 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2592 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2593 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2594 This avoids possible fingerprinting of newly generated RSA modules.
2598 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2600 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2601 while reading in libssl then we would report an error back to the
2602 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2603 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2604 therefore give a hint as to what went wrong.
2608 * Check that ed25519 and ed448 are allowed by the security level. Previously
2609 signature algorithms not using an MD were not being checked that they were
2610 allowed by the security level.
2614 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2615 was not quite right. The behaviour was not consistent between resumption
2616 and normal handshakes, and also not quite consistent with historical
2617 behaviour. The behaviour in various scenarios has been clarified and
2618 it has been updated to make it match historical behaviour as closely as
2623 * *[VMS only]* The header files that the VMS compilers include automatically,
2624 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2625 that the C++ compiler doesn't understand. This is a shortcoming in the
2626 compiler, but can be worked around with `__cplusplus` guards.
2628 C++ applications that use OpenSSL libraries must be compiled using the
2629 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2630 functions. Otherwise, only functions with symbols of less than 31
2631 characters can be used, as the linker will not be able to successfully
2632 resolve symbols with longer names.
2636 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2637 The presence of this system service is determined at run-time.
2641 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2646 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2648 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2649 number generator (RNG). This was intended to include protection in the
2650 event of a fork() system call in order to ensure that the parent and child
2651 processes did not share the same RNG state. However, this protection was not
2652 being used in the default case.
2654 A partial mitigation for this issue is that the output from a high
2655 precision timer is mixed into the RNG state so the likelihood of a parent
2656 and child process sharing state is significantly reduced.
2658 If an application already calls OPENSSL_init_crypto() explicitly using
2659 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2662 *Matthias St. Pierre*
2664 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2665 used even when parsing explicit parameters, when loading a encoded key
2666 or calling `EC_GROUP_new_from_ecpkparameters()`/
2667 `EC_GROUP_new_from_ecparameters()`.
2668 This prevents bypass of security hardening and performance gains,
2669 especially for curves with specialized EC_METHODs.
2670 By default, if a key encoded with explicit parameters is loaded and later
2671 encoded, the output is still encoded with explicit parameters, even if
2672 internally a "named" EC_GROUP is used for computation.
2676 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2677 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2678 NULL. After this change, only the cofactor parameter can be NULL. It also
2679 does some minimal sanity checks on the passed order.
2684 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2685 An attack is simple, if the first CMS_recipientInfo is valid but the
2686 second CMS_recipientInfo is chosen ciphertext. If the second
2687 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2688 encryption key will be replaced by garbage, and the message cannot be
2689 decoded, but if the RSA decryption fails, the correct encryption key is
2690 used and the recipient will not notice the attack.
2691 As a work around for this potential attack the length of the decrypted
2692 key must be equal to the cipher default key length, in case the
2693 certificate is not given and all recipientInfo are tried out.
2694 The old behaviour can be re-enabled in the CMS code by setting the
2695 CMS_DEBUG_DECRYPT flag.
2700 * Early start up entropy quality from the DEVRANDOM seed source has been
2701 improved for older Linux systems. The RAND subsystem will wait for
2702 /dev/random to be producing output before seeding from /dev/urandom.
2703 The seeded state is stored for future library initialisations using
2704 a system global shared memory segment. The shared memory identifier
2705 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2706 the desired value. The default identifier is 114.
2710 * Correct the extended master secret constant on EBCDIC systems. Without this
2711 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2712 negotiate EMS will fail. Unfortunately this also means that TLS connections
2713 between EBCDIC systems with this fix, and EBCDIC systems without this
2714 fix will fail if they negotiate EMS.
2718 * Use Windows installation paths in the mingw builds
2720 Mingw isn't a POSIX environment per se, which means that Windows
2721 paths should be used for installation.
2726 * Changed DH_check to accept parameters with order q and 2q subgroups.
2727 With order 2q subgroups the bit 0 of the private key is not secret
2728 but DH_generate_key works around that by clearing bit 0 of the
2729 private key for those. This avoids leaking bit 0 of the private key.
2733 * Significantly reduce secure memory usage by the randomness pools.
2737 * Revert the DEVRANDOM_WAIT feature for Linux systems
2739 The DEVRANDOM_WAIT feature added a select() call to wait for the
2740 /dev/random device to become readable before reading from the
2741 /dev/urandom device.
2743 It turned out that this change had negative side effects on
2744 performance which were not acceptable. After some discussion it
2745 was decided to revert this feature and leave it up to the OS
2746 resp. the platform maintainer to ensure a proper initialization
2747 during early boot time.
2749 *Matthias St. Pierre*
2751 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2753 * Add build tests for C++. These are generated files that only do one
2754 thing, to include one public OpenSSL head file each. This tests that
2755 the public header files can be usefully included in a C++ application.
2757 This test isn't enabled by default. It can be enabled with the option
2758 'enable-buildtest-c++'.
2762 * Enable SHA3 pre-hashing for ECDSA and DSA.
2766 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2767 This changes the size when using the `genpkey` command when no size is given.
2768 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2769 generation commands to use 2048 bits by default.
2773 * Reorganize the manual pages to consistently have RETURN VALUES,
2774 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2775 util/fix-doc-nits accordingly.
2777 *Paul Yang, Joshua Lock*
2779 * Add the missing accessor EVP_PKEY_get0_engine()
2783 * Have commands like `s_client` and `s_server` output the signature scheme
2784 along with other cipher suite parameters when debugging.
2788 * Make OPENSSL_config() error agnostic again.
2792 * Do the error handling in RSA decryption constant time.
2796 * Prevent over long nonces in ChaCha20-Poly1305.
2798 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2799 for every encryption operation. RFC 7539 specifies that the nonce value
2800 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2801 and front pads the nonce with 0 bytes if it is less than 12
2802 bytes. However it also incorrectly allows a nonce to be set of up to 16
2803 bytes. In this case only the last 12 bytes are significant and any
2804 additional leading bytes are ignored.
2806 It is a requirement of using this cipher that nonce values are
2807 unique. Messages encrypted using a reused nonce value are susceptible to
2808 serious confidentiality and integrity attacks. If an application changes
2809 the default nonce length to be longer than 12 bytes and then makes a
2810 change to the leading bytes of the nonce expecting the new value to be a
2811 new unique nonce then such an application could inadvertently encrypt
2812 messages with a reused nonce.
2814 Additionally the ignored bytes in a long nonce are not covered by the
2815 integrity guarantee of this cipher. Any application that relies on the
2816 integrity of these ignored leading bytes of a long nonce may be further
2817 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2818 is safe because no such use sets such a long nonce value. However user
2819 applications that use this cipher directly and set a non-default nonce
2820 length to be longer than 12 bytes may be vulnerable.
2822 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2828 * Add DEVRANDOM_WAIT feature for Linux systems
2830 On older Linux systems where the getrandom() system call is not available,
2831 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2832 Contrary to getrandom(), the /dev/urandom device will not block during
2833 early boot when the kernel CSPRNG has not been seeded yet.
2835 To mitigate this known weakness, use select() to wait for /dev/random to
2836 become readable before reading from /dev/urandom.
2838 * Ensure that SM2 only uses SM3 as digest algorithm
2842 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2844 * Change the info callback signals for the start and end of a post-handshake
2845 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2846 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2847 confused by this and assume that a TLSv1.2 renegotiation has started. This
2848 can break KeyUpdate handling. Instead we no longer signal the start and end
2849 of a post handshake message exchange (although the messages themselves are
2850 still signalled). This could break some applications that were expecting
2851 the old signals. However without this KeyUpdate is not usable for many
2856 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2858 * Timing vulnerability in DSA signature generation
2860 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2861 timing side channel attack. An attacker could use variations in the signing
2862 algorithm to recover the private key.
2864 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2869 * Timing vulnerability in ECDSA signature generation
2871 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2872 timing side channel attack. An attacker could use variations in the signing
2873 algorithm to recover the private key.
2875 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2880 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2881 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2882 of two gigabytes and the error handling improved.
2884 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2885 categorized as a normal bug, not a security issue, because the DRBG reseeds
2886 automatically and is fully functional even without additional randomness
2887 provided by the application.
2889 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2891 * Add a new ClientHello callback. Provides a callback interface that gives
2892 the application the ability to adjust the nascent SSL object at the
2893 earliest stage of ClientHello processing, immediately after extensions have
2894 been collected but before they have been processed. In particular, this
2895 callback can adjust the supported TLS versions in response to the contents
2900 * Add SM2 base algorithm support.
2904 * s390x assembly pack: add (improved) hardware-support for the following
2905 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2906 aes-cfb/cfb8, aes-ecb.
2910 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2911 parameter is no longer accepted, as it leads to a corrupt table. NULL
2912 pem_str is reserved for alias entries only.
2916 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2917 step for prime curves. The new implementation is based on formulae from
2918 differential addition-and-doubling in homogeneous projective coordinates
2919 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2920 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2921 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2922 to work in projective coordinates.
2924 *Billy Bob Brumley, Nicola Tuveri*
2926 * Change generating and checking of primes so that the error rate of not
2927 being prime depends on the intended use based on the size of the input.
2928 For larger primes this will result in more rounds of Miller-Rabin.
2929 The maximal error rate for primes with more than 1080 bits is lowered
2932 *Kurt Roeckx, Annie Yousar*
2934 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2938 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2939 moving between systems, and to avoid confusion when a Windows build is
2940 done with mingw vs with MSVC. For POSIX installs, there's still a
2941 symlink or copy named 'tsget' to avoid that confusion as well.
2945 * Revert blinding in ECDSA sign and instead make problematic addition
2946 length-invariant. Switch even to fixed-length Montgomery multiplication.
2950 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2951 step for binary curves. The new implementation is based on formulae from
2952 differential addition-and-doubling in mixed Lopez-Dahab projective
2953 coordinates, modified to independently blind the operands.
2955 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2957 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2958 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2959 EC_METHODs to implement their own specialized "ladder step", to take
2960 advantage of more favorable coordinate systems or more efficient
2961 differential addition-and-doubling algorithms.
2963 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2965 * Modified the random device based seed sources to keep the relevant
2966 file descriptors open rather than reopening them on each access.
2967 This allows such sources to operate in a chroot() jail without
2968 the associated device nodes being available. This behaviour can be
2969 controlled using RAND_keep_random_devices_open().
2973 * Numerous side-channel attack mitigations have been applied. This may have
2974 performance impacts for some algorithms for the benefit of improved
2975 security. Specific changes are noted in this change log by their respective
2980 * AIX shared library support overhaul. Switch to AIX "natural" way of
2981 handling shared libraries, which means collecting shared objects of
2982 different versions and bitnesses in one common archive. This allows to
2983 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2984 doesn't affect the way 3rd party applications are linked, only how
2985 multi-version installation is managed.
2989 * Make ec_group_do_inverse_ord() more robust and available to other
2990 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2991 mitigations are applied to the fallback BN_mod_inverse().
2992 When using this function rather than BN_mod_inverse() directly, new
2993 EC cryptosystem implementations are then safer-by-default.
2997 * Add coordinate blinding for EC_POINT and implement projective
2998 coordinate blinding for generic prime curves as a countermeasure to
2999 chosen point SCA attacks.
3001 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3003 * Add blinding to ECDSA and DSA signatures to protect against side channel
3004 attacks discovered by Keegan Ryan (NCC Group).
3008 * Enforce checking in the `pkeyutl` command to ensure that the input
3009 length does not exceed the maximum supported digest length when performing
3010 a sign, verify or verifyrecover operation.
3014 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3015 I/O in combination with something like select() or poll() will hang. This
3016 can be turned off again using SSL_CTX_clear_mode().
3017 Many applications do not properly handle non-application data records, and
3018 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3019 around the problems in those applications, but can also break some.
3020 It's recommended to read the manpages about SSL_read(), SSL_write(),
3021 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3022 SSL_CTX_set_read_ahead() again.
3026 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3027 now allow empty (zero character) pass phrases.
3031 * Apply blinding to binary field modular inversion and remove patent
3032 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3036 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3037 binary and prime elliptic curves.
3041 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3042 constant time fixed point multiplication.
3046 * Revise elliptic curve scalar multiplication with timing attack
3047 defenses: ec_wNAF_mul redirects to a constant time implementation
3048 when computing fixed point and variable point multiplication (which
3049 in OpenSSL are mostly used with secret scalars in keygen, sign,
3050 ECDH derive operations).
3051 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3054 * Updated CONTRIBUTING
3058 * Updated DRBG / RAND to request nonce and additional low entropy
3059 randomness from the system.
3061 *Matthias St. Pierre*
3063 * Updated 'openssl rehash' to use OpenSSL consistent default.
3067 * Moved the load of the ssl_conf module to libcrypto, which helps
3068 loading engines that libssl uses before libssl is initialised.
3072 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3076 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3078 *Ingo Schwarze, Rich Salz*
3080 * Added output of accepting IP address and port for 'openssl s_server'
3084 * Added a new API for TLSv1.3 ciphersuites:
3085 SSL_CTX_set_ciphersuites()
3086 SSL_set_ciphersuites()
3090 * Memory allocation failures consistently add an error to the error
3095 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3096 in libcrypto when run as setuid/setgid.
3100 * Load any config file by default when libssl is used.
3104 * Added new public header file <openssl/rand_drbg.h> and documentation
3105 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3107 *Matthias St. Pierre*
3109 * QNX support removed (cannot find contributors to get their approval
3110 for the license change).
3114 * TLSv1.3 replay protection for early data has been implemented. See the
3115 SSL_read_early_data() man page for further details.
3119 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3120 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3121 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3122 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3123 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3124 configuration has been separated out. See the ciphers man page or the
3125 SSL_CTX_set_ciphersuites() man page for more information.
3129 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3130 in responder mode now supports the new "-multi" option, which
3131 spawns the specified number of child processes to handle OCSP
3132 requests. The "-timeout" option now also limits the OCSP
3133 responder's patience to wait to receive the full client request
3134 on a newly accepted connection. Child processes are respawned
3135 as needed, and the CA index file is automatically reloaded
3136 when changed. This makes it possible to run the "ocsp" responder
3137 as a long-running service, making the OpenSSL CA somewhat more
3138 feature-complete. In this mode, most diagnostic messages logged
3139 after entering the event loop are logged via syslog(3) rather than
3144 * Added support for X448 and Ed448. Heavily based on original work by
3149 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3150 objects loaded. This adds the functions OSSL_STORE_expect() and
3151 OSSL_STORE_find() as well as needed tools to construct searches and
3152 get the search data out of them.
3156 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3157 version of OpenSSL should review their configuration settings to ensure
3158 that they are still appropriate for TLSv1.3. For further information see:
3159 <https://wiki.openssl.org/index.php/TLS1.3>
3163 * Grand redesign of the OpenSSL random generator
3165 The default RAND method now utilizes an AES-CTR DRBG according to
3166 NIST standard SP 800-90Ar1. The new random generator is essentially
3167 a port of the default random generator from the OpenSSL FIPS 2.0
3168 object module. It is a hybrid deterministic random bit generator
3169 using an AES-CTR bit stream and which seeds and reseeds itself
3170 automatically using trusted system entropy sources.
3172 Some of its new features are:
3173 - Support for multiple DRBG instances with seed chaining.
3174 - The default RAND method makes use of a DRBG.
3175 - There is a public and private DRBG instance.
3176 - The DRBG instances are fork-safe.
3177 - Keep all global DRBG instances on the secure heap if it is enabled.
3178 - The public and private DRBG instance are per thread for lock free
3181 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3183 * Changed Configure so it only says what it does and doesn't dump
3184 so much data. Instead, ./configdata.pm should be used as a script
3185 to display all sorts of configuration data.
3189 * Added processing of "make variables" to Configure.
3193 * Added SHA512/224 and SHA512/256 algorithm support.
3197 * The last traces of Netware support, first removed in 1.1.0, have
3202 * Get rid of Makefile.shared, and in the process, make the processing
3203 of certain files (rc.obj, or the .def/.map/.opt files produced from
3204 the ordinal files) more visible and hopefully easier to trace and
3205 debug (or make silent).
3209 * Make it possible to have environment variable assignments as
3210 arguments to config / Configure.
3214 * Add multi-prime RSA (RFC 8017) support.
3218 * Add SM3 implemented according to GB/T 32905-2016
3219 *Jack Lloyd <jack.lloyd@ribose.com>,*
3220 *Ronald Tse <ronald.tse@ribose.com>,*
3221 *Erick Borsboom <erick.borsboom@ribose.com>*
3223 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3224 as documented in RFC6066.
3225 Based on a patch from Tomasz Moń
3227 *Filipe Raimundo da Silva*
3229 * Add SM4 implemented according to GB/T 32907-2016.
3230 *Jack Lloyd <jack.lloyd@ribose.com>,*
3231 *Ronald Tse <ronald.tse@ribose.com>,*
3232 *Erick Borsboom <erick.borsboom@ribose.com>*
3234 * Reimplement -newreq-nodes and ERR_error_string_n; the
3235 original author does not agree with the license change.
3239 * Add ARIA AEAD TLS support.
3243 * Some macro definitions to support VS6 have been removed. Visual
3244 Studio 6 has not worked since 1.1.0
3248 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3249 without clearing the errors.
3253 * Add "atfork" functions. If building on a system that without
3254 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3255 requirements. The RAND facility now uses/requires this.
3263 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3264 not possible to disable entirely. However, it's still possible to
3265 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3268 To disable, configure with 'no-ui-console'. 'no-ui' is still
3269 possible to use as an alias. Check at compile time with the
3270 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3271 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3275 * Add a STORE module, which implements a uniform and URI based reader of
3276 stores that can contain keys, certificates, CRLs and numerous other
3277 objects. The main API is loosely based on a few stdio functions,
3278 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3279 OSSL_STORE_error and OSSL_STORE_close.
3280 The implementation uses backends called "loaders" to implement arbitrary
3281 URI schemes. There is one built in "loader" for the 'file' scheme.
3285 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3286 then adjusted to work on FreeBSD 8.4 as well.
3287 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3288 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3292 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3293 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3294 error code calls like this:
3296 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3298 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3299 that can be encoded in C. For the foreseeable future, this will only
3302 *Richard Levitte and Tim Hudson*
3304 * Removed BSD cryptodev engine.
3308 * Add a build target 'build_all_generated', to build all generated files
3309 and only that. This can be used to prepare everything that requires
3310 things like perl for a system that lacks perl and then move everything
3311 to that system and do the rest of the build there.
3315 * In the UI interface, make it possible to duplicate the user data. This
3316 can be used by engines that need to retain the data for a longer time
3317 than just the call where this user data is passed.
3321 * Ignore the '-named_curve auto' value for compatibility of applications
3324 *Tomáš Mráz <tmraz@fedoraproject.org>*
3326 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3327 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3328 alerts across multiple records (some of which could be empty). In practice
3329 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3330 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3331 support this at all. Supporting it adds significant complexity to the
3332 record layer, and its removal is unlikely to cause interoperability
3337 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3338 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3339 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3344 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3345 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3347 *Richard Levitte, Andy Polyakov*
3349 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3354 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3355 platform rather than 'mingw'.
3359 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3360 success if they are asked to add an object which already exists
3361 in the store. This change cascades to other functions which load
3362 certificates and CRLs.
3366 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3367 facilitate stack unwinding even from assembly subroutines.
3371 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3372 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3376 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3377 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3378 which is the minimum version we support.
3382 * Certificate time validation (X509_cmp_time) enforces stricter
3383 compliance with RFC 5280. Fractional seconds and timezone offsets
3384 are no longer allowed.
3388 * Add support for ARIA
3392 * s_client will now send the Server Name Indication (SNI) extension by
3393 default unless the new "-noservername" option is used. The server name is
3394 based on the host provided to the "-connect" option unless overridden by
3395 using "-servername".
3399 * Add support for SipHash
3403 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3404 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3405 prevent issues where no progress is being made and the peer continually
3406 sends unrecognised record types, using up resources processing them.
3410 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3411 using the algorithm defined in
3412 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3416 * Heartbeat support has been removed; the ABI is changed for now.
3418 *Richard Levitte, Rich Salz*
3420 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3424 * The RSA "null" method, which was partially supported to avoid patent
3425 issues, has been replaced to always returns NULL.
3432 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3434 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3435 used even when parsing explicit parameters, when loading a encoded key
3436 or calling `EC_GROUP_new_from_ecpkparameters()`/
3437 `EC_GROUP_new_from_ecparameters()`.
3438 This prevents bypass of security hardening and performance gains,
3439 especially for curves with specialized EC_METHODs.
3440 By default, if a key encoded with explicit parameters is loaded and later
3441 encoded, the output is still encoded with explicit parameters, even if
3442 internally a "named" EC_GROUP is used for computation.
3446 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3447 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3448 NULL. After this change, only the cofactor parameter can be NULL. It also
3449 does some minimal sanity checks on the passed order.
3454 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3455 An attack is simple, if the first CMS_recipientInfo is valid but the
3456 second CMS_recipientInfo is chosen ciphertext. If the second
3457 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3458 encryption key will be replaced by garbage, and the message cannot be
3459 decoded, but if the RSA decryption fails, the correct encryption key is
3460 used and the recipient will not notice the attack.
3461 As a work around for this potential attack the length of the decrypted
3462 key must be equal to the cipher default key length, in case the
3463 certificate is not given and all recipientInfo are tried out.
3464 The old behaviour can be re-enabled in the CMS code by setting the
3465 CMS_DEBUG_DECRYPT flag.
3470 * Use Windows installation paths in the mingw builds
3472 Mingw isn't a POSIX environment per se, which means that Windows
3473 paths should be used for installation.
3478 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3480 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3481 This changes the size when using the `genpkey` command when no size is given.
3482 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3483 generation commands to use 2048 bits by default.
3487 * Prevent over long nonces in ChaCha20-Poly1305.
3489 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3490 for every encryption operation. RFC 7539 specifies that the nonce value
3491 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3492 and front pads the nonce with 0 bytes if it is less than 12
3493 bytes. However it also incorrectly allows a nonce to be set of up to 16
3494 bytes. In this case only the last 12 bytes are significant and any
3495 additional leading bytes are ignored.
3497 It is a requirement of using this cipher that nonce values are
3498 unique. Messages encrypted using a reused nonce value are susceptible to
3499 serious confidentiality and integrity attacks. If an application changes
3500 the default nonce length to be longer than 12 bytes and then makes a
3501 change to the leading bytes of the nonce expecting the new value to be a
3502 new unique nonce then such an application could inadvertently encrypt
3503 messages with a reused nonce.
3505 Additionally the ignored bytes in a long nonce are not covered by the
3506 integrity guarantee of this cipher. Any application that relies on the
3507 integrity of these ignored leading bytes of a long nonce may be further
3508 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3509 is safe because no such use sets such a long nonce value. However user
3510 applications that use this cipher directly and set a non-default nonce
3511 length to be longer than 12 bytes may be vulnerable.
3513 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3519 * Added SCA hardening for modular field inversion in EC_GROUP through
3520 a new dedicated field_inv() pointer in EC_METHOD.
3521 This also addresses a leakage affecting conversions from projective
3522 to affine coordinates.
3524 *Billy Bob Brumley, Nicola Tuveri*
3526 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3527 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3531 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3535 * Remove the 'dist' target and add a tarball building script. The
3536 'dist' target has fallen out of use, and it shouldn't be
3537 necessary to configure just to create a source distribution.
3541 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3543 * Timing vulnerability in DSA signature generation
3545 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3546 timing side channel attack. An attacker could use variations in the signing
3547 algorithm to recover the private key.
3549 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3554 * Timing vulnerability in ECDSA signature generation
3556 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3557 timing side channel attack. An attacker could use variations in the signing
3558 algorithm to recover the private key.
3560 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3565 * Add coordinate blinding for EC_POINT and implement projective
3566 coordinate blinding for generic prime curves as a countermeasure to
3567 chosen point SCA attacks.
3569 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3571 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3573 * Client DoS due to large DH parameter
3575 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3576 malicious server can send a very large prime value to the client. This will
3577 cause the client to spend an unreasonably long period of time generating a
3578 key for this prime resulting in a hang until the client has finished. This
3579 could be exploited in a Denial Of Service attack.
3581 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3586 * Cache timing vulnerability in RSA Key Generation
3588 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3589 a cache timing side channel attack. An attacker with sufficient access to
3590 mount cache timing attacks during the RSA key generation process could
3591 recover the private key.
3593 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3594 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3599 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3600 parameter is no longer accepted, as it leads to a corrupt table. NULL
3601 pem_str is reserved for alias entries only.
3605 * Revert blinding in ECDSA sign and instead make problematic addition
3606 length-invariant. Switch even to fixed-length Montgomery multiplication.
3610 * Change generating and checking of primes so that the error rate of not
3611 being prime depends on the intended use based on the size of the input.
3612 For larger primes this will result in more rounds of Miller-Rabin.
3613 The maximal error rate for primes with more than 1080 bits is lowered
3616 *Kurt Roeckx, Annie Yousar*
3618 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3622 * Add blinding to ECDSA and DSA signatures to protect against side channel
3623 attacks discovered by Keegan Ryan (NCC Group).
3627 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3628 now allow empty (zero character) pass phrases.
3632 * Certificate time validation (X509_cmp_time) enforces stricter
3633 compliance with RFC 5280. Fractional seconds and timezone offsets
3634 are no longer allowed.
3638 * Fixed a text canonicalisation bug in CMS
3640 Where a CMS detached signature is used with text content the text goes
3641 through a canonicalisation process first prior to signing or verifying a
3642 signature. This process strips trailing space at the end of lines, converts
3643 line terminators to CRLF and removes additional trailing line terminators
3644 at the end of a file. A bug in the canonicalisation process meant that
3645 some characters, such as form-feed, were incorrectly treated as whitespace
3646 and removed. This is contrary to the specification (RFC5485). This fix
3647 could mean that detached text data signed with an earlier version of
3648 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3649 signed with a fixed OpenSSL may fail to verify with an earlier version of
3650 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3651 and use the "-binary" flag (for the "cms" command line application) or set
3652 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3656 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3658 * Constructed ASN.1 types with a recursive definition could exceed the stack
3660 Constructed ASN.1 types with a recursive definition (such as can be found
3661 in PKCS7) could eventually exceed the stack given malicious input with
3662 excessive recursion. This could result in a Denial Of Service attack. There
3663 are no such structures used within SSL/TLS that come from untrusted sources
3664 so this is considered safe.
3666 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3672 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3674 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3675 effectively reduced to only comparing the least significant bit of each
3676 byte. This allows an attacker to forge messages that would be considered as
3677 authenticated in an amount of tries lower than that guaranteed by the
3678 security claims of the scheme. The module can only be compiled by the
3679 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3681 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3687 * Add a build target 'build_all_generated', to build all generated files
3688 and only that. This can be used to prepare everything that requires
3689 things like perl for a system that lacks perl and then move everything
3690 to that system and do the rest of the build there.
3694 * Backport SSL_OP_NO_RENGOTIATION
3696 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3697 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3698 changes this is no longer possible in 1.1.0. Therefore, the new
3699 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3700 1.1.0 to provide equivalent functionality.
3702 Note that if an application built against 1.1.0h headers (or above) is run
3703 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3704 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3708 * Removed the OS390-Unix config target. It relied on a script that doesn't
3713 * rsaz_1024_mul_avx2 overflow bug on x86_64
3715 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3716 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3717 Analysis suggests that attacks against RSA and DSA as a result of this
3718 defect would be very difficult to perform and are not believed likely.
3719 Attacks against DH1024 are considered just feasible, because most of the
3720 work necessary to deduce information about a private key may be performed
3721 offline. The amount of resources required for such an attack would be
3722 significant. However, for an attack on TLS to be meaningful, the server
3723 would have to share the DH1024 private key among multiple clients, which is
3724 no longer an option since CVE-2016-0701.
3726 This only affects processors that support the AVX2 but not ADX extensions
3727 like Intel Haswell (4th generation).
3729 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3730 was originally found via the OSS-Fuzz project.
3735 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3737 * bn_sqrx8x_internal carry bug on x86_64
3739 There is a carry propagating bug in the x86_64 Montgomery squaring
3740 procedure. No EC algorithms are affected. Analysis suggests that attacks
3741 against RSA and DSA as a result of this defect would be very difficult to
3742 perform and are not believed likely. Attacks against DH are considered just
3743 feasible (although very difficult) because most of the work necessary to
3744 deduce information about a private key may be performed offline. The amount
3745 of resources required for such an attack would be very significant and
3746 likely only accessible to a limited number of attackers. An attacker would
3747 additionally need online access to an unpatched system using the target
3748 private key in a scenario with persistent DH parameters and a private
3749 key that is shared between multiple clients.
3751 This only affects processors that support the BMI1, BMI2 and ADX extensions
3752 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3754 This issue was reported to OpenSSL by the OSS-Fuzz project.
3759 * Malformed X.509 IPAddressFamily could cause OOB read
3761 If an X.509 certificate has a malformed IPAddressFamily extension,
3762 OpenSSL could do a one-byte buffer overread. The most likely result
3763 would be an erroneous display of the certificate in text format.
3765 This issue was reported to OpenSSL by the OSS-Fuzz project.
3770 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3772 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3773 platform rather than 'mingw'.
3777 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3778 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3779 which is the minimum version we support.
3783 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3785 * Encrypt-Then-Mac renegotiation crash
3787 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3788 negotiated where it was not in the original handshake (or vice-versa) then
3789 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3790 and servers are affected.
3792 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3797 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3799 * Truncated packet could crash via OOB read
3801 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3802 cipher is being used, then a truncated packet can cause that host to
3803 perform an out-of-bounds read, usually resulting in a crash.
3805 This issue was reported to OpenSSL by Robert Święcki of Google.
3810 * Bad (EC)DHE parameters cause a client crash
3812 If a malicious server supplies bad parameters for a DHE or ECDHE key
3813 exchange then this can result in the client attempting to dereference a
3814 NULL pointer leading to a client crash. This could be exploited in a Denial
3817 This issue was reported to OpenSSL by Guido Vranken.
3822 * BN_mod_exp may produce incorrect results on x86_64
3824 There is a carry propagating bug in the x86_64 Montgomery squaring
3825 procedure. No EC algorithms are affected. Analysis suggests that attacks
3826 against RSA and DSA as a result of this defect would be very difficult to
3827 perform and are not believed likely. Attacks against DH are considered just
3828 feasible (although very difficult) because most of the work necessary to
3829 deduce information about a private key may be performed offline. The amount
3830 of resources required for such an attack would be very significant and
3831 likely only accessible to a limited number of attackers. An attacker would
3832 additionally need online access to an unpatched system using the target
3833 private key in a scenario with persistent DH parameters and a private
3834 key that is shared between multiple clients. For example this can occur by
3835 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3836 similar to CVE-2015-3193 but must be treated as a separate problem.
3838 This issue was reported to OpenSSL by the OSS-Fuzz project.
3843 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3845 * ChaCha20/Poly1305 heap-buffer-overflow
3847 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3848 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3849 crash. This issue is not considered to be exploitable beyond a DoS.
3851 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3856 * CMS Null dereference
3858 Applications parsing invalid CMS structures can crash with a NULL pointer
3859 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3860 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3861 structure callback if an attempt is made to free certain invalid encodings.
3862 Only CHOICE structures using a callback which do not handle NULL value are
3865 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3870 * Montgomery multiplication may produce incorrect results
3872 There is a carry propagating bug in the Broadwell-specific Montgomery
3873 multiplication procedure that handles input lengths divisible by, but
3874 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3875 and DH private keys are impossible. This is because the subroutine in
3876 question is not used in operations with the private key itself and an input
3877 of the attacker's direct choice. Otherwise the bug can manifest itself as
3878 transient authentication and key negotiation failures or reproducible
3879 erroneous outcome of public-key operations with specially crafted input.
3880 Among EC algorithms only Brainpool P-512 curves are affected and one
3881 presumably can attack ECDH key negotiation. Impact was not analyzed in
3882 detail, because pre-requisites for attack are considered unlikely. Namely
3883 multiple clients have to choose the curve in question and the server has to
3884 share the private key among them, neither of which is default behaviour.
3885 Even then only clients that chose the curve will be affected.
3887 This issue was publicly reported as transient failures and was not
3888 initially recognized as a security issue. Thanks to Richard Morgan for
3889 providing reproducible case.
3894 * Removed automatic addition of RPATH in shared libraries and executables,
3895 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3899 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3901 * Fix Use After Free for large message sizes
3903 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3904 message larger than approx 16k is received then the underlying buffer to
3905 store the incoming message is reallocated and moved. Unfortunately a
3906 dangling pointer to the old location is left which results in an attempt to
3907 write to the previously freed location. This is likely to result in a
3908 crash, however it could potentially lead to execution of arbitrary code.
3910 This issue only affects OpenSSL 1.1.0a.
3912 This issue was reported to OpenSSL by Robert Święcki.
3917 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3919 * OCSP Status Request extension unbounded memory growth
3921 A malicious client can send an excessively large OCSP Status Request
3922 extension. If that client continually requests renegotiation, sending a
3923 large OCSP Status Request extension each time, then there will be unbounded
3924 memory growth on the server. This will eventually lead to a Denial Of
3925 Service attack through memory exhaustion. Servers with a default
3926 configuration are vulnerable even if they do not support OCSP. Builds using
3927 the "no-ocsp" build time option are not affected.
3929 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3934 * SSL_peek() hang on empty record
3936 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3937 sends an empty record. This could be exploited by a malicious peer in a
3938 Denial Of Service attack.
3940 This issue was reported to OpenSSL by Alex Gaynor.
3945 * Excessive allocation of memory in tls_get_message_header() and
3946 dtls1_preprocess_fragment()
3948 A (D)TLS message includes 3 bytes for its length in the header for the
3949 message. This would allow for messages up to 16Mb in length. Messages of
3950 this length are excessive and OpenSSL includes a check to ensure that a
3951 peer is sending reasonably sized messages in order to avoid too much memory
3952 being consumed to service a connection. A flaw in the logic of version
3953 1.1.0 means that memory for the message is allocated too early, prior to
3954 the excessive message length check. Due to way memory is allocated in
3955 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3956 to service a connection. This could lead to a Denial of Service through
3957 memory exhaustion. However, the excessive message length check still takes
3958 place, and this would cause the connection to immediately fail. Assuming
3959 that the application calls SSL_free() on the failed connection in a timely
3960 manner then the 21Mb of allocated memory will then be immediately freed
3961 again. Therefore, the excessive memory allocation will be transitory in
3962 nature. This then means that there is only a security impact if:
3964 1) The application does not call SSL_free() in a timely manner in the event
3965 that the connection fails
3967 2) The application is working in a constrained environment where there is
3968 very little free memory
3970 3) The attacker initiates multiple connection attempts such that there are
3971 multiple connections in a state where memory has been allocated for the
3972 connection; SSL_free() has not yet been called; and there is insufficient
3973 memory to service the multiple requests.
3975 Except in the instance of (1) above any Denial Of Service is likely to be
3976 transitory because as soon as the connection fails the memory is
3977 subsequently freed again in the SSL_free() call. However there is an
3978 increased risk during this period of application crashes due to the lack of
3979 memory - which would then mean a more serious Denial of Service.
3981 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3982 (CVE-2016-6307 and CVE-2016-6308)
3986 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3987 had to be removed. Primary reason is that vendor assembler can't
3988 assemble our modules with -KPIC flag. As result it, assembly
3989 support, was not even available as option. But its lack means
3990 lack of side-channel resistant code, which is incompatible with
3991 security by todays standards. Fortunately gcc is readily available
3992 prepackaged option, which we firmly point at...
3996 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3998 * Windows command-line tool supports UTF-8 opt-in option for arguments
3999 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4000 (to any value) allows Windows user to access PKCS#12 file generated
4001 with Windows CryptoAPI and protected with non-ASCII password, as well
4002 as files generated under UTF-8 locale on Linux also protected with
4007 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4008 have been disabled by default and removed from DEFAULT, just like RC4.
4009 See the RC4 item below to re-enable both.
4013 * The method for finding the storage location for the Windows RAND seed file
4014 has changed. First we check %RANDFILE%. If that is not set then we check
4015 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4016 all else fails we fall back to C:\.
4020 * The EVP_EncryptUpdate() function has had its return type changed from void
4021 to int. A return of 0 indicates and error while a return of 1 indicates
4026 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4027 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4028 off the constant time implementation for RSA, DSA and DH have been made
4029 no-ops and deprecated.
4033 * Windows RAND implementation was simplified to only get entropy by
4034 calling CryptGenRandom(). Various other RAND-related tickets
4037 *Joseph Wylie Yandle, Rich Salz*
4039 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4040 and `OPENSSL_LH_`, respectively. The old names are available
4041 with API compatibility. They new names are now completely documented.
4045 * Unify TYPE_up_ref(obj) methods signature.
4046 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4047 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4048 int (instead of void) like all others TYPE_up_ref() methods.
4049 So now these methods also check the return value of CRYPTO_atomic_add(),
4050 and the validity of object reference counter.
4052 *fdasilvayy@gmail.com*
4054 * With Windows Visual Studio builds, the .pdb files are installed
4055 alongside the installed libraries and executables. For a static
4056 library installation, ossl_static.pdb is the associate compiler
4057 generated .pdb file to be used when linking programs.
4061 * Remove openssl.spec. Packaging files belong with the packagers.
4065 * Automatic Darwin/OSX configuration has had a refresh, it will now
4066 recognise x86_64 architectures automatically. You can still decide
4067 to build for a different bitness with the environment variable
4068 KERNEL_BITS (can be 32 or 64), for example:
4070 KERNEL_BITS=32 ./config
4074 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4075 256 bit AES and HMAC with SHA256.
4079 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4083 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4087 * To enable users to have their own config files and build file templates,
4088 Configure looks in the directory indicated by the environment variable
4089 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4090 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4091 name and is used as is.
4095 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4096 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4097 X509_CERT_FILE_CTX was removed.
4101 * "shared" builds are now the default. To create only static libraries use
4102 the "no-shared" Configure option.
4106 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4107 All of these option have not worked for some while and are fundamental
4112 * Make various cleanup routines no-ops and mark them as deprecated. Most
4113 global cleanup functions are no longer required because they are handled
4114 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4115 Explicitly de-initing can cause problems (e.g. where a library that uses
4116 OpenSSL de-inits, but an application is still using it). The affected
4117 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4118 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4119 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4120 COMP_zlib_cleanup().
4124 * --strict-warnings no longer enables runtime debugging options
4125 such as REF_DEBUG. Instead, debug options are automatically
4126 enabled with '--debug' builds.
4128 *Andy Polyakov, Emilia Käsper*
4130 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4131 have been moved out of the public header files. New functions for managing
4132 these have been added.
4136 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4137 objects have been moved out of the public header files. New
4138 functions for managing these have been added.
4142 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4143 have been moved out of the public header files. New functions for managing
4144 these have been added.
4148 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4149 moved out of the public header files. New functions for managing these
4154 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4158 * Removed the mk1mf build scripts.
4162 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4163 it is always safe to #include a header now.
4167 * Removed the aged BC-32 config and all its supporting scripts
4171 * Removed support for Ultrix, Netware, and OS/2.
4175 * Add support for HKDF.
4177 *Alessandro Ghedini*
4179 * Add support for blake2b and blake2s
4183 * Added support for "pipelining". Ciphers that have the
4184 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4185 encryptions/decryptions simultaneously. There are currently no built-in
4186 ciphers with this property but the expectation is that engines will be able
4187 to offer it to significantly improve throughput. Support has been extended
4188 into libssl so that multiple records for a single connection can be
4189 processed in one go (for >=TLS 1.1).
4193 * Added the AFALG engine. This is an async capable engine which is able to
4194 offload work to the Linux kernel. In this initial version it only supports
4195 AES128-CBC. The kernel must be version 4.1.0 or greater.
4199 * OpenSSL now uses a new threading API. It is no longer necessary to
4200 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4201 are two supported threading models: pthreads and windows threads. It is
4202 also possible to configure OpenSSL at compile time for "no-threads". The
4203 old threading API should no longer be used. The functions have been
4204 replaced with "no-op" compatibility macros.
4206 *Alessandro Ghedini, Matt Caswell*
4208 * Modify behavior of ALPN to invoke callback after SNI/servername
4209 callback, such that updates to the SSL_CTX affect ALPN.
4213 * Add SSL_CIPHER queries for authentication and key-exchange.
4217 * Changes to the DEFAULT cipherlist:
4218 - Prefer (EC)DHE handshakes over plain RSA.
4219 - Prefer AEAD ciphers over legacy ciphers.
4220 - Prefer ECDSA over RSA when both certificates are available.
4221 - Prefer TLSv1.2 ciphers/PRF.
4222 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4227 * Change the ECC default curve list to be this, in order: x25519,
4228 secp256r1, secp521r1, secp384r1.
4232 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4233 disabled by default. They can be re-enabled using the
4234 enable-weak-ssl-ciphers option to Configure.
4238 * If the server has ALPN configured, but supports no protocols that the
4239 client advertises, send a fatal "no_application_protocol" alert.
4240 This behaviour is SHALL in RFC 7301, though it isn't universally
4241 implemented by other servers.
4245 * Add X25519 support.
4246 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4247 for public and private key encoding using the format documented in
4248 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4249 key generation and key derivation.
4251 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4256 * Deprecate SRP_VBASE_get_by_user.
4257 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4258 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4259 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4260 seed, even if the seed is configured.
4262 Users should use SRP_VBASE_get1_by_user instead. Note that in
4263 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4264 also that even though configuring the SRP seed attempts to hide
4265 invalid usernames by continuing the handshake with fake
4266 credentials, this behaviour is not constant time and no strong
4267 guarantees are made that the handshake is indistinguishable from
4268 that of a valid user.
4272 * Configuration change; it's now possible to build dynamic engines
4273 without having to build shared libraries and vice versa. This
4274 only applies to the engines in `engines/`, those in `crypto/engine/`
4275 will always be built into libcrypto (i.e. "static").
4277 Building dynamic engines is enabled by default; to disable, use
4278 the configuration option "disable-dynamic-engine".
4280 The only requirements for building dynamic engines are the
4281 presence of the DSO module and building with position independent
4282 code, so they will also automatically be disabled if configuring
4283 with "disable-dso" or "disable-pic".
4285 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4286 are also taken away from openssl/opensslconf.h, as they are
4291 * Configuration change; if there is a known flag to compile
4292 position independent code, it will always be applied on the
4293 libcrypto and libssl object files, and never on the application
4294 object files. This means other libraries that use routines from
4295 libcrypto / libssl can be made into shared libraries regardless
4296 of how OpenSSL was configured.
4298 If this isn't desirable, the configuration options "disable-pic"
4299 or "no-pic" can be used to disable the use of PIC. This will
4300 also disable building shared libraries and dynamic engines.
4304 * Removed JPAKE code. It was experimental and has no wide use.
4308 * The INSTALL_PREFIX Makefile variable has been renamed to
4309 DESTDIR. That makes for less confusion on what this variable
4310 is for. Also, the configuration option --install_prefix is
4315 * Heartbeat for TLS has been removed and is disabled by default
4316 for DTLS; configure with enable-heartbeats. Code that uses the
4317 old #define's might need to be updated.
4319 *Emilia Käsper, Rich Salz*
4321 * Rename REF_CHECK to REF_DEBUG.
4325 * New "unified" build system
4327 The "unified" build system is aimed to be a common system for all
4328 platforms we support. With it comes new support for VMS.
4330 This system builds supports building in a different directory tree
4331 than the source tree. It produces one Makefile (for unix family
4332 or lookalikes), or one descrip.mms (for VMS).
4334 The source of information to make the Makefile / descrip.mms is
4335 small files called 'build.info', holding the necessary
4336 information for each directory with source to compile, and a
4337 template in Configurations, like unix-Makefile.tmpl or
4340 With this change, the library names were also renamed on Windows
4341 and on VMS. They now have names that are closer to the standard
4342 on Unix, and include the major version number, and in certain
4343 cases, the architecture they are built for. See "Notes on shared
4344 libraries" in INSTALL.
4346 We rely heavily on the perl module Text::Template.
4350 * Added support for auto-initialisation and de-initialisation of the library.
4351 OpenSSL no longer requires explicit init or deinit routines to be called,
4352 except in certain circumstances. See the OPENSSL_init_crypto() and
4353 OPENSSL_init_ssl() man pages for further information.
4357 * The arguments to the DTLSv1_listen function have changed. Specifically the
4358 "peer" argument is now expected to be a BIO_ADDR object.
4360 * Rewrite of BIO networking library. The BIO library lacked consistent
4361 support of IPv6, and adding it required some more extensive
4362 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4363 which hold all types of addresses and chains of address information.
4364 It also introduces a new API, with functions like BIO_socket,
4365 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4366 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4367 have been adapted accordingly.
4371 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4376 * CRIME protection: disable compression by default, even if OpenSSL is
4377 compiled with zlib enabled. Applications can still enable compression
4378 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4379 using the SSL_CONF library to configure compression.
4383 * The signature of the session callback configured with
4384 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4385 was explicitly marked as `const unsigned char*` instead of
4390 * Always DPURIFY. Remove the use of uninitialized memory in the
4391 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4395 * Removed many obsolete configuration items, including
4396 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4397 MD2_CHAR, MD2_INT, MD2_LONG
4399 IDEA_SHORT, IDEA_LONG
4400 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4402 *Rich Salz, with advice from Andy Polyakov*
4404 * Many BN internals have been moved to an internal header file.
4406 *Rich Salz with help from Andy Polyakov*
4408 * Configuration and writing out the results from it has changed.
4409 Files such as Makefile include/openssl/opensslconf.h and are now
4410 produced through general templates, such as Makefile.in and
4411 crypto/opensslconf.h.in and some help from the perl module
4414 Also, the center of configuration information is no longer
4415 Makefile. Instead, Configure produces a perl module in
4416 configdata.pm which holds most of the config data (in the hash
4417 table %config), the target data that comes from the target
4418 configuration in one of the `Configurations/*.conf` files (in
4423 * To clarify their intended purposes, the Configure options
4424 --prefix and --openssldir change their semantics, and become more
4425 straightforward and less interdependent.
4427 --prefix shall be used exclusively to give the location INSTALLTOP
4428 where programs, scripts, libraries, include files and manuals are
4429 going to be installed. The default is now /usr/local.
4431 --openssldir shall be used exclusively to give the default
4432 location OPENSSLDIR where certificates, private keys, CRLs are
4433 managed. This is also where the default openssl.cnf gets
4435 If the directory given with this option is a relative path, the
4436 values of both the --prefix value and the --openssldir value will
4437 be combined to become OPENSSLDIR.
4438 The default for --openssldir is INSTALLTOP/ssl.
4440 Anyone who uses --openssldir to specify where OpenSSL is to be
4441 installed MUST change to use --prefix instead.
4445 * The GOST engine was out of date and therefore it has been removed. An up
4446 to date GOST engine is now being maintained in an external repository.
4447 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4448 support for GOST ciphersuites (these are only activated if a GOST engine
4453 * EGD is no longer supported by default; use enable-egd when
4456 *Ben Kaduk and Rich Salz*
4458 * The distribution now has Makefile.in files, which are used to
4459 create Makefile's when Configure is run. *Configure must be run
4460 before trying to build now.*
4464 * The return value for SSL_CIPHER_description() for error conditions
4469 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4471 Obtaining and performing DNSSEC validation of TLSA records is
4472 the application's responsibility. The application provides
4473 the TLSA records of its choice to OpenSSL, and these are then
4474 used to authenticate the peer.
4476 The TLSA records need not even come from DNS. They can, for
4477 example, be used to implement local end-entity certificate or
4478 trust-anchor "pinning", where the "pin" data takes the form
4479 of TLSA records, which can augment or replace verification
4480 based on the usual WebPKI public certification authorities.
4484 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4485 continues to support deprecated interfaces in default builds.
4486 However, applications are strongly advised to compile their
4487 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4488 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4489 or the 1.1.0 releases.
4491 In environments in which all applications have been ported to
4492 not use any deprecated interfaces OpenSSL's Configure script
4493 should be used with the --api=1.1.0 option to entirely remove
4494 support for the deprecated features from the library and
4495 unconditionally disable them in the installed headers.
4496 Essentially the same effect can be achieved with the "no-deprecated"
4497 argument to Configure, except that this will always restrict
4498 the build to just the latest API, rather than a fixed API
4501 As applications are ported to future revisions of the API,
4502 they should update their compile-time OPENSSL_API_COMPAT define
4503 accordingly, but in most cases should be able to continue to
4504 compile with later releases.
4506 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4507 0x10000000L and 0x00908000L, respectively. However those
4508 versions did not support the OPENSSL_API_COMPAT feature, and
4509 so applications are not typically tested for explicit support
4510 of just the undeprecated features of either release.
4514 * Add support for setting the minimum and maximum supported protocol.
4515 It can bet set via the SSL_set_min_proto_version() and
4516 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4517 MaxProtocol. It's recommended to use the new APIs to disable
4518 protocols instead of disabling individual protocols using
4519 SSL_set_options() or SSL_CONF's Protocol. This change also
4520 removes support for disabling TLS 1.2 in the OpenSSL TLS
4521 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4525 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4529 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4530 and integrates ECDSA and ECDH functionality into EC. Implementations can
4531 now redirect key generation and no longer need to convert to or from
4534 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4535 include the ec.h header file instead.
4539 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4540 ciphers who are no longer supported and drops support the ephemeral RSA key
4541 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4545 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4546 opaque. For HMAC_CTX, the following constructors and destructors
4549 HMAC_CTX *HMAC_CTX_new(void);
4550 void HMAC_CTX_free(HMAC_CTX *ctx);
4552 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4553 destroy such methods has been added. See EVP_MD_meth_new(3) and
4554 EVP_CIPHER_meth_new(3) for documentation.
4557 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4558 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4559 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4560 an already created structure.
4561 2) For consistency with the majority of our object creators and
4562 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4563 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4564 for deprecated builds.
4568 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4569 cryptographic operations to be performed asynchronously as long as an
4570 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4571 further details. Libssl has also had this capability integrated with the
4572 introduction of the new mode SSL_MODE_ASYNC and associated error
4573 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4574 pages. This work was developed in partnership with Intel Corp.
4578 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4579 always enabled now. If you want to disable the support you should
4580 exclude it using the list of supported ciphers. This also means that the
4581 "-no_ecdhe" option has been removed from s_server.
4585 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4586 SSL_{CTX_}set1_curves() which can set a list.
4590 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4591 curve you want to support using SSL_{CTX_}set1_curves().
4595 * State machine rewrite. The state machine code has been significantly
4596 refactored in order to remove much duplication of code and solve issues
4597 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4598 further details). This change does have some associated API changes.
4599 Notably the SSL_state() function has been removed and replaced by
4600 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4601 SSL_set_state() has been removed altogether. The previous handshake states
4602 defined in ssl.h and ssl3.h have also been removed.
4606 * All instances of the string "ssleay" in the public API were replaced
4607 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4608 Some error codes related to internal RSA_eay API's were renamed.
4612 * The demo files in crypto/threads were moved to demo/threads.
4616 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4619 *Matt Caswell, Rich Salz*
4621 * New ASN.1 embed macro.
4623 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4624 structure is not allocated: it is part of the parent. That is instead of
4632 This reduces memory fragmentation and make it impossible to accidentally
4633 set a mandatory field to NULL.
4635 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4636 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4637 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4642 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4646 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4647 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4648 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4649 DES and RC4 ciphersuites.
4653 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4654 This changes the decoding behaviour for some invalid messages,
4655 though the change is mostly in the more lenient direction, and
4656 legacy behaviour is preserved as much as possible.
4660 * Fix no-stdio build.
4661 *David Woodhouse <David.Woodhouse@intel.com> and also*
4662 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4664 * New testing framework
4665 The testing framework has been largely rewritten and is now using
4666 perl and the perl modules Test::Harness and an extended variant of
4667 Test::More called OpenSSL::Test to do its work. All test scripts in
4668 test/ have been rewritten into test recipes, and all direct calls to
4669 executables in test/Makefile have become individual recipes using the
4670 simplified testing OpenSSL::Test::Simple.
4672 For documentation on our testing modules, do:
4674 perldoc test/testlib/OpenSSL/Test/Simple.pm
4675 perldoc test/testlib/OpenSSL/Test.pm
4679 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4680 are used; the latter aborts on memory leaks (usually checked on exit).
4681 Some undocumented "set malloc, etc., hooks" functions were removed
4682 and others were changed. All are now documented.
4686 * In DSA_generate_parameters_ex, if the provided seed is too short,
4689 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4691 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4692 from RFC4279, RFC4785, RFC5487, RFC5489.
4694 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4695 original RSA_PSK patch.
4699 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4700 era flag was never set throughout the codebase (only read). Also removed
4701 SSL3_FLAGS_POP_BUFFER which was only used if
4702 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4706 * Changed the default name options in the "ca", "crl", "req" and "x509"
4707 to be "oneline" instead of "compat".
4711 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4712 not aware of clients that still exhibit this bug, and the workaround
4713 hasn't been working properly for a while.
4717 * The return type of BIO_number_read() and BIO_number_written() as well as
4718 the corresponding num_read and num_write members in the BIO structure has
4719 changed from unsigned long to uint64_t. On platforms where an unsigned
4720 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4725 * Given the pervasive nature of TLS extensions it is inadvisable to run
4726 OpenSSL without support for them. It also means that maintaining
4727 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4728 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4732 * Removed support for the two export grade static DH ciphersuites
4733 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4734 were newly added (along with a number of other static DH ciphersuites) to
4735 1.0.2. However the two export ones have *never* worked since they were
4736 introduced. It seems strange in any case to be adding new export
4737 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4741 * Version negotiation has been rewritten. In particular SSLv23_method(),
4742 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4743 and turned into macros which simply call the new preferred function names
4744 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4745 should use the new names instead. Also as part of this change the ssl23.h
4746 header file has been removed.
4750 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4751 code and the associated standard is no longer considered fit-for-purpose.
4755 * RT2547 was closed. When generating a private key, try to make the
4756 output file readable only by the owner. This behavior change might
4757 be noticeable when interacting with other software.
4759 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4764 * Added HTTP GET support to the ocsp command.
4768 * Changed default digest for the dgst and enc commands from MD5 to
4773 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4777 * Added support for TLS extended master secret from
4778 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4779 initial patch which was a great help during development.
4783 * All libssl internal structures have been removed from the public header
4784 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4785 now redundant). Users should not attempt to access internal structures
4786 directly. Instead they should use the provided API functions.
4790 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4791 Access to deprecated functions can be re-enabled by running config with
4792 "enable-deprecated". In addition applications wishing to use deprecated
4793 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4794 will, by default, disable some transitive includes that previously existed
4795 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4799 * Added support for OCB mode. OpenSSL has been granted a patent license
4800 compatible with the OpenSSL license for use of OCB. Details are available
4801 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4802 for OCB can be removed by calling config with no-ocb.
4806 * SSLv2 support has been removed. It still supports receiving an SSLv2
4807 compatible client hello.
4811 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4812 done while fixing the error code for the key-too-small case.
4814 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4816 * CA.sh has been removed; use CA.pl instead.
4820 * Removed old DES API.
4824 * Remove various unsupported platforms:
4830 Sinix/ReliantUNIX RM400
4835 16-bit platforms such as WIN16
4839 * Clean up OPENSSL_NO_xxx #define's
4840 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4841 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4842 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4843 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4844 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4845 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4846 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4847 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4848 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4849 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4853 * Cleaned up dead code
4854 Remove all but one '#ifdef undef' which is to be looked at.
4858 * Clean up calling of xxx_free routines.
4859 Just like free(), fix most of the xxx_free routines to accept
4860 NULL. Remove the non-null checks from callers. Save much code.
4864 * Add secure heap for storage of private keys (when possible).
4865 Add BIO_s_secmem(), CBIGNUM, etc.
4866 Contributed by Akamai Technologies under our Corporate CLA.
4870 * Experimental support for a new, fast, unbiased prime candidate generator,
4871 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4873 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4875 * New output format NSS in the sess_id command line tool. This allows
4876 exporting the session id and the master key in NSS keylog format.
4878 *Martin Kaiser <martin@kaiser.cx>*
4880 * Harmonize version and its documentation. -f flag is used to display
4883 *mancha <mancha1@zoho.com>*
4885 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4886 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4888 *mancha <mancha1@zoho.com>*
4890 * Fix some double frees. These are not thought to be exploitable.
4892 *mancha <mancha1@zoho.com>*
4894 * A missing bounds check in the handling of the TLS heartbeat extension
4895 can be used to reveal up to 64k of memory to a connected client or
4898 Thanks for Neel Mehta of Google Security for discovering this bug and to
4899 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4900 preparing the fix ([CVE-2014-0160])
4902 *Adam Langley, Bodo Moeller*
4904 * Fix for the attack described in the paper "Recovering OpenSSL
4905 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4906 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4907 <http://eprint.iacr.org/2014/140>
4909 Thanks to Yuval Yarom and Naomi Benger for discovering this
4910 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4912 *Yuval Yarom and Naomi Benger*
4914 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4915 this fixes a limitation in previous versions of OpenSSL.
4919 * Experimental encrypt-then-mac support.
4921 Experimental support for encrypt then mac from
4922 draft-gutmann-tls-encrypt-then-mac-02.txt
4924 To enable it set the appropriate extension number (0x42 for the test
4925 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4927 For non-compliant peers (i.e. just about everything) this should have no
4930 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4934 * Add EVP support for key wrapping algorithms, to avoid problems with
4935 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4936 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4937 algorithms and include tests cases.
4941 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4946 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4947 MGF1 digest and OAEP label.
4951 * Make openssl verify return errors.
4953 *Chris Palmer <palmer@google.com> and Ben Laurie*
4955 * New function ASN1_TIME_diff to calculate the difference between two
4956 ASN1_TIME structures or one structure and the current time.
4960 * Update fips_test_suite to support multiple command line options. New
4961 test to induce all self test errors in sequence and check expected
4966 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4967 sign or verify all in one operation.
4971 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4972 test programs and fips_test_suite. Includes functionality to parse
4973 the minimal script output of fipsalgest.pl directly.
4977 * Add authorisation parameter to FIPS_module_mode_set().
4981 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4985 * Use separate DRBG fields for internal and external flags. New function
4986 FIPS_drbg_health_check() to perform on demand health checking. Add
4987 generation tests to fips_test_suite with reduced health check interval to
4988 demonstrate periodic health checking. Add "nodh" option to
4989 fips_test_suite to skip very slow DH test.
4993 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4998 * More extensive health check for DRBG checking many more failure modes.
4999 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5000 combination: call this in fips_test_suite.
5004 * Add support for canonical generation of DSA parameter 'g'. See
5007 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5008 POST to handle HMAC cases.
5012 * Add functions FIPS_module_version() and FIPS_module_version_text()
5013 to return numerical and string versions of the FIPS module number.
5017 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5018 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5019 outside the validated module in the FIPS capable OpenSSL.
5023 * Minor change to DRBG entropy callback semantics. In some cases
5024 there is no multiple of the block length between min_len and
5025 max_len. Allow the callback to return more than max_len bytes
5026 of entropy but discard any extra: it is the callback's responsibility
5027 to ensure that the extra data discarded does not impact the
5028 requested amount of entropy.
5032 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5033 information in FIPS186-3, SP800-57 and SP800-131A.
5037 * CCM support via EVP. Interface is very similar to GCM case except we
5038 must supply all data in one chunk (i.e. no update, final) and the
5039 message length must be supplied if AAD is used. Add algorithm test
5044 * Initial version of POST overhaul. Add POST callback to allow the status
5045 of POST to be monitored and/or failures induced. Modify fips_test_suite
5046 to use callback. Always run all selftests even if one fails.
5050 * XTS support including algorithm test driver in the fips_gcmtest program.
5051 Note: this does increase the maximum key length from 32 to 64 bytes but
5052 there should be no binary compatibility issues as existing applications
5053 will never use XTS mode.
5057 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5058 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5059 performs algorithm blocking for unapproved PRNG types. Also do not
5060 set PRNG type in FIPS_mode_set(): leave this to the application.
5061 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5062 the standard OpenSSL PRNG: set additional data to a date time vector.
5066 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5067 This shouldn't present any incompatibility problems because applications
5068 shouldn't be using these directly and any that are will need to rethink
5069 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5073 * Extensive self tests and health checking required by SP800-90 DRBG.
5074 Remove strength parameter from FIPS_drbg_instantiate and always
5075 instantiate at maximum supported strength.
5079 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5083 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5087 * New function DH_compute_key_padded() to compute a DH key and pad with
5088 leading zeroes if needed: this complies with SP800-56A et al.
5092 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5093 anything, incomplete, subject to change and largely untested at present.
5097 * Modify fipscanisteronly build option to only build the necessary object
5098 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5102 * Add experimental option FIPSSYMS to give all symbols in
5103 fipscanister.o and FIPS or fips prefix. This will avoid
5104 conflicts with future versions of OpenSSL. Add perl script
5105 util/fipsas.pl to preprocess assembly language source files
5106 and rename any affected symbols.
5110 * Add selftest checks and algorithm block of non-fips algorithms in
5111 FIPS mode. Remove DES2 from selftests.
5115 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5116 return internal method without any ENGINE dependencies. Add new
5117 tiny fips sign and verify functions.
5121 * New build option no-ec2m to disable characteristic 2 code.
5125 * New build option "fipscanisteronly". This only builds fipscanister.o
5126 and (currently) associated fips utilities. Uses the file Makefile.fips
5127 instead of Makefile.org as the prototype.
5131 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5132 Update fips_gcmtest to use IV generator.
5136 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5137 setting output buffer to NULL. The `*Final` function must be
5138 called although it will not retrieve any additional data. The tag
5139 can be set or retrieved with a ctrl. The IV length is by default 12
5140 bytes (96 bits) but can be set to an alternative value. If the IV
5141 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5146 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5147 underlying do_cipher function handles all cipher semantics itself
5148 including padding and finalisation. This is useful if (for example)
5149 an ENGINE cipher handles block padding itself. The behaviour of
5150 do_cipher is subtly changed if this flag is set: the return value
5151 is the number of characters written to the output buffer (zero is
5152 no longer an error code) or a negative error code. Also if the
5153 input buffer is NULL and length 0 finalisation should be performed.
5157 * If a candidate issuer certificate is already part of the constructed
5158 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5162 * Improve forward-security support: add functions
5164 void SSL_CTX_set_not_resumable_session_callback(
5165 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5166 void SSL_set_not_resumable_session_callback(
5167 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5169 for use by SSL/TLS servers; the callback function will be called whenever a
5170 new session is created, and gets to decide whether the session may be
5171 cached to make it resumable (return 0) or not (return 1). (As by the
5172 SSL/TLS protocol specifications, the session_id sent by the server will be
5173 empty to indicate that the session is not resumable; also, the server will
5174 not generate RFC 4507 (RFC 5077) session tickets.)
5176 A simple reasonable callback implementation is to return is_forward_secure.
5177 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5178 by the SSL/TLS server library, indicating whether it can provide forward
5181 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5183 * New -verify_name option in command line utilities to set verification
5188 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5189 Add CMAC pkey methods.
5193 * Experimental renegotiation in s_server -www mode. If the client
5194 browses /reneg connection is renegotiated. If /renegcert it is
5195 renegotiated requesting a certificate.
5199 * Add an "external" session cache for debugging purposes to s_server. This
5200 should help trace issues which normally are only apparent in deployed
5201 multi-process servers.
5205 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5206 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5207 BIO_set_cipher() and some obscure PEM functions were changed so they
5208 can now return an error. The RAND changes required a change to the
5209 RAND_METHOD structure.
5213 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5214 a gcc attribute to warn if the result of a function is ignored. This
5215 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5216 whose return value is often ignored.
5220 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5221 These allow SCTs (signed certificate timestamps) to be requested and
5222 validated when establishing a connection.
5224 *Rob Percival <robpercival@google.com>*
5229 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5231 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5232 used even when parsing explicit parameters, when loading a encoded key
5233 or calling `EC_GROUP_new_from_ecpkparameters()`/
5234 `EC_GROUP_new_from_ecparameters()`.
5235 This prevents bypass of security hardening and performance gains,
5236 especially for curves with specialized EC_METHODs.
5237 By default, if a key encoded with explicit parameters is loaded and later
5238 encoded, the output is still encoded with explicit parameters, even if
5239 internally a "named" EC_GROUP is used for computation.
5243 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5244 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5245 NULL. After this change, only the cofactor parameter can be NULL. It also
5246 does some minimal sanity checks on the passed order.
5251 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5252 An attack is simple, if the first CMS_recipientInfo is valid but the
5253 second CMS_recipientInfo is chosen ciphertext. If the second
5254 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5255 encryption key will be replaced by garbage, and the message cannot be
5256 decoded, but if the RSA decryption fails, the correct encryption key is
5257 used and the recipient will not notice the attack.
5258 As a work around for this potential attack the length of the decrypted
5259 key must be equal to the cipher default key length, in case the
5260 certificate is not given and all recipientInfo are tried out.
5261 The old behaviour can be re-enabled in the CMS code by setting the
5262 CMS_DEBUG_DECRYPT flag.
5267 * Document issue with installation paths in diverse Windows builds
5269 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5270 binaries and run-time config file.
5275 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5277 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5278 This changes the size when using the `genpkey` command when no size is given.
5279 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5280 generation commands to use 2048 bits by default.
5284 * Add FIPS support for Android Arm 64-bit
5286 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5287 Module in Version 2.0.10. For some reason, the corresponding target
5288 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5289 built with FIPS support on Android Arm 64-bit. This omission has been
5292 *Matthias St. Pierre*
5294 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5296 * 0-byte record padding oracle
5298 If an application encounters a fatal protocol error and then calls
5299 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5300 then OpenSSL can respond differently to the calling application if a 0 byte
5301 record is received with invalid padding compared to if a 0 byte record is
5302 received with an invalid MAC. If the application then behaves differently
5303 based on that in a way that is detectable to the remote peer, then this
5304 amounts to a padding oracle that could be used to decrypt data.
5306 In order for this to be exploitable "non-stitched" ciphersuites must be in
5307 use. Stitched ciphersuites are optimised implementations of certain
5308 commonly used ciphersuites. Also the application must call SSL_shutdown()
5309 twice even if a protocol error has occurred (applications should not do
5310 this but some do anyway).
5312 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5313 Aviram, with additional investigation by Steven Collison and Andrew
5314 Hourselt. It was reported to OpenSSL on 10th December 2018.
5319 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5323 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5325 * Microarchitecture timing vulnerability in ECC scalar multiplication
5327 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5328 shown to be vulnerable to a microarchitecture timing side channel attack.
5329 An attacker with sufficient access to mount local timing attacks during
5330 ECDSA signature generation could recover the private key.
5332 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5333 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5339 * Timing vulnerability in DSA signature generation
5341 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5342 timing side channel attack. An attacker could use variations in the signing
5343 algorithm to recover the private key.
5345 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5350 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5351 Module, accidentally introduced while backporting security fixes from the
5352 development branch and hindering the use of ECC in FIPS mode.
5356 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5358 * Client DoS due to large DH parameter
5360 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5361 malicious server can send a very large prime value to the client. This will
5362 cause the client to spend an unreasonably long period of time generating a
5363 key for this prime resulting in a hang until the client has finished. This
5364 could be exploited in a Denial Of Service attack.
5366 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5371 * Cache timing vulnerability in RSA Key Generation
5373 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5374 a cache timing side channel attack. An attacker with sufficient access to
5375 mount cache timing attacks during the RSA key generation process could
5376 recover the private key.
5378 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5379 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5384 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5385 parameter is no longer accepted, as it leads to a corrupt table. NULL
5386 pem_str is reserved for alias entries only.
5390 * Revert blinding in ECDSA sign and instead make problematic addition
5391 length-invariant. Switch even to fixed-length Montgomery multiplication.
5395 * Change generating and checking of primes so that the error rate of not
5396 being prime depends on the intended use based on the size of the input.
5397 For larger primes this will result in more rounds of Miller-Rabin.
5398 The maximal error rate for primes with more than 1080 bits is lowered
5401 *Kurt Roeckx, Annie Yousar*
5403 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5407 * Add blinding to ECDSA and DSA signatures to protect against side channel
5408 attacks discovered by Keegan Ryan (NCC Group).
5412 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5413 now allow empty (zero character) pass phrases.
5417 * Certificate time validation (X509_cmp_time) enforces stricter
5418 compliance with RFC 5280. Fractional seconds and timezone offsets
5419 are no longer allowed.
5423 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5425 * Constructed ASN.1 types with a recursive definition could exceed the stack
5427 Constructed ASN.1 types with a recursive definition (such as can be found
5428 in PKCS7) could eventually exceed the stack given malicious input with
5429 excessive recursion. This could result in a Denial Of Service attack. There
5430 are no such structures used within SSL/TLS that come from untrusted sources
5431 so this is considered safe.
5433 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5439 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5441 * Read/write after SSL object in error state
5443 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5444 mechanism. The intent was that if a fatal error occurred during a handshake
5445 then OpenSSL would move into the error state and would immediately fail if
5446 you attempted to continue the handshake. This works as designed for the
5447 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5448 SSL_connect()), however due to a bug it does not work correctly if
5449 SSL_read() or SSL_write() is called directly. In that scenario, if the
5450 handshake fails then a fatal error will be returned in the initial function
5451 call. If SSL_read()/SSL_write() is subsequently called by the application
5452 for the same SSL object then it will succeed and the data is passed without
5453 being decrypted/encrypted directly from the SSL/TLS record layer.
5455 In order to exploit this issue an application bug would have to be present
5456 that resulted in a call to SSL_read()/SSL_write() being issued after having
5457 already received a fatal error.
5459 This issue was reported to OpenSSL by David Benjamin (Google).
5464 * rsaz_1024_mul_avx2 overflow bug on x86_64
5466 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5467 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5468 Analysis suggests that attacks against RSA and DSA as a result of this
5469 defect would be very difficult to perform and are not believed likely.
5470 Attacks against DH1024 are considered just feasible, because most of the
5471 work necessary to deduce information about a private key may be performed
5472 offline. The amount of resources required for such an attack would be
5473 significant. However, for an attack on TLS to be meaningful, the server
5474 would have to share the DH1024 private key among multiple clients, which is
5475 no longer an option since CVE-2016-0701.
5477 This only affects processors that support the AVX2 but not ADX extensions
5478 like Intel Haswell (4th generation).
5480 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5481 was originally found via the OSS-Fuzz project.
5486 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5488 * bn_sqrx8x_internal carry bug on x86_64
5490 There is a carry propagating bug in the x86_64 Montgomery squaring
5491 procedure. No EC algorithms are affected. Analysis suggests that attacks
5492 against RSA and DSA as a result of this defect would be very difficult to
5493 perform and are not believed likely. Attacks against DH are considered just
5494 feasible (although very difficult) because most of the work necessary to
5495 deduce information about a private key may be performed offline. The amount
5496 of resources required for such an attack would be very significant and
5497 likely only accessible to a limited number of attackers. An attacker would
5498 additionally need online access to an unpatched system using the target
5499 private key in a scenario with persistent DH parameters and a private
5500 key that is shared between multiple clients.
5502 This only affects processors that support the BMI1, BMI2 and ADX extensions
5503 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5505 This issue was reported to OpenSSL by the OSS-Fuzz project.
5510 * Malformed X.509 IPAddressFamily could cause OOB read
5512 If an X.509 certificate has a malformed IPAddressFamily extension,
5513 OpenSSL could do a one-byte buffer overread. The most likely result
5514 would be an erroneous display of the certificate in text format.
5516 This issue was reported to OpenSSL by the OSS-Fuzz project.
5520 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5522 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5523 platform rather than 'mingw'.
5527 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5529 * Truncated packet could crash via OOB read
5531 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5532 cipher is being used, then a truncated packet can cause that host to
5533 perform an out-of-bounds read, usually resulting in a crash.
5535 This issue was reported to OpenSSL by Robert Święcki of Google.
5540 * BN_mod_exp may produce incorrect results on x86_64
5542 There is a carry propagating bug in the x86_64 Montgomery squaring
5543 procedure. No EC algorithms are affected. Analysis suggests that attacks
5544 against RSA and DSA as a result of this defect would be very difficult to
5545 perform and are not believed likely. Attacks against DH are considered just
5546 feasible (although very difficult) because most of the work necessary to
5547 deduce information about a private key may be performed offline. The amount
5548 of resources required for such an attack would be very significant and
5549 likely only accessible to a limited number of attackers. An attacker would
5550 additionally need online access to an unpatched system using the target
5551 private key in a scenario with persistent DH parameters and a private
5552 key that is shared between multiple clients. For example this can occur by
5553 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5554 similar to CVE-2015-3193 but must be treated as a separate problem.
5556 This issue was reported to OpenSSL by the OSS-Fuzz project.
5561 * Montgomery multiplication may produce incorrect results
5563 There is a carry propagating bug in the Broadwell-specific Montgomery
5564 multiplication procedure that handles input lengths divisible by, but
5565 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5566 and DH private keys are impossible. This is because the subroutine in
5567 question is not used in operations with the private key itself and an input
5568 of the attacker's direct choice. Otherwise the bug can manifest itself as
5569 transient authentication and key negotiation failures or reproducible
5570 erroneous outcome of public-key operations with specially crafted input.
5571 Among EC algorithms only Brainpool P-512 curves are affected and one
5572 presumably can attack ECDH key negotiation. Impact was not analyzed in
5573 detail, because pre-requisites for attack are considered unlikely. Namely
5574 multiple clients have to choose the curve in question and the server has to
5575 share the private key among them, neither of which is default behaviour.
5576 Even then only clients that chose the curve will be affected.
5578 This issue was publicly reported as transient failures and was not
5579 initially recognized as a security issue. Thanks to Richard Morgan for
5580 providing reproducible case.
5585 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5586 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5587 prevent issues where no progress is being made and the peer continually
5588 sends unrecognised record types, using up resources processing them.
5592 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5594 * Missing CRL sanity check
5596 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5597 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5598 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5600 This issue only affects the OpenSSL 1.0.2i
5605 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5607 * OCSP Status Request extension unbounded memory growth
5609 A malicious client can send an excessively large OCSP Status Request
5610 extension. If that client continually requests renegotiation, sending a
5611 large OCSP Status Request extension each time, then there will be unbounded
5612 memory growth on the server. This will eventually lead to a Denial Of
5613 Service attack through memory exhaustion. Servers with a default
5614 configuration are vulnerable even if they do not support OCSP. Builds using
5615 the "no-ocsp" build time option are not affected.
5617 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5622 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5625 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5631 * OOB write in MDC2_Update()
5633 An overflow can occur in MDC2_Update() either if called directly or
5634 through the EVP_DigestUpdate() function using MDC2. If an attacker
5635 is able to supply very large amounts of input data after a previous
5636 call to EVP_EncryptUpdate() with a partial block then a length check
5637 can overflow resulting in a heap corruption.
5639 The amount of data needed is comparable to SIZE_MAX which is impractical
5642 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5647 * Malformed SHA512 ticket DoS
5649 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5650 DoS attack where a malformed ticket will result in an OOB read which will
5653 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5654 a custom server callback and ticket lookup mechanism.
5656 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5661 * OOB write in BN_bn2dec()
5663 The function BN_bn2dec() does not check the return value of BN_div_word().
5664 This can cause an OOB write if an application uses this function with an
5665 overly large BIGNUM. This could be a problem if an overly large certificate
5666 or CRL is printed out from an untrusted source. TLS is not affected because
5667 record limits will reject an oversized certificate before it is parsed.
5669 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5674 * OOB read in TS_OBJ_print_bio()
5676 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5677 the total length the OID text representation would use and not the amount
5678 of data written. This will result in OOB reads when large OIDs are
5681 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5686 * Pointer arithmetic undefined behaviour
5688 Avoid some undefined pointer arithmetic
5690 A common idiom in the codebase is to check limits in the following manner:
5693 Where "p" points to some malloc'd data of SIZE bytes and
5696 "len" here could be from some externally supplied data (e.g. from a TLS
5699 The rules of C pointer arithmetic are such that "p + len" is only well
5700 defined where len <= SIZE. Therefore the above idiom is actually
5701 undefined behaviour.
5703 For example this could cause problems if some malloc implementation
5704 provides an address for "p" such that "p + len" actually overflows for
5705 values of len that are too big and therefore p + len < limit.
5707 This issue was reported to OpenSSL by Guido Vranken
5712 * Constant time flag not preserved in DSA signing
5714 Operations in the DSA signing algorithm should run in constant time in
5715 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5716 implementation means that a non-constant time codepath is followed for
5717 certain operations. This has been demonstrated through a cache-timing
5718 attack to be sufficient for an attacker to recover the private DSA key.
5720 This issue was reported by César Pereida (Aalto University), Billy Brumley
5721 (Tampere University of Technology), and Yuval Yarom (The University of
5722 Adelaide and NICTA).
5727 * DTLS buffered message DoS
5729 In a DTLS connection where handshake messages are delivered out-of-order
5730 those messages that OpenSSL is not yet ready to process will be buffered
5731 for later use. Under certain circumstances, a flaw in the logic means that
5732 those messages do not get removed from the buffer even though the handshake
5733 has been completed. An attacker could force up to approx. 15 messages to
5734 remain in the buffer when they are no longer required. These messages will
5735 be cleared when the DTLS connection is closed. The default maximum size for
5736 a message is 100k. Therefore, the attacker could force an additional 1500k
5737 to be consumed per connection. By opening many simultaneous connections an
5738 attacker could cause a DoS attack through memory exhaustion.
5740 This issue was reported to OpenSSL by Quan Luo.
5745 * DTLS replay protection DoS
5747 A flaw in the DTLS replay attack protection mechanism means that records
5748 that arrive for future epochs update the replay protection "window" before
5749 the MAC for the record has been validated. This could be exploited by an
5750 attacker by sending a record for the next epoch (which does not have to
5751 decrypt or have a valid MAC), with a very large sequence number. This means
5752 that all subsequent legitimate packets are dropped causing a denial of
5753 service for a specific DTLS connection.
5755 This issue was reported to OpenSSL by the OCAP audit team.
5760 * Certificate message OOB reads
5762 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5763 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5764 theoretical DoS risk but this has not been observed in practice on common
5767 The messages affected are client certificate, client certificate request
5768 and server certificate. As a result the attack can only be performed
5769 against a client or a server which enables client authentication.
5771 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5776 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5778 * Prevent padding oracle in AES-NI CBC MAC check
5780 A MITM attacker can use a padding oracle attack to decrypt traffic
5781 when the connection uses an AES CBC cipher and the server support
5784 This issue was introduced as part of the fix for Lucky 13 padding
5785 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5786 constant time by making sure that always the same bytes are read and
5787 compared against either the MAC or padding bytes. But it no longer
5788 checked that there was enough data to have both the MAC and padding
5791 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5795 * Fix EVP_EncodeUpdate overflow
5797 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5798 Base64 encoding of binary data. If an attacker is able to supply very large
5799 amounts of input data then a length check can overflow resulting in a heap
5802 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5803 the `PEM_write_bio*` family of functions. These are mainly used within the
5804 OpenSSL command line applications, so any application which processes data
5805 from an untrusted source and outputs it as a PEM file should be considered
5806 vulnerable to this issue. User applications that call these APIs directly
5807 with large amounts of untrusted data may also be vulnerable.
5809 This issue was reported by Guido Vranken.
5814 * Fix EVP_EncryptUpdate overflow
5816 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5817 is able to supply very large amounts of input data after a previous call to
5818 EVP_EncryptUpdate() with a partial block then a length check can overflow
5819 resulting in a heap corruption. Following an analysis of all OpenSSL
5820 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5821 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5822 the first called function after an EVP_EncryptInit(), and therefore that
5823 specific call must be safe. The second form is where the length passed to
5824 EVP_EncryptUpdate() can be seen from the code to be some small value and
5825 therefore there is no possibility of an overflow. Since all instances are
5826 one of these two forms, it is believed that there can be no overflows in
5827 internal code due to this problem. It should be noted that
5828 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5829 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5830 of these calls have also been analysed too and it is believed there are no
5831 instances in internal usage where an overflow could occur.
5833 This issue was reported by Guido Vranken.
5838 * Prevent ASN.1 BIO excessive memory allocation
5840 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5841 a short invalid encoding can cause allocation of large amounts of memory
5842 potentially consuming excessive resources or exhausting memory.
5844 Any application parsing untrusted data through d2i BIO functions is
5845 affected. The memory based functions such as d2i_X509() are *not* affected.
5846 Since the memory based functions are used by the TLS library, TLS
5847 applications are not affected.
5849 This issue was reported by Brian Carpenter.
5856 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5857 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5858 in arbitrary stack data being returned in the buffer.
5860 This issue was reported by Guido Vranken.
5865 * Modify behavior of ALPN to invoke callback after SNI/servername
5866 callback, such that updates to the SSL_CTX affect ALPN.
5870 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5875 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5876 methods are enabled and ssl2 is disabled the methods return NULL.
5880 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5882 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5883 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5884 provide any "EXPORT" or "LOW" strength ciphers.
5888 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5889 is by default disabled at build-time. Builds that are not configured with
5890 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5891 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5892 will need to explicitly call either of:
5894 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5896 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5898 as appropriate. Even if either of those is used, or the application
5899 explicitly uses the version-specific SSLv2_method() or its client and
5900 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5901 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5902 ciphers, and SSLv2 56-bit DES are no longer available.
5907 * Fix a double-free in DSA code
5909 A double free bug was discovered when OpenSSL parses malformed DSA private
5910 keys and could lead to a DoS attack or memory corruption for applications
5911 that receive DSA private keys from untrusted sources. This scenario is
5914 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5920 * Disable SRP fake user seed to address a server memory leak.
5922 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5924 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5925 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5926 was changed to ignore the "fake user" SRP seed, even if the seed
5929 Users should use SRP_VBASE_get1_by_user instead. Note that in
5930 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5931 also that even though configuring the SRP seed attempts to hide
5932 invalid usernames by continuing the handshake with fake
5933 credentials, this behaviour is not constant time and no strong
5934 guarantees are made that the handshake is indistinguishable from
5935 that of a valid user.
5940 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5942 In the BN_hex2bn function the number of hex digits is calculated using an
5943 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5944 large values of `i` this can result in `bn_expand` not allocating any
5945 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5946 field as NULL leading to a subsequent NULL ptr deref. For very large values
5947 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5948 In this case memory is allocated to the internal BIGNUM data field, but it
5949 is insufficiently sized leading to heap corruption. A similar issue exists
5950 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5951 is ever called by user applications with very large untrusted hex/dec data.
5952 This is anticipated to be a rare occurrence.
5954 All OpenSSL internal usage of these functions use data that is not expected
5955 to be untrusted, e.g. config file data or application command line
5956 arguments. If user developed applications generate config file data based
5957 on untrusted data then it is possible that this could also lead to security
5958 consequences. This is also anticipated to be rare.
5960 This issue was reported to OpenSSL by Guido Vranken.
5965 * Fix memory issues in `BIO_*printf` functions
5967 The internal `fmtstr` function used in processing a "%s" format string in
5968 the `BIO_*printf` functions could overflow while calculating the length of a
5969 string and cause an OOB read when printing very long strings.
5971 Additionally the internal `doapr_outch` function can attempt to write to an
5972 OOB memory location (at an offset from the NULL pointer) in the event of a
5973 memory allocation failure. In 1.0.2 and below this could be caused where
5974 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5975 could be in processing a very long "%s" format string. Memory leaks can
5978 The first issue may mask the second issue dependent on compiler behaviour.
5979 These problems could enable attacks where large amounts of untrusted data
5980 is passed to the `BIO_*printf` functions. If applications use these functions
5981 in this way then they could be vulnerable. OpenSSL itself uses these
5982 functions when printing out human-readable dumps of ASN.1 data. Therefore
5983 applications that print this data could be vulnerable if the data is from
5984 untrusted sources. OpenSSL command line applications could also be
5985 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5986 as command line arguments.
5988 Libssl is not considered directly vulnerable. Additionally certificates etc
5989 received via remote connections via libssl are also unlikely to be able to
5990 trigger these issues because of message size limits enforced within libssl.
5992 This issue was reported to OpenSSL Guido Vranken.
5997 * Side channel attack on modular exponentiation
5999 A side-channel attack was found which makes use of cache-bank conflicts on
6000 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6001 of RSA keys. The ability to exploit this issue is limited as it relies on
6002 an attacker who has control of code in a thread running on the same
6003 hyper-threaded core as the victim thread which is performing decryptions.
6005 This issue was reported to OpenSSL by Yuval Yarom, The University of
6006 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6007 Nadia Heninger, University of Pennsylvania with more information at
6008 <http://cachebleed.info>.
6013 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6014 if no keysize is specified with default_bits. This fixes an
6015 omission in an earlier change that changed all RSA/DSA key generation
6016 commands to use 2048 bits by default.
6020 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6022 * DH small subgroups
6024 Historically OpenSSL only ever generated DH parameters based on "safe"
6025 primes. More recently (in version 1.0.2) support was provided for
6026 generating X9.42 style parameter files such as those required for RFC 5114
6027 support. The primes used in such files may not be "safe". Where an
6028 application is using DH configured with parameters based on primes that are
6029 not "safe" then an attacker could use this fact to find a peer's private
6030 DH exponent. This attack requires that the attacker complete multiple
6031 handshakes in which the peer uses the same private DH exponent. For example
6032 this could be used to discover a TLS server's private DH exponent if it's
6033 reusing the private DH exponent or it's using a static DH ciphersuite.
6035 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6036 TLS. It is not on by default. If the option is not set then the server
6037 reuses the same private DH exponent for the life of the server process and
6038 would be vulnerable to this attack. It is believed that many popular
6039 applications do set this option and would therefore not be at risk.
6041 The fix for this issue adds an additional check where a "q" parameter is
6042 available (as is the case in X9.42 based parameters). This detects the
6043 only known attack, and is the only possible defense for static DH
6044 ciphersuites. This could have some performance impact.
6046 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6047 default and cannot be disabled. This could have some performance impact.
6049 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6054 * SSLv2 doesn't block disabled ciphers
6056 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6057 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6058 been disabled, provided that the SSLv2 protocol was not also disabled via
6061 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6062 and Sebastian Schinzel.
6067 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6069 * BN_mod_exp may produce incorrect results on x86_64
6071 There is a carry propagating bug in the x86_64 Montgomery squaring
6072 procedure. No EC algorithms are affected. Analysis suggests that attacks
6073 against RSA and DSA as a result of this defect would be very difficult to
6074 perform and are not believed likely. Attacks against DH are considered just
6075 feasible (although very difficult) because most of the work necessary to
6076 deduce information about a private key may be performed offline. The amount
6077 of resources required for such an attack would be very significant and
6078 likely only accessible to a limited number of attackers. An attacker would
6079 additionally need online access to an unpatched system using the target
6080 private key in a scenario with persistent DH parameters and a private
6081 key that is shared between multiple clients. For example this can occur by
6082 default in OpenSSL DHE based SSL/TLS ciphersuites.
6084 This issue was reported to OpenSSL by Hanno Böck.
6089 * Certificate verify crash with missing PSS parameter
6091 The signature verification routines will crash with a NULL pointer
6092 dereference if presented with an ASN.1 signature using the RSA PSS
6093 algorithm and absent mask generation function parameter. Since these
6094 routines are used to verify certificate signature algorithms this can be
6095 used to crash any certificate verification operation and exploited in a
6096 DoS attack. Any application which performs certificate verification is
6097 vulnerable including OpenSSL clients and servers which enable client
6100 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6105 * X509_ATTRIBUTE memory leak
6107 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6108 memory. This structure is used by the PKCS#7 and CMS routines so any
6109 application which reads PKCS#7 or CMS data from untrusted sources is
6110 affected. SSL/TLS is not affected.
6112 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6118 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6119 This changes the decoding behaviour for some invalid messages,
6120 though the change is mostly in the more lenient direction, and
6121 legacy behaviour is preserved as much as possible.
6125 * In DSA_generate_parameters_ex, if the provided seed is too short,
6128 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6130 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6132 * Alternate chains certificate forgery
6134 During certificate verification, OpenSSL will attempt to find an
6135 alternative certificate chain if the first attempt to build such a chain
6136 fails. An error in the implementation of this logic can mean that an
6137 attacker could cause certain checks on untrusted certificates to be
6138 bypassed, such as the CA flag, enabling them to use a valid leaf
6139 certificate to act as a CA and "issue" an invalid certificate.
6141 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6146 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6148 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6149 incompatibility in the handling of HMAC. The previous ABI has now been
6154 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6156 * Malformed ECParameters causes infinite loop
6158 When processing an ECParameters structure OpenSSL enters an infinite loop
6159 if the curve specified is over a specially malformed binary polynomial
6162 This can be used to perform denial of service against any
6163 system which processes public keys, certificate requests or
6164 certificates. This includes TLS clients and TLS servers with
6165 client authentication enabled.
6167 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6172 * Exploitable out-of-bounds read in X509_cmp_time
6174 X509_cmp_time does not properly check the length of the ASN1_TIME
6175 string and can read a few bytes out of bounds. In addition,
6176 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6179 An attacker can use this to craft malformed certificates and CRLs of
6180 various sizes and potentially cause a segmentation fault, resulting in
6181 a DoS on applications that verify certificates or CRLs. TLS clients
6182 that verify CRLs are affected. TLS clients and servers with client
6183 authentication enabled may be affected if they use custom verification
6186 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6187 independently by Hanno Böck.
6192 * PKCS7 crash with missing EnvelopedContent
6194 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6195 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6196 with missing content and trigger a NULL pointer dereference on parsing.
6198 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6199 structures from untrusted sources are affected. OpenSSL clients and
6200 servers are not affected.
6202 This issue was reported to OpenSSL by Michal Zalewski (Google).
6207 * CMS verify infinite loop with unknown hash function
6209 When verifying a signedData message the CMS code can enter an infinite loop
6210 if presented with an unknown hash function OID. This can be used to perform
6211 denial of service against any system which verifies signedData messages using
6213 This issue was reported to OpenSSL by Johannes Bauer.
6218 * Race condition handling NewSessionTicket
6220 If a NewSessionTicket is received by a multi-threaded client when attempting to
6221 reuse a previous ticket then a race condition can occur potentially leading to
6222 a double free of the ticket data.
6227 * Only support 256-bit or stronger elliptic curves with the
6228 'ecdh_auto' setting (server) or by default (client). Of supported
6229 curves, prefer P-256 (both).
6233 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6235 * ClientHello sigalgs DoS fix
6237 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6238 invalid signature algorithms extension a NULL pointer dereference will
6239 occur. This can be exploited in a DoS attack against the server.
6241 This issue was was reported to OpenSSL by David Ramos of Stanford
6245 *Stephen Henson and Matt Caswell*
6247 * Multiblock corrupted pointer fix
6249 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6250 feature only applies on 64 bit x86 architecture platforms that support AES
6251 NI instructions. A defect in the implementation of "multiblock" can cause
6252 OpenSSL's internal write buffer to become incorrectly set to NULL when
6253 using non-blocking IO. Typically, when the user application is using a
6254 socket BIO for writing, this will only result in a failed connection.
6255 However if some other BIO is used then it is likely that a segmentation
6256 fault will be triggered, thus enabling a potential DoS attack.
6258 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6263 * Segmentation fault in DTLSv1_listen fix
6265 The DTLSv1_listen function is intended to be stateless and processes the
6266 initial ClientHello from many peers. It is common for user code to loop
6267 over the call to DTLSv1_listen until a valid ClientHello is received with
6268 an associated cookie. A defect in the implementation of DTLSv1_listen means
6269 that state is preserved in the SSL object from one invocation to the next
6270 that can lead to a segmentation fault. Errors processing the initial
6271 ClientHello can trigger this scenario. An example of such an error could be
6272 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6275 This issue was reported to OpenSSL by Per Allansson.
6280 * Segmentation fault in ASN1_TYPE_cmp fix
6282 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6283 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6284 certificate signature algorithm consistency this can be used to crash any
6285 certificate verification operation and exploited in a DoS attack. Any
6286 application which performs certificate verification is vulnerable including
6287 OpenSSL clients and servers which enable client authentication.
6292 * Segmentation fault for invalid PSS parameters fix
6294 The signature verification routines will crash with a NULL pointer
6295 dereference if presented with an ASN.1 signature using the RSA PSS
6296 algorithm and invalid parameters. Since these routines are used to verify
6297 certificate signature algorithms this can be used to crash any
6298 certificate verification operation and exploited in a DoS attack. Any
6299 application which performs certificate verification is vulnerable including
6300 OpenSSL clients and servers which enable client authentication.
6302 This issue was was reported to OpenSSL by Brian Carpenter.
6307 * ASN.1 structure reuse memory corruption fix
6309 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6310 memory corruption via an invalid write. Such reuse is and has been
6311 strongly discouraged and is believed to be rare.
6313 Applications that parse structures containing CHOICE or ANY DEFINED BY
6314 components may be affected. Certificate parsing (d2i_X509 and related
6315 functions) are however not affected. OpenSSL clients and servers are
6321 * PKCS7 NULL pointer dereferences fix
6323 The PKCS#7 parsing code does not handle missing outer ContentInfo
6324 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6325 missing content and trigger a NULL pointer dereference on parsing.
6327 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6328 otherwise parse PKCS#7 structures from untrusted sources are
6329 affected. OpenSSL clients and servers are not affected.
6331 This issue was reported to OpenSSL by Michal Zalewski (Google).
6336 * DoS via reachable assert in SSLv2 servers fix
6338 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6339 servers that both support SSLv2 and enable export cipher suites by sending
6340 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6342 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6343 (OpenSSL development team).
6348 * Empty CKE with client auth and DHE fix
6350 If client auth is used then a server can seg fault in the event of a DHE
6351 ciphersuite being selected and a zero length ClientKeyExchange message
6352 being sent by the client. This could be exploited in a DoS attack.
6357 * Handshake with unseeded PRNG fix
6359 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6360 with an unseeded PRNG. The conditions are:
6361 - The client is on a platform where the PRNG has not been seeded
6362 automatically, and the user has not seeded manually
6363 - A protocol specific client method version has been used (i.e. not
6364 SSL_client_methodv23)
6365 - A ciphersuite is used that does not require additional random data from
6366 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6368 If the handshake succeeds then the client random that has been used will
6369 have been generated from a PRNG with insufficient entropy and therefore the
6370 output may be predictable.
6372 For example using the following command with an unseeded openssl will
6373 succeed on an unpatched platform:
6375 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6380 * Use After Free following d2i_ECPrivatekey error fix
6382 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6383 could cause a use after free condition. This, in turn, could cause a double
6384 free in several private key parsing functions (such as d2i_PrivateKey
6385 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6386 for applications that receive EC private keys from untrusted
6387 sources. This scenario is considered rare.
6389 This issue was discovered by the BoringSSL project and fixed in their
6395 * X509_to_X509_REQ NULL pointer deref fix
6397 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6398 the certificate key is invalid. This function is rarely used in practice.
6400 This issue was discovered by Brian Carpenter.
6405 * Removed the export ciphers from the DEFAULT ciphers
6409 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6411 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6412 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6413 So far those who have to target multiple platforms would compromise
6414 and argue that binary targeting say ARMv5 would still execute on
6415 ARMv8. "Universal" build resolves this compromise by providing
6416 near-optimal performance even on newer platforms.
6420 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6421 (other platforms pending).
6423 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6425 * Add support for the SignedCertificateTimestampList certificate and
6426 OCSP response extensions from RFC6962.
6430 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6431 for corner cases. (Certain input points at infinity could lead to
6432 bogus results, with non-infinity inputs mapped to infinity too.)
6436 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6437 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6438 common cases are optimized and there still is room for further
6439 improvements. Vector Permutation AES for Altivec is also added.
6443 * Add support for little-endian ppc64 Linux target.
6445 *Marcelo Cerri (IBM)*
6447 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6448 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6449 are optimized and there still is room for further improvements.
6450 Both 32- and 64-bit modes are supported.
6452 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6454 * Improved ARMv7 NEON support.
6458 * Support for SPARC Architecture 2011 crypto extensions, first
6459 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6460 SHA256/512, MD5, GHASH and modular exponentiation.
6462 *Andy Polyakov, David Miller*
6464 * Accelerated modular exponentiation for Intel processors, a.k.a.
6467 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6469 * Support for new and upcoming Intel processors, including AVX2,
6470 BMI and SHA ISA extensions. This includes additional "stitched"
6471 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6474 This work was sponsored by Intel Corp.
6478 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6479 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6480 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6484 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6485 this fixes a limitation in previous versions of OpenSSL.
6489 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6490 MGF1 digest and OAEP label.
6494 * Add EVP support for key wrapping algorithms, to avoid problems with
6495 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6496 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6497 algorithms and include tests cases.
6501 * Add functions to allocate and set the fields of an ECDSA_METHOD
6504 *Douglas E. Engert, Steve Henson*
6506 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6507 difference in days and seconds between two tm or ASN1_TIME structures.
6511 * Add -rev test option to s_server to just reverse order of characters
6512 received by client and send back to server. Also prints an abbreviated
6513 summary of the connection parameters.
6517 * New option -brief for s_client and s_server to print out a brief summary
6518 of connection parameters.
6522 * Add callbacks for arbitrary TLS extensions.
6524 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6526 * New option -crl_download in several openssl utilities to download CRLs
6527 from CRLDP extension in certificates.
6531 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6535 * New function X509_CRL_diff to generate a delta CRL from the difference
6536 of two full CRLs. Add support to "crl" utility.
6540 * New functions to set lookup_crls function and to retrieve
6541 X509_STORE from X509_STORE_CTX.
6545 * Print out deprecated issuer and subject unique ID fields in
6550 * Extend OCSP I/O functions so they can be used for simple general purpose
6551 HTTP as well as OCSP. New wrapper function which can be used to download
6552 CRLs using the OCSP API.
6556 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6560 * `SSL_CONF*` functions. These provide a common framework for application
6561 configuration using configuration files or command lines.
6565 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6566 message callback and prints the results. Needs compile time option
6567 "enable-ssl-trace". New options to s_client and s_server to enable
6572 * New ctrl and macro to retrieve supported points extensions.
6573 Print out extension in s_server and s_client.
6577 * New functions to retrieve certificate signature and signature
6582 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6587 * New Suite B modes for TLS code. These use and enforce the requirements
6588 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6589 only use Suite B curves. The Suite B modes can be set by using the
6590 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6594 * New chain verification flags for Suite B levels of security. Check
6595 algorithms are acceptable when flags are set in X509_verify_cert.
6599 * Make tls1_check_chain return a set of flags indicating checks passed
6600 by a certificate chain. Add additional tests to handle client
6601 certificates: checks for matching certificate type and issuer name
6606 * If an attempt is made to use a signature algorithm not in the peer
6607 preference list abort the handshake. If client has no suitable
6608 signature algorithms in response to a certificate request do not
6609 use the certificate.
6613 * If server EC tmp key is not in client preference list abort handshake.
6617 * Add support for certificate stores in CERT structure. This makes it
6618 possible to have different stores per SSL structure or one store in
6619 the parent SSL_CTX. Include distinct stores for certificate chain
6620 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6621 to build and store a certificate chain in CERT structure: returning
6622 an error if the chain cannot be built: this will allow applications
6623 to test if a chain is correctly configured.
6625 Note: if the CERT based stores are not set then the parent SSL_CTX
6626 store is used to retain compatibility with existing behaviour.
6630 * New function ssl_set_client_disabled to set a ciphersuite disabled
6631 mask based on the current session, check mask when sending client
6632 hello and checking the requested ciphersuite.
6636 * New ctrls to retrieve and set certificate types in a certificate
6637 request message. Print out received values in s_client. If certificate
6638 types is not set with custom values set sensible values based on
6639 supported signature algorithms.
6643 * Support for distinct client and server supported signature algorithms.
6647 * Add certificate callback. If set this is called whenever a certificate
6648 is required by client or server. An application can decide which
6649 certificate chain to present based on arbitrary criteria: for example
6650 supported signature algorithms. Add very simple example to s_server.
6651 This fixes many of the problems and restrictions of the existing client
6652 certificate callback: for example you can now clear an existing
6653 certificate and specify the whole chain.
6657 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6658 the certificate can be used for (if anything). Set valid_flags field
6659 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6660 to have similar checks in it.
6662 Add new "cert_flags" field to CERT structure and include a "strict mode".
6663 This enforces some TLS certificate requirements (such as only permitting
6664 certificate signature algorithms contained in the supported algorithms
6665 extension) which some implementations ignore: this option should be used
6666 with caution as it could cause interoperability issues.
6670 * Update and tidy signature algorithm extension processing. Work out
6671 shared signature algorithms based on preferences and peer algorithms
6672 and print them out in s_client and s_server. Abort handshake if no
6673 shared signature algorithms.
6677 * Add new functions to allow customised supported signature algorithms
6678 for SSL and SSL_CTX structures. Add options to s_client and s_server
6683 * New function SSL_certs_clear() to delete all references to certificates
6684 from an SSL structure. Before this once a certificate had been added
6685 it couldn't be removed.
6689 * Integrate hostname, email address and IP address checking with certificate
6690 verification. New verify options supporting checking in openssl utility.
6694 * Fixes and wildcard matching support to hostname and email checking
6695 functions. Add manual page.
6697 *Florian Weimer (Red Hat Product Security Team)*
6699 * New functions to check a hostname email or IP address against a
6700 certificate. Add options x509 utility to print results of checks against
6705 * Fix OCSP checking.
6707 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6709 * Initial experimental support for explicitly trusted non-root CAs.
6710 OpenSSL still tries to build a complete chain to a root but if an
6711 intermediate CA has a trust setting included that is used. The first
6712 setting is used: whether to trust (e.g., -addtrust option to the x509
6717 * Add -trusted_first option which attempts to find certificates in the
6718 trusted store even if an untrusted chain is also supplied.
6722 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6723 platform support for Linux and Android.
6727 * Support for linux-x32, ILP32 environment in x86_64 framework.
6731 * Experimental multi-implementation support for FIPS capable OpenSSL.
6732 When in FIPS mode the approved implementations are used as normal,
6733 when not in FIPS mode the internal unapproved versions are used instead.
6734 This means that the FIPS capable OpenSSL isn't forced to use the
6735 (often lower performance) FIPS implementations outside FIPS mode.
6739 * Transparently support X9.42 DH parameters when calling
6740 PEM_read_bio_DHparameters. This means existing applications can handle
6741 the new parameter format automatically.
6745 * Initial experimental support for X9.42 DH parameter format: mainly
6746 to support use of 'q' parameter for RFC5114 parameters.
6750 * Add DH parameters from RFC5114 including test data to dhtest.
6754 * Support for automatic EC temporary key parameter selection. If enabled
6755 the most preferred EC parameters are automatically used instead of
6756 hardcoded fixed parameters. Now a server just has to call:
6757 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6758 support ECDH and use the most appropriate parameters.
6762 * Enhance and tidy EC curve and point format TLS extension code. Use
6763 static structures instead of allocation if default values are used.
6764 New ctrls to set curves we wish to support and to retrieve shared curves.
6765 Print out shared curves in s_server. New options to s_server and s_client
6766 to set list of supported curves.
6770 * New ctrls to retrieve supported signature algorithms and
6771 supported curve values as an array of NIDs. Extend openssl utility
6772 to print out received values.
6776 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6777 between NIDs and the more common NIST names such as "P-256". Enhance
6778 ecparam utility and ECC method to recognise the NIST names for curves.
6782 * Enhance SSL/TLS certificate chain handling to support different
6783 chains for each certificate instead of one chain in the parent SSL_CTX.
6787 * Support for fixed DH ciphersuite client authentication: where both
6788 server and client use DH certificates with common parameters.
6792 * Support for fixed DH ciphersuites: those requiring DH server
6797 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6799 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6800 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6801 X509_CINF_get_signature were reverted post internal team review.
6806 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6808 * OCSP Status Request extension unbounded memory growth
6810 A malicious client can send an excessively large OCSP Status Request
6811 extension. If that client continually requests renegotiation, sending a
6812 large OCSP Status Request extension each time, then there will be unbounded
6813 memory growth on the server. This will eventually lead to a Denial Of
6814 Service attack through memory exhaustion. Servers with a default
6815 configuration are vulnerable even if they do not support OCSP. Builds using
6816 the "no-ocsp" build time option are not affected.
6818 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6823 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6826 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6832 * OOB write in MDC2_Update()
6834 An overflow can occur in MDC2_Update() either if called directly or
6835 through the EVP_DigestUpdate() function using MDC2. If an attacker
6836 is able to supply very large amounts of input data after a previous
6837 call to EVP_EncryptUpdate() with a partial block then a length check
6838 can overflow resulting in a heap corruption.
6840 The amount of data needed is comparable to SIZE_MAX which is impractical
6843 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6848 * Malformed SHA512 ticket DoS
6850 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6851 DoS attack where a malformed ticket will result in an OOB read which will
6854 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6855 a custom server callback and ticket lookup mechanism.
6857 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6862 * OOB write in BN_bn2dec()
6864 The function BN_bn2dec() does not check the return value of BN_div_word().
6865 This can cause an OOB write if an application uses this function with an
6866 overly large BIGNUM. This could be a problem if an overly large certificate
6867 or CRL is printed out from an untrusted source. TLS is not affected because
6868 record limits will reject an oversized certificate before it is parsed.
6870 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6875 * OOB read in TS_OBJ_print_bio()
6877 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6878 the total length the OID text representation would use and not the amount
6879 of data written. This will result in OOB reads when large OIDs are
6882 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6887 * Pointer arithmetic undefined behaviour
6889 Avoid some undefined pointer arithmetic
6891 A common idiom in the codebase is to check limits in the following manner:
6894 Where "p" points to some malloc'd data of SIZE bytes and
6897 "len" here could be from some externally supplied data (e.g. from a TLS
6900 The rules of C pointer arithmetic are such that "p + len" is only well
6901 defined where len <= SIZE. Therefore, the above idiom is actually
6902 undefined behaviour.
6904 For example this could cause problems if some malloc implementation
6905 provides an address for "p" such that "p + len" actually overflows for
6906 values of len that are too big and therefore p + len < limit.
6908 This issue was reported to OpenSSL by Guido Vranken
6913 * Constant time flag not preserved in DSA signing
6915 Operations in the DSA signing algorithm should run in constant time in
6916 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6917 implementation means that a non-constant time codepath is followed for
6918 certain operations. This has been demonstrated through a cache-timing
6919 attack to be sufficient for an attacker to recover the private DSA key.
6921 This issue was reported by César Pereida (Aalto University), Billy Brumley
6922 (Tampere University of Technology), and Yuval Yarom (The University of
6923 Adelaide and NICTA).
6928 * DTLS buffered message DoS
6930 In a DTLS connection where handshake messages are delivered out-of-order
6931 those messages that OpenSSL is not yet ready to process will be buffered
6932 for later use. Under certain circumstances, a flaw in the logic means that
6933 those messages do not get removed from the buffer even though the handshake
6934 has been completed. An attacker could force up to approx. 15 messages to
6935 remain in the buffer when they are no longer required. These messages will
6936 be cleared when the DTLS connection is closed. The default maximum size for
6937 a message is 100k. Therefore, the attacker could force an additional 1500k
6938 to be consumed per connection. By opening many simultaneous connections an
6939 attacker could cause a DoS attack through memory exhaustion.
6941 This issue was reported to OpenSSL by Quan Luo.
6946 * DTLS replay protection DoS
6948 A flaw in the DTLS replay attack protection mechanism means that records
6949 that arrive for future epochs update the replay protection "window" before
6950 the MAC for the record has been validated. This could be exploited by an
6951 attacker by sending a record for the next epoch (which does not have to
6952 decrypt or have a valid MAC), with a very large sequence number. This means
6953 that all subsequent legitimate packets are dropped causing a denial of
6954 service for a specific DTLS connection.
6956 This issue was reported to OpenSSL by the OCAP audit team.
6961 * Certificate message OOB reads
6963 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6964 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6965 theoretical DoS risk but this has not been observed in practice on common
6968 The messages affected are client certificate, client certificate request
6969 and server certificate. As a result the attack can only be performed
6970 against a client or a server which enables client authentication.
6972 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6977 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6979 * Prevent padding oracle in AES-NI CBC MAC check
6981 A MITM attacker can use a padding oracle attack to decrypt traffic
6982 when the connection uses an AES CBC cipher and the server support
6985 This issue was introduced as part of the fix for Lucky 13 padding
6986 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6987 constant time by making sure that always the same bytes are read and
6988 compared against either the MAC or padding bytes. But it no longer
6989 checked that there was enough data to have both the MAC and padding
6992 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6997 * Fix EVP_EncodeUpdate overflow
6999 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7000 Base64 encoding of binary data. If an attacker is able to supply very large
7001 amounts of input data then a length check can overflow resulting in a heap
7004 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7005 the `PEM_write_bio*` family of functions. These are mainly used within the
7006 OpenSSL command line applications, so any application which processes data
7007 from an untrusted source and outputs it as a PEM file should be considered
7008 vulnerable to this issue. User applications that call these APIs directly
7009 with large amounts of untrusted data may also be vulnerable.
7011 This issue was reported by Guido Vranken.
7016 * Fix EVP_EncryptUpdate overflow
7018 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7019 is able to supply very large amounts of input data after a previous call to
7020 EVP_EncryptUpdate() with a partial block then a length check can overflow
7021 resulting in a heap corruption. Following an analysis of all OpenSSL
7022 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7023 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7024 the first called function after an EVP_EncryptInit(), and therefore that
7025 specific call must be safe. The second form is where the length passed to
7026 EVP_EncryptUpdate() can be seen from the code to be some small value and
7027 therefore there is no possibility of an overflow. Since all instances are
7028 one of these two forms, it is believed that there can be no overflows in
7029 internal code due to this problem. It should be noted that
7030 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7031 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7032 of these calls have also been analysed too and it is believed there are no
7033 instances in internal usage where an overflow could occur.
7035 This issue was reported by Guido Vranken.
7040 * Prevent ASN.1 BIO excessive memory allocation
7042 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7043 a short invalid encoding can casuse allocation of large amounts of memory
7044 potentially consuming excessive resources or exhausting memory.
7046 Any application parsing untrusted data through d2i BIO functions is
7047 affected. The memory based functions such as d2i_X509() are *not* affected.
7048 Since the memory based functions are used by the TLS library, TLS
7049 applications are not affected.
7051 This issue was reported by Brian Carpenter.
7058 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7059 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7060 in arbitrary stack data being returned in the buffer.
7062 This issue was reported by Guido Vranken.
7067 * Modify behavior of ALPN to invoke callback after SNI/servername
7068 callback, such that updates to the SSL_CTX affect ALPN.
7072 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7077 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7078 methods are enabled and ssl2 is disabled the methods return NULL.
7082 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7084 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7085 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7086 provide any "EXPORT" or "LOW" strength ciphers.
7090 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7091 is by default disabled at build-time. Builds that are not configured with
7092 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7093 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7094 will need to explicitly call either of:
7096 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7098 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7100 as appropriate. Even if either of those is used, or the application
7101 explicitly uses the version-specific SSLv2_method() or its client and
7102 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7103 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7104 ciphers, and SSLv2 56-bit DES are no longer available.
7109 * Fix a double-free in DSA code
7111 A double free bug was discovered when OpenSSL parses malformed DSA private
7112 keys and could lead to a DoS attack or memory corruption for applications
7113 that receive DSA private keys from untrusted sources. This scenario is
7116 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7122 * Disable SRP fake user seed to address a server memory leak.
7124 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7126 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7127 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7128 was changed to ignore the "fake user" SRP seed, even if the seed
7131 Users should use SRP_VBASE_get1_by_user instead. Note that in
7132 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7133 also that even though configuring the SRP seed attempts to hide
7134 invalid usernames by continuing the handshake with fake
7135 credentials, this behaviour is not constant time and no strong
7136 guarantees are made that the handshake is indistinguishable from
7137 that of a valid user.
7142 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7144 In the BN_hex2bn function the number of hex digits is calculated using an
7145 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7146 large values of `i` this can result in `bn_expand` not allocating any
7147 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7148 field as NULL leading to a subsequent NULL ptr deref. For very large values
7149 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7150 In this case memory is allocated to the internal BIGNUM data field, but it
7151 is insufficiently sized leading to heap corruption. A similar issue exists
7152 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7153 is ever called by user applications with very large untrusted hex/dec data.
7154 This is anticipated to be a rare occurrence.
7156 All OpenSSL internal usage of these functions use data that is not expected
7157 to be untrusted, e.g. config file data or application command line
7158 arguments. If user developed applications generate config file data based
7159 on untrusted data then it is possible that this could also lead to security
7160 consequences. This is also anticipated to be rare.
7162 This issue was reported to OpenSSL by Guido Vranken.
7167 * Fix memory issues in `BIO_*printf` functions
7169 The internal `fmtstr` function used in processing a "%s" format string in
7170 the `BIO_*printf` functions could overflow while calculating the length of a
7171 string and cause an OOB read when printing very long strings.
7173 Additionally the internal `doapr_outch` function can attempt to write to an
7174 OOB memory location (at an offset from the NULL pointer) in the event of a
7175 memory allocation failure. In 1.0.2 and below this could be caused where
7176 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7177 could be in processing a very long "%s" format string. Memory leaks can
7180 The first issue may mask the second issue dependent on compiler behaviour.
7181 These problems could enable attacks where large amounts of untrusted data
7182 is passed to the `BIO_*printf` functions. If applications use these functions
7183 in this way then they could be vulnerable. OpenSSL itself uses these
7184 functions when printing out human-readable dumps of ASN.1 data. Therefore
7185 applications that print this data could be vulnerable if the data is from
7186 untrusted sources. OpenSSL command line applications could also be
7187 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7188 as command line arguments.
7190 Libssl is not considered directly vulnerable. Additionally certificates etc
7191 received via remote connections via libssl are also unlikely to be able to
7192 trigger these issues because of message size limits enforced within libssl.
7194 This issue was reported to OpenSSL Guido Vranken.
7199 * Side channel attack on modular exponentiation
7201 A side-channel attack was found which makes use of cache-bank conflicts on
7202 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7203 of RSA keys. The ability to exploit this issue is limited as it relies on
7204 an attacker who has control of code in a thread running on the same
7205 hyper-threaded core as the victim thread which is performing decryptions.
7207 This issue was reported to OpenSSL by Yuval Yarom, The University of
7208 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7209 Nadia Heninger, University of Pennsylvania with more information at
7210 <http://cachebleed.info>.
7215 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7216 if no keysize is specified with default_bits. This fixes an
7217 omission in an earlier change that changed all RSA/DSA key generation
7218 commands to use 2048 bits by default.
7222 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7224 * Protection for DH small subgroup attacks
7226 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7227 switched on by default and cannot be disabled. This could have some
7232 * SSLv2 doesn't block disabled ciphers
7234 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7235 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7236 been disabled, provided that the SSLv2 protocol was not also disabled via
7239 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7240 and Sebastian Schinzel.
7245 * Reject DH handshakes with parameters shorter than 1024 bits.
7249 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7251 * Certificate verify crash with missing PSS parameter
7253 The signature verification routines will crash with a NULL pointer
7254 dereference if presented with an ASN.1 signature using the RSA PSS
7255 algorithm and absent mask generation function parameter. Since these
7256 routines are used to verify certificate signature algorithms this can be
7257 used to crash any certificate verification operation and exploited in a
7258 DoS attack. Any application which performs certificate verification is
7259 vulnerable including OpenSSL clients and servers which enable client
7262 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7267 * X509_ATTRIBUTE memory leak
7269 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7270 memory. This structure is used by the PKCS#7 and CMS routines so any
7271 application which reads PKCS#7 or CMS data from untrusted sources is
7272 affected. SSL/TLS is not affected.
7274 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7280 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7281 This changes the decoding behaviour for some invalid messages,
7282 though the change is mostly in the more lenient direction, and
7283 legacy behaviour is preserved as much as possible.
7287 * In DSA_generate_parameters_ex, if the provided seed is too short,
7288 use a random seed, as already documented.
7290 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7292 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7294 * Alternate chains certificate forgery
7296 During certificate verfification, OpenSSL will attempt to find an
7297 alternative certificate chain if the first attempt to build such a chain
7298 fails. An error in the implementation of this logic can mean that an
7299 attacker could cause certain checks on untrusted certificates to be
7300 bypassed, such as the CA flag, enabling them to use a valid leaf
7301 certificate to act as a CA and "issue" an invalid certificate.
7303 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7309 * Race condition handling PSK identify hint
7311 If PSK identity hints are received by a multi-threaded client then
7312 the values are wrongly updated in the parent SSL_CTX structure. This can
7313 result in a race condition potentially leading to a double free of the
7319 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7321 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7322 incompatibility in the handling of HMAC. The previous ABI has now been
7325 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7327 * Malformed ECParameters causes infinite loop
7329 When processing an ECParameters structure OpenSSL enters an infinite loop
7330 if the curve specified is over a specially malformed binary polynomial
7333 This can be used to perform denial of service against any
7334 system which processes public keys, certificate requests or
7335 certificates. This includes TLS clients and TLS servers with
7336 client authentication enabled.
7338 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7343 * Exploitable out-of-bounds read in X509_cmp_time
7345 X509_cmp_time does not properly check the length of the ASN1_TIME
7346 string and can read a few bytes out of bounds. In addition,
7347 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7350 An attacker can use this to craft malformed certificates and CRLs of
7351 various sizes and potentially cause a segmentation fault, resulting in
7352 a DoS on applications that verify certificates or CRLs. TLS clients
7353 that verify CRLs are affected. TLS clients and servers with client
7354 authentication enabled may be affected if they use custom verification
7357 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7358 independently by Hanno Böck.
7363 * PKCS7 crash with missing EnvelopedContent
7365 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7366 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7367 with missing content and trigger a NULL pointer dereference on parsing.
7369 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7370 structures from untrusted sources are affected. OpenSSL clients and
7371 servers are not affected.
7373 This issue was reported to OpenSSL by Michal Zalewski (Google).
7378 * CMS verify infinite loop with unknown hash function
7380 When verifying a signedData message the CMS code can enter an infinite loop
7381 if presented with an unknown hash function OID. This can be used to perform
7382 denial of service against any system which verifies signedData messages using
7384 This issue was reported to OpenSSL by Johannes Bauer.
7389 * Race condition handling NewSessionTicket
7391 If a NewSessionTicket is received by a multi-threaded client when attempting to
7392 reuse a previous ticket then a race condition can occur potentially leading to
7393 a double free of the ticket data.
7398 * Reject DH handshakes with parameters shorter than 768 bits.
7400 *Kurt Roeckx and Emilia Kasper*
7402 * dhparam: generate 2048-bit parameters by default.
7404 *Kurt Roeckx and Emilia Kasper*
7406 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7408 * Segmentation fault in ASN1_TYPE_cmp fix
7410 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7411 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7412 certificate signature algorithm consistency this can be used to crash any
7413 certificate verification operation and exploited in a DoS attack. Any
7414 application which performs certificate verification is vulnerable including
7415 OpenSSL clients and servers which enable client authentication.
7420 * ASN.1 structure reuse memory corruption fix
7422 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7423 memory corruption via an invalid write. Such reuse is and has been
7424 strongly discouraged and is believed to be rare.
7426 Applications that parse structures containing CHOICE or ANY DEFINED BY
7427 components may be affected. Certificate parsing (d2i_X509 and related
7428 functions) are however not affected. OpenSSL clients and servers are
7434 * PKCS7 NULL pointer dereferences fix
7436 The PKCS#7 parsing code does not handle missing outer ContentInfo
7437 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7438 missing content and trigger a NULL pointer dereference on parsing.
7440 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7441 otherwise parse PKCS#7 structures from untrusted sources are
7442 affected. OpenSSL clients and servers are not affected.
7444 This issue was reported to OpenSSL by Michal Zalewski (Google).
7449 * DoS via reachable assert in SSLv2 servers fix
7451 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7452 servers that both support SSLv2 and enable export cipher suites by sending
7453 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7455 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7456 (OpenSSL development team).
7461 * Use After Free following d2i_ECPrivatekey error fix
7463 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7464 could cause a use after free condition. This, in turn, could cause a double
7465 free in several private key parsing functions (such as d2i_PrivateKey
7466 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7467 for applications that receive EC private keys from untrusted
7468 sources. This scenario is considered rare.
7470 This issue was discovered by the BoringSSL project and fixed in their
7476 * X509_to_X509_REQ NULL pointer deref fix
7478 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7479 the certificate key is invalid. This function is rarely used in practice.
7481 This issue was discovered by Brian Carpenter.
7486 * Removed the export ciphers from the DEFAULT ciphers
7490 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7492 * Build fixes for the Windows and OpenVMS platforms
7494 *Matt Caswell and Richard Levitte*
7496 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7498 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7499 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7500 dereference. This could lead to a Denial Of Service attack. Thanks to
7501 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7506 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7507 dtls1_buffer_record function under certain conditions. In particular this
7508 could occur if an attacker sent repeated DTLS records with the same
7509 sequence number but for the next epoch. The memory leak could be exploited
7510 by an attacker in a Denial of Service attack through memory exhaustion.
7511 Thanks to Chris Mueller for reporting this issue.
7516 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7517 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7518 method would be set to NULL which could later result in a NULL pointer
7519 dereference. Thanks to Frank Schmirler for reporting this issue.
7524 * Abort handshake if server key exchange message is omitted for ephemeral
7527 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7528 reporting this issue.
7533 * Remove non-export ephemeral RSA code on client and server. This code
7534 violated the TLS standard by allowing the use of temporary RSA keys in
7535 non-export ciphersuites and could be used by a server to effectively
7536 downgrade the RSA key length used to a value smaller than the server
7537 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7538 INRIA or reporting this issue.
7543 * Fixed issue where DH client certificates are accepted without verification.
7544 An OpenSSL server will accept a DH certificate for client authentication
7545 without the certificate verify message. This effectively allows a client to
7546 authenticate without the use of a private key. This only affects servers
7547 which trust a client certificate authority which issues certificates
7548 containing DH keys: these are extremely rare and hardly ever encountered.
7549 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7555 * Ensure that the session ID context of an SSL is updated when its
7556 SSL_CTX is updated via SSL_set_SSL_CTX.
7558 The session ID context is typically set from the parent SSL_CTX,
7559 and can vary with the CTX.
7563 * Fix various certificate fingerprint issues.
7565 By using non-DER or invalid encodings outside the signed portion of a
7566 certificate the fingerprint can be changed without breaking the signature.
7567 Although no details of the signed portion of the certificate can be changed
7568 this can cause problems with some applications: e.g. those using the
7569 certificate fingerprint for blacklists.
7571 1. Reject signatures with non zero unused bits.
7573 If the BIT STRING containing the signature has non zero unused bits reject
7574 the signature. All current signature algorithms require zero unused bits.
7576 2. Check certificate algorithm consistency.
7578 Check the AlgorithmIdentifier inside TBS matches the one in the
7579 certificate signature. NB: this will result in signature failure
7580 errors for some broken certificates.
7582 Thanks to Konrad Kraszewski from Google for reporting this issue.
7584 3. Check DSA/ECDSA signatures use DER.
7586 Re-encode DSA/ECDSA signatures and compare with the original received
7587 signature. Return an error if there is a mismatch.
7589 This will reject various cases including garbage after signature
7590 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7591 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7592 (negative or with leading zeroes).
7594 Further analysis was conducted and fixes were developed by Stephen Henson
7595 of the OpenSSL core team.
7601 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7602 results on some platforms, including x86_64. This bug occurs at random
7603 with a very low probability, and is not known to be exploitable in any
7604 way, though its exact impact is difficult to determine. Thanks to Pieter
7605 Wuille (Blockstream) who reported this issue and also suggested an initial
7606 fix. Further analysis was conducted by the OpenSSL development team and
7607 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7608 the OpenSSL core team.
7613 * Do not resume sessions on the server if the negotiated protocol
7614 version does not match the session's version. Resuming with a different
7615 version, while not strictly forbidden by the RFC, is of questionable
7616 sanity and breaks all known clients.
7618 *David Benjamin, Emilia Käsper*
7620 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7621 early CCS messages during renegotiation. (Note that because
7622 renegotiation is encrypted, this early CCS was not exploitable.)
7626 * Tighten client-side session ticket handling during renegotiation:
7627 ensure that the client only accepts a session ticket if the server sends
7628 the extension anew in the ServerHello. Previously, a TLS client would
7629 reuse the old extension state and thus accept a session ticket if one was
7630 announced in the initial ServerHello.
7632 Similarly, ensure that the client requires a session ticket if one
7633 was advertised in the ServerHello. Previously, a TLS client would
7634 ignore a missing NewSessionTicket message.
7638 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7642 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7643 sends a carefully crafted handshake message, to cause OpenSSL to fail
7644 to free up to 64k of memory causing a memory leak. This could be
7645 exploited in a Denial Of Service attack. This issue affects OpenSSL
7646 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7647 whether SRTP is used or configured. Implementations of OpenSSL that
7648 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7650 The fix was developed by the OpenSSL team.
7655 * Session Ticket Memory Leak.
7657 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7658 integrity of that ticket is first verified. In the event of a session
7659 ticket integrity check failing, OpenSSL will fail to free memory
7660 causing a memory leak. By sending a large number of invalid session
7661 tickets an attacker could exploit this issue in a Denial Of Service
7667 * Build option no-ssl3 is incomplete.
7669 When OpenSSL is configured with "no-ssl3" as a build option, servers
7670 could accept and complete an SSL 3.0 handshake, and clients could be
7671 configured to send them.
7674 *Akamai and the OpenSSL team*
7676 * Add support for TLS_FALLBACK_SCSV.
7677 Client applications doing fallback retries should call
7678 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7681 *Adam Langley, Bodo Moeller*
7683 * Add additional DigestInfo checks.
7685 Re-encode DigestInto in DER and check against the original when
7686 verifying RSA signature: this will reject any improperly encoded
7687 DigestInfo structures.
7689 Note: this is a precautionary measure and no attacks are currently known.
7693 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7695 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7696 SRP code can be overrun an internal buffer. Add sanity check that
7697 g, A, B < N to SRP code.
7699 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7700 Group for discovering this issue.
7705 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7706 TLS 1.0 instead of higher protocol versions when the ClientHello message
7707 is badly fragmented. This allows a man-in-the-middle attacker to force a
7708 downgrade to TLS 1.0 even if both the server and the client support a
7709 higher protocol version, by modifying the client's TLS records.
7711 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7712 researching this issue.
7717 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7718 to a denial of service attack. A malicious server can crash the client
7719 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7720 ciphersuite and sending carefully crafted handshake messages.
7722 Thanks to Felix Gröbert (Google) for discovering and researching this
7728 * By sending carefully crafted DTLS packets an attacker could cause openssl
7729 to leak memory. This can be exploited through a Denial of Service attack.
7730 Thanks to Adam Langley for discovering and researching this issue.
7735 * An attacker can force openssl to consume large amounts of memory whilst
7736 processing DTLS handshake messages. This can be exploited through a
7737 Denial of Service attack.
7738 Thanks to Adam Langley for discovering and researching this issue.
7743 * An attacker can force an error condition which causes openssl to crash
7744 whilst processing DTLS packets due to memory being freed twice. This
7745 can be exploited through a Denial of Service attack.
7746 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7752 * If a multithreaded client connects to a malicious server using a resumed
7753 session and the server sends an ec point format extension it could write
7754 up to 255 bytes to freed memory.
7756 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7762 * A malicious server can crash an OpenSSL client with a null pointer
7763 dereference (read) by specifying an SRP ciphersuite even though it was not
7764 properly negotiated with the client. This can be exploited through a
7765 Denial of Service attack.
7767 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7768 discovering and researching this issue.
7773 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7774 X509_name_oneline, X509_name_print_ex et al. to leak some information
7775 from the stack. Applications may be affected if they echo pretty printing
7776 output to the attacker.
7778 Thanks to Ivan Fratric (Google) for discovering this issue.
7781 *Emilia Käsper, and Steve Henson*
7783 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7784 for corner cases. (Certain input points at infinity could lead to
7785 bogus results, with non-infinity inputs mapped to infinity too.)
7789 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7791 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7792 handshake can force the use of weak keying material in OpenSSL
7793 SSL/TLS clients and servers.
7795 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7796 researching this issue. ([CVE-2014-0224])
7798 *KIKUCHI Masashi, Steve Henson*
7800 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7801 OpenSSL DTLS client the code can be made to recurse eventually crashing
7804 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7807 *Imre Rad, Steve Henson*
7809 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7810 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7811 client or server. This is potentially exploitable to run arbitrary
7812 code on a vulnerable client or server.
7814 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7816 *Jüri Aedla, Steve Henson*
7818 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7819 are subject to a denial of service attack.
7821 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7822 this issue. ([CVE-2014-3470])
7824 *Felix Gröbert, Ivan Fratric, Steve Henson*
7826 * Harmonize version and its documentation. -f flag is used to display
7829 *mancha <mancha1@zoho.com>*
7831 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7832 in i2d_ECPrivateKey.
7834 *mancha <mancha1@zoho.com>*
7836 * Fix some double frees. These are not thought to be exploitable.
7838 *mancha <mancha1@zoho.com>*
7840 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7842 * A missing bounds check in the handling of the TLS heartbeat extension
7843 can be used to reveal up to 64k of memory to a connected client or
7846 Thanks for Neel Mehta of Google Security for discovering this bug and to
7847 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7848 preparing the fix ([CVE-2014-0160])
7850 *Adam Langley, Bodo Moeller*
7852 * Fix for the attack described in the paper "Recovering OpenSSL
7853 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7854 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7855 <http://eprint.iacr.org/2014/140>
7857 Thanks to Yuval Yarom and Naomi Benger for discovering this
7858 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7860 *Yuval Yarom and Naomi Benger*
7862 * TLS pad extension: draft-agl-tls-padding-03
7864 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7865 TLS client Hello record length value would otherwise be > 255 and
7866 less that 512 pad with a dummy extension containing zeroes so it
7867 is at least 512 bytes long.
7869 *Adam Langley, Steve Henson*
7871 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7873 * Fix for TLS record tampering bug. A carefully crafted invalid
7874 handshake could crash OpenSSL with a NULL pointer exception.
7875 Thanks to Anton Johansson for reporting this issues.
7878 * Keep original DTLS digest and encryption contexts in retransmission
7879 structures so we can use the previous session parameters if they need
7880 to be resent. ([CVE-2013-6450])
7884 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7885 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7886 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7887 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7888 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7889 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7891 *Rob Stradling, Adam Langley*
7893 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7895 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7896 supporting platforms or when small records were transferred.
7898 *Andy Polyakov, Steve Henson*
7900 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7902 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7904 This addresses the flaw in CBC record processing discovered by
7905 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7906 at: <http://www.isg.rhul.ac.uk/tls/>
7908 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7909 Security Group at Royal Holloway, University of London
7910 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7911 Emilia Käsper for the initial patch.
7914 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7916 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7917 ciphersuites which can be exploited in a denial of service attack.
7918 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7919 and detecting this bug and to Wolfgang Ettlinger
7920 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7925 * Return an error when checking OCSP signatures when key is NULL.
7926 This fixes a DoS attack. ([CVE-2013-0166])
7930 * Make openssl verify return errors.
7932 *Chris Palmer <palmer@google.com> and Ben Laurie*
7934 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7935 the right response is stapled. Also change SSL_get_certificate()
7936 so it returns the certificate actually sent.
7937 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7939 *Rob Stradling <rob.stradling@comodo.com>*
7941 * Fix possible deadlock when decoding public keys.
7945 * Don't use TLS 1.0 record version number in initial client hello
7950 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7952 * Sanity check record length before skipping explicit IV in TLS
7953 1.2, 1.1 and DTLS to fix DoS attack.
7955 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7956 fuzzing as a service testing platform.
7961 * Initialise tkeylen properly when encrypting CMS messages.
7962 Thanks to Solar Designer of Openwall for reporting this issue.
7966 * In FIPS mode don't try to use composite ciphers as they are not
7971 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7973 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7974 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7975 mean any application compiled against OpenSSL 1.0.0 headers setting
7976 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7977 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7978 0x10000000L Any application which was previously compiled against
7979 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7980 will need to be recompiled as a result. Letting be results in
7981 inability to disable specifically TLS 1.1 and in client context,
7982 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7986 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7987 disable just protocol X, but all protocols above X *if* there are
7988 protocols *below* X still enabled. In more practical terms it means
7989 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7990 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7991 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7996 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7998 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7999 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8000 in CRYPTO_realloc_clean.
8002 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8003 issue and to Adam Langley <agl@chromium.org> for fixing it.
8006 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8008 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8012 * Workarounds for some broken servers that "hang" if a client hello
8013 record length exceeds 255 bytes.
8015 1. Do not use record version number > TLS 1.0 in initial client
8016 hello: some (but not all) hanging servers will now work.
8017 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8018 the number of ciphers sent in the client hello. This should be
8019 set to an even number, such as 50, for example by passing:
8020 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8021 Most broken servers should now work.
8022 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8023 TLS 1.2 client support entirely.
8027 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8031 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8033 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8034 STRING form instead of a DigestInfo.
8038 * The format used for MDC2 RSA signatures is inconsistent between EVP
8039 and the RSA_sign/RSA_verify functions. This was made more apparent when
8040 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8041 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8042 the correct format in RSA_verify so both forms transparently work.
8046 * Some servers which support TLS 1.0 can choke if we initially indicate
8047 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8048 encrypted premaster secret. As a workaround use the maximum permitted
8049 client version in client hello, this should keep such servers happy
8050 and still work with previous versions of OpenSSL.
8054 * Add support for TLS/DTLS heartbeats.
8056 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8058 * Add support for SCTP.
8060 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8062 * Improved PRNG seeding for VOS.
8064 *Paul Green <Paul.Green@stratus.com>*
8066 * Extensive assembler packs updates, most notably:
8068 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8069 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8070 - x86_64: bit-sliced AES implementation;
8071 - ARM: NEON support, contemporary platforms optimizations;
8072 - s390x: z196 support;
8073 - `*`: GHASH and GF(2^m) multiplication implementations;
8077 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8078 (removal of unnecessary code)
8080 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8082 * Add TLS key material exporter from RFC 5705.
8086 * Add DTLS-SRTP negotiation from RFC 5764.
8090 * Add Next Protocol Negotiation,
8091 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8092 disabled with a no-npn flag to config or Configure. Code donated
8095 *Adam Langley <agl@google.com> and Ben Laurie*
8097 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8098 NIST-P256, NIST-P521, with constant-time single point multiplication on
8099 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8100 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8101 Code made available under Apache License version 2.0.
8103 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8104 line to include this in your build of OpenSSL, and run "make depend" (or
8105 "make update"). This enables the following EC_METHODs:
8107 EC_GFp_nistp224_method()
8108 EC_GFp_nistp256_method()
8109 EC_GFp_nistp521_method()
8111 EC_GROUP_new_by_curve_name() will automatically use these (while
8112 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8115 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8117 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8118 all platforms. Move ssize_t definition from e_os.h to the public
8119 header file e_os2.h as it now appears in public header file cms.h
8123 * New -sigopt option to the ca, req and x509 utilities. Additional
8124 signature parameters can be passed using this option and in
8129 * Add RSA PSS signing function. This will generate and set the
8130 appropriate AlgorithmIdentifiers for PSS based on those in the
8131 corresponding EVP_MD_CTX structure. No application support yet.
8135 * Support for companion algorithm specific ASN1 signing routines.
8136 New function ASN1_item_sign_ctx() signs a pre-initialised
8137 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8138 the appropriate parameters.
8142 * Add new algorithm specific ASN1 verification initialisation function
8143 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8144 handling will be the same no matter what EVP_PKEY_METHOD is used.
8145 Add a PSS handler to support verification of PSS signatures: checked
8146 against a number of sample certificates.
8150 * Add signature printing for PSS. Add PSS OIDs.
8152 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8154 * Add algorithm specific signature printing. An individual ASN1 method
8155 can now print out signatures instead of the standard hex dump.
8157 More complex signatures (e.g. PSS) can print out more meaningful
8158 information. Include DSA version that prints out the signature
8163 * Password based recipient info support for CMS library: implementing
8168 * Split password based encryption into PBES2 and PBKDF2 functions. This
8169 neatly separates the code into cipher and PBE sections and is required
8170 for some algorithms that split PBES2 into separate pieces (such as
8171 password based CMS).
8175 * Session-handling fixes:
8176 - Fix handling of connections that are resuming with a session ID,
8177 but also support Session Tickets.
8178 - Fix a bug that suppressed issuing of a new ticket if the client
8179 presented a ticket with an expired session.
8180 - Try to set the ticket lifetime hint to something reasonable.
8181 - Make tickets shorter by excluding irrelevant information.
8182 - On the client side, don't ignore renewed tickets.
8184 *Adam Langley, Bodo Moeller (Google)*
8186 * Fix PSK session representation.
8190 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8192 This work was sponsored by Intel.
8196 * Add GCM support to TLS library. Some custom code is needed to split
8197 the IV between the fixed (from PRF) and explicit (from TLS record)
8198 portions. This adds all GCM ciphersuites supported by RFC5288 and
8199 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8200 add a special AESGCM string for GCM only.
8204 * Expand range of ctrls for AES GCM. Permit setting invocation
8205 field on decrypt and retrieval of invocation field only on encrypt.
8209 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8210 As required by RFC5289 these ciphersuites cannot be used if for
8211 versions of TLS earlier than 1.2.
8215 * For FIPS capable OpenSSL interpret a NULL default public key method
8216 as unset and return the appropriate default but do *not* set the default.
8217 This means we can return the appropriate method in applications that
8218 switch between FIPS and non-FIPS modes.
8222 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8223 ENGINE is used then we cannot handle that in the FIPS module so we
8224 keep original code iff non-FIPS operations are allowed.
8228 * Add -attime option to openssl utilities.
8230 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8232 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8236 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8237 FIPS EC methods unconditionally for now.
8241 * New build option no-ec2m to disable characteristic 2 code.
8245 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8246 all cases can be covered as some introduce binary incompatibilities.
8250 * Redirect RSA operations to FIPS module including keygen,
8251 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8255 * Add similar low-level API blocking to ciphers.
8259 * low-level digest APIs are not approved in FIPS mode: any attempt
8260 to use these will cause a fatal error. Applications that *really* want
8261 to use them can use the `private_*` version instead.
8265 * Redirect cipher operations to FIPS module for FIPS builds.
8269 * Redirect digest operations to FIPS module for FIPS builds.
8273 * Update build system to add "fips" flag which will link in fipscanister.o
8274 for static and shared library builds embedding a signature if needed.
8278 * Output TLS supported curves in preference order instead of numerical
8279 order. This is currently hardcoded for the highest order curves first.
8280 This should be configurable so applications can judge speed vs strength.
8284 * Add TLS v1.2 server support for client authentication.
8288 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8293 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8294 FIPS modules versions.
8298 * Add TLS v1.2 client side support for client authentication. Keep cache
8299 of handshake records longer as we don't know the hash algorithm to use
8300 until after the certificate request message is received.
8304 * Initial TLS v1.2 client support. Add a default signature algorithms
8305 extension including all the algorithms we support. Parse new signature
8306 format in client key exchange. Relax some ECC signing restrictions for
8307 TLS v1.2 as indicated in RFC5246.
8311 * Add server support for TLS v1.2 signature algorithms extension. Switch
8312 to new signature format when needed using client digest preference.
8313 All server ciphersuites should now work correctly in TLS v1.2. No client
8314 support yet and no support for client certificates.
8318 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8319 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8320 ciphersuites. At present only RSA key exchange ciphersuites work with
8321 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8322 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8323 and version checking.
8327 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8328 with this defined it will not be affected by any changes to ssl internal
8329 structures. Add several utility functions to allow openssl application
8330 to work with OPENSSL_NO_SSL_INTERN defined.
8334 * A long standing patch to add support for SRP from EdelWeb (Peter
8335 Sylvester and Christophe Renou) was integrated.
8336 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8337 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8340 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8344 * Permit abbreviated handshakes when renegotiating using the function
8345 SSL_renegotiate_abbreviated().
8347 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8349 * Add call to ENGINE_register_all_complete() to
8350 ENGINE_load_builtin_engines(), so some implementations get used
8351 automatically instead of needing explicit application support.
8355 * Add support for TLS key exporter as described in RFC5705.
8357 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8359 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8360 a few changes are required:
8362 Add SSL_OP_NO_TLSv1_1 flag.
8363 Add TLSv1_1 methods.
8364 Update version checking logic to handle version 1.1.
8365 Add explicit IV handling (ported from DTLS code).
8366 Add command line options to s_client/s_server.
8373 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8375 * X509_ATTRIBUTE memory leak
8377 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8378 memory. This structure is used by the PKCS#7 and CMS routines so any
8379 application which reads PKCS#7 or CMS data from untrusted sources is
8380 affected. SSL/TLS is not affected.
8382 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8388 * Race condition handling PSK identify hint
8390 If PSK identity hints are received by a multi-threaded client then
8391 the values are wrongly updated in the parent SSL_CTX structure. This can
8392 result in a race condition potentially leading to a double free of the
8398 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8400 * Malformed ECParameters causes infinite loop
8402 When processing an ECParameters structure OpenSSL enters an infinite loop
8403 if the curve specified is over a specially malformed binary polynomial
8406 This can be used to perform denial of service against any
8407 system which processes public keys, certificate requests or
8408 certificates. This includes TLS clients and TLS servers with
8409 client authentication enabled.
8411 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8416 * Exploitable out-of-bounds read in X509_cmp_time
8418 X509_cmp_time does not properly check the length of the ASN1_TIME
8419 string and can read a few bytes out of bounds. In addition,
8420 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8423 An attacker can use this to craft malformed certificates and CRLs of
8424 various sizes and potentially cause a segmentation fault, resulting in
8425 a DoS on applications that verify certificates or CRLs. TLS clients
8426 that verify CRLs are affected. TLS clients and servers with client
8427 authentication enabled may be affected if they use custom verification
8430 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8431 independently by Hanno Böck.
8436 * PKCS7 crash with missing EnvelopedContent
8438 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8439 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8440 with missing content and trigger a NULL pointer dereference on parsing.
8442 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8443 structures from untrusted sources are affected. OpenSSL clients and
8444 servers are not affected.
8446 This issue was reported to OpenSSL by Michal Zalewski (Google).
8451 * CMS verify infinite loop with unknown hash function
8453 When verifying a signedData message the CMS code can enter an infinite loop
8454 if presented with an unknown hash function OID. This can be used to perform
8455 denial of service against any system which verifies signedData messages using
8457 This issue was reported to OpenSSL by Johannes Bauer.
8462 * Race condition handling NewSessionTicket
8464 If a NewSessionTicket is received by a multi-threaded client when attempting to
8465 reuse a previous ticket then a race condition can occur potentially leading to
8466 a double free of the ticket data.
8471 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8473 * Segmentation fault in ASN1_TYPE_cmp fix
8475 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8476 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8477 certificate signature algorithm consistency this can be used to crash any
8478 certificate verification operation and exploited in a DoS attack. Any
8479 application which performs certificate verification is vulnerable including
8480 OpenSSL clients and servers which enable client authentication.
8485 * ASN.1 structure reuse memory corruption fix
8487 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8488 memory corruption via an invalid write. Such reuse is and has been
8489 strongly discouraged and is believed to be rare.
8491 Applications that parse structures containing CHOICE or ANY DEFINED BY
8492 components may be affected. Certificate parsing (d2i_X509 and related
8493 functions) are however not affected. OpenSSL clients and servers are
8499 * PKCS7 NULL pointer dereferences fix
8501 The PKCS#7 parsing code does not handle missing outer ContentInfo
8502 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8503 missing content and trigger a NULL pointer dereference on parsing.
8505 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8506 otherwise parse PKCS#7 structures from untrusted sources are
8507 affected. OpenSSL clients and servers are not affected.
8509 This issue was reported to OpenSSL by Michal Zalewski (Google).
8514 * DoS via reachable assert in SSLv2 servers fix
8516 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8517 servers that both support SSLv2 and enable export cipher suites by sending
8518 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8520 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8521 (OpenSSL development team).
8526 * Use After Free following d2i_ECPrivatekey error fix
8528 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8529 could cause a use after free condition. This, in turn, could cause a double
8530 free in several private key parsing functions (such as d2i_PrivateKey
8531 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8532 for applications that receive EC private keys from untrusted
8533 sources. This scenario is considered rare.
8535 This issue was discovered by the BoringSSL project and fixed in their
8541 * X509_to_X509_REQ NULL pointer deref fix
8543 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8544 the certificate key is invalid. This function is rarely used in practice.
8546 This issue was discovered by Brian Carpenter.
8551 * Removed the export ciphers from the DEFAULT ciphers
8555 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8557 * Build fixes for the Windows and OpenVMS platforms
8559 *Matt Caswell and Richard Levitte*
8561 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8563 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8564 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8565 dereference. This could lead to a Denial Of Service attack. Thanks to
8566 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8571 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8572 dtls1_buffer_record function under certain conditions. In particular this
8573 could occur if an attacker sent repeated DTLS records with the same
8574 sequence number but for the next epoch. The memory leak could be exploited
8575 by an attacker in a Denial of Service attack through memory exhaustion.
8576 Thanks to Chris Mueller for reporting this issue.
8581 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8582 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8583 method would be set to NULL which could later result in a NULL pointer
8584 dereference. Thanks to Frank Schmirler for reporting this issue.
8589 * Abort handshake if server key exchange message is omitted for ephemeral
8592 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8593 reporting this issue.
8598 * Remove non-export ephemeral RSA code on client and server. This code
8599 violated the TLS standard by allowing the use of temporary RSA keys in
8600 non-export ciphersuites and could be used by a server to effectively
8601 downgrade the RSA key length used to a value smaller than the server
8602 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8603 INRIA or reporting this issue.
8608 * Fixed issue where DH client certificates are accepted without verification.
8609 An OpenSSL server will accept a DH certificate for client authentication
8610 without the certificate verify message. This effectively allows a client to
8611 authenticate without the use of a private key. This only affects servers
8612 which trust a client certificate authority which issues certificates
8613 containing DH keys: these are extremely rare and hardly ever encountered.
8614 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8620 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8621 results on some platforms, including x86_64. This bug occurs at random
8622 with a very low probability, and is not known to be exploitable in any
8623 way, though its exact impact is difficult to determine. Thanks to Pieter
8624 Wuille (Blockstream) who reported this issue and also suggested an initial
8625 fix. Further analysis was conducted by the OpenSSL development team and
8626 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8627 the OpenSSL core team.
8632 * Fix various certificate fingerprint issues.
8634 By using non-DER or invalid encodings outside the signed portion of a
8635 certificate the fingerprint can be changed without breaking the signature.
8636 Although no details of the signed portion of the certificate can be changed
8637 this can cause problems with some applications: e.g. those using the
8638 certificate fingerprint for blacklists.
8640 1. Reject signatures with non zero unused bits.
8642 If the BIT STRING containing the signature has non zero unused bits reject
8643 the signature. All current signature algorithms require zero unused bits.
8645 2. Check certificate algorithm consistency.
8647 Check the AlgorithmIdentifier inside TBS matches the one in the
8648 certificate signature. NB: this will result in signature failure
8649 errors for some broken certificates.
8651 Thanks to Konrad Kraszewski from Google for reporting this issue.
8653 3. Check DSA/ECDSA signatures use DER.
8655 Reencode DSA/ECDSA signatures and compare with the original received
8656 signature. Return an error if there is a mismatch.
8658 This will reject various cases including garbage after signature
8659 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8660 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8661 (negative or with leading zeroes).
8663 Further analysis was conducted and fixes were developed by Stephen Henson
8664 of the OpenSSL core team.
8670 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8672 * Session Ticket Memory Leak.
8674 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8675 integrity of that ticket is first verified. In the event of a session
8676 ticket integrity check failing, OpenSSL will fail to free memory
8677 causing a memory leak. By sending a large number of invalid session
8678 tickets an attacker could exploit this issue in a Denial Of Service
8684 * Build option no-ssl3 is incomplete.
8686 When OpenSSL is configured with "no-ssl3" as a build option, servers
8687 could accept and complete an SSL 3.0 handshake, and clients could be
8688 configured to send them.
8691 *Akamai and the OpenSSL team*
8693 * Add support for TLS_FALLBACK_SCSV.
8694 Client applications doing fallback retries should call
8695 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8698 *Adam Langley, Bodo Moeller*
8700 * Add additional DigestInfo checks.
8702 Reencode DigestInto in DER and check against the original when
8703 verifying RSA signature: this will reject any improperly encoded
8704 DigestInfo structures.
8706 Note: this is a precautionary measure and no attacks are currently known.
8710 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8712 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8713 to a denial of service attack. A malicious server can crash the client
8714 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8715 ciphersuite and sending carefully crafted handshake messages.
8717 Thanks to Felix Gröbert (Google) for discovering and researching this
8723 * By sending carefully crafted DTLS packets an attacker could cause openssl
8724 to leak memory. This can be exploited through a Denial of Service attack.
8725 Thanks to Adam Langley for discovering and researching this issue.
8730 * An attacker can force openssl to consume large amounts of memory whilst
8731 processing DTLS handshake messages. This can be exploited through a
8732 Denial of Service attack.
8733 Thanks to Adam Langley for discovering and researching this issue.
8738 * An attacker can force an error condition which causes openssl to crash
8739 whilst processing DTLS packets due to memory being freed twice. This
8740 can be exploited through a Denial of Service attack.
8741 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8747 * If a multithreaded client connects to a malicious server using a resumed
8748 session and the server sends an ec point format extension it could write
8749 up to 255 bytes to freed memory.
8751 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8757 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8758 X509_name_oneline, X509_name_print_ex et al. to leak some information
8759 from the stack. Applications may be affected if they echo pretty printing
8760 output to the attacker.
8762 Thanks to Ivan Fratric (Google) for discovering this issue.
8765 *Emilia Käsper, and Steve Henson*
8767 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8768 for corner cases. (Certain input points at infinity could lead to
8769 bogus results, with non-infinity inputs mapped to infinity too.)
8773 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8775 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8776 handshake can force the use of weak keying material in OpenSSL
8777 SSL/TLS clients and servers.
8779 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8780 researching this issue. ([CVE-2014-0224])
8782 *KIKUCHI Masashi, Steve Henson*
8784 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8785 OpenSSL DTLS client the code can be made to recurse eventually crashing
8788 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8791 *Imre Rad, Steve Henson*
8793 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8794 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8795 client or server. This is potentially exploitable to run arbitrary
8796 code on a vulnerable client or server.
8798 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8800 *Jüri Aedla, Steve Henson*
8802 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8803 are subject to a denial of service attack.
8805 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8806 this issue. ([CVE-2014-3470])
8808 *Felix Gröbert, Ivan Fratric, Steve Henson*
8810 * Harmonize version and its documentation. -f flag is used to display
8813 *mancha <mancha1@zoho.com>*
8815 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8816 in i2d_ECPrivateKey.
8818 *mancha <mancha1@zoho.com>*
8820 * Fix some double frees. These are not thought to be exploitable.
8822 *mancha <mancha1@zoho.com>*
8824 * Fix for the attack described in the paper "Recovering OpenSSL
8825 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8826 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8827 <http://eprint.iacr.org/2014/140>
8829 Thanks to Yuval Yarom and Naomi Benger for discovering this
8830 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8832 *Yuval Yarom and Naomi Benger*
8834 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8836 * Keep original DTLS digest and encryption contexts in retransmission
8837 structures so we can use the previous session parameters if they need
8838 to be resent. ([CVE-2013-6450])
8842 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8843 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8844 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8845 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8846 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8847 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8849 *Rob Stradling, Adam Langley*
8851 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8853 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8855 This addresses the flaw in CBC record processing discovered by
8856 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8857 at: <http://www.isg.rhul.ac.uk/tls/>
8859 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8860 Security Group at Royal Holloway, University of London
8861 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8862 Emilia Käsper for the initial patch.
8865 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8867 * Return an error when checking OCSP signatures when key is NULL.
8868 This fixes a DoS attack. ([CVE-2013-0166])
8872 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8873 the right response is stapled. Also change SSL_get_certificate()
8874 so it returns the certificate actually sent.
8875 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8876 (This is a backport)
8878 *Rob Stradling <rob.stradling@comodo.com>*
8880 * Fix possible deadlock when decoding public keys.
8884 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8886 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8889 * Sanity check record length before skipping explicit IV in DTLS
8892 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8893 fuzzing as a service testing platform.
8898 * Initialise tkeylen properly when encrypting CMS messages.
8899 Thanks to Solar Designer of Openwall for reporting this issue.
8903 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8905 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8906 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8907 in CRYPTO_realloc_clean.
8909 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8910 issue and to Adam Langley <agl@chromium.org> for fixing it.
8913 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8915 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8917 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8918 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8919 content decryption and always return the same error. Note: this attack
8920 needs on average 2^20 messages so it only affects automated senders. The
8921 old behaviour can be re-enabled in the CMS code by setting the
8922 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8923 an MMA defence is not necessary.
8924 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8925 this issue. ([CVE-2012-0884])
8929 * Fix CVE-2011-4619: make sure we really are receiving a
8930 client hello before rejecting multiple SGC restarts. Thanks to
8931 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8935 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8937 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8938 Thanks to Antonio Martin, Enterprise Secure Access Research and
8939 Development, Cisco Systems, Inc. for discovering this bug and
8940 preparing a fix. ([CVE-2012-0050])
8944 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8946 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8947 of the Vaudenay padding oracle attack on CBC mode encryption
8948 which enables an efficient plaintext recovery attack against
8949 the OpenSSL implementation of DTLS. Their attack exploits timing
8950 differences arising during decryption processing. A research
8951 paper describing this attack can be found at:
8952 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8953 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8954 Security Group at Royal Holloway, University of London
8955 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8956 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8957 for preparing the fix. ([CVE-2011-4108])
8959 *Robin Seggelmann, Michael Tuexen*
8961 * Clear bytes used for block padding of SSL 3.0 records.
8964 *Adam Langley (Google)*
8966 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8967 Kadianakis <desnacked@gmail.com> for discovering this issue and
8968 Adam Langley for preparing the fix. ([CVE-2011-4619])
8970 *Adam Langley (Google)*
8972 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8974 *Andrey Kulikov <amdeich@gmail.com>*
8976 * Prevent malformed RFC3779 data triggering an assertion failure.
8977 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8978 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8980 *Rob Austein <sra@hactrn.net>*
8982 * Improved PRNG seeding for VOS.
8984 *Paul Green <Paul.Green@stratus.com>*
8986 * Fix ssl_ciph.c set-up race.
8988 *Adam Langley (Google)*
8990 * Fix spurious failures in ecdsatest.c.
8992 *Emilia Käsper (Google)*
8994 * Fix the BIO_f_buffer() implementation (which was mixing different
8995 interpretations of the `..._len` fields).
8997 *Adam Langley (Google)*
8999 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9000 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9001 threads won't reuse the same blinding coefficients.
9003 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9004 lock to call BN_BLINDING_invert_ex, and avoids one use of
9005 BN_BLINDING_update for each BN_BLINDING structure (previously,
9006 the last update always remained unused).
9008 *Emilia Käsper (Google)*
9010 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9012 *Bob Buckholz (Google)*
9014 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9016 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9017 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9019 *Kaspar Brand <ossl@velox.ch>*
9021 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9022 for multi-threaded use of ECDH. ([CVE-2011-3210])
9024 *Adam Langley (Google)*
9026 * Fix x509_name_ex_d2i memory leak on bad inputs.
9030 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9031 signature public key algorithm by using OID xref utilities instead.
9032 Before this you could only use some ECC ciphersuites with SHA1 only.
9036 * Add protection against ECDSA timing attacks as mentioned in the paper
9037 by Billy Bob Brumley and Nicola Tuveri, see:
9038 <http://eprint.iacr.org/2011/232.pdf>
9040 *Billy Bob Brumley and Nicola Tuveri*
9042 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9044 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9046 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9048 * Fix bug in string printing code: if *any* escaping is enabled we must
9049 escape the escape character (backslash) or the resulting string is
9054 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9056 * Disable code workaround for ancient and obsolete Netscape browsers
9057 and servers: an attacker can use it in a ciphersuite downgrade attack.
9058 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9062 * Fixed J-PAKE implementation error, originally discovered by
9063 Sebastien Martini, further info and confirmation from Stefan
9064 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9068 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9070 * Fix extension code to avoid race conditions which can result in a buffer
9071 overrun vulnerability: resumed sessions must not be modified as they can
9072 be shared by multiple threads. CVE-2010-3864
9076 * Fix WIN32 build system to correctly link an ENGINE directory into
9081 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9083 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9086 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9088 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9090 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9091 context. The operation can be customised via the ctrl mechanism in
9092 case ENGINEs want to include additional functionality.
9096 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9100 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9101 output hashes compatible with older versions of OpenSSL.
9103 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9105 * Fix compression algorithm handling: if resuming a session use the
9106 compression algorithm of the resumed session instead of determining
9107 it from client hello again. Don't allow server to change algorithm.
9111 * Add load_crls() function to commands tidying load_certs() too. Add option
9112 to verify utility to allow additional CRLs to be included.
9116 * Update OCSP request code to permit adding custom headers to the request:
9117 some responders need this.
9121 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9124 *Julia Lawall <julia@diku.dk>*
9126 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9127 needlessly dereferenced structures, used obsolete functions and
9128 didn't handle all updated verify codes correctly.
9132 * Disable MD2 in the default configuration.
9136 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9137 indicate the initial BIO being pushed or popped. This makes it possible
9138 to determine whether the BIO is the one explicitly called or as a result
9139 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9140 it handles reference counts correctly and doesn't zero out the I/O bio
9141 when it is not being explicitly popped. WARNING: applications which
9142 included workarounds for the old buggy behaviour will need to be modified
9143 or they could free up already freed BIOs.
9147 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9148 renaming to all platforms (within the 0.9.8 branch, this was
9149 done conditionally on Netware platforms to avoid a name clash).
9151 *Guenter <lists@gknw.net>*
9153 * Add ECDHE and PSK support to DTLS.
9155 *Michael Tuexen <tuexen@fh-muenster.de>*
9157 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9162 * Add "missing" function EVP_MD_flags() (without this the only way to
9163 retrieve a digest flags is by accessing the structure directly. Update
9164 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9165 or cipher is registered as in the "from" argument. Print out all
9166 registered digests in the dgst usage message instead of manually
9167 attempting to work them out.
9171 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9172 this allows the use of compression and extensions. Change default cipher
9173 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9174 by default unless an application cipher string requests it.
9178 * Alter match criteria in PKCS12_parse(). It used to try to use local
9179 key ids to find matching certificates and keys but some PKCS#12 files
9180 don't follow the (somewhat unwritten) rules and this strategy fails.
9181 Now just gather all certificates together and the first private key
9182 then look for the first certificate that matches the key.
9186 * Support use of registered digest and cipher names for dgst and cipher
9187 commands instead of having to add each one as a special case. So now
9194 openssl dgst -sha256 foo
9196 and this works for ENGINE based algorithms too.
9200 * Update Gost ENGINE to support parameter files.
9202 *Victor B. Wagner <vitus@cryptocom.ru>*
9204 * Support GeneralizedTime in ca utility.
9206 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9208 * Enhance the hash format used for certificate directory links. The new
9209 form uses the canonical encoding (meaning equivalent names will work
9210 even if they aren't identical) and uses SHA1 instead of MD5. This form
9211 is incompatible with the older format and as a result c_rehash should
9212 be used to rebuild symbolic links.
9216 * Make PKCS#8 the default write format for private keys, replacing the
9217 traditional format. This form is standardised, more secure and doesn't
9218 include an implicit MD5 dependency.
9222 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9223 committed to OpenSSL should pass this lot as a minimum.
9227 * Add session ticket override functionality for use by EAP-FAST.
9229 *Jouni Malinen <j@w1.fi>*
9231 * Modify HMAC functions to return a value. Since these can be implemented
9232 in an ENGINE errors can occur.
9236 * Type-checked OBJ_bsearch_ex.
9240 * Type-checked OBJ_bsearch. Also some constification necessitated
9241 by type-checking. Still to come: TXT_DB, bsearch(?),
9242 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9247 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9248 seconds to a tm structure directly, instead of going through OS
9249 specific date routines. This avoids any issues with OS routines such
9250 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9251 and X509_time_adj_ex() to cover the extended range. The existing
9252 X509_time_adj() is still usable and will no longer have any date issues.
9256 * Delta CRL support. New use deltas option which will attempt to locate
9257 and search any appropriate delta CRLs available.
9259 This work was sponsored by Google.
9263 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9264 code and add additional score elements. Validate alternate CRL paths
9265 as part of the CRL checking and indicate a new error "CRL path validation
9266 error" in this case. Applications wanting additional details can use
9267 the verify callback and check the new "parent" field. If this is not
9268 NULL CRL path validation is taking place. Existing applications won't
9269 see this because it requires extended CRL support which is off by
9272 This work was sponsored by Google.
9276 * Support for freshest CRL extension.
9278 This work was sponsored by Google.
9282 * Initial indirect CRL support. Currently only supported in the CRLs
9283 passed directly and not via lookup. Process certificate issuer
9284 CRL entry extension and lookup CRL entries by bother issuer name
9285 and serial number. Check and process CRL issuer entry in IDP extension.
9287 This work was sponsored by Google.
9291 * Add support for distinct certificate and CRL paths. The CRL issuer
9292 certificate is validated separately in this case. Only enabled if
9293 an extended CRL support flag is set: this flag will enable additional
9294 CRL functionality in future.
9296 This work was sponsored by Google.
9300 * Add support for policy mappings extension.
9302 This work was sponsored by Google.
9306 * Fixes to pathlength constraint, self issued certificate handling,
9307 policy processing to align with RFC3280 and PKITS tests.
9309 This work was sponsored by Google.
9313 * Support for name constraints certificate extension. DN, email, DNS
9314 and URI types are currently supported.
9316 This work was sponsored by Google.
9320 * To cater for systems that provide a pointer-based thread ID rather
9321 than numeric, deprecate the current numeric thread ID mechanism and
9322 replace it with a structure and associated callback type. This
9323 mechanism allows a numeric "hash" to be extracted from a thread ID in
9324 either case, and on platforms where pointers are larger than 'long',
9325 mixing is done to help ensure the numeric 'hash' is usable even if it
9326 can't be guaranteed unique. The default mechanism is to use "&errno"
9327 as a pointer-based thread ID to distinguish between threads.
9329 Applications that want to provide their own thread IDs should now use
9330 CRYPTO_THREADID_set_callback() to register a callback that will call
9331 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9333 Note that ERR_remove_state() is now deprecated, because it is tied
9334 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9335 to free the current thread's error state should be replaced by
9336 ERR_remove_thread_state(NULL).
9338 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9339 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9340 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9341 application was previously providing a numeric thread callback that
9342 was inappropriate for distinguishing threads, then uniqueness might
9343 have been obtained with &errno that happened immediately in the
9344 intermediate development versions of OpenSSL; this is no longer the
9345 case, the numeric thread callback will now override the automatic use
9348 *Geoff Thorpe, with help from Bodo Moeller*
9350 * Initial support for different CRL issuing certificates. This covers a
9351 simple case where the self issued certificates in the chain exist and
9352 the real CRL issuer is higher in the existing chain.
9354 This work was sponsored by Google.
9358 * Removed effectively defunct crypto/store from the build.
9362 * Revamp of STACK to provide stronger type-checking. Still to come:
9363 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9364 ASN1_STRING, CONF_VALUE.
9368 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9369 RAM on SSL connections. This option can save about 34k per idle SSL.
9373 * Revamp of LHASH to provide stronger type-checking. Still to come:
9374 STACK, TXT_DB, bsearch, qsort.
9378 * Initial support for Cryptographic Message Syntax (aka CMS) based
9379 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9380 support for data, signedData, compressedData, digestedData and
9381 encryptedData, envelopedData types included. Scripts to check against
9382 RFC4134 examples draft and interop and consistency checks of many
9383 content types and variants.
9387 * Add options to enc utility to support use of zlib compression BIO.
9391 * Extend mk1mf to support importing of options and assembly language
9392 files from Configure script, currently only included in VC-WIN32.
9393 The assembly language rules can now optionally generate the source
9394 files from the associated perl scripts.
9398 * Implement remaining functionality needed to support GOST ciphersuites.
9399 Interop testing has been performed using CryptoPro implementations.
9401 *Victor B. Wagner <vitus@cryptocom.ru>*
9403 * s390x assembler pack.
9407 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9412 * Implement Opaque PRF Input TLS extension as specified in
9413 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9414 official specification yet and no extension type assignment by
9415 IANA exists, this extension (for now) will have to be explicitly
9416 enabled when building OpenSSL by providing the extension number
9417 to use. For example, specify an option
9419 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9421 to the "config" or "Configure" script to enable the extension,
9422 assuming extension number 0x9527 (which is a completely arbitrary
9423 and unofficial assignment based on the MD5 hash of the Internet
9424 Draft). Note that by doing so, you potentially lose
9425 interoperability with other TLS implementations since these might
9426 be using the same extension number for other purposes.
9428 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9429 opaque PRF input value to use in the handshake. This will create
9430 an internal copy of the length-'len' string at 'src', and will
9431 return non-zero for success.
9433 To get more control and flexibility, provide a callback function
9436 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9437 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9441 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9444 Callback function 'cb' will be called in handshakes, and is
9445 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9446 Argument 'arg' is for application purposes (the value as given to
9447 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9448 be provided to the callback function). The callback function
9449 has to return non-zero to report success: usually 1 to use opaque
9450 PRF input just if possible, or 2 to enforce use of the opaque PRF
9451 input. In the latter case, the library will abort the handshake
9452 if opaque PRF input is not successfully negotiated.
9454 Arguments 'peerinput' and 'len' given to the callback function
9455 will always be NULL and 0 in the case of a client. A server will
9456 see the client's opaque PRF input through these variables if
9457 available (NULL and 0 otherwise). Note that if the server
9458 provides an opaque PRF input, the length must be the same as the
9459 length of the client's opaque PRF input.
9461 Note that the callback function will only be called when creating
9462 a new session (session resumption can resume whatever was
9463 previously negotiated), and will not be called in SSL 2.0
9464 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9465 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9466 for applications that need to enforce opaque PRF input.
9470 * Update ssl code to support digests other than SHA1+MD5 for handshake
9473 *Victor B. Wagner <vitus@cryptocom.ru>*
9475 * Add RFC4507 support to OpenSSL. This includes the corrections in
9476 RFC4507bis. The encrypted ticket format is an encrypted encoded
9477 SSL_SESSION structure, that way new session features are automatically
9480 If a client application caches session in an SSL_SESSION structure
9481 support is transparent because tickets are now stored in the encoded
9484 The SSL_CTX structure automatically generates keys for ticket
9485 protection in servers so again support should be possible
9486 with no application modification.
9488 If a client or server wishes to disable RFC4507 support then the option
9489 SSL_OP_NO_TICKET can be set.
9491 Add a TLS extension debugging callback to allow the contents of any client
9492 or server extensions to be examined.
9494 This work was sponsored by Google.
9498 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9499 OpenSSL should now compile cleanly on gcc 4.2
9501 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9503 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9504 support including streaming MAC support: this is required for GOST
9505 ciphersuite support.
9507 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9509 * Add option -stream to use PKCS#7 streaming in smime utility. New
9510 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9511 to output in BER and PEM format.
9515 * Experimental support for use of HMAC via EVP_PKEY interface. This
9516 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9517 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9518 ENGINE support for HMAC keys which are unextractable. New -mac and
9519 -macopt options to dgst utility.
9523 * New option -sigopt to dgst utility. Update dgst to use
9524 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9525 alternative signing parameters such as X9.31 or PSS in the dgst
9530 * Change ssl_cipher_apply_rule(), the internal function that does
9531 the work each time a ciphersuite string requests enabling
9532 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9533 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9534 the order of disabled ciphersuites such that those ciphersuites
9535 that most recently went from enabled to disabled not only stay
9536 in order with respect to each other, but also have higher priority
9537 than other disabled ciphersuites the next time ciphersuites are
9540 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9541 the same ciphersuites as with "HIGH" alone, but in a specific
9542 order where the PSK ciphersuites come first (since they are the
9543 most recently disabled ciphersuites when "HIGH" is parsed).
9545 Also, change ssl_create_cipher_list() (using this new
9546 functionality) such that between otherwise identical
9547 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9552 * Change ssl_create_cipher_list() so that it automatically
9553 arranges the ciphersuites in reasonable order before starting
9554 to process the rule string. Thus, the definition for "DEFAULT"
9555 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9556 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9557 This makes it much easier to arrive at a reasonable default order
9558 in applications for which anonymous ciphers are OK (meaning
9559 that you can't actually use DEFAULT).
9561 *Bodo Moeller; suggested by Victor Duchovni*
9563 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9564 processing) into multiple integers instead of setting
9565 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9566 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9567 (These masks as well as the individual bit definitions are hidden
9568 away into the non-exported interface ssl/ssl_locl.h, so this
9569 change to the definition of the SSL_CIPHER structure shouldn't
9570 affect applications.) This give us more bits for each of these
9571 categories, so there is no longer a need to coagulate AES128 and
9572 AES256 into a single algorithm bit, and to coagulate Camellia128
9573 and Camellia256 into a single algorithm bit, which has led to all
9576 Thus, among other things, the kludge introduced in 0.9.7m and
9577 0.9.8e for masking out AES256 independently of AES128 or masking
9578 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9580 With the change, we also introduce new ciphersuite aliases that
9581 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9586 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9587 Use the leftmost N bytes of the signature input if the input is
9588 larger than the prime q (with N being the size in bytes of q).
9592 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9593 it yet and it is largely untested.
9597 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9601 * Initial incomplete changes to avoid need for function casts in OpenSSL
9602 some compilers (gcc 4.2 and later) reject their use. Safestack is
9603 reimplemented. Update ASN1 to avoid use of legacy functions.
9607 * Win32/64 targets are linked with Winsock2.
9611 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9612 to external functions. This can be used to increase CRL handling
9613 efficiency especially when CRLs are very large by (for example) storing
9614 the CRL revoked certificates in a database.
9618 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9619 new CRLs added to a directory can be used. New command line option
9620 -verify_return_error to s_client and s_server. This causes real errors
9621 to be returned by the verify callback instead of carrying on no matter
9622 what. This reflects the way a "real world" verify callback would behave.
9626 * GOST engine, supporting several GOST algorithms and public key formats.
9627 Kindly donated by Cryptocom.
9631 * Partial support for Issuing Distribution Point CRL extension. CRLs
9632 partitioned by DP are handled but no indirect CRL or reason partitioning
9633 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9634 selected via a scoring technique which handles IDP and AKID in CRLs.
9638 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9639 will ultimately be used for all verify operations: this will remove the
9640 X509_STORE dependency on certificate verification and allow alternative
9641 lookup methods. X509_STORE based implementations of these two callbacks.
9645 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9646 Modify get_crl() to find a valid (unexpired) CRL if possible.
9650 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9651 this would be called X509_CRL_cmp() but that name is already used by
9652 a function that just compares CRL issuer names. Cache several CRL
9653 extensions in X509_CRL structure and cache CRLDP in X509.
9657 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9658 this maps equivalent X509_NAME structures into a consistent structure.
9659 Name comparison can then be performed rapidly using memcmp().
9663 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9668 * Allow digests to supply their own micalg string for S/MIME type using
9669 the ctrl EVP_MD_CTRL_MICALG.
9673 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9674 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9675 ctrl. It can then customise the structure before and/or after signing
9680 * New function OBJ_add_sigid() to allow application defined signature OIDs
9681 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9682 to free up any added signature OIDs.
9686 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9687 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9688 digest and cipher tables. New options added to openssl utility:
9689 list-message-digest-algorithms and list-cipher-algorithms.
9693 * Change the array representation of binary polynomials: the list
9694 of degrees of non-zero coefficients is now terminated with -1.
9695 Previously it was terminated with 0, which was also part of the
9696 value; thus, the array representation was not applicable to
9697 polynomials where t^0 has coefficient zero. This change makes
9698 the array representation useful in a more general context.
9702 * Various modifications and fixes to SSL/TLS cipher string
9703 handling. For ECC, the code now distinguishes between fixed ECDH
9704 with RSA certificates on the one hand and with ECDSA certificates
9705 on the other hand, since these are separate ciphersuites. The
9706 unused code for Fortezza ciphersuites has been removed.
9708 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9709 (not "ECDHE"). For consistency with the code for DH
9710 certificates, use of ECDH certificates is now considered ECDH
9711 authentication, not RSA or ECDSA authentication (the latter is
9712 merely the CA's signing algorithm and not actively used in the
9715 The temporary ciphersuite alias "ECCdraft" is no longer
9716 available, and ECC ciphersuites are no longer excluded from "ALL"
9717 and "DEFAULT". The following aliases now exist for RFC 4492
9718 ciphersuites, most of these by analogy with the DH case:
9720 kECDHr - ECDH cert, signed with RSA
9721 kECDHe - ECDH cert, signed with ECDSA
9722 kECDH - ECDH cert (signed with either RSA or ECDSA)
9723 kEECDH - ephemeral ECDH
9724 ECDH - ECDH cert or ephemeral ECDH
9730 AECDH - anonymous ECDH
9731 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9735 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9736 Use correct micalg parameters depending on digest(s) in signed message.
9740 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9741 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9745 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9746 an engine to register a method. Add ENGINE lookups for methods and
9747 functional reference processing.
9751 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9752 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9757 * New -resign option to smime utility. This adds one or more signers
9758 to an existing PKCS#7 signedData structure. Also -md option to use an
9759 alternative message digest algorithm for signing.
9763 * Tidy up PKCS#7 routines and add new functions to make it easier to
9764 create PKCS7 structures containing multiple signers. Update smime
9765 application to support multiple signers.
9769 * New -macalg option to pkcs12 utility to allow setting of an alternative
9774 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9775 Reorganize PBE internals to lookup from a static table using NIDs,
9776 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9777 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9778 PRF which will be automatically used with PBES2.
9782 * Replace the algorithm specific calls to generate keys in "req" with the
9787 * Update PKCS#7 enveloped data routines to use new API. This is now
9788 supported by any public key method supporting the encrypt operation. A
9789 ctrl is added to allow the public key algorithm to examine or modify
9790 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9795 * Add a ctrl to asn1 method to allow a public key algorithm to express
9796 a default digest type to use. In most cases this will be SHA1 but some
9797 algorithms (such as GOST) need to specify an alternative digest. The
9798 return value indicates how strong the preference is 1 means optional and
9799 2 is mandatory (that is it is the only supported type). Modify
9800 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9801 use the default md. Update openssl utilities to use the default digest
9802 type for signing if it is not explicitly indicated.
9806 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9807 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9808 signing method from the key type. This effectively removes the link
9809 between digests and public key types.
9813 * Add an OID cross reference table and utility functions. Its purpose is to
9814 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9815 rsaEncryption. This will allow some of the algorithm specific hackery
9816 needed to use the correct OID to be removed.
9820 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9821 structures for PKCS7_sign(). They are now set up by the relevant public
9826 * Add provisional EC pkey method with support for ECDSA and ECDH.
9830 * Add support for key derivation (agreement) in the API, DH method and
9835 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9836 public and private key formats. As a side effect these add additional
9837 command line functionality not previously available: DSA signatures can be
9838 generated and verified using pkeyutl and DH key support and generation in
9845 *Oliver Tappe <zooey@hirschkaefer.de>*
9847 * New make target "install_html_docs" installs HTML renditions of the
9850 *Oliver Tappe <zooey@hirschkaefer.de>*
9852 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9853 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9854 support key and parameter generation and add initial key generation
9855 functionality for RSA.
9859 * Add functions for main EVP_PKEY_method operations. The undocumented
9860 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9861 `EVP_PKEY_{encrypt,decrypt}_old`.
9865 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9866 key API, doesn't do much yet.
9870 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9871 public key algorithms. New option to openssl utility:
9872 "list-public-key-algorithms" to print out info.
9876 * Implement the Supported Elliptic Curves Extension for
9877 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9881 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9882 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9886 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9887 utilities such as rsa, dsa, dsaparam etc except they process any key
9892 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9893 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9894 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9899 * Initial support for pluggable public key ASN1.
9900 De-spaghettify the public key ASN1 handling. Move public and private
9901 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9902 algorithm specific handling to a single module within the relevant
9903 algorithm directory. Add functions to allow (near) opaque processing
9904 of public and private key structures.
9908 * Implement the Supported Point Formats Extension for
9909 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9913 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9914 for the psk identity [hint] and the psk callback functions to the
9915 SSL_SESSION, SSL and SSL_CTX structure.
9918 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9922 SSL_CTX_use_psk_identity_hint
9923 SSL_get_psk_identity_hint
9924 SSL_get_psk_identity
9925 SSL_use_psk_identity_hint
9927 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9929 * Add RFC 3161 compliant time stamp request creation, response generation
9930 and response verification functionality.
9932 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9934 * Add initial support for TLS extensions, specifically for the server_name
9935 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9936 have new members for a hostname. The SSL data structure has an
9937 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9938 stored in that context to allow for session resumption, even after the
9939 SSL has been switched to a new SSL_CTX in reaction to a client's
9940 server_name extension.
9942 New functions (subject to change):
9944 SSL_get_servername()
9945 SSL_get_servername_type()
9948 New CTRL codes and macros (subject to change):
9950 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9951 - SSL_CTX_set_tlsext_servername_callback()
9952 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9953 - SSL_CTX_set_tlsext_servername_arg()
9954 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9956 openssl s_client has a new '-servername ...' option.
9958 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9959 '-key2 ...', '-servername_fatal' (subject to change). This allows
9960 testing the HostName extension for a specific single hostname ('-cert'
9961 and '-key' remain fallbacks for handshakes without HostName
9962 negotiation). If the unrecognized_name alert has to be sent, this by
9963 default is a warning; it becomes fatal with the '-servername_fatal'
9966 *Peter Sylvester, Remy Allais, Christophe Renou*
9968 * Whirlpool hash implementation is added.
9972 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9973 bn(64,32). Because of instruction set limitations it doesn't have
9974 any negative impact on performance. This was done mostly in order
9975 to make it possible to share assembler modules, such as bn_mul_mont
9976 implementations, between 32- and 64-bit builds without hassle.
9980 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9981 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9986 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9987 dedicated Montgomery multiplication procedure, is introduced.
9988 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9989 "64-bit" performance on certain 32-bit targets.
9993 * New option SSL_OP_NO_COMP to disable use of compression selectively
9994 in SSL structures. New SSL ctrl to set maximum send fragment size.
9995 Save memory by setting the I/O buffer sizes dynamically instead of
9996 using the maximum available value.
10000 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10001 in addition to the text details.
10005 * Very, very preliminary EXPERIMENTAL support for printing of general
10006 ASN1 structures. This currently produces rather ugly output and doesn't
10007 handle several customised structures at all.
10011 * Integrated support for PVK file format and some related formats such
10012 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10013 these in the 'rsa' and 'dsa' utilities.
10017 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10021 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10022 place for the (very old) "NETSCAPE" format certificates which are now
10023 handled using new ASN1 code equivalents.
10027 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10028 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10029 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10033 * Modify CRL distribution points extension code to print out previously
10034 unsupported fields. Enhance extension setting code to allow setting of
10039 * Add print and set support for Issuing Distribution Point CRL extension.
10043 * Change 'Configure' script to enable Camellia by default.
10050 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10052 * When rejecting SSL/TLS records due to an incorrect version number, never
10053 update s->server with a new major version number. As of
10054 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10055 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10056 the previous behavior could result in a read attempt at NULL when
10057 receiving specific incorrect SSL/TLS records once record payload
10058 protection is active. ([CVE-2010-0740])
10060 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10062 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10063 could be crashed if the relevant tables were not present (e.g. chrooted).
10065 *Tomas Hoger <thoger@redhat.com>*
10067 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10069 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10071 *Martin Olsson, Neel Mehta*
10073 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10074 accommodate for stack sorting, always a write lock!).
10078 * On some versions of WIN32 Heap32Next is very slow. This can cause
10079 excessive delays in the RAND_poll(): over a minute. As a workaround
10080 include a time check in the inner Heap32Next loop too.
10084 * The code that handled flushing of data in SSL/TLS originally used the
10085 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10086 the problem outlined in PR#1949. The fix suggested there however can
10087 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10088 of Apache). So instead simplify the code to flush unconditionally.
10089 This should be fine since flushing with no data to flush is a no op.
10093 * Handle TLS versions 2.0 and later properly and correctly use the
10094 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10095 off ancient servers have a habit of sticking around for a while...
10099 * Modify compression code so it frees up structures without using the
10100 ex_data callbacks. This works around a problem where some applications
10101 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10102 restarting) then use compression (e.g. SSL with compression) later.
10103 This results in significant per-connection memory leaks and
10104 has caused some security issues including CVE-2008-1678 and
10109 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10110 change when encrypting or decrypting.
10114 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10115 connect and renegotiate with servers which do not support RI.
10116 Until RI is more widely deployed this option is enabled by default.
10120 * Add "missing" ssl ctrls to clear options and mode.
10124 * If client attempts to renegotiate and doesn't support RI respond with
10125 a no_renegotiation alert as required by RFC5746. Some renegotiating
10126 TLS clients will continue a connection gracefully when they receive
10127 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10128 waiting for a server hello which it will never receive. Now we treat a
10129 received no_renegotiation alert as a fatal error. This is because
10130 applications requesting a renegotiation might well expect it to succeed
10131 and would have no code in place to handle the server denying it so the
10132 only safe thing to do is to terminate the connection.
10136 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10137 peer supports secure renegotiation and 0 otherwise. Print out peer
10138 renegotiation support in s_client/s_server.
10142 * Replace the highly broken and deprecated SPKAC certification method with
10143 the updated NID creation version. This should correctly handle UTF8.
10147 * Implement RFC5746. Re-enable renegotiation but require the extension
10148 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10149 turns out to be a bad idea. It has been replaced by
10150 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10151 SSL_CTX_set_options(). This is really not recommended unless you
10152 know what you are doing.
10154 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10156 * Fixes to stateless session resumption handling. Use initial_ctx when
10157 issuing and attempting to decrypt tickets in case it has changed during
10158 servername handling. Use a non-zero length session ID when attempting
10159 stateless session resumption: this makes it possible to determine if
10160 a resumption has occurred immediately after receiving server hello
10161 (several places in OpenSSL subtly assume this) instead of later in
10166 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10167 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10168 fixes for a few places where the return code is not checked
10171 *Julia Lawall <julia@diku.dk>*
10173 * Add --strict-warnings option to Configure script to include devteam
10174 warnings in other configurations.
10178 * Add support for --libdir option and LIBDIR variable in makefiles. This
10179 makes it possible to install openssl libraries in locations which
10180 have names other than "lib", for example "/usr/lib64" which some
10183 *Steve Henson, based on patch from Jeremy Utley*
10185 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10186 X690 8.9.12 and can produce some misleading textual output of OIDs.
10188 *Steve Henson, reported by Dan Kaminsky*
10190 * Delete MD2 from algorithm tables. This follows the recommendation in
10191 several standards that it is not used in new applications due to
10192 several cryptographic weaknesses. For binary compatibility reasons
10193 the MD2 API is still compiled in by default.
10197 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10202 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10203 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10206 *Guenter <lists@gknw.net>*
10208 * Fix the server certificate chain building code to use X509_verify_cert(),
10209 it used to have an ad-hoc builder which was unable to cope with anything
10210 other than a simple chain.
10212 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10214 * Don't check self signed certificate signatures in X509_verify_cert()
10215 by default (a flag can override this): it just wastes time without
10216 adding any security. As a useful side effect self signed root CAs
10217 with non-FIPS digests are now usable in FIPS mode.
10221 * In dtls1_process_out_of_seq_message() the check if the current message
10222 is already buffered was missing. For every new message was memory
10223 allocated, allowing an attacker to perform an denial of service attack
10224 with sending out of seq handshake messages until there is no memory
10225 left. Additionally every future message was buffered, even if the
10226 sequence number made no sense and would be part of another handshake.
10227 So only messages with sequence numbers less than 10 in advance will be
10228 buffered. ([CVE-2009-1378])
10230 *Robin Seggelmann, discovered by Daniel Mentz*
10232 * Records are buffered if they arrive with a future epoch to be
10233 processed after finishing the corresponding handshake. There is
10234 currently no limitation to this buffer allowing an attacker to perform
10235 a DOS attack with sending records with future epochs until there is no
10236 memory left. This patch adds the pqueue_size() function to determine
10237 the size of a buffer and limits the record buffer to 100 entries.
10240 *Robin Seggelmann, discovered by Daniel Mentz*
10242 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10243 parent structure is freed. ([CVE-2009-1379])
10247 * Handle non-blocking I/O properly in SSL_shutdown() call.
10249 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10251 * Add `2.5.4.*` OIDs
10253 *Ilya O. <vrghost@gmail.com>*
10255 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10257 * Disable renegotiation completely - this fixes a severe security
10258 problem ([CVE-2009-3555]) at the cost of breaking all
10259 renegotiation. Renegotiation can be re-enabled by setting
10260 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10261 run-time. This is really not recommended unless you know what
10266 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10268 * Don't set val to NULL when freeing up structures, it is freed up by
10269 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10270 zeroing past the valid field. ([CVE-2009-0789])
10272 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10274 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10275 checked correctly. This would allow some invalid signed attributes to
10276 appear to verify correctly. ([CVE-2009-0591])
10278 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10280 * Reject UniversalString and BMPString types with invalid lengths. This
10281 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10282 a legal length. ([CVE-2009-0590])
10286 * Set S/MIME signing as the default purpose rather than setting it
10287 unconditionally. This allows applications to override it at the store
10292 * Permit restricted recursion of ASN1 strings. This is needed in practice
10293 to handle some structures.
10297 * Improve efficiency of mem_gets: don't search whole buffer each time
10300 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10302 * New -hex option for openssl rand.
10306 * Print out UTF8String and NumericString when parsing ASN1.
10310 * Support NumericString type for name components.
10314 * Allow CC in the environment to override the automatically chosen
10315 compiler. Note that nothing is done to ensure flags work with the
10320 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10322 * Properly check EVP_VerifyFinal() and similar return values
10325 *Ben Laurie, Bodo Moeller, Google Security Team*
10327 * Enable TLS extensions by default.
10331 * Allow the CHIL engine to be loaded, whether the application is
10332 multithreaded or not. (This does not release the developer from the
10333 obligation to set up the dynamic locking callbacks.)
10335 *Sander Temme <sander@temme.net>*
10337 * Use correct exit code if there is an error in dgst command.
10339 *Steve Henson; problem pointed out by Roland Dirlewanger*
10341 * Tweak Configure so that you need to say "experimental-jpake" to enable
10342 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10346 * Add experimental JPAKE support, including demo authentication in
10347 s_client and s_server.
10351 * Set the comparison function in v3_addr_canonize().
10353 *Rob Austein <sra@hactrn.net>*
10355 * Add support for XMPP STARTTLS in s_client.
10357 *Philip Paeps <philip@freebsd.org>*
10359 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10360 to ensure that even with this option, only ciphersuites in the
10361 server's preference list will be accepted. (Note that the option
10362 applies only when resuming a session, so the earlier behavior was
10363 just about the algorithm choice for symmetric cryptography.)
10367 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10369 * Fix NULL pointer dereference if a DTLS server received
10370 ChangeCipherSpec as first record ([CVE-2009-1386]).
10374 * Fix a state transition in s3_srvr.c and d1_srvr.c
10375 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10377 *Nagendra Modadugu*
10379 * The fix in 0.9.8c that supposedly got rid of unsafe
10380 double-checked locking was incomplete for RSA blinding,
10381 addressing just one layer of what turns out to have been
10382 doubly unsafe triple-checked locking.
10384 So now fix this for real by retiring the MONT_HELPER macro
10385 in crypto/rsa/rsa_eay.c.
10387 *Bodo Moeller; problem pointed out by Marius Schilder*
10389 * Various precautionary measures:
10391 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10393 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10394 (NB: This would require knowledge of the secret session ticket key
10395 to exploit, in which case you'd be SOL either way.)
10397 - Change bn_nist.c so that it will properly handle input BIGNUMs
10398 outside the expected range.
10400 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10403 *Neel Mehta, Bodo Moeller*
10405 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10406 the load fails. Useful for distros.
10408 *Ben Laurie and the FreeBSD team*
10410 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10414 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10418 * Expand ENGINE to support engine supplied SSL client certificate functions.
10420 This work was sponsored by Logica.
10424 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10425 keystores. Support for SSL/TLS client authentication too.
10426 Not compiled unless enable-capieng specified to Configure.
10428 This work was sponsored by Logica.
10432 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10433 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10434 attribute creation routines such as certificate requests and PKCS#12
10439 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10441 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10442 handshake which could lead to a client crash as found using the
10443 Codenomicon TLS test suite ([CVE-2008-1672])
10445 *Steve Henson, Mark Cox*
10447 * Fix double free in TLS server name extensions which could lead to
10448 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10452 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10454 Clear the error queue to ensure that error entries left from
10455 older function calls do not interfere with the correct operation.
10457 *Lutz Jaenicke, Erik de Castro Lopo*
10459 * Remove root CA certificates of commercial CAs:
10461 The OpenSSL project does not recommend any specific CA and does not
10462 have any policy with respect to including or excluding any CA.
10463 Therefore, it does not make any sense to ship an arbitrary selection
10464 of root CA certificates with the OpenSSL software.
10468 * RSA OAEP patches to fix two separate invalid memory reads.
10469 The first one involves inputs when 'lzero' is greater than
10470 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10471 before the beginning of from). The second one involves inputs where
10472 the 'db' section contains nothing but zeroes (there is a one-byte
10473 invalid read after the end of 'db').
10475 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10477 * Partial backport from 0.9.9-dev:
10479 Introduce bn_mul_mont (dedicated Montgomery multiplication
10480 procedure) as a candidate for BIGNUM assembler implementation.
10481 While 0.9.9-dev uses assembler for various architectures, only
10482 x86_64 is available by default here in the 0.9.8 branch, and
10483 32-bit x86 is available through a compile-time setting.
10485 To try the 32-bit x86 assembler implementation, use Configure
10486 option "enable-montasm" (which exists only for this backport).
10488 As "enable-montasm" for 32-bit x86 disclaims code stability
10489 anyway, in this constellation we activate additional code
10490 backported from 0.9.9-dev for further performance improvements,
10491 namely BN_from_montgomery_word. (To enable this otherwise,
10492 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10494 *Andy Polyakov (backport partially by Bodo Moeller)*
10496 * Add TLS session ticket callback. This allows an application to set
10497 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10498 values. This is useful for key rollover for example where several key
10499 sets may exist with different names.
10503 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10504 This was broken until now in 0.9.8 releases, such that the only way
10505 a registered ENGINE could be used (assuming it initialises
10506 successfully on the host) was to explicitly set it as the default
10507 for the relevant algorithms. This is in contradiction with 0.9.7
10508 behaviour and the documentation. With this fix, when an ENGINE is
10509 registered into a given algorithm's table of implementations, the
10510 'uptodate' flag is reset so that auto-discovery will be used next
10511 time a new context for that algorithm attempts to select an
10514 *Ian Lister (tweaked by Geoff Thorpe)*
10516 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10517 implementation in the following ways:
10519 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10522 Lack of BER streaming support means one pass streaming processing is
10523 only supported if data is detached: setting the streaming flag is
10524 ignored for embedded content.
10526 CMS support is disabled by default and must be explicitly enabled
10527 with the enable-cms configuration option.
10531 * Update the GMP engine glue to do direct copies between BIGNUM and
10532 mpz_t when openssl and GMP use the same limb size. Otherwise the
10533 existing "conversion via a text string export" trick is still used.
10535 *Paul Sheer <paulsheer@gmail.com>*
10537 * Zlib compression BIO. This is a filter BIO which compressed and
10538 uncompresses any data passed through it.
10542 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10543 RFC3394 compatible AES key wrapping.
10547 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10548 sets string data without copying. X509_ALGOR_set0() and
10549 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10550 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10551 from an X509_ATTRIBUTE structure optionally checking it occurs only
10552 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10557 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10558 to get the expected BN_FLG_CONSTTIME behavior.
10560 *Bodo Moeller (Google)*
10564 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10565 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10566 - added some more tests to do_tests.pl
10567 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10568 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10569 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10570 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10571 - various changes to netware.pl to enable gcc-cross builds on Win32
10573 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10574 - various changes to fix missing prototype warnings
10575 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10576 - added AES, WHIRLPOOL and CPUID assembler code to build files
10577 - added missing AES assembler make rules to mk1mf.pl
10578 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10580 *Guenter Knauf <eflash@gmx.net>*
10582 * Implement certificate status request TLS extension defined in RFC3546.
10583 A client can set the appropriate parameters and receive the encoded
10584 OCSP response via a callback. A server can query the supplied parameters
10585 and set the encoded OCSP response in the callback. Add simplified examples
10586 to s_client and s_server.
10590 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10592 * Fix various bugs:
10593 + Binary incompatibility of ssl_ctx_st structure
10594 + DTLS interoperation with non-compliant servers
10595 + Don't call get_session_cb() without proposed session
10596 + Fix ia64 assembler code
10598 *Andy Polyakov, Steve Henson*
10600 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10602 * DTLS Handshake overhaul. There were longstanding issues with
10603 OpenSSL DTLS implementation, which were making it impossible for
10604 RFC 4347 compliant client to communicate with OpenSSL server.
10605 Unfortunately just fixing these incompatibilities would "cut off"
10606 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10607 server keeps tolerating non RFC compliant syntax. The opposite is
10608 not true, 0.9.8f client can not communicate with earlier server.
10609 This update even addresses CVE-2007-4995.
10613 * Changes to avoid need for function casts in OpenSSL: some compilers
10614 (gcc 4.2 and later) reject their use.
10615 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10618 * Add RFC4507 support to OpenSSL. This includes the corrections in
10619 RFC4507bis. The encrypted ticket format is an encrypted encoded
10620 SSL_SESSION structure, that way new session features are automatically
10623 If a client application caches session in an SSL_SESSION structure
10624 support is transparent because tickets are now stored in the encoded
10627 The SSL_CTX structure automatically generates keys for ticket
10628 protection in servers so again support should be possible
10629 with no application modification.
10631 If a client or server wishes to disable RFC4507 support then the option
10632 SSL_OP_NO_TICKET can be set.
10634 Add a TLS extension debugging callback to allow the contents of any client
10635 or server extensions to be examined.
10637 This work was sponsored by Google.
10641 * Add initial support for TLS extensions, specifically for the server_name
10642 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10643 have new members for a hostname. The SSL data structure has an
10644 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10645 stored in that context to allow for session resumption, even after the
10646 SSL has been switched to a new SSL_CTX in reaction to a client's
10647 server_name extension.
10649 New functions (subject to change):
10651 SSL_get_servername()
10652 SSL_get_servername_type()
10655 New CTRL codes and macros (subject to change):
10657 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10658 - SSL_CTX_set_tlsext_servername_callback()
10659 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10660 - SSL_CTX_set_tlsext_servername_arg()
10661 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10663 openssl s_client has a new '-servername ...' option.
10665 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10666 '-key2 ...', '-servername_fatal' (subject to change). This allows
10667 testing the HostName extension for a specific single hostname ('-cert'
10668 and '-key' remain fallbacks for handshakes without HostName
10669 negotiation). If the unrecognized_name alert has to be sent, this by
10670 default is a warning; it becomes fatal with the '-servername_fatal'
10673 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10675 * Add AES and SSE2 assembly language support to VC++ build.
10679 * Mitigate attack on final subtraction in Montgomery reduction.
10683 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10684 (which previously caused an internal error).
10688 * Squeeze another 10% out of IGE mode when in != out.
10692 * AES IGE mode speedup.
10694 *Dean Gaudet (Google)*
10696 * Add the Korean symmetric 128-bit cipher SEED (see
10697 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10698 add SEED ciphersuites from RFC 4162:
10700 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10701 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10702 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10703 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10705 To minimize changes between patchlevels in the OpenSSL 0.9.8
10706 series, SEED remains excluded from compilation unless OpenSSL
10707 is configured with 'enable-seed'.
10709 *KISA, Bodo Moeller*
10711 * Mitigate branch prediction attacks, which can be practical if a
10712 single processor is shared, allowing a spy process to extract
10713 information. For detailed background information, see
10714 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10715 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10716 and Necessary Software Countermeasures"). The core of the change
10717 are new versions BN_div_no_branch() and
10718 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10719 respectively, which are slower, but avoid the security-relevant
10720 conditional branches. These are automatically called by BN_div()
10721 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10722 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10723 remove a conditional branch.
10725 BN_FLG_CONSTTIME is the new name for the previous
10726 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10727 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10728 in the exponent causes BN_mod_exp_mont() to use the alternative
10729 implementation in BN_mod_exp_mont_consttime().) The old name
10730 remains as a deprecated alias.
10732 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10733 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10734 constant-time implementations for more than just exponentiation.
10735 Here too the old name is kept as a deprecated alias.
10737 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10738 the BN_BLINDING structure gets an independent copy of the
10739 modulus. This means that the previous `BIGNUM *m` argument to
10740 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10741 essentially becomes `const BIGNUM *m`, although we can't actually
10742 change this in the header file before 0.9.9. It allows
10743 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10744 enable BN_FLG_CONSTTIME.
10746 *Matthew D Wood (Intel Corp)*
10748 * In the SSL/TLS server implementation, be strict about session ID
10749 context matching (which matters if an application uses a single
10750 external cache for different purposes). Previously,
10751 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10752 set. This did ensure strict client verification, but meant that,
10753 with applications using a single external cache for quite
10754 different requirements, clients could circumvent ciphersuite
10755 restrictions for a given session ID context by starting a session
10756 in a different context.
10760 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10761 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10762 authentication-only ciphersuites.
10766 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10767 not complete and could lead to a possible single byte overflow
10768 ([CVE-2007-5135]) [Ben Laurie]
10770 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10772 * Since AES128 and AES256 (and similarly Camellia128 and
10773 Camellia256) share a single mask bit in the logic of
10774 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10775 kludge to work properly if AES128 is available and AES256 isn't
10776 (or if Camellia128 is available and Camellia256 isn't).
10780 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10781 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10782 When a point or a seed is encoded in a BIT STRING, we need to
10783 prevent the removal of trailing zero bits to get the proper DER
10784 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10785 of a NamedBitList, for which trailing 0 bits need to be removed.)
10789 * Have SSL/TLS server implementation tolerate "mismatched" record
10790 protocol version while receiving ClientHello even if the
10791 ClientHello is fragmented. (The server can't insist on the
10792 particular protocol version it has chosen before the ServerHello
10793 message has informed the client about his choice.)
10797 * Add RFC 3779 support.
10799 *Rob Austein for ARIN, Ben Laurie*
10801 * Load error codes if they are not already present instead of using a
10802 static variable. This allows them to be cleanly unloaded and reloaded.
10803 Improve header file function name parsing.
10807 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10808 or CAPABILITY handshake as required by RFCs.
10810 *Goetz Babin-Ebell*
10812 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10814 * Introduce limits to prevent malicious keys being able to
10815 cause a denial of service. ([CVE-2006-2940])
10817 *Steve Henson, Bodo Moeller*
10819 * Fix ASN.1 parsing of certain invalid structures that can result
10820 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10822 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10823 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10825 * Fix SSL client code which could crash if connecting to a
10826 malicious SSLv2 server. ([CVE-2006-4343])
10828 *Tavis Ormandy and Will Drewry, Google Security Team*
10830 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10831 match only those. Before that, "AES256-SHA" would be interpreted
10832 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10833 the same strength classification in 0.9.7h) as we currently only
10834 have a single AES bit in the ciphersuite description bitmap.
10835 That change, however, also applied to ciphersuite strings such as
10836 "RC4-MD5" that intentionally matched multiple ciphersuites --
10837 namely, SSL 2.0 ciphersuites in addition to the more common ones
10838 from SSL 3.0/TLS 1.0.
10840 So we change the selection algorithm again: Naming an explicit
10841 ciphersuite selects this one ciphersuite, and any other similar
10842 ciphersuite (same bitmap) from *other* protocol versions.
10843 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10844 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10846 Since SSL 2.0 does not have any ciphersuites for which the
10847 128/256 bit distinction would be relevant, this works for now.
10848 The proper fix will be to use different bits for AES128 and
10849 AES256, which would have avoided the problems from the beginning;
10850 however, bits are scarce, so we can only do this in a new release
10851 (not just a patchlevel) when we can change the SSL_CIPHER
10852 definition to split the single 'unsigned long mask' bitmap into
10853 multiple values to extend the available space.
10857 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10859 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10860 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10862 * Add AES IGE and biIGE modes.
10866 * Change the Unix randomness entropy gathering to use poll() when
10867 possible instead of select(), since the latter has some
10868 undesirable limitations.
10870 *Darryl Miles via Richard Levitte and Bodo Moeller*
10872 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10873 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10874 cannot be implicitly activated as part of, e.g., the "AES" alias.
10875 However, please upgrade to OpenSSL 0.9.9[-dev] for
10876 non-experimental use of the ECC ciphersuites to get TLS extension
10877 support, which is required for curve and point format negotiation
10878 to avoid potential handshake problems.
10882 * Disable rogue ciphersuites:
10884 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10885 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10886 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10888 The latter two were purportedly from
10889 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10892 Also deactivate the remaining ciphersuites from
10893 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10894 unofficial, and the ID has long expired.
10898 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10899 dual-core machines) and other potential thread-safety issues.
10903 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10904 versions), which is now available for royalty-free use
10905 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10906 Also, add Camellia TLS ciphersuites from RFC 4132.
10908 To minimize changes between patchlevels in the OpenSSL 0.9.8
10909 series, Camellia remains excluded from compilation unless OpenSSL
10910 is configured with 'enable-camellia'.
10914 * Disable the padding bug check when compression is in use. The padding
10915 bug check assumes the first packet is of even length, this is not
10916 necessarily true if compression is enabled and can result in false
10917 positives causing handshake failure. The actual bug test is ancient
10918 code so it is hoped that implementations will either have fixed it by
10919 now or any which still have the bug do not support compression.
10923 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10925 * When applying a cipher rule check to see if string match is an explicit
10926 cipher suite and only match that one cipher suite if it is.
10930 * Link in manifests for VC++ if needed.
10932 *Austin Ziegler <halostatue@gmail.com>*
10934 * Update support for ECC-based TLS ciphersuites according to
10935 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10936 TLS extensions, which are supported starting with the 0.9.9
10937 branch, not in the OpenSSL 0.9.8 branch).
10941 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10942 opaque EVP_CIPHER_CTX handling.
10946 * Fixes and enhancements to zlib compression code. We now only use
10947 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10948 to conform with the standards mentioned here:
10949 <http://www.zlib.net/DLL_FAQ.txt>
10950 Static zlib linking now works on Windows and the new --with-zlib-include
10951 --with-zlib-lib options to Configure can be used to supply the location
10952 of the headers and library. Gracefully handle case where zlib library
10957 * Several fixes and enhancements to the OID generation code. The old code
10958 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10959 handle numbers larger than ULONG_MAX, truncated printing and had a
10960 non standard OBJ_obj2txt() behaviour.
10964 * Add support for building of engines under engine/ as shared libraries
10965 under VC++ build system.
10969 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10970 Hopefully, we will not see any false combination of paths any more.
10974 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10976 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10977 (part of SSL_OP_ALL). This option used to disable the
10978 countermeasure against man-in-the-middle protocol-version
10979 rollback in the SSL 2.0 server implementation, which is a bad
10980 idea. ([CVE-2005-2969])
10982 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10983 for Information Security, National Institute of Advanced Industrial
10984 Science and Technology [AIST], Japan)*
10986 * Add two function to clear and return the verify parameter flags.
10990 * Keep cipherlists sorted in the source instead of sorting them at
10991 runtime, thus removing the need for a lock.
10995 * Avoid some small subgroup attacks in Diffie-Hellman.
10997 *Nick Mathewson and Ben Laurie*
10999 * Add functions for well-known primes.
11003 * Extended Windows CE support.
11005 *Satoshi Nakamura and Andy Polyakov*
11007 * Initialize SSL_METHOD structures at compile time instead of during
11008 runtime, thus removing the need for a lock.
11012 * Make PKCS7_decrypt() work even if no certificate is supplied by
11013 attempting to decrypt each encrypted key in turn. Add support to
11018 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11020 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11023 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11027 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11028 key into the same file any more.
11032 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11036 * Add -utf8 command line and config file option to 'ca'.
11038 *Stefan <stf@udoma.org*
11040 * Removed the macro des_crypt(), as it seems to conflict with some
11041 libraries. Use DES_crypt().
11045 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11046 involves renaming the source and generated shared-libs for
11047 both. The engines will accept the corrected or legacy ids
11048 ('ncipher' and '4758_cca' respectively) when binding. NB,
11049 this only applies when building 'shared'.
11051 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11053 * Add attribute functions to EVP_PKEY structure. Modify
11054 PKCS12_create() to recognize a CSP name attribute and
11055 use it. Make -CSP option work again in pkcs12 utility.
11059 * Add new functionality to the bn blinding code:
11060 - automatic re-creation of the BN_BLINDING parameters after
11061 a fixed number of uses (currently 32)
11062 - add new function for parameter creation
11063 - introduce flags to control the update behaviour of the
11064 BN_BLINDING parameters
11065 - hide BN_BLINDING structure
11066 Add a second BN_BLINDING slot to the RSA structure to improve
11067 performance when a single RSA object is shared among several
11072 * Add support for DTLS.
11074 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11076 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11077 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11081 * Remove buggy and incomplete DH cert support from
11082 ssl/ssl_rsa.c and ssl/s3_both.c
11086 * Use SHA-1 instead of MD5 as the default digest algorithm for
11087 the `apps/openssl` commands.
11091 * Compile clean with "-Wall -Wmissing-prototypes
11092 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11093 DEBUG_SAFESTACK must also be set.
11097 * Change ./Configure so that certain algorithms can be disabled by default.
11098 The new counterpiece to "no-xxx" is "enable-xxx".
11100 The patented RC5 and MDC2 algorithms will now be disabled unless
11101 "enable-rc5" and "enable-mdc2", respectively, are specified.
11103 (IDEA remains enabled despite being patented. This is because IDEA
11104 is frequently required for interoperability, and there is no license
11105 fee for non-commercial use. As before, "no-idea" can be used to
11106 avoid this algorithm.)
11110 * Add processing of proxy certificates (see RFC 3820). This work was
11111 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11112 EGEE (Enabling Grids for E-science in Europe).
11116 * RC4 performance overhaul on modern architectures/implementations, such
11117 as Intel P4, IA-64 and AMD64.
11121 * New utility extract-section.pl. This can be used specify an alternative
11122 section number in a pod file instead of having to treat each file as
11123 a separate case in Makefile. This can be done by adding two lines to the
11126 =for comment openssl_section:XXX
11128 The blank line is mandatory.
11132 * New arguments -certform, -keyform and -pass for s_client and s_server
11133 to allow alternative format key and certificate files and passphrase
11138 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11139 update associated structures and add various utility functions.
11141 Add new policy related verify parameters, include policy checking in
11142 standard verify code. Enhance 'smime' application with extra parameters
11143 to support policy checking and print out.
11147 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11148 Nehemiah processors. These extensions support AES encryption in hardware
11149 as well as RNG (though RNG support is currently disabled).
11151 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11153 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11157 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11159 *Andy Polyakov and a number of other people*
11161 * Improved PowerPC platform support. Most notably BIGNUM assembler
11162 implementation contributed by IBM.
11164 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11166 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11167 exponent rather than 'unsigned long'. There is a corresponding change to
11168 the new 'rsa_keygen' element of the RSA_METHOD structure.
11170 *Jelte Jansen, Geoff Thorpe*
11172 * Functionality for creating the initial serial number file is now
11173 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11175 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11176 number file to 1, which is bound to cause problems. To avoid
11177 the problems while respecting compatibility between different 0.9.7
11178 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11179 CA.pl for serial number initialization. With the new release 0.9.8,
11180 we can fix the problem directly in the 'ca' utility.)
11184 * Reduced header interdependencies by declaring more opaque objects in
11185 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11186 give fewer recursive includes, which could break lazy source code - so
11187 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11188 developers should define this symbol when building and using openssl to
11189 ensure they track the recommended behaviour, interfaces, [etc], but
11190 backwards-compatible behaviour prevails when this isn't defined.
11194 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11198 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11199 This will generate a random key of the appropriate length based on the
11200 cipher context. The EVP_CIPHER can provide its own random key generation
11201 routine to support keys of a specific form. This is used in the des and
11202 3des routines to generate a key of the correct parity. Update S/MIME
11203 code to use new functions and hence generate correct parity DES keys.
11204 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11205 valid (weak or incorrect parity).
11209 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11210 as looking them up. This is useful when the verified structure may contain
11211 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11212 present unless the new PKCS7_NO_CRL flag is asserted.
11216 * Extend ASN1 oid configuration module. It now additionally accepts the
11219 shortName = some long name, 1.2.3.4
11223 * Reimplemented the BN_CTX implementation. There is now no more static
11224 limitation on the number of variables it can handle nor the depth of the
11225 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11226 information can now expand as required, and rather than having a single
11227 static array of bignums, BN_CTX now uses a linked-list of such arrays
11228 allowing it to expand on demand whilst maintaining the usefulness of
11229 BN_CTX's "bundling".
11233 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11234 to allow all RSA operations to function using a single BN_CTX.
11238 * Preliminary support for certificate policy evaluation and checking. This
11239 is initially intended to pass the tests outlined in "Conformance Testing
11240 of Relying Party Client Certificate Path Processing Logic" v1.07.
11244 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11245 remained unused and not that useful. A variety of other little bignum
11246 tweaks and fixes have also been made continuing on from the audit (see
11251 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11252 associated ASN1, EVP and SSL functions and old ASN1 macros.
11256 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11257 and this should never fail. So the return value from the use of
11258 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11259 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11263 * BN_CTX_get() should return zero-valued bignums, providing the same
11264 initialised value as BN_new().
11266 *Geoff Thorpe, suggested by Ulf Möller*
11268 * Support for inhibitAnyPolicy certificate extension.
11272 * An audit of the BIGNUM code is underway, for which debugging code is
11273 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11274 is considered valid when processing BIGNUMs, and causes execution to
11275 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11276 further steps are taken to deliberately pollute unused data in BIGNUM
11277 structures to try and expose faulty code further on. For now, openssl will
11278 (in its default mode of operation) continue to tolerate the inconsistent
11279 forms that it has tolerated in the past, but authors and packagers should
11280 consider trying openssl and their own applications when compiled with
11281 these debugging symbols defined. It will help highlight potential bugs in
11282 their own code, and will improve the test coverage for OpenSSL itself. At
11283 some point, these tighter rules will become openssl's default to improve
11284 maintainability, though the assert()s and other overheads will remain only
11285 in debugging configurations. See bn.h for more details.
11287 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11289 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11290 that can only be obtained through BN_CTX_new() (which implicitly
11291 initialises it). The presence of this function only made it possible
11292 to overwrite an existing structure (and cause memory leaks).
11296 * Because of the callback-based approach for implementing LHASH as a
11297 template type, lh_insert() adds opaque objects to hash-tables and
11298 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11299 to clean up those corresponding objects before destroying the hash table
11300 (and losing the object pointers). So some over-zealous constifications in
11301 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11302 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11303 prototyped to have "const" restrictions on the object pointers they are
11304 given (and so aren't required to cast them away any more).
11308 * The tmdiff.h API was so ugly and minimal that our own timing utility
11309 (speed) prefers to use its own implementation. The two implementations
11310 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11311 its object type properly exposed (MS_TM) instead of casting to/from
11312 `char *`. This may still change yet if someone realises MS_TM and
11314 aren't necessarily the greatest nomenclatures - but this is what was used
11315 internally to the implementation so I've used that for now.
11319 * Ensure that deprecated functions do not get compiled when
11320 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11321 the self-tests were still using deprecated key-generation functions so
11322 these have been updated also.
11326 * Reorganise PKCS#7 code to separate the digest location functionality
11327 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11328 New function PKCS7_set_digest() to set the digest type for PKCS#7
11329 digestedData type. Add additional code to correctly generate the
11330 digestedData type and add support for this type in PKCS7 initialization
11335 * New function PKCS7_set0_type_other() this initializes a PKCS7
11336 structure of type "other".
11340 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11341 sure the loop does correctly stop and breaking ("division by zero")
11342 modulus operations are not performed. The (pre-generated) prime
11343 table crypto/bn/bn_prime.h was already correct, but it could not be
11344 re-generated on some platforms because of the "division by zero"
11345 situation in the script.
11347 *Ralf S. Engelschall*
11349 * Update support for ECC-based TLS ciphersuites according to
11350 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11351 SHA-1 now is only used for "small" curves (where the
11352 representation of a field element takes up to 24 bytes); for
11353 larger curves, the field element resulting from ECDH is directly
11354 used as premaster secret.
11356 *Douglas Stebila (Sun Microsystems Laboratories)*
11358 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11359 curve secp160r1 to the tests.
11361 *Douglas Stebila (Sun Microsystems Laboratories)*
11363 * Add the possibility to load symbols globally with DSO.
11365 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11367 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11368 control of the error stack.
11372 * Add support for STORE in ENGINE.
11376 * Add the STORE type. The intention is to provide a common interface
11377 to certificate and key stores, be they simple file-based stores, or
11378 HSM-type store, or LDAP stores, or...
11379 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11383 * Add a generic structure called OPENSSL_ITEM. This can be used to
11384 pass a list of arguments to any function as well as provide a way
11385 for a function to pass data back to the caller.
11389 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11390 works like BUF_strdup() but can be used to duplicate a portion of
11391 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11396 * Add the function sk_find_ex() which works like sk_find(), but will
11397 return an index to an element even if an exact match couldn't be
11398 found. The index is guaranteed to point at the element where the
11399 searched-for key would be inserted to preserve sorting order.
11403 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11404 takes an extra flags argument for optional functionality. Currently,
11405 the following flags are defined:
11407 OBJ_BSEARCH_VALUE_ON_NOMATCH
11408 This one gets OBJ_bsearch_ex() to return a pointer to the first
11409 element where the comparing function returns a negative or zero
11412 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11413 This one gets OBJ_bsearch_ex() to return a pointer to the first
11414 element where the comparing function returns zero. This is useful
11415 if there are more than one element where the comparing function
11420 * Make it possible to create self-signed certificates with 'openssl ca'
11421 in such a way that the self-signed certificate becomes part of the
11422 CA database and uses the same mechanisms for serial number generation
11423 as all other certificate signing. The new flag '-selfsign' enables
11424 this functionality. Adapt CA.sh and CA.pl.in.
11428 * Add functionality to check the public key of a certificate request
11429 against a given private. This is useful to check that a certificate
11430 request can be signed by that key (self-signing).
11434 * Make it possible to have multiple active certificates with the same
11435 subject in the CA index file. This is done only if the keyword
11436 'unique_subject' is set to 'no' in the main CA section (default
11437 if 'CA_default') of the configuration file. The value is saved
11438 with the database itself in a separate index attribute file,
11439 named like the index file with '.attr' appended to the name.
11443 * Generate multi-valued AVAs using '+' notation in config files for
11448 * Support for nameConstraints certificate extension.
11452 * Support for policyConstraints certificate extension.
11456 * Support for policyMappings certificate extension.
11460 * Make sure the default DSA_METHOD implementation only uses its
11461 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11462 and change its own handlers to be NULL so as to remove unnecessary
11463 indirection. This lets alternative implementations fallback to the
11464 default implementation more easily.
11468 * Support for directoryName in GeneralName related extensions
11473 * Make it possible to link applications using Makefile.shared.
11474 Make that possible even when linking against static libraries!
11478 * Support for single pass processing for S/MIME signing. This now
11479 means that S/MIME signing can be done from a pipe, in addition
11480 cleartext signing (multipart/signed type) is effectively streaming
11481 and the signed data does not need to be all held in memory.
11483 This is done with a new flag PKCS7_STREAM. When this flag is set
11484 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11485 is done after the data is output (and digests calculated) in
11486 SMIME_write_PKCS7().
11490 * Add full support for -rpath/-R, both in shared libraries and
11491 applications, at least on the platforms where it's known how
11496 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11497 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11498 will now compute a table of multiples of the generator that
11499 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11500 faster (notably in the case of a single point multiplication,
11501 scalar * generator).
11503 *Nils Larsch, Bodo Moeller*
11505 * IPv6 support for certificate extensions. The various extensions
11506 which use the IP:a.b.c.d can now take IPv6 addresses using the
11507 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11512 * Added an ENGINE that implements RSA by performing private key
11513 exponentiations with the GMP library. The conversions to and from
11514 GMP's mpz_t format aren't optimised nor are any montgomery forms
11515 cached, and on x86 it appears OpenSSL's own performance has caught up.
11516 However there are likely to be other architectures where GMP could
11517 provide a boost. This ENGINE is not built in by default, but it can be
11518 specified at Configure time and should be accompanied by the necessary
11519 linker additions, eg;
11520 ./config -DOPENSSL_USE_GMP -lgmp
11524 * "openssl engine" will not display ENGINE/DSO load failure errors when
11525 testing availability of engines with "-t" - the old behaviour is
11526 produced by increasing the feature's verbosity with "-tt".
11530 * ECDSA routines: under certain error conditions uninitialized BN objects
11531 could be freed. Solution: make sure initialization is performed early
11532 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11537 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11538 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11539 software implementations. For DSA and DH, parameter generation can
11540 also be overridden by providing the appropriate method callbacks.
11544 * Change the "progress" mechanism used in key-generation and
11545 primality testing to functions that take a new BN_GENCB pointer in
11546 place of callback/argument pairs. The new API functions have `_ex`
11547 postfixes and the older functions are reimplemented as wrappers for
11548 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11549 declarations of the old functions to help (graceful) attempts to
11550 migrate to the new functions. Also, the new key-generation API
11551 functions operate on a caller-supplied key-structure and return
11552 success/failure rather than returning a key or NULL - this is to
11553 help make "keygen" another member function of RSA_METHOD etc.
11555 Example for using the new callback interface:
11557 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11558 void *my_arg = ...;
11561 BN_GENCB_set(&my_cb, my_callback, my_arg);
11563 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11564 /* For the meaning of a, b in calls to my_callback(), see the
11565 * documentation of the function that calls the callback.
11566 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11567 * my_callback should return 1 if it wants BN_is_prime_ex()
11568 * to continue, or 0 to stop.
11573 * Change the ZLIB compression method to be stateful, and make it
11574 available to TLS with the number defined in
11575 draft-ietf-tls-compression-04.txt.
11579 * Add the ASN.1 structures and functions for CertificatePair, which
11580 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11582 CertificatePair ::= SEQUENCE {
11583 forward [0] Certificate OPTIONAL,
11584 reverse [1] Certificate OPTIONAL,
11585 -- at least one of the pair shall be present -- }
11587 Also implement the PEM functions to read and write certificate
11588 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11590 This needed to be defined, mostly for the sake of the LDAP
11591 attribute crossCertificatePair, but may prove useful elsewhere as
11596 * Make it possible to inhibit symlinking of shared libraries in
11597 Makefile.shared, for Cygwin's sake.
11601 * Extend the BIGNUM API by creating a function
11602 void BN_set_negative(BIGNUM *a, int neg);
11603 and a macro that behave like
11604 int BN_is_negative(const BIGNUM *a);
11606 to avoid the need to access 'a->neg' directly in applications.
11610 * Implement fast modular reduction for pseudo-Mersenne primes
11611 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11612 EC_GROUP_new_curve_GFp() will now automatically use this
11615 *Nils Larsch <nla@trustcenter.de>*
11617 * Add new lock type (CRYPTO_LOCK_BN).
11621 * Change the ENGINE framework to automatically load engines
11622 dynamically from specific directories unless they could be
11623 found to already be built in or loaded. Move all the
11624 current engines except for the cryptodev one to a new
11625 directory engines/.
11626 The engines in engines/ are built as shared libraries if
11627 the "shared" options was given to ./Configure or ./config.
11628 Otherwise, they are inserted in libcrypto.a.
11629 /usr/local/ssl/engines is the default directory for dynamic
11630 engines, but that can be overridden at configure time through
11631 the usual use of --prefix and/or --openssldir, and at run
11632 time with the environment variable OPENSSL_ENGINES.
11634 *Geoff Thorpe and Richard Levitte*
11636 * Add Makefile.shared, a helper makefile to build shared
11637 libraries. Adapt Makefile.org.
11641 * Add version info to Win32 DLLs.
11643 *Peter 'Luna' Runestig" <peter@runestig.com>*
11645 * Add new 'medium level' PKCS#12 API. Certificates and keys
11646 can be added using this API to created arbitrary PKCS#12
11647 files while avoiding the low-level API.
11649 New options to PKCS12_create(), key or cert can be NULL and
11650 will then be omitted from the output file. The encryption
11651 algorithm NIDs can be set to -1 for no encryption, the mac
11652 iteration count can be set to 0 to omit the mac.
11654 Enhance pkcs12 utility by making the -nokeys and -nocerts
11655 options work when creating a PKCS#12 file. New option -nomac
11656 to omit the mac, NONE can be set for an encryption algorithm.
11657 New code is modified to use the enhanced PKCS12_create()
11658 instead of the low-level API.
11662 * Extend ASN1 encoder to support indefinite length constructed
11663 encoding. This can output sequences tags and octet strings in
11664 this form. Modify pk7_asn1.c to support indefinite length
11665 encoding. This is experimental and needs additional code to
11666 be useful, such as an ASN1 bio and some enhanced streaming
11669 Extend template encode functionality so that tagging is passed
11670 down to the template encoder.
11674 * Let 'openssl req' fail if an argument to '-newkey' is not
11675 recognized instead of using RSA as a default.
11679 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11680 As these are not official, they are not included in "ALL";
11681 the "ECCdraft" ciphersuite group alias can be used to select them.
11683 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11685 * Add ECDH engine support.
11687 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11689 * Add ECDH in new directory crypto/ecdh/.
11691 *Douglas Stebila (Sun Microsystems Laboratories)*
11693 * Let BN_rand_range() abort with an error after 100 iterations
11694 without success (which indicates a broken PRNG).
11698 * Change BN_mod_sqrt() so that it verifies that the input value
11699 is really the square of the return value. (Previously,
11700 BN_mod_sqrt would show GIGO behaviour.)
11704 * Add named elliptic curves over binary fields from X9.62, SECG,
11705 and WAP/WTLS; add OIDs that were still missing.
11707 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11709 * Extend the EC library for elliptic curves over binary fields
11710 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11713 EC_GF2m_simple_method
11717 EC_GROUP_new_curve_GF2m
11718 EC_GROUP_set_curve_GF2m
11719 EC_GROUP_get_curve_GF2m
11720 EC_POINT_set_affine_coordinates_GF2m
11721 EC_POINT_get_affine_coordinates_GF2m
11722 EC_POINT_set_compressed_coordinates_GF2m
11724 Point compression for binary fields is disabled by default for
11725 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11728 As binary polynomials are represented as BIGNUMs, various members
11729 of the EC_GROUP and EC_POINT data structures can be shared
11730 between the implementations for prime fields and binary fields;
11731 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11732 are essentially identical to their `..._GFp` counterparts.
11733 (For simplicity, the `..._GFp` prefix has been dropped from
11734 various internal method names.)
11736 An internal 'field_div' method (similar to 'field_mul' and
11737 'field_sqr') has been added; this is used only for binary fields.
11739 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11741 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11742 through methods ('mul', 'precompute_mult').
11744 The generic implementations (now internally called 'ec_wNAF_mul'
11745 and 'ec_wNAF_precomputed_mult') remain the default if these
11746 methods are undefined.
11748 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11750 * New function EC_GROUP_get_degree, which is defined through
11751 EC_METHOD. For curves over prime fields, this returns the bit
11752 length of the modulus.
11754 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11756 * New functions EC_GROUP_dup, EC_POINT_dup.
11757 (These simply call ..._new and ..._copy).
11759 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11761 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11762 Polynomials are represented as BIGNUMs (where the sign bit is not
11763 used) in the following functions [macros]:
11766 BN_GF2m_sub [= BN_GF2m_add]
11767 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11768 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11769 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11771 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11772 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11773 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11774 BN_GF2m_cmp [= BN_ucmp]
11776 (Note that only the 'mod' functions are actually for fields GF(2^m).
11777 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11779 For some functions, an the irreducible polynomial defining a
11780 field can be given as an 'unsigned int[]' with strictly
11781 decreasing elements giving the indices of those bits that are set;
11782 i.e., p[] represents the polynomial
11783 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11785 p[0] > p[1] > ... > p[k] = 0.
11786 This applies to the following functions:
11789 BN_GF2m_mod_mul_arr
11790 BN_GF2m_mod_sqr_arr
11791 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11792 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11793 BN_GF2m_mod_exp_arr
11794 BN_GF2m_mod_sqrt_arr
11795 BN_GF2m_mod_solve_quad_arr
11799 Conversion can be performed by the following functions:
11804 bntest.c has additional tests for binary polynomial arithmetic.
11806 Two implementations for BN_GF2m_mod_div() are available.
11807 The default algorithm simply uses BN_GF2m_mod_inv() and
11808 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11809 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11810 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11812 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11814 * Add new error code 'ERR_R_DISABLED' that can be used when some
11815 functionality is disabled at compile-time.
11817 *Douglas Stebila <douglas.stebila@sun.com>*
11819 * Change default behaviour of 'openssl asn1parse' so that more
11820 information is visible when viewing, e.g., a certificate:
11822 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11823 mode the content of non-printable OCTET STRINGs is output in a
11824 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11825 avoid the appearance of a printable string.
11827 *Nils Larsch <nla@trustcenter.de>*
11829 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11831 EC_GROUP_set_asn1_flag()
11832 EC_GROUP_get_asn1_flag()
11833 EC_GROUP_set_point_conversion_form()
11834 EC_GROUP_get_point_conversion_form()
11835 These control ASN1 encoding details:
11836 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11837 has been set to OPENSSL_EC_NAMED_CURVE.
11838 - Points are encoded in uncompressed form by default; options for
11839 asn1_for are as for point2oct, namely
11840 POINT_CONVERSION_COMPRESSED
11841 POINT_CONVERSION_UNCOMPRESSED
11842 POINT_CONVERSION_HYBRID
11844 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11846 EC_GROUP_set_seed()
11847 EC_GROUP_get0_seed()
11848 EC_GROUP_get_seed_len()
11849 This is used only for ASN1 purposes (so far).
11851 *Nils Larsch <nla@trustcenter.de>*
11853 * Add 'field_type' member to EC_METHOD, which holds the NID
11854 of the appropriate field type OID. The new function
11855 EC_METHOD_get_field_type() returns this value.
11857 *Nils Larsch <nla@trustcenter.de>*
11860 EC_POINT_point2bn()
11861 EC_POINT_bn2point()
11862 EC_POINT_point2hex()
11863 EC_POINT_hex2point()
11864 providing useful interfaces to EC_POINT_point2oct() and
11865 EC_POINT_oct2point().
11867 *Nils Larsch <nla@trustcenter.de>*
11869 * Change internals of the EC library so that the functions
11870 EC_GROUP_set_generator()
11871 EC_GROUP_get_generator()
11872 EC_GROUP_get_order()
11873 EC_GROUP_get_cofactor()
11874 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11875 to methods, which would lead to unnecessary code duplication when
11876 adding different types of curves.
11878 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11880 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11881 arithmetic, and such that modified wNAFs are generated
11882 (which avoid length expansion in many cases).
11886 * Add a function EC_GROUP_check_discriminant() (defined via
11887 EC_METHOD) that verifies that the curve discriminant is non-zero.
11889 Add a function EC_GROUP_check() that makes some sanity tests
11890 on a EC_GROUP, its generator and order. This includes
11891 EC_GROUP_check_discriminant().
11893 *Nils Larsch <nla@trustcenter.de>*
11895 * Add ECDSA in new directory crypto/ecdsa/.
11897 Add applications 'openssl ecparam' and 'openssl ecdsa'
11898 (these are based on 'openssl dsaparam' and 'openssl dsa').
11900 ECDSA support is also included in various other files across the
11901 library. Most notably,
11902 - 'openssl req' now has a '-newkey ecdsa:file' option;
11903 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11904 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11905 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11906 them suitable for ECDSA where domain parameters must be
11907 extracted before the specific public key;
11908 - ECDSA engine support has been added.
11910 *Nils Larsch <nla@trustcenter.de>*
11912 * Include some named elliptic curves, and add OIDs from X9.62,
11913 SECG, and WAP/WTLS. Each curve can be obtained from the new
11915 EC_GROUP_new_by_curve_name(),
11916 and the list of available named curves can be obtained with
11917 EC_get_builtin_curves().
11918 Also add a 'curve_name' member to EC_GROUP objects, which can be
11920 EC_GROUP_set_curve_name()
11921 EC_GROUP_get_curve_name()
11923 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11925 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11926 was actually never needed) and in BN_mul(). The removal in BN_mul()
11927 required a small change in bn_mul_part_recursive() and the addition
11928 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11929 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11930 bn_sub_words() and bn_add_words() except they take arrays with
11935 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11937 * Cleanse PEM buffers before freeing them since they may contain
11940 *Benjamin Bennett <ben@psc.edu>*
11942 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11943 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11944 authentication-only ciphersuites.
11948 * Since AES128 and AES256 share a single mask bit in the logic of
11949 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11950 kludge to work properly if AES128 is available and AES256 isn't.
11954 * Expand security boundary to match 1.1.1 module.
11958 * Remove redundant features: hash file source, editing of test vectors
11959 modify fipsld to use external fips_premain.c signature.
11963 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11964 run algorithm test programs.
11968 * Make algorithm test programs more tolerant of whitespace.
11972 * Have SSL/TLS server implementation tolerate "mismatched" record
11973 protocol version while receiving ClientHello even if the
11974 ClientHello is fragmented. (The server can't insist on the
11975 particular protocol version it has chosen before the ServerHello
11976 message has informed the client about his choice.)
11980 * Load error codes if they are not already present instead of using a
11981 static variable. This allows them to be cleanly unloaded and reloaded.
11985 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11987 * Introduce limits to prevent malicious keys being able to
11988 cause a denial of service. ([CVE-2006-2940])
11990 *Steve Henson, Bodo Moeller*
11992 * Fix ASN.1 parsing of certain invalid structures that can result
11993 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11995 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11996 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11998 * Fix SSL client code which could crash if connecting to a
11999 malicious SSLv2 server. ([CVE-2006-4343])
12001 *Tavis Ormandy and Will Drewry, Google Security Team*
12003 * Change ciphersuite string processing so that an explicit
12004 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12005 will no longer include "AES128-SHA"), and any other similar
12006 ciphersuite (same bitmap) from *other* protocol versions (so that
12007 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12008 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12009 changes from 0.9.8b and 0.9.8d.
12013 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12015 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12016 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12018 * Change the Unix randomness entropy gathering to use poll() when
12019 possible instead of select(), since the latter has some
12020 undesirable limitations.
12022 *Darryl Miles via Richard Levitte and Bodo Moeller*
12024 * Disable rogue ciphersuites:
12026 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12027 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12028 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12030 The latter two were purportedly from
12031 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12034 Also deactivate the remaining ciphersuites from
12035 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12036 unofficial, and the ID has long expired.
12040 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12041 dual-core machines) and other potential thread-safety issues.
12045 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12047 * Adapt fipsld and the build system to link against the validated FIPS
12048 module in FIPS mode.
12052 * Fixes for VC++ 2005 build under Windows.
12056 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12057 from a Windows bash shell such as MSYS. It is autodetected from the
12058 "config" script when run from a VC++ environment. Modify standard VC++
12059 build to use fipscanister.o from the GNU make build.
12063 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12065 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12066 The value now differs depending on if you build for FIPS or not.
12067 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12068 safely run with a non-FIPSed libcrypto, as it may crash because of
12069 the difference induced by this change.
12073 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12075 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12076 (part of SSL_OP_ALL). This option used to disable the
12077 countermeasure against man-in-the-middle protocol-version
12078 rollback in the SSL 2.0 server implementation, which is a bad
12079 idea. ([CVE-2005-2969])
12081 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12082 for Information Security, National Institute of Advanced Industrial
12083 Science and Technology [AIST, Japan)]*
12085 * Minimal support for X9.31 signatures and PSS padding modes. This is
12086 mainly for FIPS compliance and not fully integrated at this stage.
12090 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12091 the exponentiation using a fixed-length exponent. (Otherwise,
12092 the information leaked through timing could expose the secret key
12093 after many signatures; cf. Bleichenbacher's attack on DSA with
12098 * Make a new fixed-window mod_exp implementation the default for
12099 RSA, DSA, and DH private-key operations so that the sequence of
12100 squares and multiplies and the memory access pattern are
12101 independent of the particular secret key. This will mitigate
12102 cache-timing and potential related attacks.
12104 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12105 and this is automatically used by BN_mod_exp_mont() if the new flag
12106 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12107 will use this BN flag for private exponents unless the flag
12108 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12109 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12111 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12113 * Change the client implementation for SSLv23_method() and
12114 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12115 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12116 (Previously, the SSL 2.0 backwards compatible Client Hello
12117 message format would be used even with SSL_OP_NO_SSLv2.)
12121 * Add support for smime-type MIME parameter in S/MIME messages which some
12126 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12127 a threadsafe manner. Modify rsa code to use new function and add calls
12128 to dsa and dh code (which had race conditions before).
12132 * Include the fixed error library code in the C error file definitions
12133 instead of fixing them up at runtime. This keeps the error code
12134 structures constant.
12138 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12140 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12143 * Fixes for newer kerberos headers. NB: the casts are needed because
12144 the 'length' field is signed on one version and unsigned on another
12145 with no (?) obvious way to tell the difference, without these VC++
12146 complains. Also the "definition" of FAR (blank) is no longer included
12147 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12148 some needed definitions.
12152 * Undo Cygwin change.
12156 * Added support for proxy certificates according to RFC 3820.
12157 Because they may be a security thread to unaware applications,
12158 they must be explicitly allowed in run-time. See
12159 docs/HOWTO/proxy_certificates.txt for further information.
12163 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12165 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12166 server and client random values. Previously
12167 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12168 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12170 This change has negligible security impact because:
12172 1. Server and client random values still have 24 bytes of pseudo random
12175 2. Server and client random values are sent in the clear in the initial
12178 3. The master secret is derived using the premaster secret (48 bytes in
12179 size for static RSA ciphersuites) as well as client server and random
12182 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12185 *Stephen Henson, reported by UK NISCC*
12187 * Use Windows randomness collection on Cygwin.
12191 * Fix hang in EGD/PRNGD query when communication socket is closed
12192 prematurely by EGD/PRNGD.
12194 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12196 * Prompt for pass phrases when appropriate for PKCS12 input format.
12200 * Back-port of selected performance improvements from development
12201 branch, as well as improved support for PowerPC platforms.
12205 * Add lots of checks for memory allocation failure, error codes to indicate
12206 failure and freeing up memory if a failure occurs.
12208 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12210 * Add new -passin argument to dgst.
12214 * Perform some character comparisons of different types in X509_NAME_cmp:
12215 this is needed for some certificates that re-encode DNs into UTF8Strings
12216 (in violation of RFC3280) and can't or won't issue name rollover
12221 * Make an explicit check during certificate validation to see that
12222 the CA setting in each certificate on the chain is correct. As a
12223 side effect always do the following basic checks on extensions,
12224 not just when there's an associated purpose to the check:
12226 - if there is an unhandled critical extension (unless the user
12227 has chosen to ignore this fault)
12228 - if the path length has been exceeded (if one is set at all)
12229 - that certain extensions fit the associated purpose (if one has
12234 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12236 * Avoid a race condition when CRLs are checked in a multi threaded
12237 environment. This would happen due to the reordering of the revoked
12238 entries during signature checking and serial number lookup. Now the
12239 encoding is cached and the serial number sort performed under a lock.
12240 Add new STACK function sk_is_sorted().
12244 * Add Delta CRL to the extension code.
12248 * Various fixes to s3_pkt.c so alerts are sent properly.
12250 *David Holmes <d.holmes@f5.com>*
12252 * Reduce the chances of duplicate issuer name and serial numbers (in
12253 violation of RFC3280) using the OpenSSL certificate creation utilities.
12254 This is done by creating a random 64 bit value for the initial serial
12255 number when a serial number file is created or when a self signed
12256 certificate is created using 'openssl req -x509'. The initial serial
12257 number file is created using 'openssl x509 -next_serial' in CA.pl
12258 rather than being initialized to 1.
12262 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12264 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12265 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12267 *Joe Orton, Steve Henson*
12269 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12272 *Joe Orton, Steve Henson*
12274 * Make it possible to have multiple active certificates with the same
12275 subject in the CA index file. This is done only if the keyword
12276 'unique_subject' is set to 'no' in the main CA section (default
12277 if 'CA_default') of the configuration file. The value is saved
12278 with the database itself in a separate index attribute file,
12279 named like the index file with '.attr' appended to the name.
12283 * X509 verify fixes. Disable broken certificate workarounds when
12284 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12285 keyUsage extension present. Don't accept CRLs with unhandled critical
12286 extensions: since verify currently doesn't process CRL extensions this
12287 rejects a CRL with *any* critical extensions. Add new verify error codes
12292 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12293 A clarification of RFC2560 will require the use of OCTET STRINGs and
12294 some implementations cannot handle the current raw format. Since OpenSSL
12295 copies and compares OCSP nonces as opaque blobs without any attempt at
12296 parsing them this should not create any compatibility issues.
12300 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12301 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12302 this HMAC (and other) operations are several times slower than OpenSSL
12307 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12309 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12311 * Use the correct content when signing type "other".
12315 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12317 * Fix various bugs revealed by running the NISCC test suite:
12319 Stop out of bounds reads in the ASN1 code when presented with
12320 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12322 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12324 If verify callback ignores invalid public key errors don't try to check
12325 certificate signature with the NULL public key.
12329 * New -ignore_err option in ocsp application to stop the server
12330 exiting on the first error in a request.
12334 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12335 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12340 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12341 extra data after the compression methods not only for TLS 1.0
12342 but also for SSL 3.0 (as required by the specification).
12344 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12346 * Change X509_certificate_type() to mark the key as exported/exportable
12347 when it's 512 *bits* long, not 512 bytes.
12351 * Change AES_cbc_encrypt() so it outputs exact multiple of
12352 blocks during encryption.
12356 * Various fixes to base64 BIO and non blocking I/O. On write
12357 flushes were not handled properly if the BIO retried. On read
12358 data was not being buffered properly and had various logic bugs.
12359 This also affects blocking I/O when the data being decoded is a
12364 * Various S/MIME bugfixes and compatibility changes:
12365 output correct application/pkcs7 MIME type if
12366 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12367 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12368 of files as .eml work). Correctly handle very long lines in MIME
12373 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12375 * Countermeasure against the Klima-Pokorny-Rosa extension of
12376 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12377 a protocol version number mismatch like a decryption error
12378 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12382 * Turn on RSA blinding by default in the default implementation
12383 to avoid a timing attack. Applications that don't want it can call
12384 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12385 They would be ill-advised to do so in most cases.
12387 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12389 * Change RSA blinding code so that it works when the PRNG is not
12390 seeded (in this case, the secret RSA exponent is abused as
12391 an unpredictable seed -- if it is not unpredictable, there
12392 is no point in blinding anyway). Make RSA blinding thread-safe
12393 by remembering the creator's thread ID in rsa->blinding and
12394 having all other threads use local one-time blinding factors
12395 (this requires more computation than sharing rsa->blinding, but
12396 avoids excessive locking; and if an RSA object is not shared
12397 between threads, blinding will still be very fast).
12401 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12402 ENGINE as defaults for all supported algorithms irrespective of
12403 the 'flags' parameter. 'flags' is now honoured, so applications
12404 should make sure they are passing it correctly.
12408 * Target "mingw" now allows native Windows code to be generated in
12409 the Cygwin environment as well as with the MinGW compiler.
12413 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12415 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12416 via timing by performing a MAC computation even if incorrect
12417 block cipher padding has been found. This is a countermeasure
12418 against active attacks where the attacker has to distinguish
12419 between bad padding and a MAC verification error. ([CVE-2003-0078])
12421 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12422 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12423 Martin Vuagnoux (EPFL, Ilion)*
12425 * Make the no-err option work as intended. The intention with no-err
12426 is not to have the whole error stack handling routines removed from
12427 libcrypto, it's only intended to remove all the function name and
12428 reason texts, thereby removing some of the footprint that may not
12429 be interesting if those errors aren't displayed anyway.
12431 NOTE: it's still possible for any application or module to have its
12432 own set of error texts inserted. The routines are there, just not
12433 used by default when no-err is given.
12437 * Add support for FreeBSD on IA64.
12439 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12441 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12442 Kerberos function mit_des_cbc_cksum(). Before this change,
12443 the value returned by DES_cbc_cksum() was like the one from
12444 mit_des_cbc_cksum(), except the bytes were swapped.
12446 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12448 * Allow an application to disable the automatic SSL chain building.
12449 Before this a rather primitive chain build was always performed in
12450 ssl3_output_cert_chain(): an application had no way to send the
12451 correct chain if the automatic operation produced an incorrect result.
12453 Now the chain builder is disabled if either:
12455 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12457 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12459 The reasoning behind this is that an application would not want the
12460 auto chain building to take place if extra chain certificates are
12461 present and it might also want a means of sending no additional
12462 certificates (for example the chain has two certificates and the
12467 * Add the possibility to build without the ENGINE framework.
12469 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12471 * Under Win32 gmtime() can return NULL: check return value in
12472 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12476 * DSA routines: under certain error conditions uninitialized BN objects
12477 could be freed. Solution: make sure initialization is performed early
12478 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12479 Nils Larsch <nla@trustcenter.de> via PR#459)
12483 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12484 checked on reconnect on the client side, therefore session resumption
12485 could still fail with a "ssl session id is different" error. This
12486 behaviour is masked when SSL_OP_ALL is used due to
12487 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12488 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12489 followup to PR #377.
12493 * IA-32 assembler support enhancements: unified ELF targets, support
12494 for SCO/Caldera platforms, fix for Cygwin shared build.
12498 * Add support for FreeBSD on sparc64. As a consequence, support for
12499 FreeBSD on non-x86 processors is separate from x86 processors on
12500 the config script, much like the NetBSD support.
12502 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12504 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12506 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12509 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12510 code (06) was taken as the first octet of the session ID and the last
12511 octet was ignored consequently. As a result SSLv2 client side session
12512 caching could not have worked due to the session ID mismatch between
12514 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12519 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12520 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12525 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12526 seems that in spite of existing for more than a year, many application
12527 author have done nothing to provide the necessary callbacks, which
12528 means that this particular engine will not work properly anywhere.
12529 This is a very unfortunate situation which forces us, in the name
12530 of usability, to give the hw_ncipher.c a static lock, which is part
12532 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12533 appear in 0.9.8 or later. We EXPECT application authors to have
12534 dealt properly with this when 0.9.8 is released (unless we actually
12535 make such changes in the libcrypto locking code that changes will
12536 have to be made anyway).
12540 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12541 octets have been read, EOF or an error occurs. Without this change
12542 some truncated ASN1 structures will not produce an error.
12546 * Disable Heimdal support, since it hasn't been fully implemented.
12547 Still give the possibility to force the use of Heimdal, but with
12548 warnings and a request that patches get sent to openssl-dev.
12552 * Add the VC-CE target, introduce the WINCE sysname, and add
12553 INSTALL.WCE and appropriate conditionals to make it build.
12555 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12557 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12558 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12559 edit numbers of the version.
12561 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12563 * Introduce safe string copy and catenation functions
12564 (BUF_strlcpy() and BUF_strlcat()).
12566 *Ben Laurie (CHATS) and Richard Levitte*
12568 * Avoid using fixed-size buffers for one-line DNs.
12570 *Ben Laurie (CHATS)*
12572 * Add BUF_MEM_grow_clean() to avoid information leakage when
12573 resizing buffers containing secrets, and use where appropriate.
12575 *Ben Laurie (CHATS)*
12577 * Avoid using fixed size buffers for configuration file location.
12579 *Ben Laurie (CHATS)*
12581 * Avoid filename truncation for various CA files.
12583 *Ben Laurie (CHATS)*
12585 * Use sizeof in preference to magic numbers.
12587 *Ben Laurie (CHATS)*
12589 * Avoid filename truncation in cert requests.
12591 *Ben Laurie (CHATS)*
12593 * Add assertions to check for (supposedly impossible) buffer
12596 *Ben Laurie (CHATS)*
12598 * Don't cache truncated DNS entries in the local cache (this could
12599 potentially lead to a spoofing attack).
12601 *Ben Laurie (CHATS)*
12603 * Fix various buffers to be large enough for hex/decimal
12604 representations in a platform independent manner.
12606 *Ben Laurie (CHATS)*
12608 * Add CRYPTO_realloc_clean() to avoid information leakage when
12609 resizing buffers containing secrets, and use where appropriate.
12611 *Ben Laurie (CHATS)*
12613 * Add BIO_indent() to avoid much slightly worrying code to do
12616 *Ben Laurie (CHATS)*
12618 * Convert sprintf()/BIO_puts() to BIO_printf().
12620 *Ben Laurie (CHATS)*
12622 * buffer_gets() could terminate with the buffer only half
12625 *Ben Laurie (CHATS)*
12627 * Add assertions to prevent user-supplied crypto functions from
12628 overflowing internal buffers by having large block sizes, etc.
12630 *Ben Laurie (CHATS)*
12632 * New OPENSSL_assert() macro (similar to assert(), but enabled
12635 *Ben Laurie (CHATS)*
12637 * Eliminate unused copy of key in RC4.
12639 *Ben Laurie (CHATS)*
12641 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12643 *Ben Laurie (CHATS)*
12645 * Fix off-by-one error in EGD path.
12647 *Ben Laurie (CHATS)*
12649 * If RANDFILE path is too long, ignore instead of truncating.
12651 *Ben Laurie (CHATS)*
12653 * Eliminate unused and incorrectly sized X.509 structure
12656 *Ben Laurie (CHATS)*
12658 * Eliminate unused and dangerous function knumber().
12660 *Ben Laurie (CHATS)*
12662 * Eliminate unused and dangerous structure, KSSL_ERR.
12664 *Ben Laurie (CHATS)*
12666 * Protect against overlong session ID context length in an encoded
12667 session object. Since these are local, this does not appear to be
12670 *Ben Laurie (CHATS)*
12672 * Change from security patch (see 0.9.6e below) that did not affect
12673 the 0.9.6 release series:
12675 Remote buffer overflow in SSL3 protocol - an attacker could
12676 supply an oversized master key in Kerberos-enabled versions.
12679 *Ben Laurie (CHATS)*
12681 * Change the SSL kerb5 codes to match RFC 2712.
12685 * Make -nameopt work fully for req and add -reqopt switch.
12687 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12689 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12691 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12693 * Make sure tests can be performed even if the corresponding algorithms
12694 have been removed entirely. This was also the last step to make
12695 OpenSSL compilable with DJGPP under all reasonable conditions.
12697 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12699 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12700 to allow version independent disabling of normally unselected ciphers,
12701 which may be activated as a side-effect of selecting a single cipher.
12703 (E.g., cipher list string "RSA" enables ciphersuites that are left
12704 out of "ALL" because they do not provide symmetric encryption.
12705 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12707 *Lutz Jaenicke, Bodo Moeller*
12709 * Add appropriate support for separate platform-dependent build
12710 directories. The recommended way to make a platform-dependent
12711 build directory is the following (tested on Linux), maybe with
12714 # Place yourself outside of the OpenSSL source tree. In
12715 # this example, the environment variable OPENSSL_SOURCE
12716 # is assumed to contain the absolute OpenSSL source directory.
12717 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12718 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12719 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12720 mkdir -p `dirname $F`
12721 ln -s $OPENSSL_SOURCE/$F $F
12724 To be absolutely sure not to disturb the source tree, a "make clean"
12725 is a good thing. If it isn't successful, don't worry about it,
12726 it probably means the source directory is very clean.
12730 * Make sure any ENGINE control commands make local copies of string
12731 pointers passed to them whenever necessary. Otherwise it is possible
12732 the caller may have overwritten (or deallocated) the original string
12733 data when a later ENGINE operation tries to use the stored values.
12735 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12737 * Improve diagnostics in file reading and command-line digests.
12739 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12741 * Add AES modes CFB and OFB to the object database. Correct an
12742 error in AES-CFB decryption.
12746 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12747 allows existing EVP_CIPHER_CTX structures to be reused after
12748 calling `EVP_*Final()`. This behaviour is used by encryption
12749 BIOs and some applications. This has the side effect that
12750 applications must explicitly clean up cipher contexts with
12751 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12755 * Check the values of dna and dnb in bn_mul_recursive before calling
12756 bn_mul_comba (a non zero value means the a or b arrays do not contain
12757 n2 elements) and fallback to bn_mul_normal if either is not zero.
12761 * Fix escaping of non-ASCII characters when using the -subj option
12762 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12766 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12767 form for "surname", serialNumber has no short form.
12768 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12769 therefore remove "mail" short name for "internet 7".
12770 The OID for unique identifiers in X509 certificates is
12771 x500UniqueIdentifier, not uniqueIdentifier.
12772 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12776 * Add an "init" command to the ENGINE config module and auto initialize
12777 ENGINEs. Without any "init" command the ENGINE will be initialized
12778 after all ctrl commands have been executed on it. If init=1 the
12779 ENGINE is initialized at that point (ctrls before that point are run
12780 on the uninitialized ENGINE and after on the initialized one). If
12781 init=0 then the ENGINE will not be initialized at all.
12785 * Fix the 'app_verify_callback' interface so that the user-defined
12786 argument is actually passed to the callback: In the
12787 SSL_CTX_set_cert_verify_callback() prototype, the callback
12788 declaration has been changed from
12791 int (*cb)(X509_STORE_CTX *,void *);
12792 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12793 i=s->ctx->app_verify_callback(&ctx)
12794 has been changed into
12795 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12797 To update applications using SSL_CTX_set_cert_verify_callback(),
12798 a dummy argument can be added to their callback functions.
12800 *D. K. Smetters <smetters@parc.xerox.com>*
12802 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12804 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12806 * Add and OPENSSL_LOAD_CONF define which will cause
12807 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12808 This allows older applications to transparently support certain
12809 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12810 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12811 load the config file and OPENSSL_add_all_algorithms_conf() which will
12812 always load it have also been added.
12816 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12817 Adjust NIDs and EVP layer.
12819 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12821 * Config modules support in openssl utility.
12823 Most commands now load modules from the config file,
12824 though in a few (such as version) this isn't done
12825 because it couldn't be used for anything.
12827 In the case of ca and req the config file used is
12828 the same as the utility itself: that is the -config
12829 command line option can be used to specify an
12834 * Move default behaviour from OPENSSL_config(). If appname is NULL
12835 use "openssl_conf" if filename is NULL use default openssl config file.
12839 * Add an argument to OPENSSL_config() to allow the use of an alternative
12840 config section name. Add a new flag to tolerate a missing config file
12841 and move code to CONF_modules_load_file().
12845 * Support for crypto accelerator cards from Accelerated Encryption
12846 Processing, www.aep.ie. (Use engine 'aep')
12847 The support was copied from 0.9.6c [engine] and adapted/corrected
12848 to work with the new engine framework.
12850 *AEP Inc. and Richard Levitte*
12852 * Support for SureWare crypto accelerator cards from Baltimore
12853 Technologies. (Use engine 'sureware')
12854 The support was copied from 0.9.6c [engine] and adapted
12855 to work with the new engine framework.
12859 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12860 make the newer ENGINE framework commands for the CHIL engine work.
12862 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12864 * Make it possible to produce shared libraries on ReliantUNIX.
12866 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12868 * Add the configuration target debug-linux-ppro.
12869 Make 'openssl rsa' use the general key loading routines
12870 implemented in `apps.c`, and make those routines able to
12871 handle the key format FORMAT_NETSCAPE and the variant
12874 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12876 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12878 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12880 * Add -keyform to rsautl, and document -engine.
12882 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12884 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12885 BIO_R_NO_SUCH_FILE error code rather than the generic
12886 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12890 * Add new functions
12891 ERR_peek_last_error
12892 ERR_peek_last_error_line
12893 ERR_peek_last_error_line_data.
12894 These are similar to
12896 ERR_peek_error_line
12897 ERR_peek_error_line_data,
12898 but report on the latest error recorded rather than the first one
12899 still in the error queue.
12901 *Ben Laurie, Bodo Moeller*
12903 * default_algorithms option in ENGINE config module. This allows things
12905 default_algorithms = ALL
12906 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12910 * Preliminary ENGINE config module.
12914 * New experimental application configuration code.
12918 * Change the AES code to follow the same name structure as all other
12919 symmetric ciphers, and behave the same way. Move everything to
12920 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12922 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12924 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12926 *Ben Laurie and Theo de Raadt*
12928 * Add option to output public keys in req command.
12930 *Massimiliano Pala madwolf@openca.org*
12932 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12933 (up to about 10% better than before for P-192 and P-224).
12937 * New functions/macros
12939 SSL_CTX_set_msg_callback(ctx, cb)
12940 SSL_CTX_set_msg_callback_arg(ctx, arg)
12941 SSL_set_msg_callback(ssl, cb)
12942 SSL_set_msg_callback_arg(ssl, arg)
12944 to request calling a callback function
12946 void cb(int write_p, int version, int content_type,
12947 const void *buf, size_t len, SSL *ssl, void *arg)
12949 whenever a protocol message has been completely received
12950 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12951 protocol version according to which the SSL library interprets
12952 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12953 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12954 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12955 specification (change_cipher_spec(20), alert(21), handshake(22)).
12956 'buf' and 'len' point to the actual message, 'ssl' to the
12957 SSL object, and 'arg' is the application-defined value set by
12958 SSL[_CTX]_set_msg_callback_arg().
12960 'openssl s_client' and 'openssl s_server' have new '-msg' options
12961 to enable a callback that displays all protocol messages.
12965 * Change the shared library support so shared libraries are built as
12966 soon as the corresponding static library is finished, and thereby get
12967 openssl and the test programs linked against the shared library.
12968 This still only happens when the keyword "shard" has been given to
12969 the configuration scripts.
12971 NOTE: shared library support is still an experimental thing, and
12972 backward binary compatibility is still not guaranteed.
12974 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12976 * Add support for Subject Information Access extension.
12978 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12980 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12981 additional bytes when new memory had to be allocated, not just
12982 when reusing an existing buffer.
12986 * New command line and configuration option 'utf8' for the req command.
12987 This allows field values to be specified as UTF8 strings.
12991 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12992 runs for the former and machine-readable output for the latter.
12996 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12997 of the e-mail address in the DN (i.e., it will go into a certificate
12998 extension only). The new configuration file option 'email_in_dn = no'
12999 has the same effect.
13001 *Massimiliano Pala madwolf@openca.org*
13003 * Change all functions with names starting with `des_` to be starting
13004 with `DES_` instead. Add wrappers that are compatible with libdes,
13005 but are named `_ossl_old_des_*`. Finally, add macros that map the
13006 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13007 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13008 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13011 Since we provide two compatibility mappings, the user needs to
13012 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13013 compatibility is desired. The default (i.e., when that macro
13014 isn't defined) is OpenSSL 0.9.6c compatibility.
13016 There are also macros that enable and disable the support of old
13017 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13018 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13019 are defined, the default will apply: to support the old des routines.
13021 In either case, one must include openssl/des.h to get the correct
13022 definitions. Do not try to just include openssl/des_old.h, that
13025 NOTE: This is a major break of an old API into a new one. Software
13026 authors are encouraged to switch to the `DES_` style functions. Some
13027 time in the future, des_old.h and the libdes compatibility functions
13028 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13029 default), and then completely removed.
13033 * Test for certificates which contain unsupported critical extensions.
13034 If such a certificate is found during a verify operation it is
13035 rejected by default: this behaviour can be overridden by either
13036 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13037 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13038 X509_supported_extension() has also been added which returns 1 if a
13039 particular extension is supported.
13043 * Modify the behaviour of EVP cipher functions in similar way to digests
13044 to retain compatibility with existing code.
13048 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13049 compatibility with existing code. In particular the 'ctx' parameter does
13050 not have to be to be initialized before the call to EVP_DigestInit() and
13051 it is tidied up after a call to EVP_DigestFinal(). New function
13052 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13053 EVP_MD_CTX_copy() changed to not require the destination to be
13054 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13055 requires the destination to be valid.
13057 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13058 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13062 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13063 so that complete 'Handshake' protocol structures are kept in memory
13064 instead of overwriting 'msg_type' and 'length' with 'body' data.
13068 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13070 *Massimo Santin via Richard Levitte*
13072 * Major restructuring to the underlying ENGINE code. This includes
13073 reduction of linker bloat, separation of pure "ENGINE" manipulation
13074 (initialisation, etc) from functionality dealing with implementations
13075 of specific crypto interfaces. This change also introduces integrated
13076 support for symmetric ciphers and digest implementations - so ENGINEs
13077 can now accelerate these by providing EVP_CIPHER and EVP_MD
13078 implementations of their own. This is detailed in
13079 [crypto/engine/README.md](crypto/engine/README.md)
13080 as it couldn't be adequately described here. However, there are a few
13081 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13082 were changed in the original introduction of ENGINE code have now
13083 reverted back - the hooking from this code to ENGINE is now a good
13084 deal more passive and at run-time, operations deal directly with
13085 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13086 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13087 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13088 they were not being used by the framework as there is no concept of a
13089 BIGNUM_METHOD and they could not be generalised to the new
13090 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13091 ENGINE_cpy() has been removed as it cannot be consistently defined in
13096 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13100 * Change mkdef.pl to sort symbols that get the same entry number,
13101 and make sure the automatically generated functions `ERR_load_*`
13102 become part of libeay.num as well.
13106 * New function SSL_renegotiate_pending(). This returns true once
13107 renegotiation has been requested (either SSL_renegotiate() call
13108 or HelloRequest/ClientHello received from the peer) and becomes
13109 false once a handshake has been completed.
13110 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13111 sends a HelloRequest, but does not ensure that a handshake takes
13112 place. SSL_renegotiate_pending() is useful for checking if the
13113 client has followed the request.)
13117 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13118 By default, clients may request session resumption even during
13119 renegotiation (if session ID contexts permit); with this option,
13120 session resumption is possible only in the first handshake.
13122 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13123 more bits available for options that should not be part of
13124 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13128 * Add some demos for certificate and certificate request creation.
13132 * Make maximum certificate chain size accepted from the peer application
13133 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13134 "Douglas E. Engert" <deengert@anl.gov>.
13138 * Add support for shared libraries for Unixware-7
13139 (Boyd Lynn Gerber <gerberb@zenez.com>).
13143 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13144 be done prior to destruction. Use this to unload error strings from
13145 ENGINEs that load their own error strings. NB: This adds two new API
13146 functions to "get" and "set" this destroy handler in an ENGINE.
13150 * Alter all existing ENGINE implementations (except "openssl" and
13151 "openbsd") to dynamically instantiate their own error strings. This
13152 makes them more flexible to be built both as statically-linked ENGINEs
13153 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13154 Also, add stub code to each that makes building them as self-contained
13155 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13159 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13160 implementations into applications that are completely implemented in
13161 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13162 commands that can be used to configure what shared-library to load and
13163 to control aspects of the way it is handled. Also, made an update to
13164 the [README-Engine.md](README-Engine.md) file
13165 that brings its information up-to-date and
13166 provides some information and instructions on the "dynamic" ENGINE
13167 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13171 * Make it possible to unload ranges of ERR strings with a new
13172 "ERR_unload_strings" function.
13176 * Add a copy() function to EVP_MD.
13180 * Make EVP_MD routines take a context pointer instead of just the
13181 md_data void pointer.
13185 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13186 that the digest can only process a single chunk of data
13187 (typically because it is provided by a piece of
13188 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13189 is only going to provide a single chunk of data, and hence the
13190 framework needn't accumulate the data for oneshot drivers.
13194 * As with "ERR", make it possible to replace the underlying "ex_data"
13195 functions. This change also alters the storage and management of global
13196 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13197 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13198 index counters. The API functions that use this state have been changed
13199 to take a "class_index" rather than pointers to the class's local STACK
13200 and counter, and there is now an API function to dynamically create new
13201 classes. This centralisation allows us to (a) plug a lot of the
13202 thread-safety problems that existed, and (b) makes it possible to clean
13203 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13204 such data would previously have always leaked in application code and
13205 workarounds were in place to make the memory debugging turn a blind eye
13206 to it. Application code that doesn't use this new function will still
13207 leak as before, but their memory debugging output will announce it now
13208 rather than letting it slide.
13210 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13211 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13212 has a return value to indicate success or failure.
13216 * Make it possible to replace the underlying "ERR" functions such that the
13217 global state (2 LHASH tables and 2 locks) is only used by the "default"
13218 implementation. This change also adds two functions to "get" and "set"
13219 the implementation prior to it being automatically set the first time
13220 any other ERR function takes place. Ie. an application can call "get",
13221 pass the return value to a module it has just loaded, and that module
13222 can call its own "set" function using that value. This means the
13223 module's "ERR" operations will use (and modify) the error state in the
13224 application and not in its own statically linked copy of OpenSSL code.
13228 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13229 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13230 the operation, and provides a more encapsulated way for external code
13231 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13232 to use these functions rather than manually incrementing the counts.
13234 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13238 * Add EVP test program.
13242 * Add symmetric cipher support to ENGINE. Expect the API to change!
13246 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13247 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13248 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13249 These allow a CRL to be built without having to access X509_CRL fields
13250 directly. Modify 'ca' application to use new functions.
13254 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13255 bug workarounds. Rollback attack detection is a security feature.
13256 The problem will only arise on OpenSSL servers when TLSv1 is not
13257 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13258 Software authors not wanting to support TLSv1 will have special reasons
13259 for their choice and can explicitly enable this option.
13261 *Bodo Moeller, Lutz Jaenicke*
13263 * Rationalise EVP so it can be extended: don't include a union of
13264 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13265 (similar to those existing for EVP_CIPHER_CTX).
13270 EVP_MD_CTX_init(&md); /* new function call */
13271 EVP_DigestInit(&md, EVP_sha1());
13272 EVP_DigestUpdate(&md, in, len);
13273 EVP_DigestFinal(&md, out, NULL);
13274 EVP_MD_CTX_cleanup(&md); /* new function call */
13278 * Make DES key schedule conform to the usual scheme, as well as
13279 correcting its structure. This means that calls to DES functions
13280 now have to pass a pointer to a des_key_schedule instead of a
13281 plain des_key_schedule (which was actually always a pointer
13284 des_key_schedule ks;
13286 des_set_key_checked(..., &ks);
13287 des_ncbc_encrypt(..., &ks, ...);
13289 (Note that a later change renames 'des_...' into 'DES_...'.)
13293 * Initial reduction of linker bloat: the use of some functions, such as
13294 PEM causes large amounts of unused functions to be linked in due to
13295 poor organisation. For example pem_all.c contains every PEM function
13296 which has a knock on effect of linking in large amounts of (unused)
13297 ASN1 code. Grouping together similar functions and splitting unrelated
13298 functions prevents this.
13302 * Cleanup of EVP macros.
13306 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13307 correct `_ecb suffix`.
13311 * Add initial OCSP responder support to ocsp application. The
13312 revocation information is handled using the text based index
13313 use by the ca application. The responder can either handle
13314 requests generated internally, supplied in files (for example
13315 via a CGI script) or using an internal minimal server.
13319 * Add configuration choices to get zlib compression for TLS.
13323 * Changes to Kerberos SSL for RFC 2712 compliance:
13324 1. Implemented real KerberosWrapper, instead of just using
13325 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13326 2. Implemented optional authenticator field of KerberosWrapper.
13328 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13329 and authenticator structs; see crypto/krb5/.
13331 Generalized Kerberos calls to support multiple Kerberos libraries.
13332 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13333 via Richard Levitte*
13335 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13336 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13337 values for each of the key sizes rather than having just
13338 parameters (and 'speed' generating keys each time).
13342 * Speed up EVP routines.
13345 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13346 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13347 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13348 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13350 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13351 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13352 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13355 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13357 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13361 * Added the OS2-EMX target.
13363 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13365 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13366 New functions to support `NCONF` routines in extension code.
13367 New function `CONF_set_nconf()`
13368 to allow functions which take an `NCONF` to also handle the old `LHASH`
13369 structure: this means that the old `CONF` compatible routines can be
13370 retained (in particular w.rt. extensions) without having to duplicate the
13371 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13375 * Enhance the general user interface with mechanisms for inner control
13376 and with possibilities to have yes/no kind of prompts.
13380 * Change all calls to low-level digest routines in the library and
13381 applications to use EVP. Add missing calls to HMAC_cleanup() and
13382 don't assume HMAC_CTX can be copied using memcpy().
13384 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13386 * Add the possibility to control engines through control names but with
13387 arbitrary arguments instead of just a string.
13388 Change the key loaders to take a UI_METHOD instead of a callback
13389 function pointer. NOTE: this breaks binary compatibility with earlier
13390 versions of OpenSSL [engine].
13391 Adapt the nCipher code for these new conditions and add a card insertion
13396 * Enhance the general user interface with mechanisms to better support
13397 dialog box interfaces, application-defined prompts, the possibility
13398 to use defaults (for example default passwords from somewhere else)
13399 and interrupts/cancellations.
13403 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13404 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13408 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13409 tidy up some unnecessarily weird code in 'sk_new()').
13411 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13413 * Change the key loading routines for ENGINEs to use the same kind
13414 callback (pem_password_cb) as all other routines that need this
13419 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13420 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13421 than this minimum value is recommended.
13425 * New random seeder for OpenVMS, using the system process statistics
13426 that are easily reachable.
13430 * Windows apparently can't transparently handle global
13431 variables defined in DLLs. Initialisations such as:
13433 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13435 won't compile. This is used by the any applications that need to
13436 declare their own ASN1 modules. This was fixed by adding the option
13437 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13438 needed for static libraries under Win32.
13442 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13443 setting of purpose and trust fields. New X509_STORE trust and
13444 purpose functions and tidy up setting in other SSL functions.
13448 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13449 structure. These are inherited by X509_STORE_CTX when it is
13450 initialised. This allows various defaults to be set in the
13451 X509_STORE structure (such as flags for CRL checking and custom
13452 purpose or trust settings) for functions which only use X509_STORE_CTX
13453 internally such as S/MIME.
13455 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13456 trust settings if they are not set in X509_STORE. This allows X509_STORE
13457 purposes and trust (in S/MIME for example) to override any set by default.
13459 Add command line options for CRL checking to smime, s_client and s_server
13464 * Initial CRL based revocation checking. If the CRL checking flag(s)
13465 are set then the CRL is looked up in the X509_STORE structure and
13466 its validity and signature checked, then if the certificate is found
13467 in the CRL the verify fails with a revoked error.
13469 Various new CRL related callbacks added to X509_STORE_CTX structure.
13471 Command line options added to 'verify' application to support this.
13473 This needs some additional work, such as being able to handle multiple
13474 CRLs with different times, extension based lookup (rather than just
13475 by subject name) and ultimately more complete V2 CRL extension
13480 * Add a general user interface API (crypto/ui/). This is designed
13481 to replace things like des_read_password and friends (backward
13482 compatibility functions using this new API are provided).
13483 The purpose is to remove prompting functions from the DES code
13484 section as well as provide for prompting through dialog boxes in
13485 a window system and the like.
13489 * Add "ex_data" support to ENGINE so implementations can add state at a
13490 per-structure level rather than having to store it globally.
13494 * Make it possible for ENGINE structures to be copied when retrieved by
13495 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13496 This causes the "original" ENGINE structure to act like a template,
13497 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13498 operational state can be localised to each ENGINE structure, despite the
13499 fact they all share the same "methods". New ENGINE structures returned in
13500 this case have no functional references and the return value is the single
13501 structural reference. This matches the single structural reference returned
13502 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13507 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13508 needs to match any other type at all we need to manually clear the
13513 * Changes to the "openssl engine" utility to include;
13514 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13515 about an ENGINE's available control commands.
13516 - executing control commands from command line arguments using the
13517 '-pre' and '-post' switches. '-post' is only used if '-t' is
13518 specified and the ENGINE is successfully initialised. The syntax for
13519 the individual commands are colon-separated, for example;
13520 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13524 * New dynamic control command support for ENGINEs. ENGINEs can now
13525 declare their own commands (numbers), names (strings), descriptions,
13526 and input types for run-time discovery by calling applications. A
13527 subset of these commands are implicitly classed as "executable"
13528 depending on their input type, and only these can be invoked through
13529 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13530 can be based on user input, config files, etc). The distinction is
13531 that "executable" commands cannot return anything other than a boolean
13532 result and can only support numeric or string input, whereas some
13533 discoverable commands may only be for direct use through
13534 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13535 pointers, or other custom uses. The "executable" commands are to
13536 support parameterisations of ENGINE behaviour that can be
13537 unambiguously defined by ENGINEs and used consistently across any
13538 OpenSSL-based application. Commands have been added to all the
13539 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13540 control over shared-library paths without source code alterations.
13544 * Changed all ENGINE implementations to dynamically allocate their
13545 ENGINEs rather than declaring them statically. Apart from this being
13546 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13547 this also allows the implementations to compile without using the
13548 internal engine_int.h header.
13552 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13553 'const' value. Any code that should be able to modify a RAND_METHOD
13554 should already have non-const pointers to it (ie. they should only
13555 modify their own ones).
13559 * Made a variety of little tweaks to the ENGINE code.
13560 - "atalla" and "ubsec" string definitions were moved from header files
13561 to C code. "nuron" string definitions were placed in variables
13562 rather than hard-coded - allowing parameterisation of these values
13563 later on via ctrl() commands.
13564 - Removed unused "#if 0"'d code.
13565 - Fixed engine list iteration code so it uses ENGINE_free() to release
13566 structural references.
13567 - Constified the RAND_METHOD element of ENGINE structures.
13568 - Constified various get/set functions as appropriate and added
13569 missing functions (including a catch-all ENGINE_cpy that duplicates
13570 all ENGINE values onto a new ENGINE except reference counts/state).
13571 - Removed NULL parameter checks in get/set functions. Setting a method
13572 or function to NULL is a way of cancelling out a previously set
13573 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13574 and doesn't justify the extra error symbols and code.
13575 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13576 flags from engine_int.h to engine.h.
13577 - Changed prototypes for ENGINE handler functions (init(), finish(),
13578 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13582 * Implement binary inversion algorithm for BN_mod_inverse in addition
13583 to the algorithm using long division. The binary algorithm can be
13584 used only if the modulus is odd. On 32-bit systems, it is faster
13585 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13586 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13587 up to 450 bits. In 64-bit environments, the binary algorithm
13588 appears to be advantageous for much longer moduli; here we use it
13589 for moduli up to 2048 bits.
13593 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13594 could not support the combine flag in choice fields.
13598 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13599 extensions from a certificate request to the certificate.
13603 * Allow multiple 'certopt' and 'nameopt' options to be separated
13604 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13605 file: this allows the display of the certificate about to be
13606 signed to be customised, to allow certain fields to be included
13607 or excluded and extension details. The old system didn't display
13608 multicharacter strings properly, omitted fields not in the policy
13609 and couldn't display additional details such as extensions.
13613 * Function EC_POINTs_mul for multiple scalar multiplication
13614 of an arbitrary number of elliptic curve points
13615 \sum scalars[i]*points[i],
13616 optionally including the generator defined for the EC_GROUP:
13617 scalar*generator + \sum scalars[i]*points[i].
13619 EC_POINT_mul is a simple wrapper function for the typical case
13620 that the point list has just one item (besides the optional
13625 * First EC_METHODs for curves over GF(p):
13627 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13628 operations and provides various method functions that can also
13629 operate with faster implementations of modular arithmetic.
13631 EC_GFp_mont_method() reuses most functions that are part of
13632 EC_GFp_simple_method, but uses Montgomery arithmetic.
13634 *Bodo Moeller; point addition and point doubling
13635 implementation directly derived from source code provided by
13636 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13638 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13639 crypto/ec/ec_lib.c):
13641 Curves are EC_GROUP objects (with an optional group generator)
13642 based on EC_METHODs that are built into the library.
13644 Points are EC_POINT objects based on EC_GROUP objects.
13646 Most of the framework would be able to handle curves over arbitrary
13647 finite fields, but as there are no obvious types for fields other
13648 than GF(p), some functions are limited to that for now.
13652 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13653 that the file contains a complete HTTP response.
13657 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13658 change the def and num file printf format specifier from "%-40sXXX"
13659 to "%-39s XXX". The latter will always guarantee a space after the
13660 field while the former will cause them to run together if the field
13661 is 40 of more characters long.
13665 * Constify the cipher and digest 'method' functions and structures
13666 and modify related functions to take constant EVP_MD and EVP_CIPHER
13671 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13672 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13676 * Modify `EVP_Digest*()` routines so they now return values. Although the
13677 internal software routines can never fail additional hardware versions
13682 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13684 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13685 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13688 ERR_R_NESTED_ASN1_ERROR
13690 ERR_R_MISSING_ASN1_EOS
13691 were 4 .. 9, conflicting with
13692 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13694 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13695 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13697 Add new error code 'ERR_R_INTERNAL_ERROR'.
13701 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13706 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13707 sets the subject name for a new request or supersedes the
13708 subject name in a given request. Formats that can be parsed are
13709 'CN=Some Name, OU=myOU, C=IT'
13711 'CN=Some Name/OU=myOU/C=IT'.
13713 Add options '-batch' and '-verbose' to 'openssl req'.
13715 *Massimiliano Pala <madwolf@hackmasters.net>*
13717 * Introduce the possibility to access global variables through
13718 functions on platform were that's the best way to handle exporting
13719 global variables in shared libraries. To enable this functionality,
13720 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13721 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13722 is normally done by Configure or something similar).
13724 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13725 in the source file (foo.c) like this:
13727 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13728 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13730 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13731 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13733 OPENSSL_DECLARE_GLOBAL(int,foo);
13734 #define foo OPENSSL_GLOBAL_REF(foo)
13735 OPENSSL_DECLARE_GLOBAL(double,bar);
13736 #define bar OPENSSL_GLOBAL_REF(bar)
13738 The #defines are very important, and therefore so is including the
13739 header file everywhere where the defined globals are used.
13741 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13742 of ASN.1 items, but that structure is a bit different.
13744 The largest change is in util/mkdef.pl which has been enhanced with
13745 better and easier to understand logic to choose which symbols should
13746 go into the Windows .def files as well as a number of fixes and code
13747 cleanup (among others, algorithm keywords are now sorted
13748 lexicographically to avoid constant rewrites).
13752 * In BN_div() keep a copy of the sign of 'num' before writing the
13753 result to 'rm' because if rm==num the value will be overwritten
13754 and produce the wrong result if 'num' is negative: this caused
13755 problems with BN_mod() and BN_nnmod().
13759 * Function OCSP_request_verify(). This checks the signature on an
13760 OCSP request and verifies the signer certificate. The signer
13761 certificate is just checked for a generic purpose and OCSP request
13766 * Add OCSP_check_validity() function to check the validity of OCSP
13767 responses. OCSP responses are prepared in real time and may only
13768 be a few seconds old. Simply checking that the current time lies
13769 between thisUpdate and nextUpdate max reject otherwise valid responses
13770 caused by either OCSP responder or client clock inaccuracy. Instead
13771 we allow thisUpdate and nextUpdate to fall within a certain period of
13772 the current time. The age of the response can also optionally be
13773 checked. Two new options -validity_period and -status_age added to
13778 * If signature or public key algorithm is unrecognized print out its
13779 OID rather that just UNKNOWN.
13783 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13784 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13785 ID to be generated from the issuer certificate alone which can then be
13786 passed to OCSP_id_issuer_cmp().
13790 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13791 ASN1 modules to export functions returning ASN1_ITEM pointers
13792 instead of the ASN1_ITEM structures themselves. This adds several
13793 new macros which allow the underlying ASN1 function/structure to
13794 be accessed transparently. As a result code should not use ASN1_ITEM
13795 references directly (such as &X509_it) but instead use the relevant
13796 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13797 use of the new ASN1 code on platforms where exporting structures
13798 is problematical (for example in shared libraries) but exporting
13799 functions returning pointers to structures is not.
13803 * Add support for overriding the generation of SSL/TLS session IDs.
13804 These callbacks can be registered either in an SSL_CTX or per SSL.
13805 The purpose of this is to allow applications to control, if they wish,
13806 the arbitrary values chosen for use as session IDs, particularly as it
13807 can be useful for session caching in multiple-server environments. A
13808 command-line switch for testing this (and any client code that wishes
13809 to use such a feature) has been added to "s_server".
13811 *Geoff Thorpe, Lutz Jaenicke*
13813 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13814 of the form `#if defined(...) || defined(...) || ...` and
13815 `#if !defined(...) && !defined(...) && ...`. This also avoids
13816 the growing number of special cases it was previously handling.
13820 * Make all configuration macros available for application by making
13821 sure they are available in opensslconf.h, by giving them names starting
13822 with `OPENSSL_` to avoid conflicts with other packages and by making
13823 sure e_os2.h will cover all platform-specific cases together with
13825 Additionally, it is now possible to define configuration/platform-
13826 specific names (called "system identities"). In the C code, these
13827 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13828 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13829 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13834 * New option -set_serial to 'req' and 'x509' this allows the serial
13835 number to use to be specified on the command line. Previously self
13836 signed certificates were hard coded with serial number 0 and the
13837 CA options of 'x509' had to use a serial number in a file which was
13842 * New options to 'ca' utility to support V2 CRL entry extensions.
13843 Currently CRL reason, invalidity date and hold instruction are
13844 supported. Add new CRL extensions to V3 code and some new objects.
13848 * New function EVP_CIPHER_CTX_set_padding() this is used to
13849 disable standard block padding (aka PKCS#5 padding) in the EVP
13850 API, which was previously mandatory. This means that the data is
13851 not padded in any way and so the total length much be a multiple
13852 of the block size, otherwise an error occurs.
13856 * Initial (incomplete) OCSP SSL support.
13860 * New function OCSP_parse_url(). This splits up a URL into its host,
13861 port and path components: primarily to parse OCSP URLs. New -url
13862 option to ocsp utility.
13866 * New nonce behavior. The return value of OCSP_check_nonce() now
13867 reflects the various checks performed. Applications can decide
13868 whether to tolerate certain situations such as an absent nonce
13869 in a response when one was present in a request: the ocsp application
13870 just prints out a warning. New function OCSP_add1_basic_nonce()
13871 this is to allow responders to include a nonce in a response even if
13872 the request is nonce-less.
13876 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13877 skipped when using openssl x509 multiple times on a single input file,
13878 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13882 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13883 set string type: to handle setting ASN1_TIME structures. Fix ca
13884 utility to correctly initialize revocation date of CRLs.
13888 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13889 the clients preferred ciphersuites and rather use its own preferences.
13890 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13891 Internet Explorer by ensuring unchanged hash method during stepup.
13892 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13896 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13897 to aes and add a new 'exist' option to print out symbols that don't
13902 * Additional options to ocsp utility to allow flags to be set and
13903 additional certificates supplied.
13907 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13908 OCSP client a number of certificate to only verify the response
13913 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13914 handle the new API. Currently only ECB, CBC modes supported. Add new
13917 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13918 Encryption Standard (AES) Ciphersuites for Transport Layer
13919 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13920 not enabled by default and were not part of the "ALL" ciphersuite
13921 alias because they were not yet official; they could be
13922 explicitly requested by specifying the "AESdraft" ciphersuite
13923 group alias. In the final release of OpenSSL 0.9.7, the group
13924 alias is called "AES" and is part of "ALL".)
13926 *Ben Laurie, Steve Henson, Bodo Moeller*
13928 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13929 request to response.
13933 * Functions for OCSP responders. OCSP_request_onereq_count(),
13934 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13935 extract information from a certificate request. OCSP_response_create()
13936 creates a response and optionally adds a basic response structure.
13937 OCSP_basic_add1_status() adds a complete single response to a basic
13938 response and returns the OCSP_SINGLERESP structure just added (to allow
13939 extensions to be included for example). OCSP_basic_add1_cert() adds a
13940 certificate to a basic response and OCSP_basic_sign() signs a basic
13941 response with various flags. New helper functions ASN1_TIME_check()
13942 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13943 (converts ASN1_TIME to GeneralizedTime).
13947 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13948 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13949 structure from a certificate. X509_pubkey_digest() digests the public_key
13950 contents: this is used in various key identifiers.
13954 * Make sk_sort() tolerate a NULL argument.
13956 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13958 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13959 passed by the function are trusted implicitly. If any of them signed the
13960 response then it is assumed to be valid and is not verified.
13964 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13965 to data. This was previously part of the PKCS7 ASN1 code. This
13966 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13967 *Steve Henson, reported by Kenneth R. Robinette
13968 <support@securenetterm.com>*
13970 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13971 routines: without these tracing memory leaks is very painful.
13972 Fix leaks in PKCS12 and PKCS7 routines.
13976 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13977 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13978 effectively meant GeneralizedTime would never be used. Now it
13979 is initialised to -1 but X509_time_adj() now has to check the value
13980 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13981 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13982 *Steve Henson, reported by Kenneth R. Robinette
13983 <support@securenetterm.com>*
13985 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13986 result in a zero length in the ASN1_INTEGER structure which was
13987 not consistent with the structure when d2i_ASN1_INTEGER() was used
13988 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13989 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13990 where it did not print out a minus for negative ASN1_INTEGER.
13994 * Add summary printout to ocsp utility. The various functions which
13995 convert status values to strings have been renamed to:
13996 OCSP_response_status_str(), OCSP_cert_status_str() and
13997 OCSP_crl_reason_str() and are no longer static. New options
13998 to verify nonce values and to disable verification. OCSP response
13999 printout format cleaned up.
14003 * Add additional OCSP certificate checks. These are those specified
14004 in RFC2560. This consists of two separate checks: the CA of the
14005 certificate being checked must either be the OCSP signer certificate
14006 or the issuer of the OCSP signer certificate. In the latter case the
14007 OCSP signer certificate must contain the OCSP signing extended key
14008 usage. This check is performed by attempting to match the OCSP
14009 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14010 in the OCSP_CERTID structures of the response.
14014 * Initial OCSP certificate verification added to OCSP_basic_verify()
14015 and related routines. This uses the standard OpenSSL certificate
14016 verify routines to perform initial checks (just CA validity) and
14017 to obtain the certificate chain. Then additional checks will be
14018 performed on the chain. Currently the root CA is checked to see
14019 if it is explicitly trusted for OCSP signing. This is used to set
14020 a root CA as a global signing root: that is any certificate that
14021 chains to that CA is an acceptable OCSP signing certificate.
14025 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14026 extensions from a separate configuration file.
14027 As when reading extensions from the main configuration file,
14028 the '-extensions ...' option may be used for specifying the
14031 *Massimiliano Pala <madwolf@comune.modena.it>*
14033 * New OCSP utility. Allows OCSP requests to be generated or
14034 read. The request can be sent to a responder and the output
14035 parsed, outputted or printed in text form. Not complete yet:
14036 still needs to check the OCSP response validity.
14040 * New subcommands for 'openssl ca':
14041 `openssl ca -status <serial>` prints the status of the cert with
14042 the given serial number (according to the index file).
14043 `openssl ca -updatedb` updates the expiry status of certificates
14046 *Massimiliano Pala <madwolf@comune.modena.it>*
14048 * New '-newreq-nodes' command option to CA.pl. This is like
14049 '-newreq', but calls 'openssl req' with the '-nodes' option
14050 so that the resulting key is not encrypted.
14052 *Damien Miller <djm@mindrot.org>*
14054 * New configuration for the GNU Hurd.
14056 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14058 * Initial code to implement OCSP basic response verify. This
14059 is currently incomplete. Currently just finds the signer's
14060 certificate and verifies the signature on the response.
14064 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14065 value of OPENSSLDIR. This is available via the new '-d' option
14066 to 'openssl version', and is also included in 'openssl version -a'.
14070 * Allowing defining memory allocation callbacks that will be given
14071 file name and line number information in additional arguments
14072 (a `const char*` and an int). The basic functionality remains, as
14073 well as the original possibility to just replace malloc(),
14074 realloc() and free() by functions that do not know about these
14075 additional arguments. To register and find out the current
14076 settings for extended allocation functions, the following
14077 functions are provided:
14079 CRYPTO_set_mem_ex_functions
14080 CRYPTO_set_locked_mem_ex_functions
14081 CRYPTO_get_mem_ex_functions
14082 CRYPTO_get_locked_mem_ex_functions
14084 These work the same way as CRYPTO_set_mem_functions and friends.
14085 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14086 extended allocation function is enabled.
14087 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14088 a conventional allocation function is enabled.
14090 *Richard Levitte, Bodo Moeller*
14092 * Finish off removing the remaining LHASH function pointer casts.
14093 There should no longer be any prototype-casting required when using
14094 the LHASH abstraction, and any casts that remain are "bugs". See
14095 the callback types and macros at the head of lhash.h for details
14096 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14100 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14101 If /dev/[u]random devices are not available or do not return enough
14102 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14104 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14105 /etc/entropy will be queried once each in this sequence, querying stops
14106 when enough entropy was collected without querying more sockets.
14110 * Change the Unix RAND_poll() variant to be able to poll several
14111 random devices, as specified by DEVRANDOM, until a sufficient amount
14112 of data has been collected. We spend at most 10 ms on each file
14113 (select timeout) and read in non-blocking mode. DEVRANDOM now
14114 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14115 (previously it was just the string "/dev/urandom"), so on typical
14116 platforms the 10 ms delay will never occur.
14117 Also separate out the Unix variant to its own file, rand_unix.c.
14118 For VMS, there's a currently-empty rand_vms.c.
14122 * Move OCSP client related routines to ocsp_cl.c. These
14123 provide utility functions which an application needing
14124 to issue a request to an OCSP responder and analyse the
14125 response will typically need: as opposed to those which an
14126 OCSP responder itself would need which will be added later.
14128 OCSP_request_sign() signs an OCSP request with an API similar
14129 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14130 response. OCSP_response_get1_basic() extracts basic response
14131 from response. OCSP_resp_find_status(): finds and extracts status
14132 information from an OCSP_CERTID structure (which will be created
14133 when the request structure is built). These are built from lower
14134 level functions which work on OCSP_SINGLERESP structures but
14135 won't normally be used unless the application wishes to examine
14136 extensions in the OCSP response for example.
14138 Replace nonce routines with a pair of functions.
14139 OCSP_request_add1_nonce() adds a nonce value and optionally
14140 generates a random value. OCSP_check_nonce() checks the
14141 validity of the nonce in an OCSP response.
14145 * Change function OCSP_request_add() to OCSP_request_add0_id().
14146 This doesn't copy the supplied OCSP_CERTID and avoids the
14147 need to free up the newly created id. Change return type
14148 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14149 This can then be used to add extensions to the request.
14150 Deleted OCSP_request_new(), since most of its functionality
14151 is now in OCSP_REQUEST_new() (and the case insensitive name
14152 clash) apart from the ability to set the request name which
14153 will be added elsewhere.
14157 * Update OCSP API. Remove obsolete extensions argument from
14158 various functions. Extensions are now handled using the new
14159 OCSP extension code. New simple OCSP HTTP function which
14160 can be used to send requests and parse the response.
14164 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14165 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14166 uses the special reorder version of SET OF to sort the attributes
14167 and reorder them to match the encoded order. This resolves a long
14168 standing problem: a verify on a PKCS7 structure just after signing
14169 it used to fail because the attribute order did not match the
14170 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14171 it uses the received order. This is necessary to tolerate some broken
14172 software that does not order SET OF. This is handled by encoding
14173 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14174 to produce the required SET OF.
14178 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14179 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14180 files to get correct declarations of the ASN.1 item variables.
14184 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14185 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14186 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14187 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14188 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14189 ASN1_ITEM and no wrapper functions.
14193 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14194 replace the old function pointer based I/O routines. Change most of
14195 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14199 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14200 lines, recognize more "algorithms" that can be deselected, and make
14201 it complain about algorithm deselection that isn't recognised.
14205 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14206 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14207 to use new functions. Add NO_ASN1_OLD which can be set to remove
14208 some old style ASN1 functions: this can be used to determine if old
14209 code will still work when these eventually go away.
14213 * New extension functions for OCSP structures, these follow the
14214 same conventions as certificates and CRLs.
14218 * New function X509V3_add1_i2d(). This automatically encodes and
14219 adds an extension. Its behaviour can be customised with various
14220 flags to append, replace or delete. Various wrappers added for
14221 certificates and CRLs.
14225 * Fix to avoid calling the underlying ASN1 print routine when
14226 an extension cannot be parsed. Correct a typo in the
14227 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14231 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14232 entries for variables.
14236 * Add functionality to `apps/openssl.c` for detecting locking
14237 problems: As the program is single-threaded, all we have
14238 to do is register a locking callback using an array for
14239 storing which locks are currently held by the program.
14243 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14244 SSL_get_ex_data_X509_STORE_idx(), which is used in
14245 ssl_verify_cert_chain() and thus can be called at any time
14246 during TLS/SSL handshakes so that thread-safety is essential.
14247 Unfortunately, the ex_data design is not at all suited
14248 for multi-threaded use, so it probably should be abolished.
14252 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14254 *Broadcom, tweaked and integrated by Geoff Thorpe*
14256 * Move common extension printing code to new function
14257 X509V3_print_extensions(). Reorganise OCSP print routines and
14258 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14262 * New function X509_signature_print() to remove duplication in some
14267 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14268 set (this was treated exactly the same as SET OF previously). This
14269 is used to reorder the STACK representing the structure to match the
14270 encoding. This will be used to get round a problem where a PKCS7
14271 structure which was signed could not be verified because the STACK
14272 order did not reflect the encoded order.
14276 * Reimplement the OCSP ASN1 module using the new code.
14280 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14281 for its ASN1 operations. The old style function pointers still exist
14282 for now but they will eventually go away.
14286 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14287 completely replaces the old ASN1 functionality with a table driven
14288 encoder and decoder which interprets an ASN1_ITEM structure describing
14289 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14290 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14291 has also been converted to the new form.
14295 * Change BN_mod_exp_recp so that negative moduli are tolerated
14296 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14297 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14298 for negative moduli.
14302 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14303 of not touching the result's sign bit.
14307 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14312 * Changed the LHASH code to use prototypes for callbacks, and created
14313 macros to declare and implement thin (optionally static) functions
14314 that provide type-safety and avoid function pointer casting for the
14315 type-specific callbacks.
14319 * Added Kerberos Cipher Suites to be used with TLS, as written in
14321 *Veers Staats <staatsvr@asc.hpc.mil>,
14322 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14324 * Reformat the FAQ so the different questions and answers can be divided
14325 in sections depending on the subject.
14329 * Have the zlib compression code load ZLIB.DLL dynamically under
14334 * New function BN_mod_sqrt for computing square roots modulo a prime
14335 (using the probabilistic Tonelli-Shanks algorithm unless
14336 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14337 be handled deterministically).
14339 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14341 * Make BN_mod_inverse faster by explicitly handling small quotients
14342 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14343 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14347 * New function BN_kronecker.
14351 * Fix BN_gcd so that it works on negative inputs; the result is
14352 positive unless both parameters are zero.
14353 Previously something reasonably close to an infinite loop was
14354 possible because numbers could be growing instead of shrinking
14355 in the implementation of Euclid's algorithm.
14359 * Fix BN_is_word() and BN_is_one() macros to take into account the
14360 sign of the number in question.
14362 Fix BN_is_word(a,w) to work correctly for w == 0.
14364 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14365 because its test if the absolute value of 'a' equals 'w'.
14366 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14367 it exists mostly for use in the implementations of BN_is_zero(),
14368 BN_is_one(), and BN_is_word().
14372 * New function BN_swap.
14376 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14377 the exponentiation functions are more likely to produce reasonable
14378 results on negative inputs.
14382 * Change BN_mod_mul so that the result is always non-negative.
14383 Previously, it could be negative if one of the factors was negative;
14384 I don't think anyone really wanted that behaviour.
14388 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14389 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14390 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14391 and add new functions:
14400 BN_mod_lshift1_quick
14402 BN_mod_lshift_quick
14404 These functions always generate non-negative results.
14406 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14407 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14409 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14410 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14411 be reduced modulo `m`.
14413 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14416 The following entry accidentally appeared in the CHANGES file
14417 distributed with OpenSSL 0.9.7. The modifications described in
14418 it do *not* apply to OpenSSL 0.9.7.
14420 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14421 was actually never needed) and in BN_mul(). The removal in BN_mul()
14422 required a small change in bn_mul_part_recursive() and the addition
14423 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14424 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14425 bn_sub_words() and bn_add_words() except they take arrays with
14431 * In 'openssl passwd', verify passwords read from the terminal
14432 unless the '-salt' option is used (which usually means that
14433 verification would just waste user's time since the resulting
14434 hash is going to be compared with some given password hash)
14435 or the new '-noverify' option is used.
14437 This is an incompatible change, but it does not affect
14438 non-interactive use of 'openssl passwd' (passwords on the command
14439 line, '-stdin' option, '-in ...' option) and thus should not
14440 cause any problems.
14444 * Remove all references to RSAref, since there's no more need for it.
14448 * Make DSO load along a path given through an environment variable
14449 (SHLIB_PATH) with shl_load().
14453 * Constify the ENGINE code as a result of BIGNUM constification.
14454 Also constify the RSA code and most things related to it. In a
14455 few places, most notable in the depth of the ASN.1 code, ugly
14456 casts back to non-const were required (to be solved at a later
14461 * Make it so the openssl application has all engines loaded by default.
14465 * Constify the BIGNUM routines a little more.
14469 * Add the following functions:
14471 ENGINE_load_cswift()
14473 ENGINE_load_atalla()
14474 ENGINE_load_nuron()
14475 ENGINE_load_builtin_engines()
14477 That way, an application can itself choose if external engines that
14478 are built-in in OpenSSL shall ever be used or not. The benefit is
14479 that applications won't have to be linked with libdl or other dso
14480 libraries unless it's really needed.
14482 Changed 'openssl engine' to load all engines on demand.
14483 Changed the engine header files to avoid the duplication of some
14484 declarations (they differed!).
14488 * 'openssl engine' can now list capabilities.
14492 * Better error reporting in 'openssl engine'.
14496 * Never call load_dh_param(NULL) in s_server.
14500 * Add engine application. It can currently list engines by name and
14501 identity, and test if they are actually available.
14505 * Improve RPM specification file by forcing symbolic linking and making
14506 sure the installed documentation is also owned by root.root.
14508 *Damien Miller <djm@mindrot.org>*
14510 * Give the OpenSSL applications more possibilities to make use of
14511 keys (public as well as private) handled by engines.
14515 * Add OCSP code that comes from CertCo.
14519 * Add VMS support for the Rijndael code.
14523 * Added untested support for Nuron crypto accelerator.
14527 * Add support for external cryptographic devices. This code was
14528 previously distributed separately as the "engine" branch.
14530 *Geoff Thorpe, Richard Levitte*
14532 * Rework the filename-translation in the DSO code. It is now possible to
14533 have far greater control over how a "name" is turned into a filename
14534 depending on the operating environment and any oddities about the
14535 different shared library filenames on each system.
14539 * Support threads on FreeBSD-elf in Configure.
14543 * Fix for SHA1 assembly problem with MASM: it produces
14544 warnings about corrupt line number information when assembling
14545 with debugging information. This is caused by the overlapping
14548 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14551 NCONF_get_number() has no error checking at all. As a replacement,
14552 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14553 promoted strongly. The old NCONF_get_number is kept around for
14554 binary backward compatibility.
14555 Make it possible for methods to load from something other than a BIO,
14556 by providing a function pointer that is given a name instead of a BIO.
14557 For example, this could be used to load configuration data from an
14562 * Fix for non blocking accept BIOs. Added new I/O special reason
14563 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14564 with non blocking I/O was not possible because no retry code was
14565 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14570 * Added the beginnings of Rijndael support.
14574 * Fix for bug in DirectoryString mask setting. Add support for
14575 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14576 to allow certificate printing to more controllable, additional
14577 'certopt' option to 'x509' to allow new printing options to be
14582 * Clean old EAY MD5 hack from e_os.h.
14586 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14588 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14589 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14591 *Joe Orton, Steve Henson*
14593 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14595 * Fix additional bug revealed by the NISCC test suite:
14597 Stop bug triggering large recursion when presented with
14598 certain ASN.1 tags ([CVE-2003-0851])
14602 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14604 * Fix various bugs revealed by running the NISCC test suite:
14606 Stop out of bounds reads in the ASN1 code when presented with
14607 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14609 If verify callback ignores invalid public key errors don't try to check
14610 certificate signature with the NULL public key.
14614 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14615 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14620 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14621 extra data after the compression methods not only for TLS 1.0
14622 but also for SSL 3.0 (as required by the specification).
14624 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14626 * Change X509_certificate_type() to mark the key as exported/exportable
14627 when it's 512 *bits* long, not 512 bytes.
14631 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14633 * Countermeasure against the Klima-Pokorny-Rosa extension of
14634 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14635 a protocol version number mismatch like a decryption error
14636 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14640 * Turn on RSA blinding by default in the default implementation
14641 to avoid a timing attack. Applications that don't want it can call
14642 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14643 They would be ill-advised to do so in most cases.
14645 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14647 * Change RSA blinding code so that it works when the PRNG is not
14648 seeded (in this case, the secret RSA exponent is abused as
14649 an unpredictable seed -- if it is not unpredictable, there
14650 is no point in blinding anyway). Make RSA blinding thread-safe
14651 by remembering the creator's thread ID in rsa->blinding and
14652 having all other threads use local one-time blinding factors
14653 (this requires more computation than sharing rsa->blinding, but
14654 avoids excessive locking; and if an RSA object is not shared
14655 between threads, blinding will still be very fast).
14659 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14661 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14662 via timing by performing a MAC computation even if incorrect
14663 block cipher padding has been found. This is a countermeasure
14664 against active attacks where the attacker has to distinguish
14665 between bad padding and a MAC verification error. ([CVE-2003-0078])
14667 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14668 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14669 Martin Vuagnoux (EPFL, Ilion)*
14671 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14673 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14674 memory from its contents. This is done with a counter that will
14675 place alternating values in each byte. This can be used to solve
14676 two issues: 1) the removal of calls to memset() by highly optimizing
14677 compilers, and 2) cleansing with other values than 0, since those can
14678 be read through on certain media, for example a swap space on disk.
14682 * Bugfix: client side session caching did not work with external caching,
14683 because the session->cipher setting was not restored when reloading
14684 from the external cache. This problem was masked, when
14685 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14686 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14690 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14691 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14693 *Zeev Lieber <zeev-l@yahoo.com>*
14695 * Undo an undocumented change introduced in 0.9.6e which caused
14696 repeated calls to OpenSSL_add_all_ciphers() and
14697 OpenSSL_add_all_digests() to be ignored, even after calling
14702 * Change the default configuration reader to deal with last line not
14703 being properly terminated.
14707 * Change X509_NAME_cmp() so it applies the special rules on handling
14708 DN values that are of type PrintableString, as well as RDNs of type
14709 emailAddress where the value has the type ia5String.
14711 *stefank@valicert.com via Richard Levitte*
14713 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14714 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14715 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14716 the bitwise-OR of the two for use by the majority of applications
14717 wanting this behaviour, and update the docs. The documented
14718 behaviour and actual behaviour were inconsistent and had been
14719 changing anyway, so this is more a bug-fix than a behavioural
14722 *Geoff Thorpe, diagnosed by Nadav Har'El*
14724 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14725 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14729 * Fix initialization code race conditions in
14730 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14731 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14732 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14733 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14734 ssl2_get_cipher_by_char(),
14735 ssl3_get_cipher_by_char().
14737 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14739 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14740 the cached sessions are flushed, as the remove_cb() might use ex_data
14741 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14742 (see [openssl.org #212]).
14744 *Geoff Thorpe, Lutz Jaenicke*
14746 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14747 length, instead of the encoding length to d2i_ASN1_OBJECT.
14751 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14753 * [In 0.9.6g-engine release:]
14754 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14756 *Lynn Gazis <lgazis@rainbow.com>*
14758 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14760 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14761 and get fix the header length calculation.
14762 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14763 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14765 * Use proper error handling instead of 'assertions' in buffer
14766 overflow checks added in 0.9.6e. This prevents DoS (the
14767 assertions could call abort()).
14769 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14771 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14773 * Add various sanity checks to asn1_get_length() to reject
14774 the ASN1 length bytes if they exceed sizeof(long), will appear
14775 negative or the content length exceeds the length of the
14778 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14780 * Fix cipher selection routines: ciphers without encryption had no flags
14781 for the cipher strength set and where therefore not handled correctly
14782 by the selection routines (PR #130).
14786 * Fix EVP_dsa_sha macro.
14791 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14792 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14793 that was added in OpenSSL 0.9.6d.
14795 As the countermeasure turned out to be incompatible with some
14796 broken SSL implementations, the new option is part of SSL_OP_ALL.
14797 SSL_OP_ALL is usually employed when compatibility with weird SSL
14798 implementations is desired (e.g. '-bugs' option to 's_client' and
14799 's_server'), so the new option is automatically set in many
14804 * Changes in security patch:
14806 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14807 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14808 Air Force Materiel Command, USAF, under agreement number
14811 * Add various sanity checks to asn1_get_length() to reject
14812 the ASN1 length bytes if they exceed sizeof(long), will appear
14813 negative or the content length exceeds the length of the
14814 supplied buffer. ([CVE-2002-0659])
14816 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14818 * Assertions for various potential buffer overflows, not known to
14819 happen in practice.
14821 *Ben Laurie (CHATS)*
14823 * Various temporary buffers to hold ASCII versions of integers were
14824 too small for 64 bit platforms. ([CVE-2002-0655])
14825 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14827 * Remote buffer overflow in SSL3 protocol - an attacker could
14828 supply an oversized session ID to a client. ([CVE-2002-0656])
14830 *Ben Laurie (CHATS)*
14832 * Remote buffer overflow in SSL2 protocol - an attacker could
14833 supply an oversized client master key. ([CVE-2002-0656])
14835 *Ben Laurie (CHATS)*
14837 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14839 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14840 encoded as NULL) with id-dsa-with-sha1.
14842 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14844 * Check various `X509_...()` return values in `apps/req.c`.
14846 *Nils Larsch <nla@trustcenter.de>*
14848 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14849 an end-of-file condition would erroneously be flagged, when the CRLF
14850 was just at the end of a processed block. The bug was discovered when
14851 processing data through a buffering memory BIO handing the data to a
14852 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14853 <ptsekov@syntrex.com> and Nedelcho Stanev.
14857 * Implement a countermeasure against a vulnerability recently found
14858 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14859 before application data chunks to avoid the use of known IVs
14860 with data potentially chosen by the attacker.
14864 * Fix length checks in ssl3_get_client_hello().
14868 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14869 to prevent ssl3_read_internal() from incorrectly assuming that
14870 ssl3_read_bytes() found application data while handshake
14871 processing was enabled when in fact s->s3->in_read_app_data was
14872 merely automatically cleared during the initial handshake.
14874 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14876 * Fix object definitions for Private and Enterprise: they were not
14877 recognized in their shortname (=lowercase) representation. Extend
14878 obj_dat.pl to issue an error when using undefined keywords instead
14879 of silently ignoring the problem (Svenning Sorensen
14880 <sss@sss.dnsalias.net>).
14884 * Fix DH_generate_parameters() so that it works for 'non-standard'
14885 generators, i.e. generators other than 2 and 5. (Previously, the
14886 code did not properly initialise the 'add' and 'rem' values to
14887 BN_generate_prime().)
14889 In the new general case, we do not insist that 'generator' is
14890 actually a primitive root: This requirement is rather pointless;
14891 a generator of the order-q subgroup is just as good, if not
14896 * Map new X509 verification errors to alerts. Discovered and submitted by
14897 Tom Wu <tom@arcot.com>.
14901 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14902 returning non-zero before the data has been completely received
14903 when using non-blocking I/O.
14905 *Bodo Moeller; problem pointed out by John Hughes*
14907 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14909 *Ben Laurie, Lutz Jaenicke*
14911 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14912 Yoram Zahavi <YoramZ@gilian.com>).
14916 * Add information about CygWin 1.3 and on, and preserve proper
14917 configuration for the versions before that.
14919 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14921 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14922 check whether we deal with a copy of a session and do not delete from
14923 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14924 <izhar@checkpoint.com>.
14928 * Do not store session data into the internal session cache, if it
14929 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14930 flag is set). Proposed by Aslam <aslam@funk.com>.
14934 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14939 * [In 0.9.6d-engine release:]
14940 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14942 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14944 * Add the configuration target linux-s390x.
14946 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14948 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14949 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14950 variable as an indication that a ClientHello message has been
14951 received. As the flag value will be lost between multiple
14952 invocations of ssl3_accept when using non-blocking I/O, the
14953 function may not be aware that a handshake has actually taken
14954 place, thus preventing a new session from being added to the
14957 To avoid this problem, we now set s->new_session to 2 instead of
14958 using a local variable.
14960 *Lutz Jaenicke, Bodo Moeller*
14962 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14963 if the SSL_R_LENGTH_MISMATCH error is detected.
14965 *Geoff Thorpe, Bodo Moeller*
14967 * New 'shared_ldflag' column in Configure platform table.
14971 * Fix EVP_CIPHER_mode macro.
14973 *"Dan S. Camper" <dan@bti.net>*
14975 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14976 type, we must throw them away by setting rr->length to 0.
14978 *D P Chang <dpc@qualys.com>*
14980 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14982 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14983 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14984 worked incorrectly for those cases where range = `10..._2` and
14985 `3*range` is two bits longer than range.)
14989 * Only add signing time to PKCS7 structures if it is not already
14994 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14995 OBJ_ld_ce should be OBJ_id_ce.
14996 Also some ip-pda OIDs in crypto/objects/objects.txt were
14997 incorrect (cf. RFC 3039).
14999 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15001 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15002 returns early because it has nothing to do.
15004 *Andy Schneider <andy.schneider@bjss.co.uk>*
15006 * [In 0.9.6c-engine release:]
15007 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15009 *Andy Schneider <andy.schneider@bjss.co.uk>*
15011 * [In 0.9.6c-engine release:]
15012 Add support for Cryptographic Appliance's keyserver technology.
15013 (Use engine 'keyclient')
15015 *Cryptographic Appliances and Geoff Thorpe*
15017 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15018 is called via tools/c89.sh because arguments have to be
15019 rearranged (all '-L' options must appear before the first object
15022 *Richard Shapiro <rshapiro@abinitio.com>*
15024 * [In 0.9.6c-engine release:]
15025 Add support for Broadcom crypto accelerator cards, backported
15028 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15030 * [In 0.9.6c-engine release:]
15031 Add support for SureWare crypto accelerator cards from
15032 Baltimore Technologies. (Use engine 'sureware')
15034 *Baltimore Technologies and Mark Cox*
15036 * [In 0.9.6c-engine release:]
15037 Add support for crypto accelerator cards from Accelerated
15038 Encryption Processing, www.aep.ie. (Use engine 'aep')
15040 *AEP Inc. and Mark Cox*
15042 * Add a configuration entry for gcc on UnixWare.
15044 *Gary Benson <gbenson@redhat.com>*
15046 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15047 messages are stored in a single piece (fixed-length part and
15048 variable-length part combined) and fix various bugs found on the way.
15052 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15053 instead. BIO_gethostbyname() does not know what timeouts are
15054 appropriate, so entries would stay in cache even when they have
15056 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15058 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15059 faced with a pathologically small ClientHello fragment that does
15060 not contain client_version: Instead of aborting with an error,
15061 simply choose the highest available protocol version (i.e.,
15062 TLS 1.0 unless it is disabled). In practice, ClientHello
15063 messages are never sent like this, but this change gives us
15064 strictly correct behaviour at least for TLS.
15068 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15069 never resets s->method to s->ctx->method when called from within
15070 one of the SSL handshake functions.
15072 *Bodo Moeller; problem pointed out by Niko Baric*
15074 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15075 (sent using the client's version number) if client_version is
15076 smaller than the protocol version in use. Also change
15077 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15078 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15079 the client will at least see that alert.
15083 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15088 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15089 client receives HelloRequest while in a handshake.
15091 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15093 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15094 should end in 'break', not 'goto end' which circumvents various
15095 cleanups done in state SSL_ST_OK. But session related stuff
15096 must be disabled for SSL_ST_OK in the case that we just sent a
15099 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15100 before just sending a HelloRequest.
15102 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15104 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15105 reveal whether illegal block cipher padding was found or a MAC
15106 verification error occurred. (Neither SSLerr() codes nor alerts
15107 are directly visible to potential attackers, but the information
15108 may leak via logfiles.)
15110 Similar changes are not required for the SSL 2.0 implementation
15111 because the number of padding bytes is sent in clear for SSL 2.0,
15112 and the extra bytes are just ignored. However ssl/s2_pkt.c
15113 failed to verify that the purported number of padding bytes is in
15118 * Add OpenUNIX-8 support including shared libraries
15119 (Boyd Lynn Gerber <gerberb@zenez.com>).
15123 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15124 'wristwatch attack' using huge encoding parameters (cf.
15125 James H. Manger's CRYPTO 2001 paper). Note that the
15126 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15127 encoding parameters and hence was not vulnerable.
15131 * BN_sqr() bug fix.
15133 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15135 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15136 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15137 followed by modular reduction.
15139 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15141 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15142 equivalent based on BN_pseudo_rand() instead of BN_rand().
15146 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15147 This function was broken, as the check for a new client hello message
15148 to handle SGC did not allow these large messages.
15149 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15153 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15157 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15158 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15162 * Rework the configuration and shared library support for Tru64 Unix.
15163 The configuration part makes use of modern compiler features and
15164 still retains old compiler behavior for those that run older versions
15165 of the OS. The shared library support part includes a variant that
15166 uses the RPATH feature, and is available through the special
15167 configuration target "alpha-cc-rpath", which will never be selected
15170 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15172 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15173 with the same message size as in ssl3_get_certificate_request().
15174 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15175 messages might inadvertently be reject as too long.
15177 *Petr Lampa <lampa@fee.vutbr.cz>*
15179 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15183 * Modified SSL library such that the verify_callback that has been set
15184 specifically for an SSL object with SSL_set_verify() is actually being
15185 used. Before the change, a verify_callback set with this function was
15186 ignored and the verify_callback() set in the SSL_CTX at the time of
15187 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15188 to allow the necessary settings.
15192 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15193 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15194 done automatically (in contradiction to the requirements of the C
15195 standard). This made problems when used from OpenSSH.
15199 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15200 dh->length and always used
15202 BN_rand_range(priv_key, dh->p).
15204 BN_rand_range() is not necessary for Diffie-Hellman, and this
15205 specific range makes Diffie-Hellman unnecessarily inefficient if
15206 dh->length (recommended exponent length) is much smaller than the
15207 length of dh->p. We could use BN_rand_range() if the order of
15208 the subgroup was stored in the DH structure, but we only have
15213 BN_rand(priv_key, l, ...)
15215 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15222 RSA_eay_public_encrypt
15223 RSA_eay_private_decrypt
15224 RSA_eay_private_encrypt (signing)
15225 RSA_eay_public_decrypt (signature verification)
15227 (default implementations for RSA_public_encrypt,
15228 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15229 always reject numbers >= n.
15233 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15234 to synchronize access to 'locking_thread'. This is necessary on
15235 systems where access to 'locking_thread' (an 'unsigned long'
15236 variable) is not atomic.
15240 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15241 *before* setting the 'crypto_lock_rand' flag. The previous code had
15242 a race condition if 0 is a valid thread ID.
15244 *Travis Vitek <vitek@roguewave.com>*
15246 * Add support for shared libraries under Irix.
15248 *Albert Chin-A-Young <china@thewrittenword.com>*
15250 * Add configuration option to build on Linux on both big-endian and
15251 little-endian MIPS.
15253 *Ralf Baechle <ralf@uni-koblenz.de>*
15255 * Add the possibility to create shared libraries on HP-UX.
15259 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15261 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15262 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15263 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15264 PRNG state recovery was possible based on the output of
15265 one PRNG request appropriately sized to gain knowledge on
15266 'md' followed by enough consecutive 1-byte PRNG requests
15267 to traverse all of 'state'.
15269 1. When updating 'md_local' (the current thread's copy of 'md')
15270 during PRNG output generation, hash all of the previous
15271 'md_local' value, not just the half used for PRNG output.
15273 2. Make the number of bytes from 'state' included into the hash
15274 independent from the number of PRNG bytes requested.
15276 The first measure alone would be sufficient to avoid
15277 Markku-Juhani's attack. (Actually it had never occurred
15278 to me that the half of 'md_local' used for chaining was the
15279 half from which PRNG output bytes were taken -- I had always
15280 assumed that the secret half would be used.) The second
15281 measure makes sure that additional data from 'state' is never
15282 mixed into 'md_local' in small portions; this heuristically
15283 further strengthens the PRNG.
15287 * Fix crypto/bn/asm/mips3.s.
15291 * When only the key is given to "enc", the IV is undefined. Print out
15292 an error message in this case.
15296 * Handle special case when X509_NAME is empty in X509 printing routines.
15300 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15301 positive and less than q.
15305 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15306 used: it isn't thread safe and the add_lock_callback should handle
15309 *Paul Rose <Paul.Rose@bridge.com>*
15311 * Verify that incoming data obeys the block size in
15312 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15318 *Ulf Möller, Bodo Möller*
15320 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15321 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15322 when fixing the server behaviour for backwards-compatible 'client
15323 hello' messages. (Note that the attack is impractical against
15324 SSL 3.0 and TLS 1.0 anyway because length and version checking
15325 means that the probability of guessing a valid ciphertext is
15326 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15329 Before 0.9.5, the countermeasure (hide the error by generating a
15330 random 'decryption result') did not work properly because
15331 ERR_clear_error() was missing, meaning that SSL_get_error() would
15332 detect the supposedly ignored error.
15334 Both problems are now fixed.
15338 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15339 (previously it was 1024).
15343 * Fix for compatibility mode trust settings: ignore trust settings
15344 unless some valid trust or reject settings are present.
15348 * Fix for blowfish EVP: its a variable length cipher.
15352 * Fix various bugs related to DSA S/MIME verification. Handle missing
15353 parameters in DSA public key structures and return an error in the
15354 DSA routines if parameters are absent.
15358 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15359 in the current directory if neither $RANDFILE nor $HOME was set.
15360 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15361 caused some confusion to Windows users who haven't defined $HOME.
15362 Thus RAND_file_name() is changed again: e_os.h can define a
15363 DEFAULT_HOME, which will be used if $HOME is not set.
15364 For Windows, we use "C:"; on other platforms, we still require
15365 environment variables.
15367 * Move 'if (!initialized) RAND_poll()' into regions protected by
15368 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15369 having multiple threads call RAND_poll() concurrently.
15373 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15374 combination of a flag and a thread ID variable.
15375 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15376 flag), *other* threads can enter ssleay_add_bytes without obeying
15377 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15378 that they do not hold after the first thread unsets add_do_not_lock).
15382 * Change bctest again: '-x' expressions are not available in all
15383 versions of 'test'.
15387 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15389 * Fix a couple of memory leaks in PKCS7_dataDecode()
15391 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15393 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15394 the default extension for executables, if any. Also, make the perl
15395 scripts that use symlink() to test if it really exists and use "cp"
15396 if it doesn't. All this made OpenSSL compilable and installable in
15401 * Fix for asn1_GetSequence() for indefinite length constructed data.
15402 If SEQUENCE is length is indefinite just set c->slen to the total
15403 amount of data available.
15405 *Steve Henson, reported by shige@FreeBSD.org*
15407 *This change does not apply to 0.9.7.*
15409 * Change bctest to avoid here-documents inside command substitution
15410 (workaround for FreeBSD /bin/sh bug).
15411 For compatibility with Ultrix, avoid shell functions (introduced
15412 in the bctest version that searches along $PATH).
15416 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15417 with des_encrypt() defined on some operating systems, like Solaris
15422 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15423 On the Importance of Eliminating Errors in Cryptographic
15424 Computations, J. Cryptology 14 (2001) 2, 101-119,
15425 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15429 * MIPS assembler BIGNUM division bug fix.
15433 * Disabled incorrect Alpha assembler code.
15437 * Fix PKCS#7 decode routines so they correctly update the length
15438 after reading an EOC for the EXPLICIT tag.
15442 *This change does not apply to 0.9.7.*
15444 * Fix bug in PKCS#12 key generation routines. This was triggered
15445 if a 3DES key was generated with a 0 initial byte. Include
15446 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15447 (but broken) behaviour.
15451 * Enhance bctest to search for a working bc along $PATH and print
15454 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15456 * Fix memory leaks in err.c: free err_data string if necessary;
15457 don't write to the wrong index in ERR_set_error_data.
15461 * Implement ssl23_peek (analogous to ssl23_read), which previously
15466 * Replace rdtsc with `_emit` statements for VC++ version 5.
15468 *Jeremy Cooper <jeremy@baymoo.org>*
15470 * Make it possible to reuse SSLv2 sessions.
15474 * In copy_email() check for >= 0 as a return value for
15475 X509_NAME_get_index_by_NID() since 0 is a valid index.
15477 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15479 * Avoid coredump with unsupported or invalid public keys by checking if
15480 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15481 PKCS7_verify() fails with non detached data.
15485 * Don't use getenv in library functions when run as setuid/setgid.
15486 New function OPENSSL_issetugid().
15490 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15491 due to incorrect handling of multi-threading:
15493 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15495 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15497 3. Count how many times MemCheck_off() has been called so that
15498 nested use can be treated correctly. This also avoids
15499 inband-signalling in the previous code (which relied on the
15500 assumption that thread ID 0 is impossible).
15504 * Add "-rand" option also to s_client and s_server.
15508 * Fix CPU detection on Irix 6.x.
15509 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15510 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15512 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15517 *This change does not apply to 0.9.7.*
15519 * Use the cached encoding of an X509_NAME structure rather than
15520 copying it. This is apparently the reason for the libsafe "errors"
15521 but the code is actually correct.
15525 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15526 Bleichenbacher's DSA attack.
15527 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15528 to be set and top=0 forces the highest bit to be set; top=-1 is new
15529 and leaves the highest bit random.
15531 *Ulf Moeller, Bodo Moeller*
15533 * In the `NCONF_...`-based implementations for `CONF_...` queries
15534 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15535 a temporary CONF structure with the data component set to NULL
15536 (which gives segmentation faults in lh_retrieve).
15537 Instead, use NULL for the CONF pointer in CONF_get_string and
15538 CONF_get_number (which may use environment variables) and directly
15539 return NULL from CONF_get_section.
15543 * Fix potential buffer overrun for EBCDIC.
15547 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15548 keyUsage if basicConstraints absent for a CA.
15552 * Make SMIME_write_PKCS7() write mail header values with a format that
15553 is more generally accepted (no spaces before the semicolon), since
15554 some programs can't parse those values properly otherwise. Also make
15555 sure BIO's that break lines after each write do not create invalid
15560 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15561 macros previously used would not encode an empty SEQUENCE OF
15562 and break the signature.
15566 *This change does not apply to 0.9.7.*
15568 * Zero the premaster secret after deriving the master secret in
15573 * Add some EVP_add_digest_alias registrations (as found in
15574 OpenSSL_add_all_digests()) to SSL_library_init()
15575 aka OpenSSL_add_ssl_algorithms(). This provides improved
15576 compatibility with peers using X.509 certificates
15577 with unconventional AlgorithmIdentifier OIDs.
15581 * Fix for Irix with NO_ASM.
15583 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15585 * ./config script fixes.
15587 *Ulf Moeller, Richard Levitte*
15589 * Fix 'openssl passwd -1'.
15593 * Change PKCS12_key_gen_asc() so it can cope with non null
15594 terminated strings whose length is passed in the passlen
15595 parameter, for example from PEM callbacks. This was done
15596 by adding an extra length parameter to asc2uni().
15598 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15600 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15601 call failed, free the DSA structure.
15605 * Fix to uni2asc() to cope with zero length Unicode strings.
15606 These are present in some PKCS#12 files.
15610 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15611 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15612 when writing a 32767 byte record.
15614 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15616 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15617 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15619 (RSA objects have a reference count access to which is protected
15620 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15621 so they are meant to be shared between threads.)
15622 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15623 "Reddie, Steven" <Steven.Reddie@ca.com>*
15625 * Fix a deadlock in CRYPTO_mem_leaks().
15629 * Use better test patterns in bntest.
15633 * rand_win.c fix for Borland C.
15637 * BN_rshift bugfix for n == 0.
15641 * Add a 'bctest' script that checks for some known 'bc' bugs
15642 so that 'make test' does not abort just because 'bc' is broken.
15646 * Store verify_result within SSL_SESSION also for client side to
15647 avoid potential security hole. (Re-used sessions on the client side
15648 always resulted in verify_result==X509_V_OK, not using the original
15649 result of the server certificate verification.)
15653 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15654 SSL3_RT_APPLICATION_DATA, return 0.
15655 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15660 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15661 releases, have been re-implemented by renaming the previous
15662 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15663 and ssl3_read_internal, respectively, and adding 'peek' parameters
15664 to them. The new ssl[23]_{read,peek} functions are calls to
15665 ssl[23]_read_internal with the 'peek' flag set appropriately.
15666 A 'peek' parameter has also been added to ssl3_read_bytes, which
15667 does the actual work for ssl3_read_internal.
15671 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15672 the method-specific "init()" handler. Also clean up ex_data after
15673 calling the method-specific "finish()" handler. Previously, this was
15674 happening the other way round.
15678 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15679 The previous value, 12, was not always sufficient for BN_mod_exp().
15683 * Make sure that shared libraries get the internal name engine with
15684 the full version number and not just 0. This should mark the
15685 shared libraries as not backward compatible. Of course, this should
15686 be changed again when we can guarantee backward binary compatibility.
15690 * Fix typo in get_cert_by_subject() in by_dir.c
15692 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15694 * Rework the system to generate shared libraries:
15696 - Make note of the expected extension for the shared libraries and
15697 if there is a need for symbolic links from for example libcrypto.so.0
15698 to libcrypto.so.0.9.7. There is extended info in Configure for
15701 - Make as few rebuilds of the shared libraries as possible.
15703 - Still avoid linking the OpenSSL programs with the shared libraries.
15705 - When installing, install the shared libraries separately from the
15710 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15712 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15713 and not in SSL_clear because the latter is also used by the
15714 accept/connect functions; previously, the settings made by
15715 SSL_set_read_ahead would be lost during the handshake.
15717 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15719 * Correct util/mkdef.pl to be selective about disabled algorithms.
15720 Previously, it would create entries for disabled algorithms no
15725 * Added several new manual pages for SSL_* function.
15729 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15731 * In ssl23_get_client_hello, generate an error message when faced
15732 with an initial SSL 3.0/TLS record that is too small to contain the
15733 first two bytes of the ClientHello message, i.e. client_version.
15734 (Note that this is a pathologic case that probably has never happened
15735 in real life.) The previous approach was to use the version number
15736 from the record header as a substitute; but our protocol choice
15737 should not depend on that one because it is not authenticated
15738 by the Finished messages.
15742 * More robust randomness gathering functions for Windows.
15744 *Jeffrey Altman <jaltman@columbia.edu>*
15746 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15747 not set then we don't setup the error code for issuer check errors
15748 to avoid possibly overwriting other errors which the callback does
15749 handle. If an application does set the flag then we assume it knows
15750 what it is doing and can handle the new informational codes
15755 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15756 a general "ANY" type, as such it should be able to decode anything
15757 including tagged types. However it didn't check the class so it would
15758 wrongly interpret tagged types in the same way as their universal
15759 counterpart and unknown types were just rejected. Changed so that the
15760 tagged and unknown types are handled in the same way as a SEQUENCE:
15761 that is the encoding is stored intact. There is also a new type
15762 "V_ASN1_OTHER" which is used when the class is not universal, in this
15763 case we have no idea what the actual type is so we just lump them all
15768 * On VMS, stdout may very well lead to a file that is written to
15769 in a record-oriented fashion. That means that every write() will
15770 write a separate record, which will be read separately by the
15771 programs trying to read from it. This can be very confusing.
15773 The solution is to put a BIO filter in the way that will buffer
15774 text until a linefeed is reached, and then write everything a
15775 line at a time, so every record written will be an actual line,
15776 not chunks of lines and not (usually doesn't happen, but I've
15777 seen it once) several lines in one record. BIO_f_linebuffer() is
15780 Currently, it's a VMS-only method, because that's where it has
15781 been tested well enough.
15785 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15786 it can return incorrect results.
15787 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15788 but it was in 0.9.6-beta[12].)
15792 * Disable the check for content being present when verifying detached
15793 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15794 include zero length content when signing messages.
15798 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15799 BIO_ctrl (for BIO pairs).
15803 * Add DSO method for VMS.
15807 * Bug fix: Montgomery multiplication could produce results with the
15812 * Add RPM specification openssl.spec and modify it to build three
15813 packages. The default package contains applications, application
15814 documentation and run-time libraries. The devel package contains
15815 include files, static libraries and function documentation. The
15816 doc package contains the contents of the doc directory. The original
15817 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15821 * Add a large number of documentation files for many SSL routines.
15823 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15825 * Add a configuration entry for Sony News 4.
15827 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15829 * Don't set the two most significant bits to one when generating a
15830 random number < q in the DSA library.
15834 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15835 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15836 the underlying transport is blocking) if a handshake took place.
15837 (The default behaviour is needed by applications such as s_client
15838 and s_server that use select() to determine when to use SSL_read;
15839 but for applications that know in advance when to expect data, it
15840 just makes things more complicated.)
15844 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15849 * Add a few more EBCDIC conditionals that make `req` and `x509`
15850 work better on such systems.
15852 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15854 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15855 Update PKCS12_parse() so it copies the friendlyName and the
15856 keyid to the certificates aux info.
15860 * Fix bug in PKCS7_verify() which caused an infinite loop
15861 if there was more than one signature.
15863 *Sven Uszpelkat <su@celocom.de>*
15865 * Major change in util/mkdef.pl to include extra information
15866 about each symbol, as well as presenting variables as well
15867 as functions. This change means that there's n more need
15868 to rebuild the .num files when some algorithms are excluded.
15872 * Allow the verify time to be set by an application,
15873 rather than always using the current time.
15877 * Phase 2 verify code reorganisation. The certificate
15878 verify code now looks up an issuer certificate by a
15879 number of criteria: subject name, authority key id
15880 and key usage. It also verifies self signed certificates
15881 by the same criteria. The main comparison function is
15882 X509_check_issued() which performs these checks.
15884 Lot of changes were necessary in order to support this
15885 without completely rewriting the lookup code.
15887 Authority and subject key identifier are now cached.
15889 The LHASH 'certs' is X509_STORE has now been replaced
15890 by a STACK_OF(X509_OBJECT). This is mainly because an
15891 LHASH can't store or retrieve multiple objects with
15892 the same hash value.
15894 As a result various functions (which were all internal
15895 use only) have changed to handle the new X509_STORE
15896 structure. This will break anything that messed round
15897 with X509_STORE internally.
15899 The functions X509_STORE_add_cert() now checks for an
15900 exact match, rather than just subject name.
15902 The X509_STORE API doesn't directly support the retrieval
15903 of multiple certificates matching a given criteria, however
15904 this can be worked round by performing a lookup first
15905 (which will fill the cache with candidate certificates)
15906 and then examining the cache for matches. This is probably
15907 the best we can do without throwing out X509_LOOKUP
15908 entirely (maybe later...).
15910 The X509_VERIFY_CTX structure has been enhanced considerably.
15912 All certificate lookup operations now go via a get_issuer()
15913 callback. Although this currently uses an X509_STORE it
15914 can be replaced by custom lookups. This is a simple way
15915 to bypass the X509_STORE hackery necessary to make this
15916 work and makes it possible to use more efficient techniques
15917 in future. A very simple version which uses a simple
15918 STACK for its trusted certificate store is also provided
15919 using X509_STORE_CTX_trusted_stack().
15921 The verify_cb() and verify() callbacks now have equivalents
15922 in the X509_STORE_CTX structure.
15924 X509_STORE_CTX also has a 'flags' field which can be used
15925 to customise the verify behaviour.
15929 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15930 excludes S/MIME capabilities.
15934 * When a certificate request is read in keep a copy of the
15935 original encoding of the signed data and use it when outputting
15936 again. Signatures then use the original encoding rather than
15937 a decoded, encoded version which may cause problems if the
15938 request is improperly encoded.
15942 * For consistency with other BIO_puts implementations, call
15943 buffer_write(b, ...) directly in buffer_puts instead of calling
15946 In BIO_puts, increment b->num_write as in BIO_write.
15948 *Peter.Sylvester@EdelWeb.fr*
15950 * Fix BN_mul_word for the case where the word is 0. (We have to use
15951 BN_zero, we may not return a BIGNUM with an array consisting of
15952 words set to zero.)
15956 * Avoid calling abort() from within the library when problems are
15957 detected, except if preprocessor symbols have been defined
15958 (such as REF_CHECK, BN_DEBUG etc.).
15962 * New openssl application 'rsautl'. This utility can be
15963 used for low-level RSA operations. DER public key
15964 BIO/fp routines also added.
15968 * New Configure entry and patches for compiling on QNX 4.
15970 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15972 * A demo state-machine implementation was sponsored by
15973 Nuron (<http://www.nuron.com/>) and is now available in
15974 demos/state_machine.
15978 * New options added to the 'dgst' utility for signature
15979 generation and verification.
15983 * Unrecognized PKCS#7 content types are now handled via a
15984 catch all ASN1_TYPE structure. This allows unsupported
15985 types to be stored as a "blob" and an application can
15986 encode and decode it manually.
15990 * Fix various signed/unsigned issues to make a_strex.c
15991 compile under VC++.
15993 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15995 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15996 length if passed a buffer. ASN1_INTEGER_to_BN failed
15997 if passed a NULL BN and its argument was negative.
15999 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16001 * Modification to PKCS#7 encoding routines to output definite
16002 length encoding. Since currently the whole structures are in
16003 memory there's not real point in using indefinite length
16004 constructed encoding. However if OpenSSL is compiled with
16005 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16009 * Added BIO_vprintf() and BIO_vsnprintf().
16013 * Added more prefixes to parse for in the strings written
16014 through a logging bio, to cover all the levels that are available
16015 through syslog. The prefixes are now:
16017 PANIC, EMERG, EMR => LOG_EMERG
16018 ALERT, ALR => LOG_ALERT
16019 CRIT, CRI => LOG_CRIT
16020 ERROR, ERR => LOG_ERR
16021 WARNING, WARN, WAR => LOG_WARNING
16022 NOTICE, NOTE, NOT => LOG_NOTICE
16023 INFO, INF => LOG_INFO
16024 DEBUG, DBG => LOG_DEBUG
16026 and as before, if none of those prefixes are present at the
16027 beginning of the string, LOG_ERR is chosen.
16029 On Win32, the `LOG_*` levels are mapped according to this:
16031 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16032 LOG_WARNING => EVENTLOG_WARNING_TYPE
16033 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16037 * Made it possible to reconfigure with just the configuration
16038 argument "reconf" or "reconfigure". The command line arguments
16039 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16040 and are retrieved from there when reconfiguring.
16046 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16048 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16052 * The obj_dat.pl script was messing up the sorting of object
16053 names. The reason was that it compared the quoted version
16054 of strings as a result "OCSP" > "OCSP Signing" because
16055 " > SPACE. Changed script to store unquoted versions of
16056 names and add quotes on output. It was also omitting some
16057 names from the lookup table if they were given a default
16058 value (that is if SN is missing it is given the same
16059 value as LN and vice versa), these are now added on the
16060 grounds that if an object has a name we should be able to
16061 look it up. Finally added warning output when duplicate
16062 short or long names are found.
16066 * Changes needed for Tandem NSK.
16068 *Scott Uroff <scott@xypro.com>*
16070 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16071 RSA_padding_check_SSLv23(), special padding was never detected
16072 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16073 version rollback attacks was not effective.
16075 In s23_clnt.c, don't use special rollback-attack detection padding
16076 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16077 client; similarly, in s23_srvr.c, don't do the rollback check if
16078 SSL 2.0 is the only protocol enabled in the server.
16082 * Make it possible to get hexdumps of unprintable data with 'openssl
16083 asn1parse'. By implication, the functions ASN1_parse_dump() and
16084 BIO_dump_indent() are added.
16088 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16089 these print out strings and name structures based on various
16090 flags including RFC2253 support and proper handling of
16091 multibyte characters. Added options to the 'x509' utility
16092 to allow the various flags to be set.
16096 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16097 Also change the functions X509_cmp_current_time() and
16098 X509_gmtime_adj() work with an ASN1_TIME structure,
16099 this will enable certificates using GeneralizedTime in validity
16100 dates to be checked.
16104 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16105 negative public key encodings) on by default,
16106 NO_NEG_PUBKEY_BUG can be set to disable it.
16110 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16111 content octets. An i2c_ASN1_OBJECT is unnecessary because
16112 the encoding can be trivially obtained from the structure.
16116 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16117 not read locks (`CRYPTO_r_[un]lock`).
16121 * A first attempt at creating official support for shared
16122 libraries through configuration. I've kept it so the
16123 default is static libraries only, and the OpenSSL programs
16124 are always statically linked for now, but there are
16125 preparations for dynamic linking in place.
16126 This has been tested on Linux and Tru64.
16130 * Randomness polling function for Win9x, as described in:
16131 Peter Gutmann, Software Generation of Practically Strong
16136 * Fix so PRNG is seeded in req if using an already existing
16141 * New options to smime application. -inform and -outform
16142 allow alternative formats for the S/MIME message including
16143 PEM and DER. The -content option allows the content to be
16144 specified separately. This should allow things like Netscape
16145 form signing output easier to verify.
16149 * Fix the ASN1 encoding of tags using the 'long form'.
16153 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16154 STRING types. These convert content octets to and from the
16155 underlying type. The actual tag and length octets are
16156 already assumed to have been read in and checked. These
16157 are needed because all other string types have virtually
16158 identical handling apart from the tag. By having versions
16159 of the ASN1 functions that just operate on content octets
16160 IMPLICIT tagging can be handled properly. It also allows
16161 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16162 and ASN1_INTEGER are identical apart from the tag.
16166 * Change the handling of OID objects as follows:
16168 - New object identifiers are inserted in objects.txt, following
16169 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16170 - objects.pl is used to process obj_mac.num and create a new
16172 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16175 This is currently kind of a hack, and the perl code in objects.pl
16176 isn't very elegant, but it works as I intended. The simplest way
16177 to check that it worked correctly is to look in obj_dat.h and
16178 check the array nid_objs and make sure the objects haven't moved
16179 around (this is important!). Additions are OK, as well as
16180 consistent name changes.
16184 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16188 * Addition of the command line parameter '-rand file' to 'openssl req'.
16189 The given file adds to whatever has already been seeded into the
16190 random pool through the RANDFILE configuration file option or
16191 environment variable, or the default random state file.
16195 * mkstack.pl now sorts each macro group into lexical order.
16196 Previously the output order depended on the order the files
16197 appeared in the directory, resulting in needless rewriting
16202 * Patches to make OpenSSL compile under Win32 again. Mostly
16203 work arounds for the VC++ problem that it treats func() as
16204 func(void). Also stripped out the parts of mkdef.pl that
16205 added extra typesafe functions: these no longer exist.
16209 * Reorganisation of the stack code. The macros are now all
16210 collected in safestack.h . Each macro is defined in terms of
16211 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16212 DEBUG_SAFESTACK is now handled in terms of function casts,
16213 this has the advantage of retaining type safety without the
16214 use of additional functions. If DEBUG_SAFESTACK is not defined
16215 then the non typesafe macros are used instead. Also modified the
16216 mkstack.pl script to handle the new form. Needs testing to see
16217 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16218 the default if no major problems. Similar behaviour for ASN1_SET_OF
16219 and PKCS12_STACK_OF.
16223 * When some versions of IIS use the 'NET' form of private key the
16224 key derivation algorithm is different. Normally MD5(password) is
16225 used as a 128 bit RC4 key. In the modified case
16226 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16227 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16228 as the old Netscape_RSA functions except they have an additional
16229 'sgckey' parameter which uses the modified algorithm. Also added
16230 an -sgckey command line option to the rsa utility. Thanks to
16231 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16232 algorithm to openssl-dev.
16236 * The evp_local.h macros were using 'c.##kname' which resulted in
16237 invalid expansion on some systems (SCO 5.0.5 for example).
16238 Corrected to 'c.kname'.
16240 *Phillip Porch <root@theporch.com>*
16242 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16243 a STACK of email addresses from a certificate or request, these look
16244 in the subject name and the subject alternative name extensions and
16245 omit any duplicate addresses.
16249 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16250 This makes DSA verification about 2 % faster.
16254 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16255 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16256 plus overhead for 1024 bit moduli).
16257 This makes exponentiations about 0.5 % faster for 1024 bit
16258 exponents (as measured by "openssl speed rsa2048").
16262 * Rename memory handling macros to avoid conflicts with other
16264 Malloc => OPENSSL_malloc
16265 Malloc_locked => OPENSSL_malloc_locked
16266 Realloc => OPENSSL_realloc
16267 Free => OPENSSL_free
16271 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16272 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16276 * CygWin32 support.
16278 *John Jarvie <jjarvie@newsguy.com>*
16280 * The type-safe stack code has been rejigged. It is now only compiled
16281 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16282 by default all type-specific stack functions are "#define"d back to
16283 standard stack functions. This results in more streamlined output
16284 but retains the type-safety checking possibilities of the original
16289 * The STACK code has been cleaned up, and certain type declarations
16290 that didn't make a lot of sense have been brought in line. This has
16291 also involved a cleanup of sorts in safestack.h to more correctly
16292 map type-safe stack functions onto their plain stack counterparts.
16293 This work has also resulted in a variety of "const"ifications of
16294 lots of the code, especially `_cmp` operations which should normally
16295 be prototyped with "const" parameters anyway.
16299 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16300 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16301 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16302 where all of 'md' is used each time the PRNG is used, but 'state'
16303 is used only indexed by a cyclic counter. As entropy may not be
16304 well distributed from the beginning, 'md' is important as a
16305 chaining variable. However, the output function chains only half
16306 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16307 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16308 in all of 'state' being rewritten, with the new values depending
16309 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16313 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16314 the handshake is continued after ssl_verify_cert_chain();
16315 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16316 can lead to 'unexplainable' connection aborts later.
16318 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16320 * Major EVP API cipher revision.
16321 Add hooks for extra EVP features. This allows various cipher
16322 parameters to be set in the EVP interface. Support added for variable
16323 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16324 setting of RC2 and RC5 parameters.
16326 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16329 Remove lots of duplicated code from the EVP library. For example *every*
16330 cipher init() function handles the 'iv' in the same way according to the
16331 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16332 for CFB and OFB modes they zero ctx->num.
16334 New functionality allows removal of S/MIME code RC2 hack.
16336 Most of the routines have the same form and so can be declared in terms
16339 By shifting this to the top level EVP_CipherInit() it can be removed from
16340 all individual ciphers. If the cipher wants to handle IVs or keys
16341 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16344 Change lots of functions like EVP_EncryptUpdate() to now return a
16345 value: although software versions of the algorithms cannot fail
16346 any installed hardware versions can.
16350 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16351 this option is set, tolerate broken clients that send the negotiated
16352 protocol version number instead of the requested protocol version
16357 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16358 i.e. non-zero for export ciphersuites, zero otherwise.
16359 Previous versions had this flag inverted, inconsistent with
16360 rsa_tmp_cb (..._TMP_RSA_CB).
16362 *Bodo Moeller; problem reported by Amit Chopra*
16364 * Add missing DSA library text string. Work around for some IIS
16365 key files with invalid SEQUENCE encoding.
16369 * Add a document (doc/standards.txt) that list all kinds of standards
16370 and so on that are implemented in OpenSSL.
16374 * Enhance c_rehash script. Old version would mishandle certificates
16375 with the same subject name hash and wouldn't handle CRLs at all.
16376 Added -fingerprint option to crl utility, to support new c_rehash
16381 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16385 * Fix for SSL server purpose checking. Server checking was
16386 rejecting certificates which had extended key usage present
16387 but no ssl client purpose.
16389 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16391 * Make PKCS#12 code work with no password. The PKCS#12 spec
16392 is a little unclear about how a blank password is handled.
16393 Since the password in encoded as a BMPString with terminating
16394 double NULL a zero length password would end up as just the
16395 double NULL. However no password at all is different and is
16396 handled differently in the PKCS#12 key generation code. NS
16397 treats a blank password as zero length. MSIE treats it as no
16398 password on export: but it will try both on import. We now do
16399 the same: PKCS12_parse() tries zero length and no password if
16400 the password is set to "" or NULL (NULL is now a valid password:
16401 it wasn't before) as does the pkcs12 application.
16405 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16406 perror when PEM_read_bio_X509_REQ fails, the error message must
16407 be obtained from the error queue.
16411 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16412 it in ERR_remove_state if appropriate, and change ERR_get_state
16413 accordingly to avoid race conditions (this is necessary because
16414 thread_hash is no longer constant once set).
16418 * Bugfix for linux-elf makefile.one.
16422 * RSA_get_default_method() will now cause a default
16423 RSA_METHOD to be chosen if one doesn't exist already.
16424 Previously this was only set during a call to RSA_new()
16425 or RSA_new_method(NULL) meaning it was possible for
16426 RSA_get_default_method() to return NULL.
16430 * Added native name translation to the existing DSO code
16431 that will convert (if the flag to do so is set) filenames
16432 that are sufficiently small and have no path information
16433 into a canonical native form. Eg. "blah" converted to
16434 "libblah.so" or "blah.dll" etc.
16438 * New function ERR_error_string_n(e, buf, len) which is like
16439 ERR_error_string(e, buf), but writes at most 'len' bytes
16440 including the 0 terminator. For ERR_error_string_n, 'buf'
16443 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16445 * CONF library reworked to become more general. A new CONF
16446 configuration file reader "class" is implemented as well as a
16447 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16448 old `CONF_*` functions are still there, but are reimplemented to
16449 work in terms of the new functions. Also, a set of functions
16450 to handle the internal storage of the configuration data is
16451 provided to make it easier to write new configuration file
16452 reader "classes" (I can definitely see something reading a
16453 configuration file in XML format, for example), called `_CONF_*`,
16454 or "the configuration storage API"...
16456 The new configuration file reading functions are:
16458 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16459 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16461 NCONF_default, NCONF_WIN32
16463 NCONF_dump_fp, NCONF_dump_bio
16465 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16466 NCONF_new creates a new CONF object. This works in the same way
16467 as other interfaces in OpenSSL, like the BIO interface.
16468 `NCONF_dump_*` dump the internal storage of the configuration file,
16469 which is useful for debugging. All other functions take the same
16470 arguments as the old `CONF_*` functions with the exception of the
16471 first that must be a `CONF *` instead of a `LHASH *`.
16473 To make it easier to use the new classes with the old `CONF_*` functions,
16474 the function CONF_set_default_method is provided.
16478 * Add '-tls1' option to 'openssl ciphers', which was already
16479 mentioned in the documentation but had not been implemented.
16480 (This option is not yet really useful because even the additional
16481 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16485 * Initial DSO code added into libcrypto for letting OpenSSL (and
16486 OpenSSL-based applications) load shared libraries and bind to
16487 them in a portable way.
16489 *Geoff Thorpe, with contributions from Richard Levitte*
16491 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16493 * Make sure _lrotl and _lrotr are only used with MSVC.
16495 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16496 (the default implementation of RAND_status).
16498 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16499 to '-clrext' (= clear extensions), as intended and documented.
16500 *Bodo Moeller; inconsistency pointed out by Michael Attili
16501 <attili@amaxo.com>*
16503 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16504 was larger than the MD block size.
16506 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16508 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16509 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16510 using the passed key: if the passed key was a private key the result
16511 of X509_print(), for example, would be to print out all the private key
16516 * des_quad_cksum() byte order bug fix.
16517 *Ulf Möller, using the problem description in krb4-0.9.7, where
16518 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16520 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16523 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16525 * For easily testing in shell scripts whether some command
16526 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16527 returns with exit code 0 iff no command of the given name is available.
16528 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16529 the output goes to stdout and nothing is printed to stderr.
16530 Additional arguments are always ignored.
16532 Since for each cipher there is a command of the same name,
16533 the 'no-cipher' compilation switches can be tested this way.
16535 ('openssl no-XXX' is not able to detect pseudo-commands such
16536 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16540 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16544 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16545 is set; it will be thrown away anyway because each handshake creates
16547 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16548 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16549 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16550 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16554 * New s_client option -ign_eof: EOF at stdin is ignored, and
16555 'Q' and 'R' lose their special meanings (quit/renegotiate).
16556 This is part of what -quiet does; unlike -quiet, -ign_eof
16557 does not suppress any output.
16561 * Add compatibility options to the purpose and trust code. The
16562 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16563 accepts a certificate or CA, this was the previous behaviour,
16564 with all the associated security issues.
16566 X509_TRUST_COMPAT is the old trust behaviour: only and
16567 automatically trust self signed roots in certificate store. A
16568 new trust setting X509_TRUST_DEFAULT is used to specify that
16569 a purpose has no associated trust setting and it should instead
16570 use the value in the default purpose.
16574 * Fix the PKCS#8 DSA private key code so it decodes keys again
16575 and fix a memory leak.
16579 * In util/mkerr.pl (which implements 'make errors'), preserve
16580 reason strings from the previous version of the .c file, as
16581 the default to have only downcase letters (and digits) in
16582 automatically generated reasons codes is not always appropriate.
16586 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16587 using strerror. Previously, ERR_reason_error_string() returned
16588 library names as reason strings for SYSerr; but SYSerr is a special
16589 case where small numbers are errno values, not library numbers.
16593 * Add '-dsaparam' option to 'openssl dhparam' application. This
16594 converts DSA parameters into DH parameters. (When creating parameters,
16595 DSA_generate_parameters is used.)
16599 * Include 'length' (recommended exponent length) in C code generated
16600 by 'openssl dhparam -C'.
16604 * The second argument to set_label in perlasm was already being used
16605 so couldn't be used as a "file scope" flag. Moved to third argument
16610 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16611 instead of RAND_bytes for encryption IVs and salts.
16615 * Include RAND_status() into RAND_METHOD instead of implementing
16616 it only for md_rand.c Otherwise replacing the PRNG by calling
16617 RAND_set_rand_method would be impossible.
16621 * Don't let DSA_generate_key() enter an infinite loop if the random
16622 number generation fails.
16626 * New 'rand' application for creating pseudo-random output.
16630 * Added configuration support for Linux/IA64
16632 *Rolf Haberrecker <rolf@suse.de>*
16634 * Assembler module support for Mingw32.
16638 * Shared library support for HPUX (in shlib/).
16640 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16642 * Shared library support for Solaris gcc.
16644 *Lutz Behnke <behnke@trustcenter.de>*
16646 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16648 * PKCS7_encrypt() was adding text MIME headers twice because they
16649 were added manually and by SMIME_crlf_copy().
16653 * In bntest.c don't call BN_rand with zero bits argument.
16655 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16657 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16658 case was implemented. This caused BN_div_recp() to fail occasionally.
16662 * Add an optional second argument to the set_label() in the perl
16663 assembly language builder. If this argument exists and is set
16664 to 1 it signals that the assembler should use a symbol whose
16665 scope is the entire file, not just the current function. This
16666 is needed with MASM which uses the format label:: for this scope.
16668 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16670 * Change the ASN1 types so they are typedefs by default. Before
16671 almost all types were #define'd to ASN1_STRING which was causing
16672 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16677 * Change names of new functions to the new get1/get0 naming
16678 convention: After 'get1', the caller owns a reference count
16679 and has to call `..._free`; 'get0' returns a pointer to some
16680 data structure without incrementing reference counters.
16681 (Some of the existing 'get' functions increment a reference
16682 counter, some don't.)
16683 Similarly, 'set1' and 'add1' functions increase reference
16684 counters or duplicate objects.
16688 * Allow for the possibility of temp RSA key generation failure:
16689 the code used to assume it always worked and crashed on failure.
16693 * Fix potential buffer overrun problem in BIO_printf().
16694 *Ulf Möller, using public domain code by Patrick Powell; problem
16695 pointed out by David Sacerdote <das33@cornell.edu>*
16697 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16698 RAND_egd() and RAND_status(). In the command line application,
16699 the EGD socket can be specified like a seed file using RANDFILE
16704 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16705 Some CAs (e.g. Verisign) distribute certificates in this form.
16709 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16710 list to exclude them. This means that no special compilation option
16711 is needed to use anonymous DH: it just needs to be included in the
16716 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16717 EVP_MD_type. The old functionality is available in a new macro called
16718 EVP_MD_md(). Change code that uses it and update docs.
16722 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16723 where the `void *` argument is replaced by a function pointer argument.
16724 Previously `void *` was abused to point to functions, which works on
16725 many platforms, but is not correct. As these functions are usually
16726 called by macros defined in OpenSSL header files, most source code
16727 should work without changes.
16731 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16732 sections with information on -D... compiler switches used for
16733 compiling the library so that applications can see them. To enable
16734 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16735 must be defined. E.g.,
16736 #define OPENSSL_ALGORITHM_DEFINES
16737 #include <openssl/opensslconf.h>
16738 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16740 *Richard Levitte, Ulf and Bodo Möller*
16742 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16747 * Change the 'other' type in certificate aux info to a STACK_OF
16748 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16749 the required ASN1 format: arbitrary types determined by an OID.
16753 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16754 argument to 'req'. This is not because the function is newer or
16755 better than others it just uses the work 'NEW' in the certificate
16756 request header lines. Some software needs this.
16760 * Reorganise password command line arguments: now passwords can be
16761 obtained from various sources. Delete the PEM_cb function and make
16762 it the default behaviour: i.e. if the callback is NULL and the
16763 usrdata argument is not NULL interpret it as a null terminated pass
16764 phrase. If usrdata and the callback are NULL then the pass phrase
16765 is prompted for as usual.
16769 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16770 the support is automatically enabled. The resulting binaries will
16771 autodetect the card and use it if present.
16773 *Ben Laurie and Compaq Inc.*
16775 * Work around for Netscape hang bug. This sends certificate request
16776 and server done in one record. Since this is perfectly legal in the
16777 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16778 the bugs/SSLv3 entry for more info.
16782 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16786 * Add -rand argument to smime and pkcs12 applications and read/write
16791 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16795 * Add command line password options to the remaining applications.
16799 * Bug fix for BN_div_recp() for numerators with an even number of
16804 * More tests in bntest.c, and changed test_bn output.
16808 * ./config recognizes MacOS X now.
16812 * Bug fix for BN_div() when the first words of num and divisor are
16813 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16817 * Add support for various broken PKCS#8 formats, and command line
16818 options to produce them.
16822 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16823 get temporary BIGNUMs from a BN_CTX.
16827 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16832 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16833 include a #define from the old name to the new. The original intent
16834 was that statically linked binaries could for example just call
16835 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16836 link with digests. This never worked because SSLeay_add_all_digests()
16837 and SSLeay_add_all_ciphers() were in the same source file so calling
16838 one would link with the other. They are now in separate source files.
16842 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16846 * Use a less unusual form of the Miller-Rabin primality test (it used
16847 a binary algorithm for exponentiation integrated into the Miller-Rabin
16848 loop, our standard modexp algorithms are faster).
16852 * Support for the EBCDIC character set completed.
16854 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16856 * Source code cleanups: use const where appropriate, eliminate casts,
16857 use `void *` instead of `char *` in lhash.
16861 * Bugfix: ssl3_send_server_key_exchange was not restartable
16862 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16863 this the server could overwrite ephemeral keys that the client
16868 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16869 using 50 iterations of the Rabin-Miller test.
16871 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16872 iterations of the Rabin-Miller test as required by the appendix
16873 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16874 As BN_is_prime_fasttest includes trial division, DSA parameter
16875 generation becomes much faster.
16877 This implies a change for the callback functions in DSA_is_prime
16878 and DSA_generate_parameters: The callback function is called once
16879 for each positive witness in the Rabin-Miller test, not just
16880 occasionally in the inner loop; and the parameters to the
16881 callback function now provide an iteration count for the outer
16882 loop rather than for the current invocation of the inner loop.
16883 DSA_generate_parameters additionally can call the callback
16884 function with an 'iteration count' of -1, meaning that a
16885 candidate has passed the trial division test (when q is generated
16886 from an application-provided seed, trial division is skipped).
16890 * New function BN_is_prime_fasttest that optionally does trial
16891 division before starting the Rabin-Miller test and has
16892 an additional BN_CTX * argument (whereas BN_is_prime always
16893 has to allocate at least one BN_CTX).
16894 'callback(1, -1, cb_arg)' is called when a number has passed the
16895 trial division stage.
16899 * Fix for bug in CRL encoding. The validity dates weren't being handled
16904 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16908 * New function BN_pseudo_rand().
16912 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16913 bignum version of BN_from_montgomery() with the working code from
16914 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16919 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16920 made it impossible to use the same SSL_SESSION data structure in
16921 SSL2 clients in multiple threads.
16925 * The return value of RAND_load_file() no longer counts bytes obtained
16926 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16927 to seed the PRNG (previously an explicit byte count was required).
16929 *Ulf Möller, Bodo Möller*
16931 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16932 used `char *` instead of `void *` and had casts all over the place.
16936 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16940 * Retain source code compatibility for BN_prime_checks macro:
16941 BN_is_prime(..., BN_prime_checks, ...) now uses
16942 BN_prime_checks_for_size to determine the appropriate number of
16943 Rabin-Miller iterations.
16947 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16948 DH_CHECK_P_NOT_SAFE_PRIME.
16949 (Check if this is true? OpenPGP calls them "strong".)
16953 * Merge the functionality of "dh" and "gendh" programs into a new program
16954 "dhparam". The old programs are retained for now but will handle DH keys
16955 (instead of parameters) in future.
16959 * Make the ciphers, s_server and s_client programs check the return values
16960 when a new cipher list is set.
16964 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16965 ciphers. Before when the 56bit ciphers were enabled the sorting was
16968 The syntax for the cipher sorting has been extended to support sorting by
16969 cipher-strength (using the strength_bits hard coded in the tables).
16970 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16972 Fix a bug in the cipher-command parser: when supplying a cipher command
16973 string with an "undefined" symbol (neither command nor alphanumeric
16974 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16975 an error is flagged.
16977 Due to the strength-sorting extension, the code of the
16978 ssl_create_cipher_list() function was completely rearranged. I hope that
16979 the readability was also increased :-)
16981 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16983 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16984 for the first serial number and places 2 in the serial number file. This
16985 avoids problems when the root CA is created with serial number zero and
16986 the first user certificate has the same issuer name and serial number
16991 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16992 the new code. Add documentation for this stuff.
16996 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16997 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16998 structures and behave in an analogous way to the X509v3 functions:
16999 they shouldn't be called directly but wrapper functions should be used
17002 So we also now have some wrapper functions that call the X509at functions
17003 when passed certificate requests. (TO DO: similar things can be done with
17004 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17005 things. Some of these need some d2i or i2d and print functionality
17006 because they handle more complex structures.)
17010 * Add missing #ifndefs that caused missing symbols when building libssl
17011 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17012 NO_RSA in `ssl/s2*.c`.
17014 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17016 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17017 has a return value which indicates the quality of the random data
17018 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17019 error queue. New function RAND_pseudo_bytes() generates output that is
17020 guaranteed to be unique but not unpredictable. RAND_add is like
17021 RAND_seed, but takes an extra argument for an entropy estimate
17022 (RAND_seed always assumes full entropy).
17026 * Do more iterations of Rabin-Miller probable prime test (specifically,
17027 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17028 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17029 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17030 false-positive rate of at most 2^-80 for random input.
17034 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17038 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17039 in the 0.9.5 release), this returns the chain
17040 from an X509_CTX structure with a dup of the stack and all
17041 the X509 reference counts upped: so the stack will exist
17042 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17045 Also make SSL_SESSION_print() print out the verify return
17050 * Add manpage for the pkcs12 command. Also change the default
17051 behaviour so MAC iteration counts are used unless the new
17052 -nomaciter option is used. This improves file security and
17053 only older versions of MSIE (4.0 for example) need it.
17057 * Honor the no-xxx Configure options when creating .DEF files.
17061 * Add PKCS#10 attributes to field table: challengePassword,
17062 unstructuredName and unstructuredAddress. These are taken from
17063 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17064 international characters are used.
17066 More changes to X509_ATTRIBUTE code: allow the setting of types
17067 based on strings. Remove the 'loc' parameter when adding
17068 attributes because these will be a SET OF encoding which is sorted
17073 * Initial changes to the 'req' utility to allow request generation
17074 automation. This will allow an application to just generate a template
17075 file containing all the field values and have req construct the
17078 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17079 used all over the place including certificate requests and PKCS#7
17080 structures. They are currently handled manually where necessary with
17081 some primitive wrappers for PKCS#7. The new functions behave in a
17082 manner analogous to the X509 extension functions: they allow
17083 attributes to be looked up by NID and added.
17085 Later something similar to the X509V3 code would be desirable to
17086 automatically handle the encoding, decoding and printing of the
17087 more complex types. The string types like challengePassword can
17088 be handled by the string table functions.
17090 Also modified the multi byte string table handling. Now there is
17091 a 'global mask' which masks out certain types. The table itself
17092 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17093 is useful when for example there is only one permissible type
17094 (as in countryName) and using the mask might result in no valid
17099 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17100 SSL_get_peer_finished to allow applications to obtain the latest
17101 Finished messages sent to the peer or expected from the peer,
17102 respectively. (SSL_get_peer_finished is usually the Finished message
17103 actually received from the peer, otherwise the protocol will be aborted.)
17105 As the Finished message are message digests of the complete handshake
17106 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17107 be used for external authentication procedures when the authentication
17108 provided by SSL/TLS is not desired or is not enough.
17112 * Enhanced support for Alpha Linux is added. Now ./config checks if
17113 the host supports BWX extension and if Compaq C is present on the
17114 $PATH. Just exploiting of the BWX extension results in 20-30%
17115 performance kick for some algorithms, e.g. DES and RC4 to mention
17116 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17121 * Add support for MS "fast SGC". This is arguably a violation of the
17122 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17123 weak crypto and after checking the certificate is SGC a second one
17124 with strong crypto. MS SGC stops the first handshake after receiving
17125 the server certificate message and sends a second client hello. Since
17126 a server will typically do all the time consuming operations before
17127 expecting any further messages from the client (server key exchange
17128 is the most expensive) there is little difference between the two.
17130 To get OpenSSL to support MS SGC we have to permit a second client
17131 hello message after we have sent server done. In addition we have to
17132 reset the MAC if we do get this second client hello.
17136 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17137 if a DER encoded private key is RSA or DSA traditional format. Changed
17138 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17139 format DER encoded private key. Newer code should use PKCS#8 format which
17140 has the key type encoded in the ASN1 structure. Added DER private key
17141 support to pkcs8 application.
17145 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17146 ciphersuites has been selected (as required by the SSL 3/TLS 1
17147 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17148 is set, we interpret this as a request to violate the specification
17149 (the worst that can happen is a handshake failure, and 'correct'
17150 behaviour would result in a handshake failure anyway).
17154 * In SSL_CTX_add_session, take into account that there might be multiple
17155 SSL_SESSION structures with the same session ID (e.g. when two threads
17156 concurrently obtain them from an external cache).
17157 The internal cache can handle only one SSL_SESSION with a given ID,
17158 so if there's a conflict, we now throw out the old one to achieve
17163 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17164 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17165 some routines that use cipher OIDs: some ciphers do not have OIDs
17166 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17171 * Simplify the trust setting structure and code. Now we just have
17172 two sequences of OIDs for trusted and rejected settings. These will
17173 typically have values the same as the extended key usage extension
17174 and any application specific purposes.
17176 The trust checking code now has a default behaviour: it will just
17177 check for an object with the same NID as the passed id. Functions can
17178 be provided to override either the default behaviour or the behaviour
17179 for a given id. SSL client, server and email already have functions
17180 in place for compatibility: they check the NID and also return "trusted"
17181 if the certificate is self signed.
17185 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17186 traditional format into an EVP_PKEY structure.
17190 * Add a password callback function PEM_cb() which either prompts for
17191 a password if usr_data is NULL or otherwise assumes it is a null
17192 terminated password. Allow passwords to be passed on command line
17193 environment or config files in a few more utilities.
17197 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17198 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17199 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17200 Update documentation.
17204 * Support for ASN1 "NULL" type. This could be handled before by using
17205 ASN1_TYPE but there wasn't any function that would try to read a NULL
17206 and produce an error if it couldn't. For compatibility we also have
17207 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17208 don't allocate anything because they don't need to.
17212 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17215 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17217 * Rebuild of the memory allocation routines used by OpenSSL code and
17218 possibly others as well. The purpose is to make an interface that
17219 provide hooks so anyone can build a separate set of allocation and
17220 deallocation routines to be used by OpenSSL, for example memory
17221 pool implementations, or something else, which was previously hard
17222 since Malloc(), Realloc() and Free() were defined as macros having
17223 the values malloc, realloc and free, respectively (except for Win32
17224 compilations). The same is provided for memory debugging code.
17225 OpenSSL already comes with functionality to find memory leaks, but
17226 this gives people a chance to debug other memory problems.
17228 With these changes, a new set of functions and macros have appeared:
17230 CRYPTO_set_mem_debug_functions() [F]
17231 CRYPTO_get_mem_debug_functions() [F]
17232 CRYPTO_dbg_set_options() [F]
17233 CRYPTO_dbg_get_options() [F]
17234 CRYPTO_malloc_debug_init() [M]
17236 The memory debug functions are NULL by default, unless the library
17237 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17238 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17239 gives the standard debugging functions that come with OpenSSL) or
17240 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17241 provided by the library user) must be used. When the standard
17242 debugging functions are used, CRYPTO_dbg_set_options can be used to
17243 request additional information:
17244 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17245 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17247 Also, things like CRYPTO_set_mem_functions will always give the
17248 expected result (the new set of functions is used for allocation
17249 and deallocation) at all times, regardless of platform and compiler
17252 To finish it up, some functions that were never use in any other
17253 way than through macros have a new API and new semantic:
17255 CRYPTO_dbg_malloc()
17256 CRYPTO_dbg_realloc()
17259 All macros of value have retained their old syntax.
17261 *Richard Levitte and Bodo Moeller*
17263 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17264 ordering of SMIMECapabilities wasn't in "strength order" and there
17265 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17270 * Some ASN1 types with illegal zero length encoding (INTEGER,
17271 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17273 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17275 * Merge in my S/MIME library for OpenSSL. This provides a simple
17276 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17277 functionality to handle multipart/signed properly) and a utility
17278 called 'smime' to call all this stuff. This is based on code I
17279 originally wrote for Celo who have kindly allowed it to be
17280 included in OpenSSL.
17284 * Add variants des_set_key_checked and des_set_key_unchecked of
17285 des_set_key (aka des_key_sched). Global variable des_check_key
17286 decides which of these is called by des_set_key; this way
17287 des_check_key behaves as it always did, but applications and
17288 the library itself, which was buggy for des_check_key == 1,
17289 have a cleaner way to pick the version they need.
17293 * New function PKCS12_newpass() which changes the password of a
17298 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17299 dynamic mix. In both cases the ids can be used as an index into the
17300 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17301 functions so they accept a list of the field values and the
17302 application doesn't need to directly manipulate the X509_TRUST
17307 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17312 * Modify the way the V3 extension code looks up extensions. This now
17313 works in a similar way to the object code: we have some "standard"
17314 extensions in a static table which is searched with OBJ_bsearch()
17315 and the application can add dynamic ones if needed. The file
17316 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17317 updated whenever a new extension is added to the core code and kept
17318 in ext_nid order. There is a simple program 'tabtest.c' which checks
17319 this. New extensions are not added too often so this file can readily
17320 be maintained manually.
17322 There are two big advantages in doing things this way. The extensions
17323 can be looked up immediately and no longer need to be "added" using
17324 X509V3_add_standard_extensions(): this function now does nothing.
17325 Side note: I get *lots* of email saying the extension code doesn't
17326 work because people forget to call this function.
17327 Also no dynamic allocation is done unless new extensions are added:
17328 so if we don't add custom extensions there is no need to call
17329 X509V3_EXT_cleanup().
17333 * Modify enc utility's salting as follows: make salting the default. Add a
17334 magic header, so unsalted files fail gracefully instead of just decrypting
17335 to garbage. This is because not salting is a big security hole, so people
17336 should be discouraged from doing it.
17340 * Fixes and enhancements to the 'x509' utility. It allowed a message
17341 digest to be passed on the command line but it only used this
17342 parameter when signing a certificate. Modified so all relevant
17343 operations are affected by the digest parameter including the
17344 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17345 DSA key was used because it didn't fix the digest.
17349 * Initial certificate chain verify code. Currently tests the untrusted
17350 certificates for consistency with the verify purpose (which is set
17351 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17353 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17354 this is because it will reject chains with invalid extensions whereas
17355 every previous version of OpenSSL and SSLeay made no checks at all.
17357 Trust code: checks the root CA for the relevant trust settings. Trust
17358 settings have an initial value consistent with the verify purpose: e.g.
17359 if the verify purpose is for SSL client use it expects the CA to be
17360 trusted for SSL client use. However the default value can be changed to
17361 permit custom trust settings: one example of this would be to only trust
17362 certificates from a specific "secure" set of CAs.
17364 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17365 which should be used for version portability: especially since the
17366 verify structure is likely to change more often now.
17368 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17369 to set them. If not set then assume SSL clients will verify SSL servers
17372 Two new options to the verify program: -untrusted allows a set of
17373 untrusted certificates to be passed in and -purpose which sets the
17374 intended purpose of the certificate. If a purpose is set then the
17375 new chain verify code is used to check extension consistency.
17379 * Support for the authority information access extension.
17383 * Modify RSA and DSA PEM read routines to transparently handle
17384 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17385 public keys in a format compatible with certificate
17386 SubjectPublicKeyInfo structures. Unfortunately there were already
17387 functions called *_PublicKey_* which used various odd formats so
17388 these are retained for compatibility: however the DSA variants were
17389 never in a public release so they have been deleted. Changed dsa/rsa
17390 utilities to handle the new format: note no releases ever handled public
17391 keys so we should be OK.
17393 The primary motivation for this change is to avoid the same fiasco
17394 that dogs private keys: there are several incompatible private key
17395 formats some of which are standard and some OpenSSL specific and
17396 require various evil hacks to allow partial transparent handling and
17397 even then it doesn't work with DER formats. Given the option anything
17398 other than PKCS#8 should be dumped: but the other formats have to
17399 stay in the name of compatibility.
17401 With public keys and the benefit of hindsight one standard format
17402 is used which works with EVP_PKEY, RSA or DSA structures: though
17403 it clearly returns an error if you try to read the wrong kind of key.
17405 Added a -pubkey option to the 'x509' utility to output the public key.
17406 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17407 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17408 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17409 that do the same as the `EVP_PKEY_assign_*()` except they up the
17410 reference count of the added key (they don't "swallow" the
17415 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17416 CRLs would fail if the file contained no certificates or no CRLs:
17417 added a new function to read in both types and return the number
17418 read: this means that if none are read it will be an error. The
17419 DER versions of the certificate and CRL reader would always fail
17420 because it isn't possible to mix certificates and CRLs in DER format
17421 without choking one or the other routine. Changed this to just read
17422 a certificate: this is the best we can do. Also modified the code
17423 in `apps/verify.c` to take notice of return codes: it was previously
17424 attempting to read in certificates from NULL pointers and ignoring
17425 any errors: this is one reason why the cert and CRL reader seemed
17426 to work. It doesn't check return codes from the default certificate
17427 routines: these may well fail if the certificates aren't installed.
17431 * Code to support otherName option in GeneralName.
17435 * First update to verify code. Change the verify utility
17436 so it warns if it is passed a self signed certificate:
17437 for consistency with the normal behaviour. X509_verify
17438 has been modified to it will now verify a self signed
17439 certificate if *exactly* the same certificate appears
17440 in the store: it was previously impossible to trust a
17441 single self signed certificate. This means that:
17442 openssl verify ss.pem
17443 now gives a warning about a self signed certificate but
17444 openssl verify -CAfile ss.pem ss.pem
17449 * For servers, store verify_result in SSL_SESSION data structure
17450 (and add it to external session representation).
17451 This is needed when client certificate verifications fails,
17452 but an application-provided verification callback (set by
17453 SSL_CTX_set_cert_verify_callback) allows accepting the session
17454 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17455 but returns 1): When the session is reused, we have to set
17456 ssl->verify_result to the appropriate error code to avoid
17459 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17461 * Fix a bug in the new PKCS#7 code: it didn't consider the
17462 case in PKCS7_dataInit() where the signed PKCS7 structure
17463 didn't contain any existing data because it was being created.
17465 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17467 * Add a salt to the key derivation routines in enc.c. This
17468 forms the first 8 bytes of the encrypted file. Also add a
17469 -S option to allow a salt to be input on the command line.
17473 * New function X509_cmp(). Oddly enough there wasn't a function
17474 to compare two certificates. We do this by working out the SHA1
17475 hash and comparing that. X509_cmp() will be needed by the trust
17480 * SSL_get1_session() is like SSL_get_session(), but increments
17481 the reference count in the SSL_SESSION returned.
17483 *Geoff Thorpe <geoff@eu.c2.net>*
17485 * Fix for 'req': it was adding a null to request attributes.
17486 Also change the X509_LOOKUP and X509_INFO code to handle
17487 certificate auxiliary information.
17491 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17496 * Add the possibility to add extra information to the memory leak
17497 detecting output, to form tracebacks, showing from where each
17498 allocation was originated: CRYPTO_push_info("constant string") adds
17499 the string plus current file name and line number to a per-thread
17500 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17501 is like calling CYRPTO_pop_info() until the stack is empty.
17502 Also updated memory leak detection code to be multi-thread-safe.
17506 * Add options -text and -noout to pkcs7 utility and delete the
17507 encryption options which never did anything. Update docs.
17511 * Add options to some of the utilities to allow the pass phrase
17512 to be included on either the command line (not recommended on
17513 OSes like Unix) or read from the environment. Update the
17514 manpages and fix a few bugs.
17518 * Add a few manpages for some of the openssl commands.
17522 * Fix the -revoke option in ca. It was freeing up memory twice,
17523 leaking and not finding already revoked certificates.
17527 * Extensive changes to support certificate auxiliary information.
17528 This involves the use of X509_CERT_AUX structure and X509_AUX
17529 functions. An X509_AUX function such as PEM_read_X509_AUX()
17530 can still read in a certificate file in the usual way but it
17531 will also read in any additional "auxiliary information". By
17532 doing things this way a fair degree of compatibility can be
17533 retained: existing certificates can have this information added
17534 using the new 'x509' options.
17536 Current auxiliary information includes an "alias" and some trust
17537 settings. The trust settings will ultimately be used in enhanced
17538 certificate chain verification routines: currently a certificate
17539 can only be trusted if it is self signed and then it is trusted
17544 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17545 The problem was that one of the replacement routines had not been working
17546 since SSLeay releases. For now the offending routine has been replaced
17547 with non-optimised assembler. Even so, this now gives around 95%
17548 performance improvement for 1024 bit RSA signs.
17552 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17553 handling. Most clients have the effective key size in bits equal to
17554 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17555 A few however don't do this and instead use the size of the decrypted key
17556 to determine the RC2 key length and the AlgorithmIdentifier to determine
17557 the effective key length. In this case the effective key length can still
17558 be 40 bits but the key length can be 168 bits for example. This is fixed
17559 by manually forcing an RC2 key into the EVP_PKEY structure because the
17560 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17561 the key length and effective key length are equal.
17565 * Add a bunch of functions that should simplify the creation of
17566 X509_NAME structures. Now you should be able to do:
17567 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17568 and have it automatically work out the correct field type and fill in
17569 the structures. The more adventurous can try:
17570 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17571 and it will (hopefully) work out the correct multibyte encoding.
17575 * Change the 'req' utility to use the new field handling and multibyte
17576 copy routines. Before the DN field creation was handled in an ad hoc
17577 way in req, ca, and x509 which was rather broken and didn't support
17578 BMPStrings or UTF8Strings. Since some software doesn't implement
17579 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17580 using the dirstring_type option. See the new comment in the default
17581 openssl.cnf for more info.
17585 * Make crypto/rand/md_rand.c more robust:
17586 - Assure unique random numbers after fork().
17587 - Make sure that concurrent threads access the global counter and
17588 md serializably so that we never lose entropy in them
17589 or use exactly the same state in multiple threads.
17590 Access to the large state is not always serializable because
17591 the additional locking could be a performance killer, and
17592 md should be large enough anyway.
17596 * New file `apps/app_rand.c` with commonly needed functionality
17597 for handling the random seed file.
17599 Use the random seed file in some applications that previously did not:
17601 dsaparam -genkey (which also ignored its '-rand' option),
17604 x509 (when signing).
17605 Except on systems with /dev/urandom, it is crucial to have a random
17606 seed file at least for key creation, DSA signing, and for DH exchanges;
17607 for RSA signatures we could do without one.
17609 gendh and gendsa (unlike genrsa) used to read only the first byte
17610 of each file listed in the '-rand' option. The function as previously
17611 found in genrsa is now in app_rand.c and is used by all programs
17612 that support '-rand'.
17616 * In RAND_write_file, use mode 0600 for creating files;
17617 don't just chmod when it may be too late.
17621 * Report an error from X509_STORE_load_locations
17622 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17626 * New function ASN1_mbstring_copy() this copies a string in either
17627 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17628 into an ASN1_STRING type. A mask of permissible types is passed
17629 and it chooses the "minimal" type to use or an error if not type
17634 * Add function equivalents to the various macros in asn1.h. The old
17635 macros are retained with an `M_` prefix. Code inside the library can
17636 use the `M_` macros. External code (including the openssl utility)
17637 should *NOT* in order to be "shared library friendly".
17641 * Add various functions that can check a certificate's extensions
17642 to see if it usable for various purposes such as SSL client,
17643 server or S/MIME and CAs of these types. This is currently
17644 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17645 verification. Also added a -purpose flag to x509 utility to
17646 print out all the purposes.
17650 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17655 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17656 for, obtain and decode and extension and obtain its critical flag.
17657 This allows all the necessary extension code to be handled in a
17658 single function call.
17662 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17663 platforms. See crypto/rc4/rc4_enc.c for further details.
17667 * New -noout option to asn1parse. This causes no output to be produced
17668 its main use is when combined with -strparse and -out to extract data
17669 from a file (which may not be in ASN.1 format).
17673 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17674 when producing the local key id.
17676 *Richard Levitte <levitte@stacken.kth.se>*
17678 * New option -dhparam in s_server. This allows a DH parameter file to be
17679 stated explicitly. If it is not stated then it tries the first server
17680 certificate file. The previous behaviour hard coded the filename
17685 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17686 a public key to be input or output. For example:
17687 openssl rsa -in key.pem -pubout -out pubkey.pem
17688 Also added necessary DSA public key functions to handle this.
17692 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17693 in the message. This was handled by allowing
17694 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17696 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17698 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17699 to the end of the strings whereas this didn't. This would cause problems
17700 if strings read with d2i_ASN1_bytes() were later modified.
17702 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17704 * Fix for base64 decode bug. When a base64 bio reads only one line of
17705 data and it contains EOF it will end up returning an error. This is
17706 caused by input 46 bytes long. The cause is due to the way base64
17707 BIOs find the start of base64 encoded data. They do this by trying a
17708 trial decode on each line until they find one that works. When they
17709 do a flag is set and it starts again knowing it can pass all the
17710 data directly through the decoder. Unfortunately it doesn't reset
17711 the context it uses. This means that if EOF is reached an attempt
17712 is made to pass two EOFs through the context and this causes the
17713 resulting error. This can also cause other problems as well. As is
17714 usual with these problems it takes *ages* to find and the fix is
17715 trivial: move one line.
17717 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17719 * Ugly workaround to get s_client and s_server working under Windows. The
17720 old code wouldn't work because it needed to select() on sockets and the
17721 tty (for keypresses and to see if data could be written). Win32 only
17722 supports select() on sockets so we select() with a 1s timeout on the
17723 sockets and then see if any characters are waiting to be read, if none
17724 are present then we retry, we also assume we can always write data to
17725 the tty. This isn't nice because the code then blocks until we've
17726 received a complete line of data and it is effectively polling the
17727 keyboard at 1s intervals: however it's quite a bit better than not
17728 working at all :-) A dedicated Windows application might handle this
17729 with an event loop for example.
17733 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17734 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17735 will be called when RSA_sign() and RSA_verify() are used. This is useful
17736 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17737 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17738 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17739 This necessitated the support of an extra signature type NID_md5_sha1
17740 for SSL signatures and modifications to the SSL library to use it instead
17741 of calling RSA_public_decrypt() and RSA_private_encrypt().
17745 * Add new -verify -CAfile and -CApath options to the crl program, these
17746 will lookup a CRL issuers certificate and verify the signature in a
17747 similar way to the verify program. Tidy up the crl program so it
17748 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17749 less strict. It will now permit CRL extensions even if it is not
17750 a V2 CRL: this will allow it to tolerate some broken CRLs.
17754 * Initialize all non-automatic variables each time one of the openssl
17755 sub-programs is started (this is necessary as they may be started
17756 multiple times from the "OpenSSL>" prompt).
17758 *Lennart Bang, Bodo Moeller*
17760 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17761 removing all other RSA functionality (this is what NO_RSA does). This
17762 is so (for example) those in the US can disable those operations covered
17763 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17768 * Non-copying interface to BIO pairs.
17769 (still largely untested)
17773 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17774 ASCII string. This was handled independently in various places before.
17778 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17779 UTF8 strings a character at a time.
17783 * Use client_version from client hello to select the protocol
17784 (s23_srvr.c) and for RSA client key exchange verification
17785 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17789 * Add various utility functions to handle SPKACs, these were previously
17790 handled by poking round in the structure internals. Added new function
17791 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17792 print, verify and generate SPKACs. Based on an original idea from
17793 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17797 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17801 * Allow the config file extension section to be overwritten on the
17802 command line. Based on an original idea from Massimiliano Pala
17803 <madwolf@comune.modena.it>. The new option is called -extensions
17804 and can be applied to ca, req and x509. Also -reqexts to override
17805 the request extensions in req and -crlexts to override the crl extensions
17810 * Add new feature to the SPKAC handling in ca. Now you can include
17811 the same field multiple times by preceding it by "XXXX." for example:
17814 this is the same syntax as used in the req config file.
17818 * Allow certificate extensions to be added to certificate requests. These
17819 are specified in a 'req_extensions' option of the req section of the
17820 config file. They can be printed out with the -text option to req but
17821 are otherwise ignored at present.
17825 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17826 data read consists of only the final block it would not decrypted because
17827 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17828 A misplaced 'break' also meant the decrypted final block might not be
17829 copied until the next read.
17833 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17834 a few extra parameters to the DH structure: these will be useful if
17835 for example we want the value of 'q' or implement X9.42 DH.
17839 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17840 provides hooks that allow the default DSA functions or functions on a
17841 "per key" basis to be replaced. This allows hardware acceleration and
17842 hardware key storage to be handled without major modification to the
17843 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17844 associated functions.
17848 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17849 as "read only": it can't be written to and the buffer it points to will
17850 not be freed. Reading from a read only BIO is much more efficient than
17851 a normal memory BIO. This was added because there are several times when
17852 an area of memory needs to be read from a BIO. The previous method was
17853 to create a memory BIO and write the data to it, this results in two
17854 copies of the data and an O(n^2) reading algorithm. There is a new
17855 function BIO_new_mem_buf() which creates a read only memory BIO from
17856 an area of memory. Also modified the PKCS#7 routines to use read only
17861 * Bugfix: ssl23_get_client_hello did not work properly when called in
17862 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17863 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17864 but a retry condition occurred while trying to read the rest.
17868 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17869 NID_pkcs7_encrypted by default: this was wrong since this should almost
17870 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17871 the encrypted data type: this is a more sensible place to put it and it
17872 allows the PKCS#12 code to be tidied up that duplicated this
17877 * Changed obj_dat.pl script so it takes its input and output files on
17878 the command line. This should avoid shell escape redirection problems
17883 * Initial support for certificate extension requests, these are included
17884 in things like Xenroll certificate requests. Included functions to allow
17885 extensions to be obtained and added.
17889 * -crlf option to s_client and s_server for sending newlines as
17890 CRLF (as required by many protocols).
17894 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17896 * Install libRSAglue.a when OpenSSL is built with RSAref.
17898 *Ralf S. Engelschall*
17900 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17902 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17904 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17909 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17910 DH parameters/keys (q is lost during that conversion, but the resulting
17911 DH parameters contain its length).
17913 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17914 much faster than DH_generate_parameters (which creates parameters
17915 where `p = 2*q + 1`), and also the smaller q makes DH computations
17916 much more efficient (160-bit exponentiation instead of 1024-bit
17917 exponentiation); so this provides a convenient way to support DHE
17918 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17919 utter importance to use
17920 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17922 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17923 when such DH parameters are used, because otherwise small subgroup
17924 attacks may become possible!
17928 * Avoid memory leak in i2d_DHparams.
17932 * Allow the -k option to be used more than once in the enc program:
17933 this allows the same encrypted message to be read by multiple recipients.
17937 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17938 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17939 it will always use the numerical form of the OID, even if it has a short
17944 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17945 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17946 otherwise bn_mod_exp was called. In the case of hardware keys for example
17947 no private key components need be present and it might store extra data
17948 in the RSA structure, which cannot be accessed from bn_mod_exp.
17949 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17950 private key operations.
17954 * Added support for SPARC Linux.
17958 * pem_password_cb function type incompatibly changed from
17959 typedef int pem_password_cb(char *buf, int size, int rwflag);
17961 ....(char *buf, int size, int rwflag, void *userdata);
17962 so that applications can pass data to their callbacks:
17963 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17964 additional void * argument, which is just handed through whenever
17965 the password callback is called.
17967 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17969 New function SSL_CTX_set_default_passwd_cb_userdata.
17971 Compatibility note: As many C implementations push function arguments
17972 onto the stack in reverse order, the new library version is likely to
17973 interoperate with programs that have been compiled with the old
17974 pem_password_cb definition (PEM_whatever takes some data that
17975 happens to be on the stack as its last argument, and the callback
17976 just ignores this garbage); but there is no guarantee whatsoever that
17979 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17980 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17981 problems not only on Windows, but also on some Unix platforms.
17982 To avoid problematic command lines, these definitions are now in an
17983 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17984 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17988 * MIPS III/IV assembler module is reimplemented.
17992 * More DES library cleanups: remove references to srand/rand and
17993 delete an unused file.
17997 * Add support for the free Netwide assembler (NASM) under Win32,
17998 since not many people have MASM (ml) and it can be hard to obtain.
17999 This is currently experimental but it seems to work OK and pass all
18000 the tests. Check out INSTALL.W32 for info.
18004 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18005 without temporary keys kept an extra copy of the server key,
18006 and connections with temporary keys did not free everything in case
18011 * New function RSA_check_key and new openssl rsa option -check
18012 for verifying the consistency of RSA keys.
18014 *Ulf Moeller, Bodo Moeller*
18016 * Various changes to make Win32 compile work:
18017 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18018 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18019 comparison" warnings.
18020 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18024 * Add a debugging option to PKCS#5 v2 key generation function: when
18025 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18026 derived keys are printed to stderr.
18030 * Copy the flags in ASN1_STRING_dup().
18032 *Roman E. Pavlov <pre@mo.msk.ru>*
18034 * The x509 application mishandled signing requests containing DSA
18035 keys when the signing key was also DSA and the parameters didn't match.
18037 It was supposed to omit the parameters when they matched the signing key:
18038 the verifying software was then supposed to automatically use the CA's
18039 parameters if they were absent from the end user certificate.
18041 Omitting parameters is no longer recommended. The test was also
18042 the wrong way round! This was probably due to unusual behaviour in
18043 EVP_cmp_parameters() which returns 1 if the parameters match.
18044 This meant that parameters were omitted when they *didn't* match and
18045 the certificate was useless. Certificates signed with 'ca' didn't have
18048 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18050 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18051 The interface is as follows:
18052 Applications can use
18053 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18054 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18055 "off" is now the default.
18056 The library internally uses
18057 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18058 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18059 to disable memory-checking temporarily.
18061 Some inconsistent states that previously were possible (and were
18062 even the default) are now avoided.
18064 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18065 with each memory chunk allocated; this is occasionally more helpful
18066 than just having a counter.
18068 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18070 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18075 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18076 which largely parallels "options", but is for changing API behaviour,
18077 whereas "options" are about protocol behaviour.
18078 Initial "mode" flags are:
18080 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18081 a single record has been written.
18082 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18083 retries use the same buffer location.
18084 (But all of the contents must be
18089 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18092 * Fix problems with no-hmac etc.
18094 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18096 * New functions RSA_get_default_method(), RSA_set_method() and
18097 RSA_get_method(). These allows replacement of RSA_METHODs without having
18098 to mess around with the internals of an RSA structure.
18102 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18103 Also really enable memory leak checks in openssl.c and in some
18106 *Chad C. Mulligan, Bodo Moeller*
18108 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18109 up the length of negative integers. This has now been simplified to just
18110 store the length when it is first determined and use it later, rather
18111 than trying to keep track of where data is copied and updating it to
18113 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18115 * Add a new function PKCS7_signatureVerify. This allows the verification
18116 of a PKCS#7 signature but with the signing certificate passed to the
18117 function itself. This contrasts with PKCS7_dataVerify which assumes the
18118 certificate is present in the PKCS#7 structure. This isn't always the
18119 case: certificates can be omitted from a PKCS#7 structure and be
18120 distributed by "out of band" means (such as a certificate database).
18124 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18125 function prototypes in pem.h, also change util/mkdef.pl to add the
18126 necessary function names.
18130 * mk1mf.pl (used by Windows builds) did not properly read the
18131 options set by Configure in the top level Makefile, and Configure
18132 was not even able to write more than one option correctly.
18133 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18137 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18138 file to be loaded from a BIO or FILE pointer. The BIO version will
18139 for example allow memory BIOs to contain config info.
18143 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18144 Whoever hopes to achieve shared-library compatibility across versions
18145 must use this, not the compile-time macro.
18146 (Exercise 0.9.4: Which is the minimum library version required by
18148 Note: All this applies only to multi-threaded programs, others don't
18153 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18154 through a BIO pair triggered the default case, i.e.
18155 SSLerr(...,SSL_R_UNKNOWN_STATE).
18159 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18160 can use the SSL library even if none of the specific BIOs is
18165 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18166 for the encoded length.
18168 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18170 * Add initial documentation of the X509V3 functions.
18174 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18175 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18176 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18177 secure PKCS#8 private key format with a high iteration count.
18181 * Fix determination of Perl interpreter: A perl or perl5
18182 *directory* in $PATH was also accepted as the interpreter.
18184 *Ralf S. Engelschall*
18186 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18187 wrong with it but it was very old and did things like calling
18188 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18189 unusual formatting.
18193 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18194 to use the new extension code.
18198 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18199 with macros. This should make it easier to change their form, add extra
18200 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18205 * Add to configuration table a new entry that can specify an alternative
18206 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18207 according to Mark Crispin <MRC@Panda.COM>.
18211 * DES CBC did not update the IV. Weird.
18215 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18216 Changing the behaviour of the former might break existing programs --
18217 where IV updating is needed, des_ncbc_encrypt can be used.
18220 * When bntest is run from "make test" it drives bc to check its
18221 calculations, as well as internally checking them. If an internal check
18222 fails, it needs to cause bc to give a non-zero result or make test carries
18223 on without noticing the failure. Fixed.
18227 * DES library cleanups.
18231 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18232 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18233 ciphers. NOTE: although the key derivation function has been verified
18234 against some published test vectors it has not been extensively tested
18235 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18240 * Instead of "mkdir -p", which is not fully portable, use new
18241 Perl script "util/mkdir-p.pl".
18245 * Rewrite the way password based encryption (PBE) is handled. It used to
18246 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18247 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18248 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18249 the 'parameter' field of the AlgorithmIdentifier is passed to the
18250 underlying key generation function so it must do its own ASN1 parsing.
18251 This has also changed the EVP_PBE_CipherInit() function which now has a
18252 'parameter' argument instead of literal salt and iteration count values
18253 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18257 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18258 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18259 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18260 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18261 value was just used as a "magic string" and not used directly its
18262 value doesn't matter.
18266 * Introduce some semblance of const correctness to BN. Shame C doesn't
18271 * "linux-sparc64" configuration (ultrapenguin).
18273 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18274 "linux-sparc" configuration.
18276 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18278 * config now generates no-xxx options for missing ciphers.
18282 * Support the EBCDIC character set (work in progress).
18283 File ebcdic.c not yet included because it has a different license.
18285 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18287 * Support BS2000/OSD-POSIX.
18289 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18291 * Make callbacks for key generation use `void *` instead of `char *`.
18295 * Make S/MIME samples compile (not yet tested).
18299 * Additional typesafe stacks.
18303 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18307 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18309 * New configuration variant "sco5-gcc".
18311 * Updated some demos.
18313 *Sean O Riordain, Wade Scholine*
18315 * Add missing BIO_free at exit of pkcs12 application.
18319 * Fix memory leak in conf.c.
18323 * Updates for Win32 to assembler version of MD5.
18327 * Set #! path to perl in `apps/der_chop` to where we found it
18328 instead of using a fixed path.
18332 * SHA library changes for irix64-mips4-cc.
18336 * Improvements for VMS support.
18340 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18342 * Bignum library bug fix. IRIX 6 passes "make test" now!
18343 This also avoids the problems with SC4.2 and unpatched SC5.
18345 *Andy Polyakov <appro@fy.chalmers.se>*
18347 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18348 These are required because of the typesafe stack would otherwise break
18349 existing code. If old code used a structure member which used to be STACK
18350 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18351 sk_num or sk_value it would produce an error because the num, data members
18352 are not present in STACK_OF. Now it just produces a warning. sk_set
18353 replaces the old method of assigning a value to sk_value
18354 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18355 that does this will no longer work (and should use sk_set instead) but
18356 this could be regarded as a "questionable" behaviour anyway.
18360 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18361 correctly handle encrypted S/MIME data.
18365 * Change type of various DES function arguments from des_cblock
18366 (which means, in function argument declarations, pointer to char)
18367 to des_cblock * (meaning pointer to array with 8 char elements),
18368 which allows the compiler to do more typechecking; it was like
18369 that back in SSLeay, but with lots of ugly casts.
18371 Introduce new type const_des_cblock.
18375 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18376 problems: find RecipientInfo structure that matches recipient certificate
18377 and initialise the ASN1 structures properly based on passed cipher.
18381 * Belatedly make the BN tests actually check the results.
18385 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18386 to and from BNs: it was completely broken. New compilation option
18387 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18388 key elements as negative integers.
18392 * Reorganize and speed up MD5.
18394 *Andy Polyakov <appro@fy.chalmers.se>*
18398 *Richard Levitte <richard@levitte.org>*
18400 * New option -out to asn1parse to allow the parsed structure to be
18401 output to a file. This is most useful when combined with the -strparse
18402 option to examine the output of things like OCTET STRINGS.
18406 * Make SSL library a little more fool-proof by not requiring any longer
18407 that `SSL_set_{accept,connect}_state` be called before
18408 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18409 in many applications because usually everything *appeared* to work as
18410 intended anyway -- now it really works as intended).
18414 * Move openssl.cnf out of lib/.
18418 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18419 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18420 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18422 *Ralf S. Engelschall*
18424 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18425 handle PKCS#7 enveloped data properly.
18427 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18429 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18430 copying pointers. The cert_st handling is changed by this in
18431 various ways (and thus what used to be known as ctx->default_cert
18432 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18433 any longer when s->cert does not give us what we need).
18434 ssl_cert_instantiate becomes obsolete by this change.
18435 As soon as we've got the new code right (possibly it already is?),
18436 we have solved a couple of bugs of the earlier code where s->cert
18437 was used as if it could not have been shared with other SSL structures.
18439 Note that using the SSL API in certain dirty ways now will result
18440 in different behaviour than observed with earlier library versions:
18441 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18442 does not influence s as it used to.
18444 In order to clean up things more thoroughly, inside SSL_SESSION
18445 we don't use CERT any longer, but a new structure SESS_CERT
18446 that holds per-session data (if available); currently, this is
18447 the peer's certificate chain and, for clients, the server's certificate
18448 and temporary key. CERT holds only those values that can have
18449 meaningful defaults in an SSL_CTX.
18453 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18454 from the internal representation. Various PKCS#7 fixes: remove some
18455 evil casts and set the enc_dig_alg field properly based on the signing
18460 * Allow PKCS#12 password to be set from the command line or the
18461 environment. Let 'ca' get its config file name from the environment
18462 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18467 * Allow certificate policies extension to use an IA5STRING for the
18468 organization field. This is contrary to the PKIX definition but
18469 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18474 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18475 without disallowing inline assembler and the like for non-pedantic builds.
18479 * Support Borland C++ builder.
18481 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18487 * SHA-1 cleanups and performance enhancements.
18489 *Andy Polyakov <appro@fy.chalmers.se>*
18491 * Sparc v8plus assembler for the bignum library.
18493 *Andy Polyakov <appro@fy.chalmers.se>*
18495 * Accept any -xxx and +xxx compiler options in Configure.
18499 * Update HPUX configuration.
18503 * Add missing `sk_<type>_unshift()` function to safestack.h
18505 *Ralf S. Engelschall*
18507 * New function SSL_CTX_use_certificate_chain_file that sets the
18508 "extra_cert"s in addition to the certificate. (This makes sense
18509 only for "PEM" format files, as chains as a whole are not
18514 * Support verify_depth from the SSL API.
18515 x509_vfy.c had what can be considered an off-by-one-error:
18516 Its depth (which was not part of the external interface)
18517 was actually counting the number of certificates in a chain;
18518 now it really counts the depth.
18522 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18523 instead of X509err, which often resulted in confusing error
18524 messages since the error codes are not globally unique
18525 (e.g. an alleged error in ssl3_accept when a certificate
18526 didn't match the private key).
18528 * New function SSL_CTX_set_session_id_context that allows to set a default
18529 value (so that you don't need SSL_set_session_id_context for each
18530 connection using the SSL_CTX).
18534 * OAEP decoding bug fix.
18538 * Support INSTALL_PREFIX for package builders, as proposed by
18543 * New Configure options "threads" and "no-threads". For systems
18544 where the proper compiler options are known (currently Solaris
18545 and Linux), "threads" is the default.
18549 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18553 * Install various scripts to $(OPENSSLDIR)/misc, not to
18554 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18555 such as /usr/local/bin.
18559 * "make linux-shared" to build shared libraries.
18561 *Niels Poppe <niels@netbox.org>*
18563 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18567 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18568 extension adding in x509 utility.
18572 * Remove NOPROTO sections and error code comments.
18576 * Partial rewrite of the DEF file generator to now parse the ANSI
18581 * New Configure options --prefix=DIR and --openssldir=DIR.
18585 * Complete rewrite of the error code script(s). It is all now handled
18586 by one script at the top level which handles error code gathering,
18587 header rewriting and C source file generation. It should be much better
18588 than the old method: it now uses a modified version of Ulf's parser to
18589 read the ANSI prototypes in all header files (thus the old K&R definitions
18590 aren't needed for error creation any more) and do a better job of
18591 translating function codes into names. The old 'ASN1 error code embedded
18592 in a comment' is no longer necessary and it doesn't use .err files which
18593 have now been deleted. Also the error code call doesn't have to appear all
18594 on one line (which resulted in some large lines...).
18598 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18602 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18603 0 (which usually indicates a closed connection), but continue reading.
18607 * Fix some race conditions.
18611 * Add support for CRL distribution points extension. Add Certificate
18612 Policies and CRL distribution points documentation.
18616 * Move the autogenerated header file parts to crypto/opensslconf.h.
18620 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18621 8 of keying material. Merlin has also confirmed interop with this fix
18622 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18624 *Merlin Hughes <merlin@baltimore.ie>*
18626 * Fix lots of warnings.
18628 *Richard Levitte <levitte@stacken.kth.se>*
18630 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18631 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18633 *Richard Levitte <levitte@stacken.kth.se>*
18635 * Fix problems with sizeof(long) == 8.
18637 *Andy Polyakov <appro@fy.chalmers.se>*
18639 * Change functions to ANSI C.
18643 * Fix typos in error codes.
18645 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18647 * Remove defunct assembler files from Configure.
18651 * SPARC v8 assembler BIGNUM implementation.
18653 *Andy Polyakov <appro@fy.chalmers.se>*
18655 * Support for Certificate Policies extension: both print and set.
18656 Various additions to support the r2i method this uses.
18660 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18661 return a const string when you are expecting an allocated buffer.
18665 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18666 types DirectoryString and DisplayText.
18670 * Add code to allow r2i extensions to access the configuration database,
18671 add an LHASH database driver and add several ctx helper functions.
18675 * Fix an evil bug in bn_expand2() which caused various BN functions to
18676 fail when they extended the size of a BIGNUM.
18680 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18681 support typesafe stack.
18685 * Fix typo in SSL_[gs]et_options().
18687 *Nils Frostberg <nils@medcom.se>*
18689 * Delete various functions and files that belonged to the (now obsolete)
18690 old X509V3 handling code.
18694 * New Configure option "rsaref".
18698 * Don't auto-generate pem.h.
18702 * Introduce type-safe ASN.1 SETs.
18706 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18708 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18710 * Introduce type-safe STACKs. This will almost certainly break lots of code
18711 that links with OpenSSL (well at least cause lots of warnings), but fear
18712 not: the conversion is trivial, and it eliminates loads of evil casts. A
18713 few STACKed things have been converted already. Feel free to convert more.
18714 In the fullness of time, I'll do away with the STACK type altogether.
18718 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18719 specified in `<certfile>` by updating the entry in the index.txt file.
18720 This way one no longer has to edit the index.txt file manually for
18721 revoking a certificate. The -revoke option does the gory details now.
18723 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18725 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18726 `-text` option at all and this way the `-noout -text` combination was
18727 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18729 *Ralf S. Engelschall*
18731 * Make sure a corresponding plain text error message exists for the
18732 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18733 verify callback function determined that a certificate was revoked.
18735 *Ralf S. Engelschall*
18737 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18738 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18739 all available ciphers including rc5, which was forgotten until now.
18740 In order to let the testing shell script know which algorithms
18741 are available, a new (up to now undocumented) command
18742 `openssl list-cipher-commands` is used.
18746 * Bugfix: s_client occasionally would sleep in select() when
18747 it should have checked SSL_pending() first.
18751 * New functions DSA_do_sign and DSA_do_verify to provide access to
18752 the raw DSA values prior to ASN.1 encoding.
18756 * Tweaks to Configure
18758 *Niels Poppe <niels@netbox.org>*
18760 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18765 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18769 * New config option to avoid instructions that are illegal on the 80386.
18770 The default code is faster, but requires at least a 486.
18774 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18775 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18776 same as SSL2_VERSION anyway.
18780 * New "-showcerts" option for s_client.
18784 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18785 application. Various cleanups and fixes.
18789 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18790 modify error routines to work internally. Add error codes and PBE init
18791 to library startup routines.
18795 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18796 packing functions to asn1 and evp. Changed function names and error
18797 codes along the way.
18801 * PKCS12 integration: and so it begins... First of several patches to
18802 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18803 objects to objects.h
18807 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18808 and display support for Thawte strong extranet extension.
18812 * Add LinuxPPC support.
18814 *Jeff Dubrule <igor@pobox.org>*
18816 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18817 bn_div_words in alpha.s.
18819 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18821 * Make sure the RSA OAEP test is skipped under -DRSAref because
18822 OAEP isn't supported when OpenSSL is built with RSAref.
18824 *Ulf Moeller <ulf@fitug.de>*
18826 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18827 so they no longer are missing under -DNOPROTO.
18829 *Soren S. Jorvang <soren@t.dk>*
18831 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18833 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18834 doesn't work when the session is reused. Coming soon!
18838 * Fix a security hole, that allows sessions to be reused in the wrong
18839 context thus bypassing client cert protection! All software that uses
18840 client certs and session caches in multiple contexts NEEDS PATCHING to
18841 allow session reuse! A fuller solution is in the works.
18843 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18845 * Some more source tree cleanups (removed obsolete files
18846 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18847 permission on "config" script to be executable) and a fix for the INSTALL
18850 *Ulf Moeller <ulf@fitug.de>*
18852 * Remove some legacy and erroneous uses of malloc, free instead of
18855 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18857 * Make rsa_oaep_test return non-zero on error.
18859 *Ulf Moeller <ulf@fitug.de>*
18861 * Add support for native Solaris shared libraries. Configure
18862 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18863 if someone would make that last step automatic.
18865 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18867 * ctx_size was not built with the right compiler during "make links". Fixed.
18871 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18872 except NULL ciphers". This means the default cipher list will no longer
18873 enable NULL ciphers. They need to be specifically enabled e.g. with
18874 the string "DEFAULT:eNULL".
18878 * Fix to RSA private encryption routines: if p < q then it would
18879 occasionally produce an invalid result. This will only happen with
18880 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18884 * Be less restrictive and allow also `perl util/perlpath.pl
18885 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18886 because this way one can also use an interpreter named `perl5` (which is
18887 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18888 installed as `perl`).
18890 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18892 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18894 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18896 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18897 advapi32.lib to Win32 build and change the pem test comparison
18898 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18899 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18900 and crypto/des/ede_cbcm_enc.c.
18904 * DES quad checksum was broken on big-endian architectures. Fixed.
18908 * Comment out two functions in bio.h that aren't implemented. Fix up the
18909 Win32 test batch file so it (might) work again. The Win32 test batch file
18910 is horrible: I feel ill....
18914 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18915 in e_os.h. Audit of header files to check ANSI and non ANSI
18916 sections: 10 functions were absent from non ANSI section and not exported
18917 from Windows DLLs. Fixed up libeay.num for new functions.
18921 * Make `openssl version` output lines consistent.
18923 *Ralf S. Engelschall*
18925 * Fix Win32 symbol export lists for BIO functions: Added
18926 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18927 to ms/libeay{16,32}.def.
18929 *Ralf S. Engelschall*
18931 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18932 fine under Unix and passes some trivial tests I've now added. But the
18933 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18934 added to make sure no one expects that this stuff really works in the
18935 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18936 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18939 *Ralf S. Engelschall*
18941 * Fix the generation of two part addresses in perl.
18943 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18945 * Add config entry for Linux on MIPS.
18947 *John Tobey <jtobey@channel1.com>*
18949 * Make links whenever Configure is run, unless we are on Windoze.
18953 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18954 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18959 * Add a useful kludge to allow package maintainers to specify compiler and
18960 other platforms details on the command line without having to patch the
18961 Configure script every time: One now can use
18962 `perl Configure <id>:<details>`,
18963 i.e. platform ids are allowed to have details appended
18964 to them (separated by colons). This is treated as there would be a static
18965 pre-configured entry in Configure's %table under key `<id>` with value
18966 `<details>` and `perl Configure <id>` is called. So, when you want to
18967 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18968 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18969 now, which overrides the FreeBSD-elf entry on-the-fly.
18971 *Ralf S. Engelschall*
18973 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18977 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18978 on the `perl Configure ...` command line. This way one can compile
18979 OpenSSL libraries with Position Independent Code (PIC) which is needed
18980 for linking it into DSOs.
18982 *Ralf S. Engelschall*
18984 * Remarkably, export ciphers were totally broken and no-one had noticed!
18989 * Cleaned up the LICENSE document: The official contact for any license
18990 questions now is the OpenSSL core team under openssl-core@openssl.org.
18991 And add a paragraph about the dual-license situation to make sure people
18992 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18993 to the OpenSSL toolkit.
18995 *Ralf S. Engelschall*
18997 * General source tree makefile cleanups: Made `making xxx in yyy...`
18998 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18999 Additionally cleaned up the `make links` target: Remove unnecessary
19000 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19001 to speed processing and no longer clutter the display with confusing
19002 stuff. Instead only the actually done links are displayed.
19004 *Ralf S. Engelschall*
19006 * Permit null encryption ciphersuites, used for authentication only. It used
19007 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19008 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19013 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19014 signed attributes when verifying signatures (this would break them),
19015 the detached data encoding was wrong and public keys obtained using
19016 X509_get_pubkey() weren't freed.
19020 * Add text documentation for the BUFFER functions. Also added a work around
19021 to a Win95 console bug. This was triggered by the password read stuff: the
19022 last character typed gets carried over to the next fread(). If you were
19023 generating a new cert request using 'req' for example then the last
19024 character of the passphrase would be CR which would then enter the first
19029 * Added the new 'Includes OpenSSL Cryptography Software' button as
19030 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19031 button and can be used by applications based on OpenSSL to show the
19032 relationship to the OpenSSL project.
19034 *Ralf S. Engelschall*
19036 * Remove confusing variables in function signatures in files
19037 ssl/ssl_lib.c and ssl/ssl.h.
19039 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19041 * Don't install bss_file.c under PREFIX/include/
19043 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19045 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19046 functions that return function pointers and has support for NT specific
19047 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19048 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19049 unsigned to signed types: this was killing the Win32 compile.
19053 * Add new certificate file to stack functions,
19054 SSL_add_dir_cert_subjects_to_stack() and
19055 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19056 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19057 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19058 This means that Apache-SSL and similar packages don't have to mess around
19059 to add as many CAs as they want to the preferred list.
19063 * Experiment with doxygen documentation. Currently only partially applied to
19065 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19066 openssl.doxy as the configuration file.
19070 * Get rid of remaining C++-style comments which strict C compilers hate.
19072 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19074 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19075 compiled in by default: it has problems with large keys.
19079 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19080 DH private keys and/or callback functions which directly correspond to
19081 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19082 is needed for applications which have to configure certificates on a
19083 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19085 For the RSA certificate situation is makes no difference, but
19086 for the DSA certificate situation this fixes the "no shared cipher"
19087 problem where the OpenSSL cipher selection procedure failed because the
19088 temporary keys were not overtaken from the context and the API provided
19089 no way to reconfigure them.
19090 The new functions now let applications reconfigure the stuff and they
19091 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19092 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19093 non-public-API function ssl_cert_instantiate() is used as a helper
19094 function and also to reduce code redundancy inside ssl_rsa.c.
19096 *Ralf S. Engelschall*
19098 * Move s_server -dcert and -dkey options out of the undocumented feature
19099 area because they are useful for the DSA situation and should be
19100 recognized by the users.
19102 *Ralf S. Engelschall*
19104 * Fix the cipher decision scheme for export ciphers: the export bits are
19105 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19106 SSL_EXP_MASK. So, the original variable has to be used instead of the
19107 already masked variable.
19109 *Richard Levitte <levitte@stacken.kth.se>*
19111 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19113 *Richard Levitte <levitte@stacken.kth.se>*
19115 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19116 from `int` to `unsigned int` because it is a length and initialized by
19117 EVP_DigestFinal() which expects an `unsigned int *`.
19119 *Richard Levitte <levitte@stacken.kth.se>*
19121 * Don't hard-code path to Perl interpreter on shebang line of Configure
19122 script. Instead use the usual Shell->Perl transition trick.
19124 *Ralf S. Engelschall*
19126 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19127 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19128 -noout -modulus` as it's already the case for `openssl rsa -noout
19129 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19130 currently the public key is printed (a decision which was already done by
19131 `openssl dsa -modulus` in the past) which serves a similar purpose.
19132 Additionally the NO_RSA no longer completely removes the whole -modulus
19133 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19136 *Ralf S. Engelschall*
19138 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19139 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19141 *Arne Ansper <arne@ats.cyber.ee>*
19143 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19144 to be added. Now both 'req' and 'ca' can use new objects defined in the
19149 * Add cool BIO that does syslog (or event log on NT).
19151 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19153 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19154 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19155 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19156 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19160 * Add preliminary config info for new extension code.
19164 * Make RSA_NO_PADDING really use no padding.
19166 *Ulf Moeller <ulf@fitug.de>*
19168 * Generate errors when private/public key check is done.
19172 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19173 for some CRL extensions and new objects added.
19177 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19178 key usage extension and fuller support for authority key id.
19182 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19183 padding method for RSA, which is recommended for new applications in PKCS
19184 #1 v2.0 (RFC 2437, October 1998).
19185 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19186 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19187 against Bleichbacher's attack on RSA.
19188 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19191 * Updates to the new SSL compression code
19193 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19195 * Fix so that the version number in the master secret, when passed
19196 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19197 (because the server will not accept higher), that the version number
19198 is 0x03,0x01, not 0x03,0x00
19200 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19202 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19203 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19204 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19208 * Support for RAW extensions where an arbitrary extension can be
19209 created by including its DER encoding. See `apps/openssl.cnf` for
19214 * Make sure latest Perl versions don't interpret some generated C array
19215 code as Perl array code in the crypto/err/err_genc.pl script.
19217 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19219 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19220 not many people have the assembler. Various Win32 compilation fixes and
19221 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19222 build instructions.
19226 * Modify configure script 'Configure' to automatically create crypto/date.h
19227 file under Win32 and also build pem.h from pem.org. New script
19228 util/mkfiles.pl to create the MINFO file on environments that can't do a
19229 'make files': perl util/mkfiles.pl >MINFO should work.
19233 * Major rework of DES function declarations, in the pursuit of correctness
19234 and purity. As a result, many evil casts evaporated, and some weirdness,
19235 too. You may find this causes warnings in your code. Zapping your evil
19236 casts will probably fix them. Mostly.
19240 * Fix for a typo in asn1.h. Bug fix to object creation script
19241 obj_dat.pl. It considered a zero in an object definition to mean
19242 "end of object": none of the objects in objects.h have any zeros
19243 so it wasn't spotted.
19245 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19247 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19248 Masking (CBCM). In the absence of test vectors, the best I have been able
19249 to do is check that the decrypt undoes the encrypt, so far. Send me test
19250 vectors if you have them.
19254 * Correct calculation of key length for export ciphers (too much space was
19255 allocated for null ciphers). This has not been tested!
19259 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19260 message is now correct (it understands "crypto" and "ssl" on its
19261 command line). There is also now an "update" option. This will update
19262 the util/ssleay.num and util/libeay.num files with any new functions.
19264 perl util/mkdef.pl crypto ssl update
19265 it will update them.
19269 * Overhauled the Perl interface:
19270 - ported BN stuff to OpenSSL's different BN library
19271 - made the perl/ source tree CVS-aware
19272 - renamed the package from SSLeay to OpenSSL (the files still contain
19273 their history because I've copied them in the repository)
19274 - removed obsolete files (the test scripts will be replaced
19275 by better Test::Harness variants in the future)
19277 *Ralf S. Engelschall*
19279 * First cut for a very conservative source tree cleanup:
19280 1. merge various obsolete readme texts into doc/ssleay.txt
19281 where we collect the old documents and readme texts.
19282 2. remove the first part of files where I'm already sure that we no
19283 longer need them because of three reasons: either they are just temporary
19284 files which were left by Eric or they are preserved original files where
19285 I've verified that the diff is also available in the CVS via "cvs diff
19286 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19287 the crypto/md/ stuff).
19289 *Ralf S. Engelschall*
19291 * More extension code. Incomplete support for subject and issuer alt
19292 name, issuer and authority key id. Change the i2v function parameters
19293 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19294 what that's for :-) Fix to ASN1 macro which messed up
19295 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19299 * Preliminary support for ENUMERATED type. This is largely copied from the
19304 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19306 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19308 * Make sure `make rehash` target really finds the `openssl` program.
19310 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19312 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19313 like to hear about it if this slows down other processors.
19317 * Add CygWin32 platform information to Configure script.
19319 *Alan Batie <batie@aahz.jf.intel.com>*
19321 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19323 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19325 * New program nseq to manipulate netscape certificate sequences
19329 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19334 * Fixes to BN code. Previously the default was to define BN_RECURSION
19335 but the BN code had some problems that would cause failures when
19336 doing certificate verification and some other functions.
19338 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19340 * Add ASN1 and PEM code to support netscape certificate sequences.
19344 * Add ASN1 and PEM code to support netscape certificate sequences.
19348 * Add several PKIX and private extended key usage OIDs.
19352 * Modify the 'ca' program to handle the new extension code. Modify
19353 openssl.cnf for new extension format, add comments.
19357 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19358 and add a sample to openssl.cnf so req -x509 now adds appropriate
19363 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19364 error code, add initial support to X509_print() and x509 application.
19368 * Takes a deep breath and start adding X509 V3 extension support code. Add
19369 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19370 stuff is currently isolated and isn't even compiled yet.
19374 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19375 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19376 Removed the versions check from X509 routines when loading extensions:
19377 this allows certain broken certificates that don't set the version
19378 properly to be processed.
19382 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19383 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19384 can still be regenerated with "make depend".
19388 * Spelling mistake in C version of CAST-128.
19390 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19392 * Changes to the error generation code. The perl script err-code.pl
19393 now reads in the old error codes and retains the old numbers, only
19394 adding new ones if necessary. It also only changes the .err files if new
19395 codes are added. The makefiles have been modified to only insert errors
19396 when needed (to avoid needlessly modifying header files). This is done
19397 by only inserting errors if the .err file is newer than the auto generated
19398 C file. To rebuild all the error codes from scratch (the old behaviour)
19399 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19400 or delete all the .err files.
19404 * CAST-128 was incorrectly implemented for short keys. The C version has
19405 been fixed, but is untested. The assembler versions are also fixed, but
19406 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19407 to regenerate it if needed.
19408 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19409 Hagino <itojun@kame.net>*
19411 * File was opened incorrectly in randfile.c.
19413 *Ulf Möller <ulf@fitug.de>*
19415 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19416 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19417 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19418 al: it's just almost always a UTCTime. Note this patch adds new error
19419 codes so do a "make errors" if there are problems.
19423 * Correct Linux 1 recognition in config.
19425 *Ulf Möller <ulf@fitug.de>*
19427 * Remove pointless MD5 hash when using DSA keys in ca.
19429 *Anonymous <nobody@replay.com>*
19431 * Generate an error if given an empty string as a cert directory. Also
19432 generate an error if handed NULL (previously returned 0 to indicate an
19433 error, but didn't set one).
19435 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19437 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19441 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19442 parameters. This was causing a warning which killed off the Win32 compile.
19446 * Remove C++ style comments from crypto/bn/bn_local.h.
19448 *Neil Costigan <neil.costigan@celocom.com>*
19450 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19451 based on a text string, looking up short and long names and finally
19452 "dot" format. The "dot" format stuff didn't work. Added new function
19453 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19454 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19455 OID is not part of the table.
19459 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19460 X509_LOOKUP_by_alias().
19464 * Sort openssl functions by name.
19468 * Get the `gendsa` command working and add it to the `list` command. Remove
19469 encryption from sample DSA keys (in case anyone is interested the password
19474 * Make *all* `*_free` functions accept a NULL pointer.
19476 *Frans Heymans <fheymans@isaserver.be>*
19478 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19481 *Anonymous <nobody@replay.com>*
19483 * s_server should send the CAfile as acceptable CAs, not its own cert.
19485 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19487 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19489 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19491 * Temp key "for export" tests were wrong in s3_srvr.c.
19493 *Anonymous <nobody@replay.com>*
19495 * Add prototype for temp key callback functions
19496 SSL_CTX_set_tmp_{rsa,dh}_callback().
19500 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19501 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19505 * X509_name_add_entry() freed the wrong thing after an error.
19507 *Arne Ansper <arne@ats.cyber.ee>*
19509 * rsa_eay.c would attempt to free a NULL context.
19511 *Arne Ansper <arne@ats.cyber.ee>*
19513 * BIO_s_socket() had a broken should_retry() on Windoze.
19515 *Arne Ansper <arne@ats.cyber.ee>*
19517 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19519 *Arne Ansper <arne@ats.cyber.ee>*
19521 * Make sure the already existing X509_STORE->depth variable is initialized
19522 in X509_STORE_new(), but document the fact that this variable is still
19523 unused in the certificate verification process.
19525 *Ralf S. Engelschall*
19527 * Fix the various library and `apps/` files to free up pkeys obtained from
19528 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19532 * Fix reference counting in X509_PUBKEY_get(). This makes
19533 demos/maurice/example2.c work, amongst others, probably.
19535 *Steve Henson and Ben Laurie*
19537 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19538 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19539 are no longer created. This way we have a single and consistent command
19540 line interface `openssl <command>`, similar to `cvs <command>`.
19542 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19544 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19545 BIT STRING wrapper always have zero unused bits.
19549 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19553 * Make the top-level INSTALL documentation easier to understand.
19557 * Makefiles updated to exit if an error occurs in a sub-directory
19558 make (including if user presses ^C) [Paul Sutton]
19560 * Make Montgomery context stuff explicit in RSA data structure.
19564 * Fix build order of pem and err to allow for generated pem.h.
19568 * Fix renumbering bug in X509_NAME_delete_entry().
19572 * Enhanced the err-ins.pl script so it makes the error library number
19573 global and can add a library name. This is needed for external ASN1 and
19574 other error libraries.
19578 * Fixed sk_insert which never worked properly.
19582 * Fix ASN1 macros so they can handle indefinite length constructed
19583 EXPLICIT tags. Some non standard certificates use these: they can now
19588 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19589 into a single doc/ssleay.txt bundle. This way the information is still
19590 preserved but no longer messes up this directory. Now it's new room for
19591 the new set of documentation files.
19593 *Ralf S. Engelschall*
19595 * SETs were incorrectly DER encoded. This was a major pain, because they
19596 shared code with SEQUENCEs, which aren't coded the same. This means that
19597 almost everything to do with SETs or SEQUENCEs has either changed name or
19598 number of arguments.
19600 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19602 * Fix test data to work with the above.
19606 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19607 was already fixed by Eric for 0.9.1 it seems.
19609 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19611 * Autodetect FreeBSD3.
19615 * Fix various bugs in Configure. This affects the following platforms:
19619 unixware-2.0-pentium
19624 * Eliminate generated files from CVS. Reorder tests to regenerate files
19625 before they are needed.
19629 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19633 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19635 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19636 changed SSLeay to OpenSSL in version strings.
19638 *Ralf S. Engelschall*
19640 * Some fixups to the top-level documents.
19644 * Fixed the nasty bug where rsaref.h was not found under compile-time
19645 because the symlink to include/ was missing.
19647 *Ralf S. Engelschall*
19649 * Incorporated the popular no-RSA/DSA-only patches
19650 which allow to compile an RSA-free SSLeay.
19652 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19654 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19655 when "ssleay" is still not found.
19657 *Ralf S. Engelschall*
19659 * Added more platforms to Configure: Cray T3E, HPUX 11,
19661 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19663 * Updated the README file.
19665 *Ralf S. Engelschall*
19667 * Added various .cvsignore files in the CVS repository subdirs
19668 to make a "cvs update" really silent.
19670 *Ralf S. Engelschall*
19672 * Recompiled the error-definition header files and added
19673 missing symbols to the Win32 linker tables.
19675 *Ralf S. Engelschall*
19677 * Cleaned up the top-level documents;
19678 o new files: CHANGES and LICENSE
19679 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19680 o merged COPYRIGHT into LICENSE
19681 o removed obsolete TODO file
19682 o renamed MICROSOFT to INSTALL.W32
19684 *Ralf S. Engelschall*
19686 * Removed dummy files from the 0.9.1b source tree:
19687 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19688 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19689 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19690 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19691 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19693 *Ralf S. Engelschall*
19695 * Added various platform portability fixes.
19699 * The Genesis of the OpenSSL rpject:
19700 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19701 Young and Tim J. Hudson created while they were working for C2Net until
19704 *The OpenSSL Project*
19706 ### Changes between 0.9.0b and 0.9.1b [not released]
19708 * Updated a few CA certificates under certs/
19712 * Changed some BIGNUM api stuff.
19716 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19717 DGUX x86, Linux Alpha, etc.
19721 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19722 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19727 * Add -strparse option to asn1pars program which parses nested
19730 *Dr Stephen Henson <shenson@bigfoot.com>*
19732 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19736 * DSA fix for "ca" program.
19740 * Added "-genkey" option to "dsaparam" program.
19744 * Added RIPE MD160 (rmd160) message digest.
19748 * Added -a (all) option to "ssleay version" command.
19752 * Added PLATFORM define which is the id given to Configure.
19756 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19760 * Extended the ASN.1 parser routines.
19764 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19768 * Added a BN_CTX to the BN library.
19772 * Fixed the weak key values in DES library
19776 * Changed API in EVP library for cipher aliases.
19780 * Added support for RC2/64bit cipher.
19784 * Converted the lhash library to the crypto/mem.c functions.
19788 * Added more recognized ASN.1 object ids.
19792 * Added more RSA padding checks for SSL/TLS.
19796 * Added BIO proxy/filter functionality.
19800 * Added extra_certs to SSL_CTX which can be used
19801 send extra CA certificates to the client in the CA cert chain sending
19802 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19806 * Now Fortezza is denied in the authentication phase because
19807 this is key exchange mechanism is not supported by SSLeay at all.
19811 * Additional PKCS1 checks.
19815 * Support the string "TLSv1" for all TLS v1 ciphers.
19819 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19820 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19824 * Fixed a few memory leaks.
19828 * Fixed various code and comment typos.
19832 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19833 bytes sent in the client random.
19835 *Edward Bishop <ebishop@spyglass.com>*
19839 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19840 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19841 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19842 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19843 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19844 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19845 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19846 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19847 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19848 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19849 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19850 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19851 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19852 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19853 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19854 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19855 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19856 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19857 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19858 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19859 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19860 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19861 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19862 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19863 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19864 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19865 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19866 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19867 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19868 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19869 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19870 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19871 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19872 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19873 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19874 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19875 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19876 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19877 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19878 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19879 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19880 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19881 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19882 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19883 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19884 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19885 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19886 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19887 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19888 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19889 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19890 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19891 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19892 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19893 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19894 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19895 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19896 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19897 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19898 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19899 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19900 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19901 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19902 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19903 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19904 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19905 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19906 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19907 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19908 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19909 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19910 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19911 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19912 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19913 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19914 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19915 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19916 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19917 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19918 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19919 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19920 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19921 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19922 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19923 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19924 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19925 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19926 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19927 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19928 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19929 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19930 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19931 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19932 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19933 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19934 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19935 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19936 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19937 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19938 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19939 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19940 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19941 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19942 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19943 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19944 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19945 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19946 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19947 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19948 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19949 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19950 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19951 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19952 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19953 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19954 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19955 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19956 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19957 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19958 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19959 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19960 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19961 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19962 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19963 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19964 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19965 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19966 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19967 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19968 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19969 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19970 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19971 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19972 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19973 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19974 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19975 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19976 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19977 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19978 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19979 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19980 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19981 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19982 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19983 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19984 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19985 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19986 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19987 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19988 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19989 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19990 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19991 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19992 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19993 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19994 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19995 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19996 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19997 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19998 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19999 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20000 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20001 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20002 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20003 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20004 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20005 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20006 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20007 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20008 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20009 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20010 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20011 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20012 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20013 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20014 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20015 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20016 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20017 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20018 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20019 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20020 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655