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 3.1.3 and 3.1.4 [xx XXX xxxx]
27 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
28 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
29 that alter the key or IV length ([CVE-2023-5363]).
33 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
35 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
37 The POLY1305 MAC (message authentication code) implementation in OpenSSL
38 does not save the contents of non-volatile XMM registers on Windows 64
39 platform when calculating the MAC of data larger than 64 bytes. Before
40 returning to the caller all the XMM registers are set to zero rather than
41 restoring their previous content. The vulnerable code is used only on newer
42 x86_64 processors supporting the AVX512-IFMA instructions.
44 The consequences of this kind of internal application state corruption can
45 be various - from no consequences, if the calling application does not
46 depend on the contents of non-volatile XMM registers at all, to the worst
47 consequences, where the attacker could get complete control of the
48 application process. However given the contents of the registers are just
49 zeroized so the attacker cannot put arbitrary values inside, the most likely
50 consequence, if any, would be an incorrect result of some application
51 dependent calculations or a crash leading to a denial of service.
57 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
59 * Fix excessive time spent checking DH q parameter value.
61 The function DH_check() performs various checks on DH parameters. After
62 fixing CVE-2023-3446 it was discovered that a large q parameter value can
63 also trigger an overly long computation during some of these checks.
64 A correct q value, if present, cannot be larger than the modulus p
65 parameter, thus it is unnecessary to perform these checks if q is larger
68 If DH_check() is called with such q parameter value,
69 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
70 intensive checks are skipped.
76 * Fix DH_check() excessive time with over sized modulus.
78 The function DH_check() performs various checks on DH parameters. One of
79 those checks confirms that the modulus ("p" parameter) is not too large.
80 Trying to use a very large modulus is slow and OpenSSL will not normally use
81 a modulus which is over 10,000 bits in length.
83 However the DH_check() function checks numerous aspects of the key or
84 parameters that have been supplied. Some of those checks use the supplied
85 modulus value even if it has already been found to be too large.
87 A new limit has been added to DH_check of 32,768 bits. Supplying a
88 key/parameters with a modulus over this size will simply cause DH_check() to
95 * Do not ignore empty associated data entries with AES-SIV.
97 The AES-SIV algorithm allows for authentication of multiple associated
98 data entries along with the encryption. To authenticate empty data the
99 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
100 with NULL pointer as the output buffer and 0 as the input buffer length.
101 The AES-SIV implementation in OpenSSL just returns success for such call
102 instead of performing the associated data authentication operation.
103 The empty data thus will not be authenticated. ([CVE-2023-2975])
105 Thanks to Juerg Wullschleger (Google) for discovering the issue.
107 The fix changes the authentication tag value and the ciphertext for
108 applications that use empty associated data entries with AES-SIV.
109 To decrypt data encrypted with previous versions of OpenSSL the application
110 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
115 * When building with the `enable-fips` option and using the resulting
116 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
117 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
118 not operate with truncated digests (FIPS 140-3 IG G.R).
122 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
124 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
125 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
127 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
128 numeric text form. For gigantic sub-identifiers, this would take a very
129 long time, the time complexity being O(n^2) where n is the size of that
130 sub-identifier. ([CVE-2023-2650])
132 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
133 IDENTIFIER to canonical numeric text form if the size of that OBJECT
134 IDENTIFIER is 586 bytes or less, and fail otherwise.
136 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
137 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
138 most 128 sub-identifiers, and that the maximum value that each sub-
139 identifier may have is 2^32-1 (4294967295 decimal).
141 For each byte of every sub-identifier, only the 7 lower bits are part of
142 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
143 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
148 * Multiple algorithm implementation fixes for ARM BE platforms.
152 * Added a -pedantic option to fipsinstall that adjusts the various
153 settings to ensure strict FIPS compliance rather than backwards
158 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
159 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
160 trigger a crash of an application using AES-XTS decryption if the memory
161 just after the buffer being decrypted is not mapped.
162 Thanks to Anton Romanov (Amazon) for discovering the issue.
167 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
168 The previous fix for this timing side channel turned out to cause
169 a severe 2-3x performance regression in the typical use case
170 compared to 3.0.7. The new fix uses existing constant time
171 code paths, and restores the previous performance level while
172 fully eliminating all existing timing side channels.
173 The fix was developed by Bernd Edlinger with testing support
178 * Add FIPS provider configuration option to disallow the use of
179 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
180 The option '-no_drbg_truncated_digests' can optionally be
181 supplied to 'openssl fipsinstall'.
185 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
186 that it does not enable policy checking. Thanks to David Benjamin for
187 discovering this issue.
192 * Fixed an issue where invalid certificate policies in leaf certificates are
193 silently ignored by OpenSSL and other certificate policy checks are skipped
194 for that certificate. A malicious CA could use this to deliberately assert
195 invalid certificate policies in order to circumvent policy checking on the
196 certificate altogether.
201 * Limited the number of nodes created in a policy tree to mitigate
202 against CVE-2023-0464. The default limit is set to 1000 nodes, which
203 should be sufficient for most installations. If required, the limit
204 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
205 time define to a desired maximum number of nodes or zero to allow
211 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
213 * Add FIPS provider configuration option to enforce the
214 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
215 The option '-ems-check' can optionally be supplied to
216 'openssl fipsinstall'.
220 * The FIPS provider includes a few non-approved algorithms for
221 backward compatibility purposes and the "fips=yes" property query
222 must be used for all algorithm fetches to ensure FIPS compliance.
224 The algorithms that are included but not approved are Triple DES ECB,
225 Triple DES CBC and EdDSA.
229 * Added support for KMAC in KBKDF.
233 * RNDR and RNDRRS support in provider functions to provide
234 random number generation for Arm CPUs (aarch64).
238 * s_client and s_server apps now explicitly say when the TLS version
239 does not include the renegotiation mechanism. This avoids confusion
240 between that scenario versus when the TLS version includes secure
241 renegotiation but the peer lacks support for it.
245 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
247 *Tomasz Kantecki, Andrey Matyukov*
249 * The various OBJ_* functions have been made thread safe.
253 * Parallel dual-prime 1536/2048-bit modular exponentiation for
254 AVX512_IFMA capable processors.
256 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
258 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
259 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
260 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
261 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
262 `OPENSSL_NO_DEPRECATED_3_1`.
264 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
265 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
266 definitions for these functions regardless of whether
267 `OPENSSL_NO_DEPRECATED_3_1` is defined.
269 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
270 functions regardless of whether they are using them. It is recommended that
271 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
275 * When generating safe-prime DH parameters set the recommended private key
276 length equivalent to minimum key lengths as in RFC 7919.
280 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
281 maximum size that is smaller or equal to the digest length to comply with
282 FIPS 186-4 section 5. This is implemented by a new option
283 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
284 `rsa_pss_saltlen` parameter, which is now the default. Signature
285 verification is not affected by this change and continues to work as before.
292 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
293 listed here are only a brief description.
294 The migration guide contains more detailed information related to new features,
295 breaking changes, and mappings for the large list of deprecated functions.
297 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
299 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
301 * Fixed NULL dereference during PKCS7 data verification.
303 A NULL pointer can be dereferenced when signatures are being
304 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
305 algorithm used for the signature is known to the OpenSSL library but
306 the implementation of the hash algorithm is not available the digest
307 initialization will fail. There is a missing check for the return
308 value from the initialization function which later leads to invalid
309 usage of the digest API most likely leading to a crash.
312 PKCS7 data is processed by the SMIME library calls and also by the
313 time stamp (TS) library calls. The TLS implementation in OpenSSL does
314 not call these functions however third party applications would be
315 affected if they call these functions to verify signatures on untrusted
320 * Fixed X.400 address type confusion in X.509 GeneralName.
322 There is a type confusion vulnerability relating to X.400 address processing
323 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
324 but the public structure definition for GENERAL_NAME incorrectly specified
325 the type of the x400Address field as ASN1_TYPE. This field is subsequently
326 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
329 When CRL checking is enabled (i.e. the application sets the
330 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
331 pass arbitrary pointers to a memcmp call, enabling them to read memory
332 contents or enact a denial of service.
337 * Fixed NULL dereference validating DSA public key.
339 An invalid pointer dereference on read can be triggered when an
340 application tries to check a malformed DSA public key by the
341 EVP_PKEY_public_check() function. This will most likely lead
342 to an application crash. This function can be called on public
343 keys supplied from untrusted sources which could allow an attacker
344 to cause a denial of service attack.
346 The TLS implementation in OpenSSL does not call this function
347 but applications might call the function if there are additional
348 security requirements imposed by standards such as FIPS 140-3.
351 *Shane Lontis, Tomáš Mráz*
353 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
355 An invalid pointer dereference on read can be triggered when an
356 application tries to load malformed PKCS7 data with the
357 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
359 The result of the dereference is an application crash which could
360 lead to a denial of service attack. The TLS implementation in OpenSSL
361 does not call this function however third party applications might
362 call these functions on untrusted data.
367 * Fixed Use-after-free following BIO_new_NDEF.
369 The public API function BIO_new_NDEF is a helper function used for
370 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
371 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
372 be called directly by end user applications.
374 The function receives a BIO from the caller, prepends a new BIO_f_asn1
375 filter BIO onto the front of it to form a BIO chain, and then returns
376 the new head of the BIO chain to the caller. Under certain conditions,
377 for example if a CMS recipient public key is invalid, the new filter BIO
378 is freed and the function returns a NULL result indicating a failure.
379 However, in this case, the BIO chain is not properly cleaned up and the
380 BIO passed by the caller still retains internal pointers to the previously
381 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
382 then a use-after-free will occur. This will most likely result in a crash.
385 *Viktor Dukhovni, Matt Caswell*
387 * Fixed Double free after calling PEM_read_bio_ex.
389 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
390 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
391 data. If the function succeeds then the "name_out", "header" and "data"
392 arguments are populated with pointers to buffers containing the relevant
393 decoded data. The caller is responsible for freeing those buffers. It is
394 possible to construct a PEM file that results in 0 bytes of payload data.
395 In this case PEM_read_bio_ex() will return a failure code but will populate
396 the header argument with a pointer to a buffer that has already been freed.
397 If the caller also frees this buffer then a double free will occur. This
398 will most likely lead to a crash.
400 The functions PEM_read_bio() and PEM_read() are simple wrappers around
401 PEM_read_bio_ex() and therefore these functions are also directly affected.
403 These functions are also called indirectly by a number of other OpenSSL
404 functions including PEM_X509_INFO_read_bio_ex() and
405 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
406 internal uses of these functions are not vulnerable because the caller does
407 not free the header argument if PEM_read_bio_ex() returns a failure code.
410 *Kurt Roeckx, Matt Caswell*
412 * Fixed Timing Oracle in RSA Decryption.
414 A timing based side channel exists in the OpenSSL RSA Decryption
415 implementation which could be sufficient to recover a plaintext across
416 a network in a Bleichenbacher style attack. To achieve a successful
417 decryption an attacker would have to be able to send a very large number
418 of trial messages for decryption. The vulnerability affects all RSA padding
419 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
422 *Dmitry Belyavsky, Hubert Kario*
424 * Fixed X.509 Name Constraints Read Buffer Overflow.
426 A read buffer overrun can be triggered in X.509 certificate verification,
427 specifically in name constraint checking. The read buffer overrun might
428 result in a crash which could lead to a denial of service attack.
429 In a TLS client, this can be triggered by connecting to a malicious
430 server. In a TLS server, this can be triggered if the server requests
431 client authentication and a malicious client connects.
436 * Fixed X.509 Policy Constraints Double Locking security issue.
438 If an X.509 certificate contains a malformed policy constraint and
439 policy processing is enabled, then a write lock will be taken twice
440 recursively. On some operating systems (most widely: Windows) this
441 results in a denial of service when the affected process hangs. Policy
442 processing being enabled on a publicly facing server is not considered
443 to be a common setup.
448 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
449 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
450 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
451 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
452 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
453 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
454 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
455 for legacy EC and SM2 keys is also changed similarly to honor the
456 equivalent conversion format flag as specified in the underlying
457 `EC_KEY` object being exported to a provider, when this function is
458 called through `EVP_PKEY_export()`.
462 * Fixed a type confusion vulnerability relating to X.400 address processing
463 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
464 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
465 vulnerability may allow an attacker who can provide a certificate chain and
466 CRL (neither of which need have a valid signature) to pass arbitrary pointers
467 to a `memcmp` call, creating a possible read primitive, subject to some
468 constraints. Refer to the advisory for more information. Thanks to David
469 Benjamin for discovering this issue. ([CVE-2023-0286])
471 This issue has been fixed by changing the public header file definition of
472 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
473 possible for any existing application to successfully use the existing
474 definition; however, if any application references the `x400Address` field
475 (e.g. in dead code), note that the type of this field has changed. There is
480 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
482 * Fixed two buffer overflows in punycode decoding functions.
484 A buffer overrun can be triggered in X.509 certificate verification,
485 specifically in name constraint checking. Note that this occurs after
486 certificate chain signature verification and requires either a CA to
487 have signed the malicious certificate or for the application to continue
488 certificate verification despite failure to construct a path to a trusted
491 In a TLS client, this can be triggered by connecting to a malicious
492 server. In a TLS server, this can be triggered if the server requests
493 client authentication and a malicious client connects.
495 An attacker can craft a malicious email address to overflow
496 an arbitrary number of bytes containing the `.` character (decimal 46)
497 on the stack. This buffer overflow could result in a crash (causing a
501 An attacker can craft a malicious email address to overflow four
502 attacker-controlled bytes on the stack. This buffer overflow could
503 result in a crash (causing a denial of service) or potentially remote code
504 execution depending on stack layout for any given platform/compiler.
509 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
510 parameters in OpenSSL code.
511 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
512 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
513 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
514 Using these invalid names may cause algorithms to use slower methods
515 that ignore the CRT parameters.
519 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
524 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
525 data to be signed before signing the certificate.
529 * Added RIPEMD160 to the default provider.
533 * Ensured that the key share group sent or accepted for the key exchange
534 is allowed for the protocol version.
538 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
540 * OpenSSL supports creating a custom cipher via the legacy
541 EVP_CIPHER_meth_new() function and associated function calls. This function
542 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
543 to use the new provider mechanism in order to implement custom ciphers.
545 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
546 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
547 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
548 and decryption initialisation functions). Instead of using the custom cipher
549 directly it incorrectly tries to fetch an equivalent cipher from the
550 available providers. An equivalent cipher is found based on the NID passed to
551 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
552 given cipher. However it is possible for an application to incorrectly pass
553 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
554 is used in this way the OpenSSL encryption/decryption initialisation function
555 will match the NULL cipher as being equivalent and will fetch this from the
556 available providers. This will succeed if the default provider has been
557 loaded (or if a third party provider has been loaded that offers this
558 cipher). Using the NULL cipher means that the plaintext is emitted as the
561 Applications are only affected by this issue if they call
562 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
563 encryption/decryption initialisation function. Applications that only use
564 SSL/TLS are not impacted by this issue.
569 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
574 * Fixed the linux-mips64 Configure target which was missing the
575 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
580 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
585 * Correctly handle a retransmitted ClientHello in DTLS
589 * Fixed detection of ktls support in cross-compile environment on Linux
593 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
598 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
599 report correct results in some cases
603 * Fix UWP builds by defining VirtualLock
607 * For known safe primes use the minimum key length according to RFC 7919.
608 Longer private key sizes unnecessarily raise the cycles needed to compute the
609 shared secret without any increase of the real security. This fixes a
610 regression from 1.1.1 where these shorter keys were generated for the known
615 * Added the loongarch64 target
619 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
620 only passed to the FIPS provider and not to the default or legacy provider.
624 * Fixed reported performance degradation on aarch64. Restored the
625 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
626 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
627 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
628 The new algorithm is still used for 32 bit targets.
632 * Added a missing header for memcmp that caused compilation failure on some
637 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
639 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
640 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
641 This issue makes the RSA implementation with 2048 bit private keys
642 incorrect on such machines and memory corruption will happen during
643 the computation. As a consequence of the memory corruption an attacker
644 may be able to trigger a remote code execution on the machine performing
647 SSL/TLS servers or other servers using 2048 bit RSA private keys running
648 on machines supporting AVX512IFMA instructions of the X86_64 architecture
649 are affected by this issue.
654 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
655 implementation would not encrypt the entirety of the data under some
656 circumstances. This could reveal sixteen bytes of data that was
657 preexisting in the memory that wasn't written. In the special case of
658 "in place" encryption, sixteen bytes of the plaintext would be revealed.
660 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
661 they are both unaffected.
664 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
666 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
668 * In addition to the c_rehash shell command injection identified in
669 CVE-2022-1292, further bugs where the c_rehash script does not
670 properly sanitise shell metacharacters to prevent command injection have been
673 When the CVE-2022-1292 was fixed it was not discovered that there
674 are other places in the script where the file names of certificates
675 being hashed were possibly passed to a command executed through the shell.
677 This script is distributed by some operating systems in a manner where
678 it is automatically executed. On such operating systems, an attacker
679 could execute arbitrary commands with the privileges of the script.
681 Use of the c_rehash script is considered obsolete and should be replaced
682 by the OpenSSL rehash command line tool.
685 *Daniel Fiala, Tomáš Mráz*
687 * Case insensitive string comparison no longer uses locales. It has instead
688 been directly implemented.
692 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
694 * Case insensitive string comparison is reimplemented via new locale-agnostic
695 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
696 comparison. The previous implementation had problems when the Turkish locale
701 * Fixed a bug in the c_rehash script which was not properly sanitising shell
702 metacharacters to prevent command injection. This script is distributed by
703 some operating systems in a manner where it is automatically executed. On
704 such operating systems, an attacker could execute arbitrary commands with the
705 privileges of the script.
707 Use of the c_rehash script is considered obsolete and should be replaced
708 by the OpenSSL rehash command line tool.
713 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
714 certificate on an OCSP response. The bug caused the function in the case
715 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
716 response (meaning a successful verification) even in the case where the
717 response signing certificate fails to verify.
719 It is anticipated that most users of `OCSP_basic_verify` will not use the
720 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
721 a negative value (indicating a fatal error) in the case of a certificate
722 verification failure. The normal expected return value in this case would be
725 This issue also impacts the command line OpenSSL "ocsp" application. When
726 verifying an ocsp response with the "-no_cert_checks" option the command line
727 application will report that the verification is successful even though it
728 has in fact failed. In this case the incorrect successful response will also
729 be accompanied by error messages showing the failure and contradicting the
730 apparently successful result.
735 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
736 AAD data as the MAC key. This made the MAC key trivially predictable.
738 An attacker could exploit this issue by performing a man-in-the-middle attack
739 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
740 that the modified data would still pass the MAC integrity check.
742 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
743 endpoint will always be rejected by the recipient and the connection will
744 fail at that point. Many application protocols require data to be sent from
745 the client to the server first. Therefore, in such a case, only an OpenSSL
746 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
748 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
749 sent in both directions. In this case both clients and servers could be
750 affected, regardless of the application protocol.
752 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
753 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
754 the handshake when using this ciphersuite.
756 The confidentiality of data is not impacted by this issue, i.e. an attacker
757 cannot decrypt data that has been encrypted using this ciphersuite - they can
760 In order for this attack to work both endpoints must legitimately negotiate
761 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
762 OpenSSL 3.0, and is not available within the default provider or the default
763 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
764 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
765 following must have occurred:
767 1) OpenSSL must have been compiled with the (non-default) compile time option
768 enable-weak-ssl-ciphers
770 2) OpenSSL must have had the legacy provider explicitly loaded (either
771 through application code or via configuration)
773 3) The ciphersuite must have been explicitly added to the ciphersuite list
775 4) The libssl security level must have been set to 0 (default is 1)
777 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
779 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
780 others that both endpoints have in common
785 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
786 occupied by the removed hash table entries.
788 This function is used when decoding certificates or keys. If a long lived
789 process periodically decodes certificates or keys its memory usage will
790 expand without bounds and the process might be terminated by the operating
791 system causing a denial of service. Also traversing the empty hash table
792 entries will take increasingly more time.
794 Typically such long lived processes might be TLS clients or TLS servers
795 configured to accept client certificate authentication.
798 *Hugo Landau, Aliaksei Levin*
800 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
801 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
802 statistics are no longer supported. For compatibility, these statistics are
803 still listed in the output but are now always reported as zero.
807 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
809 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
810 for non-prime moduli.
812 Internally this function is used when parsing certificates that contain
813 elliptic curve public keys in compressed form or explicit elliptic curve
814 parameters with a base point encoded in compressed form.
816 It is possible to trigger the infinite loop by crafting a certificate that
817 has invalid explicit curve parameters.
819 Since certificate parsing happens prior to verification of the certificate
820 signature, any process that parses an externally supplied certificate may thus
821 be subject to a denial of service attack. The infinite loop can also be
822 reached when parsing crafted private keys as they can contain explicit
823 elliptic curve parameters.
825 Thus vulnerable situations include:
827 - TLS clients consuming server certificates
828 - TLS servers consuming client certificates
829 - Hosting providers taking certificates or private keys from customers
830 - Certificate authorities parsing certification requests from subscribers
831 - Anything else which parses ASN.1 elliptic curve parameters
833 Also any other applications that use the BN_mod_sqrt() where the attacker
834 can control the parameter values are vulnerable to this DoS issue.
839 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
840 to the list of ciphersuites providing Perfect Forward Secrecy as
841 required by SECLEVEL >= 3.
843 *Dmitry Belyavskiy, Nicola Tuveri*
845 * Made the AES constant time code for no-asm configurations
846 optional due to the resulting 95% performance degradation.
847 The AES constant time code can be enabled, for no assembly
848 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
852 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
857 * The negative return value handling of the certificate verification callback
858 was reverted. The replacement is to set the verification retry state with
859 the SSL_set_retry_verify() function.
863 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
865 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
866 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
867 verify a certificate supplied by a server. That function may return a
868 negative return value to indicate an internal error (for example out of
869 memory). Such a negative return value is mishandled by OpenSSL and will cause
870 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
871 success and a subsequent call to SSL_get_error() to return the value
872 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
873 returned by OpenSSL if the application has previously called
874 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
875 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
876 totally unexpected and applications may not behave correctly as a result. The
877 exact behaviour will depend on the application but it could result in
878 crashes, infinite loops or other similar incorrect responses.
880 This issue is made more serious in combination with a separate bug in OpenSSL
881 3.0 that will cause X509_verify_cert() to indicate an internal error when
882 processing a certificate chain. This will occur where a certificate does not
883 include the Subject Alternative Name extension but where a Certificate
884 Authority has enforced name constraints. This issue can occur even with valid
890 * Corrected a few file name and file reference bugs in the build,
891 installation and setup scripts, which lead to installation verification
892 failures. Slightly enhanced the installation verification script.
896 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
901 * Fixed PVK encoder to properly query for the passphrase.
905 * Multiple fixes in the OSSL_HTTP API functions.
909 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
910 OSSL_PARAM_INTEGER data type and return error on negative numbers
911 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
912 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
916 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
920 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
924 * Multiple threading fixes.
928 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
932 * Allow fetching an operation from the provider that owns an unexportable key
933 as a fallback if that is still allowed by the property query.
937 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
939 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
944 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
945 S390X capability vector to zero. This simplifies testing of different code
946 paths on S390X architecture.
950 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
951 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
952 SP 800-38D". The communication will fail at this point.
956 * The EC_GROUP_clear_free() function is deprecated as there is nothing
957 confidential in EC_GROUP data.
961 * The byte order mark (BOM) character is ignored if encountered at the
962 beginning of a PEM-formatted file.
966 * Added CMS support for the Russian GOST algorithms.
970 * Due to move of the implementation of cryptographic operations
971 to the providers, validation of various operation parameters can
972 be postponed until the actual operation is executed where previously
973 it happened immediately when an operation parameter was set.
975 For example when setting an unsupported curve with
976 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
977 fail but later keygen operations with the EVP_PKEY_CTX will fail.
979 *OpenSSL team members and many third party contributors*
981 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
982 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
983 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
984 instead to retrieve these algorithms from a provider.
988 * On build targets where the multilib postfix is set in the build
989 configuration the libdir directory was changing based on whether
990 the lib directory with the multilib postfix exists on the system
991 or not. This unpredictable behavior was removed and eventual
992 multilib postfix is now always added to the default libdir. Use
993 `--libdir=lib` to override the libdir if adding the postfix is
998 * The triple DES key wrap functionality now conforms to RFC 3217 but is
999 no longer interoperable with OpenSSL 1.1.1.
1003 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1004 function codes, this function can only cause problems for calling
1009 * Add a configurable flag to output date formats as ISO 8601. Does not
1010 change the default date format.
1014 * Version of MSVC earlier than 1300 could get link warnings, which could
1015 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1016 Support for this flag has been removed.
1020 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1021 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1022 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1023 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1024 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1028 * The signatures of the functions to get and set options on SSL and
1029 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1030 Some source code changes may be required.
1034 * The public definitions of conf_method_st and conf_st have been
1035 deprecated. They will be made opaque in a future release.
1037 *Rich Salz and Tomáš Mráz*
1039 * Client-initiated renegotiation is disabled by default. To allow it, use
1040 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1041 flag, or the "ClientRenegotiation" config parameter as appropriate.
1045 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1046 or modify relative pathname inclusion.
1050 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1051 validated. Please consult the README-FIPS and
1052 README-PROVIDERS files, as well as the migration guide.
1054 *OpenSSL team members and many third party contributors*
1056 * For the key types DH and DHX the allowed settable parameters are now different.
1060 * The openssl commands that read keys, certificates, and CRLs now
1061 automatically detect the PEM or DER format of the input files.
1063 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1065 * Added enhanced PKCS#12 APIs which accept a library context.
1069 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1073 * Added support for Kernel TLS (KTLS).
1075 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1077 * Support for RFC 5746 secure renegotiation is now required by default for
1078 SSL or TLS connections to succeed.
1082 * The signature of the `copy` functional parameter of the
1083 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1084 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1085 the signature of the `pub_decode` functional parameter of the
1086 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1087 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1091 * The error return values from some control calls (ctrl) have changed.
1095 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1099 * Many functions in the EVP_ namespace that are getters of values from
1100 implementations or contexts were renamed to include get or get0 in their
1101 names. Old names are provided as macro aliases for compatibility and
1106 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1107 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1108 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1113 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1116 * The output from the command line applications may have minor
1121 * The output from numerous "printing" may have minor changes.
1125 * Windows thread synchronization uses read/write primitives (SRWLock) when
1126 supported by the OS, otherwise CriticalSection continues to be used.
1130 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1131 work on read only BIO source/sinks that do not support these functions.
1132 This allows piping or redirection of a file BIO using stdin to be buffered
1133 into memory. This is used internally in OSSL_DECODER_from_bio().
1137 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1138 this function would return one of the values OSSL_STORE_INFO_NAME,
1139 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1140 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1141 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1142 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1143 using this function should be amended to handle the changed return value.
1147 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1148 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1149 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1150 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1151 contains more than one certificate identifier: This means that all
1152 certificates referenced there MUST be part of the validation chain.
1156 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1157 RC5, DESX and DES have been moved to the legacy provider.
1161 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1162 RIPEMD-160 have been moved to the legacy provider.
1166 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1171 * The deprecated functions EVP_PKEY_get0_RSA(),
1172 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1173 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1174 well as the similarly named "get1" functions behave differently in
1179 * A number of functions handling low-level keys or engines were deprecated
1180 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1181 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1182 EVP_PKEY_get0_siphash().
1186 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1187 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1188 will need to load the legacy crypto provider. This includes these PBE
1189 algorithms which use this KDF:
1190 - NID_pbeWithMD2AndDES_CBC
1191 - NID_pbeWithMD5AndDES_CBC
1192 - NID_pbeWithSHA1AndRC2_CBC
1193 - NID_pbeWithMD2AndRC2_CBC
1194 - NID_pbeWithMD5AndRC2_CBC
1195 - NID_pbeWithSHA1AndDES_CBC
1199 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1200 BIO_debug_callback() functions.
1204 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1205 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1209 * The RAND_METHOD APIs have been deprecated.
1213 * The SRP APIs have been deprecated.
1217 * Add a compile time option to prevent the caching of provider fetched
1218 algorithms. This is enabled by including the no-cached-fetch option
1219 at configuration time.
1223 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1224 count of PKCS12_DEFAULT_ITER.
1226 *Tomáš Mráz and Sahana Prasad*
1228 * The openssl speed command does not use low-level API calls anymore.
1232 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1235 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1237 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1241 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1242 providers may supply their own group implementations (using either the "key
1243 exchange" or the "key encapsulation" methods) which will automatically be
1244 detected and used by libssl.
1246 *Matt Caswell, Nicola Tuveri*
1248 * The undocumented function X509_certificate_type() has been deprecated;
1252 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1256 * Removed RSA padding mode for SSLv23 (which was only used for
1257 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1258 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1263 * Deprecated the obsolete X9.31 RSA key generation related functions.
1265 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1266 is not allowed to return a value > 1, this is no more taken as failure.
1268 *Viktor Dukhovni and David von Oheimb*
1270 * Deprecated the obsolete X9.31 RSA key generation related functions
1271 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1272 BN_X931_generate_prime_ex().
1276 * The default key generation method for the regular 2-prime RSA keys was
1277 changed to the FIPS 186-4 B.3.6 method.
1281 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1285 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1289 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1290 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1292 *Rich Salz, Richard Levitte, and David von Oheimb*
1294 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1298 * Deprecated `OCSP_parse_url()`.
1302 * Validation of SM2 keys has been separated from the validation of regular EC
1307 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1308 switches: a validation failure triggers an early exit, returning a failure
1309 exit status to the parent process.
1313 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1314 to ignore unknown ciphers.
1318 * The `-cipher-commands` and `-digest-commands` options
1319 of the command line utility `list` have been deprecated.
1320 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1324 * Added convenience functions for generating asymmetric key pairs:
1325 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1326 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1330 * All of the low-level EC_KEY functions have been deprecated.
1332 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1334 * Deprecated all the libcrypto and libssl error string loading
1339 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1340 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1345 * The `-crypt` option to the `passwd` command line tool has been removed.
1349 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1354 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1358 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1359 EVP_PKEY_get1_tls_encodedpoint().
1363 * The security callback, which can be customised by application code, supports
1364 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1365 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1369 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1370 interface. Their functionality remains unchanged.
1374 * Added new option for 'openssl list', '-providers', which will display the
1375 list of loaded providers, their names, version and status. It optionally
1376 displays their gettable parameters.
1380 * Removed EVP_PKEY_set_alias_type().
1384 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1385 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1389 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1390 parameters are still checked for type safety at compile time via helper
1395 * Remove the RAND_DRBG API
1397 *Paul Dale and Matthias St. Pierre*
1399 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1400 as well as actual hostnames.
1404 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1405 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1406 conversely, silently ignore DTLS protocol version bounds when configuring
1407 TLS-based contexts. The commands can be repeated to set bounds of both
1408 types. The same applies with the corresponding "min_protocol" and
1409 "max_protocol" command-line switches, in case some application uses both TLS
1412 SSL_CTX instances that are created for a fixed protocol version (e.g.
1413 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1414 attempts to apply bounds to these protocol versions would result in an
1415 error. Now only the "version-flexible" SSL_CTX instances are subject to
1416 limits in configuration files in command-line options.
1420 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1425 * Reworked the recorded ERR codes to make better space for system errors.
1426 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1427 given code is a system error (true) or an OpenSSL error (false).
1431 * Reworked the test perl framework to better allow parallel testing.
1433 *Nicola Tuveri and David von Oheimb*
1435 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1436 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1440 * 'Configure' has been changed to figure out the configuration target if
1441 none is given on the command line. Consequently, the 'config' script is
1442 now only a mere wrapper. All documentation is changed to only mention
1445 *Rich Salz and Richard Levitte*
1447 * Added a library context `OSSL_LIB_CTX` that applications as well as
1448 other libraries can use to form a separate context within which
1449 libcrypto operations are performed.
1453 * Added various `_ex` functions to the OpenSSL API that support using
1454 a non-default `OSSL_LIB_CTX`.
1458 * Handshake now fails if Extended Master Secret extension is dropped
1463 * Dropped interactive mode from the `openssl` program.
1467 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1469 *David von Oheimb and Shane Lontis*
1471 * Deprecated `EC_METHOD_get_field_type()`.
1475 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1476 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1477 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1481 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1485 * Add CAdES-BES signature verification support, mostly derived
1486 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1488 *Filipe Raimundo da Silva*
1490 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1494 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1495 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1499 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1503 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1504 EC_KEY_precompute_mult().
1508 * Deprecated EC_POINTs_mul().
1512 * Removed FIPS_mode() and FIPS_mode_set().
1516 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1520 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1521 EC_POINT_get_Jprojective_coordinates_GFp().
1525 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1526 arrays to be more easily constructed via a series of utility functions.
1527 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1528 the various push functions and finally convert to a passable OSSL_PARAM
1529 array using OSSL_PARAM_BLD_to_param().
1533 * The security strength of SHA1 and MD5 based signatures in TLS has been
1538 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1539 contain a provider side internal key.
1543 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1547 * Project text documents not yet having a proper file name extension
1548 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1549 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1553 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1554 have been converted to Markdown with the goal to produce documents
1555 which not only look pretty when viewed online in the browser, but
1556 remain well readable inside a plain text editor.
1558 To achieve this goal, a 'minimalistic' Markdown style has been applied
1559 which avoids formatting elements that interfere too much with the
1560 reading flow in the text file. For example, it
1562 * avoids [ATX headings][] and uses [setext headings][] instead
1563 (which works for `<h1>` and `<h2>` headings only).
1564 * avoids [inline links][] and uses [reference links][] instead.
1565 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1567 [ATX headings]: https://github.github.com/gfm/#atx-headings
1568 [setext headings]: https://github.github.com/gfm/#setext-headings
1569 [inline links]: https://github.github.com/gfm/#inline-link
1570 [reference links]: https://github.github.com/gfm/#reference-link
1571 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1572 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1574 *Matthias St. Pierre*
1576 * The test suite is changed to preserve results of each test recipe.
1577 A new directory test-runs/ with subdirectories named like the
1578 test recipes are created in the build tree for this purpose.
1582 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1583 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1584 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1586 *David von Oheimb, Martin Peylo*
1588 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1589 It supports arbitrary request and response content types, GET redirection,
1590 TLS, connections via HTTP(S) proxies, connections and exchange via
1591 user-defined BIOs (allowing implicit connections), persistent connections,
1592 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1593 The legacy OCSP-focused (and only partly documented) API
1594 is retained for backward compatibility, while most of it is deprecated.
1598 * Added `util/check-format.pl`, a tool for checking adherence to the
1599 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1600 The checks performed are incomplete and yield some false positives.
1601 Still the tool should be useful for detecting most typical glitches.
1605 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1606 If domain name resolution yields multiple IP addresses all of them are tried
1607 after `connect()` failures.
1611 * All of the low-level RSA functions have been deprecated.
1615 * X509 certificates signed using SHA1 are no longer allowed at security
1620 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1621 modified to use PKEY APIs. These commands are now in maintenance mode
1622 and no new features will be added to them.
1626 * The command line utility rsautl has been deprecated.
1630 * The command line utilities genrsa and rsa have been modified to use PKEY
1631 APIs. They now write PKCS#8 keys by default. These commands are now in
1632 maintenance mode and no new features will be added to them.
1636 * All of the low-level DH functions have been deprecated.
1638 *Paul Dale and Matt Caswell*
1640 * All of the low-level DSA functions have been deprecated.
1644 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1645 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1649 * Deprecated low-level ECDH and ECDSA functions.
1653 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1657 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1658 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1659 a new formulation to include all the things it can be used for,
1660 as well as words of caution.
1664 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1668 * All of the low-level HMAC functions have been deprecated.
1670 *Paul Dale and David von Oheimb*
1672 * Over two thousand fixes were made to the documentation, including:
1673 - Common options (such as -rand/-writerand, TLS version control, etc)
1674 were refactored and point to newly-enhanced descriptions in openssl.pod.
1675 - Added style conformance for all options (with help from Richard Levitte),
1676 documented all reported missing options, added a CI build to check
1677 that all options are documented and that no unimplemented options
1679 - Documented some internals, such as all use of environment variables.
1680 - Addressed all internal broken L<> references.
1684 * All of the low-level CMAC functions have been deprecated.
1688 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1689 functions have been deprecated.
1691 *Paul Dale and David von Oheimb*
1693 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1694 set of functions. The documentation mentioned negative values for some
1695 errors, but this was never the case, so the mention of negative values
1698 Code that followed the documentation and thereby check with something
1699 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1703 * All of the low-level cipher functions have been deprecated.
1705 *Matt Caswell and Paul Dale*
1707 * Removed include/openssl/opensslconf.h.in and replaced it with
1708 include/openssl/configuration.h.in, which differs in not including
1709 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1710 was added to include both.
1712 This allows internal hacks where one might need to modify the set
1713 of configured macros, for example this if deprecated symbols are
1714 still supposed to be available internally:
1716 #include <openssl/configuration.h>
1718 #undef OPENSSL_NO_DEPRECATED
1719 #define OPENSSL_SUPPRESS_DEPRECATED
1721 #include <openssl/macros.h>
1723 This should not be used by applications that use the exported
1724 symbols, as that will lead to linking errors.
1728 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1729 used in exponentiation with 512-bit moduli. No EC algorithms are
1730 affected. Analysis suggests that attacks against 2-prime RSA1024,
1731 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1732 difficult to perform and are not believed likely. Attacks against DH512
1733 are considered just feasible. However, for an attack the target would
1734 have to re-use the DH512 private key, which is not recommended anyway.
1735 Also applications directly using the low-level API BN_mod_exp may be
1736 affected if they use BN_FLG_CONSTTIME.
1741 * Most memory-debug features have been deprecated, and the functionality
1742 replaced with no-ops.
1746 * Added documentation for the STACK API.
1750 * Introduced a new method type and API, OSSL_ENCODER, to represent
1751 generic encoders. These do the same sort of job that PEM writers
1752 and d2i functions do, but with support for methods supplied by
1753 providers, and the possibility for providers to support other
1758 * Introduced a new method type and API, OSSL_DECODER, to represent
1759 generic decoders. These do the same sort of job that PEM readers
1760 and i2d functions do, but with support for methods supplied by
1761 providers, and the possibility for providers to support other
1766 * Added a .pragma directive to the syntax of configuration files, to
1767 allow varying behavior in a supported and predictable manner.
1768 Currently added pragma:
1772 This allows dollar signs to be a keyword character unless it's
1773 followed by a opening brace or parenthesis. This is useful for
1774 platforms where dollar signs are commonly used in names, such as
1775 volume names and system directory names on VMS.
1779 * Added functionality to create an EVP_PKEY from user data.
1783 * Change the interpretation of the '--api' configuration option to
1784 mean that this is a desired API compatibility level with no
1785 further meaning. The previous interpretation, that this would
1786 also mean to remove all deprecated symbols up to and including
1787 the given version, no requires that 'no-deprecated' is also used
1788 in the configuration.
1790 When building applications, the desired API compatibility level
1791 can be set with the OPENSSL_API_COMPAT macro like before. For
1792 API compatibility version below 3.0, the old style numerical
1793 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1794 For version 3.0 and on, the value is expected to be the decimal
1795 value calculated from the major and minor version like this:
1797 MAJOR * 10000 + MINOR * 100
1801 -DOPENSSL_API_COMPAT=30000 For 3.0
1802 -DOPENSSL_API_COMPAT=30200 For 3.2
1804 To hide declarations that are deprecated up to and including the
1805 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1806 given when building the application as well.
1810 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1811 access to certificate and CRL stores via URIs and OSSL_STORE
1814 This adds the following functions:
1816 - X509_LOOKUP_store()
1817 - X509_STORE_load_file()
1818 - X509_STORE_load_path()
1819 - X509_STORE_load_store()
1820 - SSL_add_store_cert_subjects_to_stack()
1821 - SSL_CTX_set_default_verify_store()
1822 - SSL_CTX_load_verify_file()
1823 - SSL_CTX_load_verify_dir()
1824 - SSL_CTX_load_verify_store()
1828 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1829 The presence of this system service is determined at run-time.
1833 * Added functionality to create an EVP_PKEY context based on data
1834 for methods from providers. This takes an algorithm name and a
1835 property query string and simply stores them, with the intent
1836 that any operation that uses this context will use those strings
1837 to fetch the needed methods implicitly, thereby making the port
1838 of application written for pre-3.0 OpenSSL easier.
1842 * The undocumented function NCONF_WIN32() has been deprecated; for
1843 conversion details see the HISTORY section of doc/man5/config.pod
1847 * Introduced the new functions EVP_DigestSignInit_ex() and
1848 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1849 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1850 pages for further details.
1854 * Over two thousand fixes were made to the documentation, including:
1855 adding missing command flags, better style conformance, documentation
1858 *Rich Salz, Richard Levitte*
1860 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1861 X25519, X448, Ed25519 and Ed448.
1865 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1870 * Deprecated the public definition of `ERR_STATE` as well as the function
1871 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1876 * Added ERR functionality to give callers access to the stored function
1877 names that have replaced the older function code based functions.
1879 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1880 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1881 ERR_peek_error_all() and ERR_peek_last_error_all().
1883 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1884 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1885 ERR_func_error_string().
1889 * Extended testing to be verbose for failing tests only. The make variables
1890 VERBOSE_FAILURE or VF can be used to enable this:
1892 $ make VF=1 test # Unix
1893 $ mms /macro=(VF=1) test ! OpenVMS
1894 $ nmake VF=1 test # Windows
1898 * Added the `-copy_extensions` option to the `x509` command for use with
1899 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1900 all extensions in the request are copied to the certificate or vice versa.
1902 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1904 * Added the `-copy_extensions` option to the `req` command for use with
1905 `-x509`. When given with the `copy` or `copyall` argument,
1906 all extensions in the certification request are copied to the certificate.
1910 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1911 they generate are by default RFC 5280 compliant in the following sense:
1912 There is a subjectKeyIdentifier extension with a hash value of the public key
1913 and for not self-signed certs there is an authorityKeyIdentifier extension
1914 with a keyIdentifier field or issuer information identifying the signing key.
1915 This is done unless some configuration overrides the new default behavior,
1916 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1920 * Added several checks to `X509_verify_cert()` according to requirements in
1921 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1922 (which may be done by using the CLI option `-x509_strict`):
1923 * The basicConstraints of CA certificates must be marked critical.
1924 * CA certificates must explicitly include the keyUsage extension.
1925 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1926 * The issuer name of any certificate must not be empty.
1927 * The subject name of CA certs, certs with keyUsage crlSign,
1928 and certs without subjectAlternativeName must not be empty.
1929 * If a subjectAlternativeName extension is given it must not be empty.
1930 * The signatureAlgorithm field and the cert signature must be consistent.
1931 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1932 must not be marked critical.
1933 * The authorityKeyIdentifier must be given for X.509v3 certs
1934 unless they are self-signed.
1935 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1939 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1940 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1944 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1945 used even when parsing explicit parameters, when loading a encoded key
1946 or calling `EC_GROUP_new_from_ecpkparameters()`/
1947 `EC_GROUP_new_from_ecparameters()`.
1948 This prevents bypass of security hardening and performance gains,
1949 especially for curves with specialized EC_METHODs.
1950 By default, if a key encoded with explicit parameters is loaded and later
1951 encoded, the output is still encoded with explicit parameters, even if
1952 internally a "named" EC_GROUP is used for computation.
1956 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1957 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1958 NULL. After this change, only the cofactor parameter can be NULL. It also
1959 does some minimal sanity checks on the passed order.
1964 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1965 An attack is simple, if the first CMS_recipientInfo is valid but the
1966 second CMS_recipientInfo is chosen ciphertext. If the second
1967 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1968 encryption key will be replaced by garbage, and the message cannot be
1969 decoded, but if the RSA decryption fails, the correct encryption key is
1970 used and the recipient will not notice the attack.
1971 As a work around for this potential attack the length of the decrypted
1972 key must be equal to the cipher default key length, in case the
1973 certificate is not given and all recipientInfo are tried out.
1974 The old behaviour can be re-enabled in the CMS code by setting the
1975 CMS_DEBUG_DECRYPT flag.
1979 * Early start up entropy quality from the DEVRANDOM seed source has been
1980 improved for older Linux systems. The RAND subsystem will wait for
1981 /dev/random to be producing output before seeding from /dev/urandom.
1982 The seeded state is stored for future library initialisations using
1983 a system global shared memory segment. The shared memory identifier
1984 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1985 the desired value. The default identifier is 114.
1989 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1990 when primes for RSA keys are computed.
1991 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1992 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1993 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1994 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1995 This avoids possible fingerprinting of newly generated RSA modules.
1999 * Correct the extended master secret constant on EBCDIC systems. Without this
2000 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2001 negotiate EMS will fail. Unfortunately this also means that TLS connections
2002 between EBCDIC systems with this fix, and EBCDIC systems without this
2003 fix will fail if they negotiate EMS.
2007 * Changed the library initialisation so that the config file is now loaded
2008 by default. This was already the case for libssl. It now occurs for both
2009 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2010 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2014 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2015 where the former acts as a replacement for `ERR_put_error()`, and the
2016 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2017 `ERR_raise_data()` adds more flexibility by taking a format string and
2018 an arbitrary number of arguments following it, to be processed with
2023 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2024 to check if a named provider is loaded and available. When called, it
2025 will also activate all fallback providers if such are still present.
2029 * Enforce a minimum DH modulus size of 512 bits.
2033 * Changed DH parameters to generate the order q subgroup instead of 2q.
2034 Previously generated DH parameters are still accepted by DH_check
2035 but DH_generate_key works around that by clearing bit 0 of the
2036 private key for those. This avoids leaking bit 0 of the private key.
2040 * Significantly reduce secure memory usage by the randomness pools.
2044 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2049 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2050 algorithms. An implementation of a key exchange algorithm can be obtained
2051 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2052 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2053 the older EVP_PKEY_derive_init() function. See the man pages for the new
2054 functions for further details.
2058 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2062 * Removed the function names from error messages and deprecated the
2067 * Removed NextStep support and the macro OPENSSL_UNISTD
2071 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2072 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2073 Also removed "export var as function" capability; we do not export
2074 variables, only functions.
2078 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2079 an error and 1 indicating success. In previous versions of OpenSSL this
2080 was a void type. If a key was set longer than the maximum possible this
2085 * Support SM2 signing and verification schemes with X509 certificate.
2089 * Use SHA256 as the default digest for TS query in the `ts` app.
2093 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2097 * Default cipher lists/suites are now available via a function, the
2098 #defines are deprecated.
2102 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2103 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2104 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2108 * Join the directories crypto/x509 and crypto/x509v3
2112 * Added command 'openssl kdf' that uses the EVP_KDF API.
2116 * Added command 'openssl mac' that uses the EVP_MAC API.
2120 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2121 as default directories. Also added the command 'openssl info'
2122 for scripting purposes.
2126 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2131 * Add prediction resistance to the DRBG reseeding process.
2135 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2136 mandated by IEEE Std 1619-2018.
2140 * Added newline escaping functionality to a filename when using openssl dgst.
2141 This output format is to replicate the output format found in the `*sum`
2142 checksum programs. This aims to preserve backward compatibility.
2144 *Matt Eaton, Richard Levitte, and Paul Dale*
2146 * Removed the heartbeat message in DTLS feature, as it has very
2147 little usage and doesn't seem to fulfill a valuable purpose.
2148 The configuration option is now deprecated.
2152 * Changed the output of 'openssl {digestname} < file' to display the
2153 digest name in its output.
2157 * Added a new generic trace API which provides support for enabling
2158 instrumentation through trace output.
2160 *Richard Levitte & Matthias St. Pierre*
2162 * Added build tests for C++. These are generated files that only do one
2163 thing, to include one public OpenSSL head file each. This tests that
2164 the public header files can be usefully included in a C++ application.
2166 This test isn't enabled by default. It can be enabled with the option
2167 'enable-buildtest-c++'.
2171 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2175 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2179 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2183 * Added KMAC to EVP_MAC.
2187 * Added property based algorithm implementation selection framework to
2192 * Added SCA hardening for modular field inversion in EC_GROUP through
2193 a new dedicated field_inv() pointer in EC_METHOD.
2194 This also addresses a leakage affecting conversions from projective
2195 to affine coordinates.
2197 *Billy Bob Brumley, Nicola Tuveri*
2199 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2200 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2201 those algorithms that were already supported through the EVP_PKEY API
2202 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2203 and scrypt are now wrappers that call EVP_KDF.
2207 * Build devcrypto engine as a dynamic engine.
2209 *Eneas U de Queiroz*
2211 * Add keyed BLAKE2 to EVP_MAC.
2215 * Fix a bug in the computation of the endpoint-pair shared secret used
2216 by DTLS over SCTP. This breaks interoperability with older versions
2217 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2218 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2219 interoperability with such broken implementations. However, enabling
2220 this switch breaks interoperability with correct implementations.
2222 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2223 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2227 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2231 * Changed the license to the Apache License v2.0.
2235 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2237 - Major releases (indicated by incrementing the MAJOR release number)
2238 may introduce incompatible API/ABI changes.
2239 - Minor releases (indicated by incrementing the MINOR release number)
2240 may introduce new features but retain API/ABI compatibility.
2241 - Patch releases (indicated by incrementing the PATCH number)
2242 are intended for bug fixes and other improvements of existing
2243 features only (like improving performance or adding documentation)
2244 and retain API/ABI compatibility.
2248 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2252 * Remove the 'dist' target and add a tarball building script. The
2253 'dist' target has fallen out of use, and it shouldn't be
2254 necessary to configure just to create a source distribution.
2258 * Recreate the OS390-Unix config target. It no longer relies on a
2259 special script like it did for OpenSSL pre-1.1.0.
2263 * Instead of having the source directories listed in Configure, add
2264 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2269 * Add GMAC to EVP_MAC.
2273 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2277 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2278 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2279 to facilitate the continued use of MACs through raw private keys in
2280 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2284 * Deprecate ECDH_KDF_X9_62().
2288 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2289 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2290 are retained for backwards compatibility.
2294 * AES-XTS mode now enforces that its two keys are different to mitigate
2295 the attacked described in "Efficient Instantiations of Tweakable
2296 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2297 Details of this attack can be obtained from:
2298 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2302 * Rename the object files, i.e. give them other names than in previous
2303 versions. Their names now include the name of the final product, as
2304 well as its type mnemonic (bin, lib, shlib).
2308 * Added new option for 'openssl list', '-objects', which will display the
2309 list of built in objects, i.e. OIDs with names.
2313 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2314 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2319 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2320 improves application performance by removing data copies and providing
2321 applications with zero-copy system calls such as sendfile and splice.
2325 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2329 * `PKCS12_parse` now maintains the order of the parsed certificates
2330 when outputting them via `*ca` (rather than reversing it).
2334 * Deprecated pthread fork support methods.
2338 * Added support for FFDHE key exchange in TLS 1.3.
2342 * Added a new concept for OpenSSL plugability: providers. This
2343 functionality is designed to replace the ENGINE API and ENGINE
2344 implementations, and to be much more dynamic, allowing provider
2345 authors to introduce new algorithms among other things, as long as
2346 there's an API that supports the algorithm type.
2348 With this concept comes a new core API for interaction between
2349 libcrypto and provider implementations. Public libcrypto functions
2350 that want to use providers do so through this core API.
2352 The main documentation for this core API is found in
2353 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2354 refer to other manuals describing the API specific for supported
2355 algorithm types (also called operations).
2362 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2364 * Avoid loading of a dynamic engine twice.
2368 * Prioritise DANE TLSA issuer certs over peer certs
2372 * Fixed random API for MacOS prior to 10.12
2374 These MacOS versions don't support the CommonCrypto APIs
2378 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2380 * Fixed an SM2 Decryption Buffer Overflow.
2382 In order to decrypt SM2 encrypted data an application is expected to
2383 call the API function EVP_PKEY_decrypt(). Typically an application will
2384 call this function twice. The first time, on entry, the "out" parameter
2385 can be NULL and, on exit, the "outlen" parameter is populated with the
2386 buffer size required to hold the decrypted plaintext. The application
2387 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2388 again, but this time passing a non-NULL value for the "out" parameter.
2390 A bug in the implementation of the SM2 decryption code means that the
2391 calculation of the buffer size required to hold the plaintext returned
2392 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2393 size required by the second call. This can lead to a buffer overflow
2394 when EVP_PKEY_decrypt() is called by the application a second time with
2395 a buffer that is too small.
2397 A malicious attacker who is able present SM2 content for decryption to
2398 an application could cause attacker chosen data to overflow the buffer
2399 by up to a maximum of 62 bytes altering the contents of other data held
2400 after the buffer, possibly changing application behaviour or causing
2401 the application to crash. The location of the buffer is application
2402 dependent but is typically heap allocated.
2407 * Fixed various read buffer overruns processing ASN.1 strings
2409 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2410 structure which contains a buffer holding the string data and a field
2411 holding the buffer length. This contrasts with normal C strings which
2412 are represented as a buffer for the string data which is terminated
2413 with a NUL (0) byte.
2415 Although not a strict requirement, ASN.1 strings that are parsed using
2416 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2417 well as any string whose value has been set with the ASN1_STRING_set()
2418 function will additionally NUL terminate the byte array in the
2419 ASN1_STRING structure.
2421 However, it is possible for applications to directly construct valid
2422 ASN1_STRING structures which do not NUL terminate the byte array by
2423 directly setting the "data" and "length" fields in the ASN1_STRING
2424 array. This can also happen by using the ASN1_STRING_set0() function.
2426 Numerous OpenSSL functions that print ASN.1 data have been found to
2427 assume that the ASN1_STRING byte array will be NUL terminated, even
2428 though this is not guaranteed for strings that have been directly
2429 constructed. Where an application requests an ASN.1 structure to be
2430 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2431 been directly constructed by the application without NUL terminating
2432 the "data" field, then a read buffer overrun can occur.
2434 The same thing can also occur during name constraints processing
2435 of certificates (for example if a certificate has been directly
2436 constructed by the application instead of loading it via the OpenSSL
2437 parsing functions, and the certificate contains non NUL terminated
2438 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2439 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2441 If a malicious actor can cause an application to directly construct an
2442 ASN1_STRING and then process it through one of the affected OpenSSL
2443 functions then this issue could be hit. This might result in a crash
2444 (causing a Denial of Service attack). It could also result in the
2445 disclosure of private memory contents (such as private keys, or
2446 sensitive plaintext).
2451 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2453 * Fixed a problem with verifying a certificate chain when using the
2454 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2455 the certificates present in a certificate chain. It is not set by default.
2457 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2458 the chain that have explicitly encoded elliptic curve parameters was added
2459 as an additional strict check.
2461 An error in the implementation of this check meant that the result of a
2462 previous check to confirm that certificates in the chain are valid CA
2463 certificates was overwritten. This effectively bypasses the check
2464 that non-CA certificates must not be able to issue other certificates.
2466 If a "purpose" has been configured then there is a subsequent opportunity
2467 for checks that the certificate is a valid CA. All of the named "purpose"
2468 values implemented in libcrypto perform this check. Therefore, where
2469 a purpose is set the certificate chain will still be rejected even when the
2470 strict flag has been used. A purpose is set by default in libssl client and
2471 server certificate verification routines, but it can be overridden or
2472 removed by an application.
2474 In order to be affected, an application must explicitly set the
2475 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2476 for the certificate verification or, in the case of TLS client or server
2477 applications, override the default purpose.
2482 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2483 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2484 renegotiation ClientHello omits the signature_algorithms extension (where it
2485 was present in the initial ClientHello), but includes a
2486 signature_algorithms_cert extension then a NULL pointer dereference will
2487 result, leading to a crash and a denial of service attack.
2489 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2490 (which is the default configuration). OpenSSL TLS clients are not impacted by
2494 *Peter Kästle and Samuel Sapalski*
2496 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2498 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2499 create a unique hash value based on the issuer and serial number data
2500 contained within an X509 certificate. However, it was failing to correctly
2501 handle any errors that may occur while parsing the issuer field (which might
2502 occur if the issuer field is maliciously constructed). This may subsequently
2503 result in a NULL pointer deref and a crash leading to a potential denial of
2509 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2510 padding mode to correctly check for rollback attacks. This is considered a
2511 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2516 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2517 functions. Previously they could overflow the output length argument in some
2518 cases where the input length is close to the maximum permissible length for
2519 an integer on the platform. In such cases the return value from the function
2520 call would be 1 (indicating success), but the output length value would be
2521 negative. This could cause applications to behave incorrectly or crash.
2526 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2527 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2528 could be exploited in a side channel attack to recover the password. Since
2529 the attack is local host only this is outside of the current OpenSSL
2530 threat model and therefore no CVE is assigned.
2532 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2537 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2539 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2540 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2541 If an attacker can control both items being compared then this could lead
2542 to a possible denial of service attack. OpenSSL itself uses the
2543 GENERAL_NAME_cmp function for two purposes:
2544 1) Comparing CRL distribution point names between an available CRL and a
2545 CRL distribution point embedded in an X509 certificate
2546 2) When verifying that a timestamp response token signer matches the
2547 timestamp authority name (exposed via the API functions
2548 TS_RESP_verify_response and TS_RESP_verify_token)
2553 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2555 * Certificates with explicit curve parameters are now disallowed in
2556 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2560 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2561 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2562 conversely, silently ignore DTLS protocol version bounds when configuring
2563 TLS-based contexts. The commands can be repeated to set bounds of both
2564 types. The same applies with the corresponding "min_protocol" and
2565 "max_protocol" command-line switches, in case some application uses both TLS
2568 SSL_CTX instances that are created for a fixed protocol version (e.g.
2569 TLSv1_server_method()) also silently ignore version bounds. Previously
2570 attempts to apply bounds to these protocol versions would result in an
2571 error. Now only the "version-flexible" SSL_CTX instances are subject to
2572 limits in configuration files in command-line options.
2576 * Handshake now fails if Extended Master Secret extension is dropped
2581 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2583 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2585 * Fixed segmentation fault in SSL_check_chain()
2586 Server or client applications that call the SSL_check_chain() function
2587 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2588 dereference as a result of incorrect handling of the
2589 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2590 or unrecognised signature algorithm is received from the peer. This could
2591 be exploited by a malicious peer in a Denial of Service attack.
2596 * Added AES consttime code for no-asm configurations
2597 an optional constant time support for AES was added
2598 when building openssl for no-asm.
2599 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2600 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2601 At this time this feature is by default disabled.
2602 It will be enabled by default in 3.0.
2606 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2608 * Revert the change of EOF detection while reading in libssl to avoid
2609 regressions in applications depending on the current way of reporting
2610 the EOF. As the existing method is not fully accurate the change to
2611 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2612 branch and will be present in the 3.0 release.
2616 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2617 when primes for RSA keys are computed.
2618 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2619 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2620 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2621 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2622 This avoids possible fingerprinting of newly generated RSA modules.
2626 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2628 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2629 while reading in libssl then we would report an error back to the
2630 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2631 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2632 therefore give a hint as to what went wrong.
2636 * Check that ed25519 and ed448 are allowed by the security level. Previously
2637 signature algorithms not using an MD were not being checked that they were
2638 allowed by the security level.
2642 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2643 was not quite right. The behaviour was not consistent between resumption
2644 and normal handshakes, and also not quite consistent with historical
2645 behaviour. The behaviour in various scenarios has been clarified and
2646 it has been updated to make it match historical behaviour as closely as
2651 * *[VMS only]* The header files that the VMS compilers include automatically,
2652 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2653 that the C++ compiler doesn't understand. This is a shortcoming in the
2654 compiler, but can be worked around with `__cplusplus` guards.
2656 C++ applications that use OpenSSL libraries must be compiled using the
2657 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2658 functions. Otherwise, only functions with symbols of less than 31
2659 characters can be used, as the linker will not be able to successfully
2660 resolve symbols with longer names.
2664 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2665 The presence of this system service is determined at run-time.
2669 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2674 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2676 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2677 number generator (RNG). This was intended to include protection in the
2678 event of a fork() system call in order to ensure that the parent and child
2679 processes did not share the same RNG state. However, this protection was not
2680 being used in the default case.
2682 A partial mitigation for this issue is that the output from a high
2683 precision timer is mixed into the RNG state so the likelihood of a parent
2684 and child process sharing state is significantly reduced.
2686 If an application already calls OPENSSL_init_crypto() explicitly using
2687 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2690 *Matthias St. Pierre*
2692 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2693 used even when parsing explicit parameters, when loading a encoded key
2694 or calling `EC_GROUP_new_from_ecpkparameters()`/
2695 `EC_GROUP_new_from_ecparameters()`.
2696 This prevents bypass of security hardening and performance gains,
2697 especially for curves with specialized EC_METHODs.
2698 By default, if a key encoded with explicit parameters is loaded and later
2699 encoded, the output is still encoded with explicit parameters, even if
2700 internally a "named" EC_GROUP is used for computation.
2704 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2705 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2706 NULL. After this change, only the cofactor parameter can be NULL. It also
2707 does some minimal sanity checks on the passed order.
2712 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2713 An attack is simple, if the first CMS_recipientInfo is valid but the
2714 second CMS_recipientInfo is chosen ciphertext. If the second
2715 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2716 encryption key will be replaced by garbage, and the message cannot be
2717 decoded, but if the RSA decryption fails, the correct encryption key is
2718 used and the recipient will not notice the attack.
2719 As a work around for this potential attack the length of the decrypted
2720 key must be equal to the cipher default key length, in case the
2721 certificate is not given and all recipientInfo are tried out.
2722 The old behaviour can be re-enabled in the CMS code by setting the
2723 CMS_DEBUG_DECRYPT flag.
2728 * Early start up entropy quality from the DEVRANDOM seed source has been
2729 improved for older Linux systems. The RAND subsystem will wait for
2730 /dev/random to be producing output before seeding from /dev/urandom.
2731 The seeded state is stored for future library initialisations using
2732 a system global shared memory segment. The shared memory identifier
2733 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2734 the desired value. The default identifier is 114.
2738 * Correct the extended master secret constant on EBCDIC systems. Without this
2739 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2740 negotiate EMS will fail. Unfortunately this also means that TLS connections
2741 between EBCDIC systems with this fix, and EBCDIC systems without this
2742 fix will fail if they negotiate EMS.
2746 * Use Windows installation paths in the mingw builds
2748 Mingw isn't a POSIX environment per se, which means that Windows
2749 paths should be used for installation.
2754 * Changed DH_check to accept parameters with order q and 2q subgroups.
2755 With order 2q subgroups the bit 0 of the private key is not secret
2756 but DH_generate_key works around that by clearing bit 0 of the
2757 private key for those. This avoids leaking bit 0 of the private key.
2761 * Significantly reduce secure memory usage by the randomness pools.
2765 * Revert the DEVRANDOM_WAIT feature for Linux systems
2767 The DEVRANDOM_WAIT feature added a select() call to wait for the
2768 /dev/random device to become readable before reading from the
2769 /dev/urandom device.
2771 It turned out that this change had negative side effects on
2772 performance which were not acceptable. After some discussion it
2773 was decided to revert this feature and leave it up to the OS
2774 resp. the platform maintainer to ensure a proper initialization
2775 during early boot time.
2777 *Matthias St. Pierre*
2779 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2781 * Add build tests for C++. These are generated files that only do one
2782 thing, to include one public OpenSSL head file each. This tests that
2783 the public header files can be usefully included in a C++ application.
2785 This test isn't enabled by default. It can be enabled with the option
2786 'enable-buildtest-c++'.
2790 * Enable SHA3 pre-hashing for ECDSA and DSA.
2794 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2795 This changes the size when using the `genpkey` command when no size is given.
2796 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2797 generation commands to use 2048 bits by default.
2801 * Reorganize the manual pages to consistently have RETURN VALUES,
2802 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2803 util/fix-doc-nits accordingly.
2805 *Paul Yang, Joshua Lock*
2807 * Add the missing accessor EVP_PKEY_get0_engine()
2811 * Have commands like `s_client` and `s_server` output the signature scheme
2812 along with other cipher suite parameters when debugging.
2816 * Make OPENSSL_config() error agnostic again.
2820 * Do the error handling in RSA decryption constant time.
2824 * Prevent over long nonces in ChaCha20-Poly1305.
2826 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2827 for every encryption operation. RFC 7539 specifies that the nonce value
2828 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2829 and front pads the nonce with 0 bytes if it is less than 12
2830 bytes. However it also incorrectly allows a nonce to be set of up to 16
2831 bytes. In this case only the last 12 bytes are significant and any
2832 additional leading bytes are ignored.
2834 It is a requirement of using this cipher that nonce values are
2835 unique. Messages encrypted using a reused nonce value are susceptible to
2836 serious confidentiality and integrity attacks. If an application changes
2837 the default nonce length to be longer than 12 bytes and then makes a
2838 change to the leading bytes of the nonce expecting the new value to be a
2839 new unique nonce then such an application could inadvertently encrypt
2840 messages with a reused nonce.
2842 Additionally the ignored bytes in a long nonce are not covered by the
2843 integrity guarantee of this cipher. Any application that relies on the
2844 integrity of these ignored leading bytes of a long nonce may be further
2845 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2846 is safe because no such use sets such a long nonce value. However user
2847 applications that use this cipher directly and set a non-default nonce
2848 length to be longer than 12 bytes may be vulnerable.
2850 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2856 * Add DEVRANDOM_WAIT feature for Linux systems
2858 On older Linux systems where the getrandom() system call is not available,
2859 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2860 Contrary to getrandom(), the /dev/urandom device will not block during
2861 early boot when the kernel CSPRNG has not been seeded yet.
2863 To mitigate this known weakness, use select() to wait for /dev/random to
2864 become readable before reading from /dev/urandom.
2866 * Ensure that SM2 only uses SM3 as digest algorithm
2870 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2872 * Change the info callback signals for the start and end of a post-handshake
2873 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2874 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2875 confused by this and assume that a TLSv1.2 renegotiation has started. This
2876 can break KeyUpdate handling. Instead we no longer signal the start and end
2877 of a post handshake message exchange (although the messages themselves are
2878 still signalled). This could break some applications that were expecting
2879 the old signals. However without this KeyUpdate is not usable for many
2884 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2886 * Timing vulnerability in DSA signature generation
2888 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2889 timing side channel attack. An attacker could use variations in the signing
2890 algorithm to recover the private key.
2892 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2897 * Timing vulnerability in ECDSA signature generation
2899 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2900 timing side channel attack. An attacker could use variations in the signing
2901 algorithm to recover the private key.
2903 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2908 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2909 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2910 of two gigabytes and the error handling improved.
2912 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2913 categorized as a normal bug, not a security issue, because the DRBG reseeds
2914 automatically and is fully functional even without additional randomness
2915 provided by the application.
2917 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2919 * Add a new ClientHello callback. Provides a callback interface that gives
2920 the application the ability to adjust the nascent SSL object at the
2921 earliest stage of ClientHello processing, immediately after extensions have
2922 been collected but before they have been processed. In particular, this
2923 callback can adjust the supported TLS versions in response to the contents
2928 * Add SM2 base algorithm support.
2932 * s390x assembly pack: add (improved) hardware-support for the following
2933 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2934 aes-cfb/cfb8, aes-ecb.
2938 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2939 parameter is no longer accepted, as it leads to a corrupt table. NULL
2940 pem_str is reserved for alias entries only.
2944 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2945 step for prime curves. The new implementation is based on formulae from
2946 differential addition-and-doubling in homogeneous projective coordinates
2947 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2948 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2949 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2950 to work in projective coordinates.
2952 *Billy Bob Brumley, Nicola Tuveri*
2954 * Change generating and checking of primes so that the error rate of not
2955 being prime depends on the intended use based on the size of the input.
2956 For larger primes this will result in more rounds of Miller-Rabin.
2957 The maximal error rate for primes with more than 1080 bits is lowered
2960 *Kurt Roeckx, Annie Yousar*
2962 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2966 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2967 moving between systems, and to avoid confusion when a Windows build is
2968 done with mingw vs with MSVC. For POSIX installs, there's still a
2969 symlink or copy named 'tsget' to avoid that confusion as well.
2973 * Revert blinding in ECDSA sign and instead make problematic addition
2974 length-invariant. Switch even to fixed-length Montgomery multiplication.
2978 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2979 step for binary curves. The new implementation is based on formulae from
2980 differential addition-and-doubling in mixed Lopez-Dahab projective
2981 coordinates, modified to independently blind the operands.
2983 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2985 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2986 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2987 EC_METHODs to implement their own specialized "ladder step", to take
2988 advantage of more favorable coordinate systems or more efficient
2989 differential addition-and-doubling algorithms.
2991 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2993 * Modified the random device based seed sources to keep the relevant
2994 file descriptors open rather than reopening them on each access.
2995 This allows such sources to operate in a chroot() jail without
2996 the associated device nodes being available. This behaviour can be
2997 controlled using RAND_keep_random_devices_open().
3001 * Numerous side-channel attack mitigations have been applied. This may have
3002 performance impacts for some algorithms for the benefit of improved
3003 security. Specific changes are noted in this change log by their respective
3008 * AIX shared library support overhaul. Switch to AIX "natural" way of
3009 handling shared libraries, which means collecting shared objects of
3010 different versions and bitnesses in one common archive. This allows to
3011 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3012 doesn't affect the way 3rd party applications are linked, only how
3013 multi-version installation is managed.
3017 * Make ec_group_do_inverse_ord() more robust and available to other
3018 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3019 mitigations are applied to the fallback BN_mod_inverse().
3020 When using this function rather than BN_mod_inverse() directly, new
3021 EC cryptosystem implementations are then safer-by-default.
3025 * Add coordinate blinding for EC_POINT and implement projective
3026 coordinate blinding for generic prime curves as a countermeasure to
3027 chosen point SCA attacks.
3029 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3031 * Add blinding to ECDSA and DSA signatures to protect against side channel
3032 attacks discovered by Keegan Ryan (NCC Group).
3036 * Enforce checking in the `pkeyutl` command to ensure that the input
3037 length does not exceed the maximum supported digest length when performing
3038 a sign, verify or verifyrecover operation.
3042 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3043 I/O in combination with something like select() or poll() will hang. This
3044 can be turned off again using SSL_CTX_clear_mode().
3045 Many applications do not properly handle non-application data records, and
3046 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3047 around the problems in those applications, but can also break some.
3048 It's recommended to read the manpages about SSL_read(), SSL_write(),
3049 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3050 SSL_CTX_set_read_ahead() again.
3054 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3055 now allow empty (zero character) pass phrases.
3059 * Apply blinding to binary field modular inversion and remove patent
3060 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3064 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3065 binary and prime elliptic curves.
3069 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3070 constant time fixed point multiplication.
3074 * Revise elliptic curve scalar multiplication with timing attack
3075 defenses: ec_wNAF_mul redirects to a constant time implementation
3076 when computing fixed point and variable point multiplication (which
3077 in OpenSSL are mostly used with secret scalars in keygen, sign,
3078 ECDH derive operations).
3079 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3082 * Updated CONTRIBUTING
3086 * Updated DRBG / RAND to request nonce and additional low entropy
3087 randomness from the system.
3089 *Matthias St. Pierre*
3091 * Updated 'openssl rehash' to use OpenSSL consistent default.
3095 * Moved the load of the ssl_conf module to libcrypto, which helps
3096 loading engines that libssl uses before libssl is initialised.
3100 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3104 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3106 *Ingo Schwarze, Rich Salz*
3108 * Added output of accepting IP address and port for 'openssl s_server'
3112 * Added a new API for TLSv1.3 ciphersuites:
3113 SSL_CTX_set_ciphersuites()
3114 SSL_set_ciphersuites()
3118 * Memory allocation failures consistently add an error to the error
3123 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3124 in libcrypto when run as setuid/setgid.
3128 * Load any config file by default when libssl is used.
3132 * Added new public header file <openssl/rand_drbg.h> and documentation
3133 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3135 *Matthias St. Pierre*
3137 * QNX support removed (cannot find contributors to get their approval
3138 for the license change).
3142 * TLSv1.3 replay protection for early data has been implemented. See the
3143 SSL_read_early_data() man page for further details.
3147 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3148 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3149 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3150 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3151 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3152 configuration has been separated out. See the ciphers man page or the
3153 SSL_CTX_set_ciphersuites() man page for more information.
3157 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3158 in responder mode now supports the new "-multi" option, which
3159 spawns the specified number of child processes to handle OCSP
3160 requests. The "-timeout" option now also limits the OCSP
3161 responder's patience to wait to receive the full client request
3162 on a newly accepted connection. Child processes are respawned
3163 as needed, and the CA index file is automatically reloaded
3164 when changed. This makes it possible to run the "ocsp" responder
3165 as a long-running service, making the OpenSSL CA somewhat more
3166 feature-complete. In this mode, most diagnostic messages logged
3167 after entering the event loop are logged via syslog(3) rather than
3172 * Added support for X448 and Ed448. Heavily based on original work by
3177 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3178 objects loaded. This adds the functions OSSL_STORE_expect() and
3179 OSSL_STORE_find() as well as needed tools to construct searches and
3180 get the search data out of them.
3184 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3185 version of OpenSSL should review their configuration settings to ensure
3186 that they are still appropriate for TLSv1.3. For further information see:
3187 <https://wiki.openssl.org/index.php/TLS1.3>
3191 * Grand redesign of the OpenSSL random generator
3193 The default RAND method now utilizes an AES-CTR DRBG according to
3194 NIST standard SP 800-90Ar1. The new random generator is essentially
3195 a port of the default random generator from the OpenSSL FIPS 2.0
3196 object module. It is a hybrid deterministic random bit generator
3197 using an AES-CTR bit stream and which seeds and reseeds itself
3198 automatically using trusted system entropy sources.
3200 Some of its new features are:
3201 - Support for multiple DRBG instances with seed chaining.
3202 - The default RAND method makes use of a DRBG.
3203 - There is a public and private DRBG instance.
3204 - The DRBG instances are fork-safe.
3205 - Keep all global DRBG instances on the secure heap if it is enabled.
3206 - The public and private DRBG instance are per thread for lock free
3209 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3211 * Changed Configure so it only says what it does and doesn't dump
3212 so much data. Instead, ./configdata.pm should be used as a script
3213 to display all sorts of configuration data.
3217 * Added processing of "make variables" to Configure.
3221 * Added SHA512/224 and SHA512/256 algorithm support.
3225 * The last traces of Netware support, first removed in 1.1.0, have
3230 * Get rid of Makefile.shared, and in the process, make the processing
3231 of certain files (rc.obj, or the .def/.map/.opt files produced from
3232 the ordinal files) more visible and hopefully easier to trace and
3233 debug (or make silent).
3237 * Make it possible to have environment variable assignments as
3238 arguments to config / Configure.
3242 * Add multi-prime RSA (RFC 8017) support.
3246 * Add SM3 implemented according to GB/T 32905-2016
3247 *Jack Lloyd <jack.lloyd@ribose.com>,*
3248 *Ronald Tse <ronald.tse@ribose.com>,*
3249 *Erick Borsboom <erick.borsboom@ribose.com>*
3251 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3252 as documented in RFC6066.
3253 Based on a patch from Tomasz Moń
3255 *Filipe Raimundo da Silva*
3257 * Add SM4 implemented according to GB/T 32907-2016.
3258 *Jack Lloyd <jack.lloyd@ribose.com>,*
3259 *Ronald Tse <ronald.tse@ribose.com>,*
3260 *Erick Borsboom <erick.borsboom@ribose.com>*
3262 * Reimplement -newreq-nodes and ERR_error_string_n; the
3263 original author does not agree with the license change.
3267 * Add ARIA AEAD TLS support.
3271 * Some macro definitions to support VS6 have been removed. Visual
3272 Studio 6 has not worked since 1.1.0
3276 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3277 without clearing the errors.
3281 * Add "atfork" functions. If building on a system that without
3282 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3283 requirements. The RAND facility now uses/requires this.
3291 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3292 not possible to disable entirely. However, it's still possible to
3293 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3296 To disable, configure with 'no-ui-console'. 'no-ui' is still
3297 possible to use as an alias. Check at compile time with the
3298 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3299 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3303 * Add a STORE module, which implements a uniform and URI based reader of
3304 stores that can contain keys, certificates, CRLs and numerous other
3305 objects. The main API is loosely based on a few stdio functions,
3306 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3307 OSSL_STORE_error and OSSL_STORE_close.
3308 The implementation uses backends called "loaders" to implement arbitrary
3309 URI schemes. There is one built in "loader" for the 'file' scheme.
3313 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3314 then adjusted to work on FreeBSD 8.4 as well.
3315 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3316 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3320 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3321 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3322 error code calls like this:
3324 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3326 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3327 that can be encoded in C. For the foreseeable future, this will only
3330 *Richard Levitte and Tim Hudson*
3332 * Removed BSD cryptodev engine.
3336 * Add a build target 'build_all_generated', to build all generated files
3337 and only that. This can be used to prepare everything that requires
3338 things like perl for a system that lacks perl and then move everything
3339 to that system and do the rest of the build there.
3343 * In the UI interface, make it possible to duplicate the user data. This
3344 can be used by engines that need to retain the data for a longer time
3345 than just the call where this user data is passed.
3349 * Ignore the '-named_curve auto' value for compatibility of applications
3352 *Tomáš Mráz <tmraz@fedoraproject.org>*
3354 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3355 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3356 alerts across multiple records (some of which could be empty). In practice
3357 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3358 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3359 support this at all. Supporting it adds significant complexity to the
3360 record layer, and its removal is unlikely to cause interoperability
3365 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3366 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3367 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3372 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3373 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3375 *Richard Levitte, Andy Polyakov*
3377 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3382 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3383 platform rather than 'mingw'.
3387 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3388 success if they are asked to add an object which already exists
3389 in the store. This change cascades to other functions which load
3390 certificates and CRLs.
3394 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3395 facilitate stack unwinding even from assembly subroutines.
3399 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3400 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3404 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3405 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3406 which is the minimum version we support.
3410 * Certificate time validation (X509_cmp_time) enforces stricter
3411 compliance with RFC 5280. Fractional seconds and timezone offsets
3412 are no longer allowed.
3416 * Add support for ARIA
3420 * s_client will now send the Server Name Indication (SNI) extension by
3421 default unless the new "-noservername" option is used. The server name is
3422 based on the host provided to the "-connect" option unless overridden by
3423 using "-servername".
3427 * Add support for SipHash
3431 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3432 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3433 prevent issues where no progress is being made and the peer continually
3434 sends unrecognised record types, using up resources processing them.
3438 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3439 using the algorithm defined in
3440 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3444 * Heartbeat support has been removed; the ABI is changed for now.
3446 *Richard Levitte, Rich Salz*
3448 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3452 * The RSA "null" method, which was partially supported to avoid patent
3453 issues, has been replaced to always returns NULL.
3460 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3462 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3463 used even when parsing explicit parameters, when loading a encoded key
3464 or calling `EC_GROUP_new_from_ecpkparameters()`/
3465 `EC_GROUP_new_from_ecparameters()`.
3466 This prevents bypass of security hardening and performance gains,
3467 especially for curves with specialized EC_METHODs.
3468 By default, if a key encoded with explicit parameters is loaded and later
3469 encoded, the output is still encoded with explicit parameters, even if
3470 internally a "named" EC_GROUP is used for computation.
3474 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3475 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3476 NULL. After this change, only the cofactor parameter can be NULL. It also
3477 does some minimal sanity checks on the passed order.
3482 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3483 An attack is simple, if the first CMS_recipientInfo is valid but the
3484 second CMS_recipientInfo is chosen ciphertext. If the second
3485 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3486 encryption key will be replaced by garbage, and the message cannot be
3487 decoded, but if the RSA decryption fails, the correct encryption key is
3488 used and the recipient will not notice the attack.
3489 As a work around for this potential attack the length of the decrypted
3490 key must be equal to the cipher default key length, in case the
3491 certificate is not given and all recipientInfo are tried out.
3492 The old behaviour can be re-enabled in the CMS code by setting the
3493 CMS_DEBUG_DECRYPT flag.
3498 * Use Windows installation paths in the mingw builds
3500 Mingw isn't a POSIX environment per se, which means that Windows
3501 paths should be used for installation.
3506 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3508 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3509 This changes the size when using the `genpkey` command when no size is given.
3510 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3511 generation commands to use 2048 bits by default.
3515 * Prevent over long nonces in ChaCha20-Poly1305.
3517 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3518 for every encryption operation. RFC 7539 specifies that the nonce value
3519 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3520 and front pads the nonce with 0 bytes if it is less than 12
3521 bytes. However it also incorrectly allows a nonce to be set of up to 16
3522 bytes. In this case only the last 12 bytes are significant and any
3523 additional leading bytes are ignored.
3525 It is a requirement of using this cipher that nonce values are
3526 unique. Messages encrypted using a reused nonce value are susceptible to
3527 serious confidentiality and integrity attacks. If an application changes
3528 the default nonce length to be longer than 12 bytes and then makes a
3529 change to the leading bytes of the nonce expecting the new value to be a
3530 new unique nonce then such an application could inadvertently encrypt
3531 messages with a reused nonce.
3533 Additionally the ignored bytes in a long nonce are not covered by the
3534 integrity guarantee of this cipher. Any application that relies on the
3535 integrity of these ignored leading bytes of a long nonce may be further
3536 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3537 is safe because no such use sets such a long nonce value. However user
3538 applications that use this cipher directly and set a non-default nonce
3539 length to be longer than 12 bytes may be vulnerable.
3541 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3547 * Added SCA hardening for modular field inversion in EC_GROUP through
3548 a new dedicated field_inv() pointer in EC_METHOD.
3549 This also addresses a leakage affecting conversions from projective
3550 to affine coordinates.
3552 *Billy Bob Brumley, Nicola Tuveri*
3554 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3555 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3559 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3563 * Remove the 'dist' target and add a tarball building script. The
3564 'dist' target has fallen out of use, and it shouldn't be
3565 necessary to configure just to create a source distribution.
3569 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3571 * Timing vulnerability in DSA signature generation
3573 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3574 timing side channel attack. An attacker could use variations in the signing
3575 algorithm to recover the private key.
3577 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3582 * Timing vulnerability in ECDSA signature generation
3584 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3585 timing side channel attack. An attacker could use variations in the signing
3586 algorithm to recover the private key.
3588 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3593 * Add coordinate blinding for EC_POINT and implement projective
3594 coordinate blinding for generic prime curves as a countermeasure to
3595 chosen point SCA attacks.
3597 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3599 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3601 * Client DoS due to large DH parameter
3603 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3604 malicious server can send a very large prime value to the client. This will
3605 cause the client to spend an unreasonably long period of time generating a
3606 key for this prime resulting in a hang until the client has finished. This
3607 could be exploited in a Denial Of Service attack.
3609 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3614 * Cache timing vulnerability in RSA Key Generation
3616 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3617 a cache timing side channel attack. An attacker with sufficient access to
3618 mount cache timing attacks during the RSA key generation process could
3619 recover the private key.
3621 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3622 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3627 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3628 parameter is no longer accepted, as it leads to a corrupt table. NULL
3629 pem_str is reserved for alias entries only.
3633 * Revert blinding in ECDSA sign and instead make problematic addition
3634 length-invariant. Switch even to fixed-length Montgomery multiplication.
3638 * Change generating and checking of primes so that the error rate of not
3639 being prime depends on the intended use based on the size of the input.
3640 For larger primes this will result in more rounds of Miller-Rabin.
3641 The maximal error rate for primes with more than 1080 bits is lowered
3644 *Kurt Roeckx, Annie Yousar*
3646 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3650 * Add blinding to ECDSA and DSA signatures to protect against side channel
3651 attacks discovered by Keegan Ryan (NCC Group).
3655 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3656 now allow empty (zero character) pass phrases.
3660 * Certificate time validation (X509_cmp_time) enforces stricter
3661 compliance with RFC 5280. Fractional seconds and timezone offsets
3662 are no longer allowed.
3666 * Fixed a text canonicalisation bug in CMS
3668 Where a CMS detached signature is used with text content the text goes
3669 through a canonicalisation process first prior to signing or verifying a
3670 signature. This process strips trailing space at the end of lines, converts
3671 line terminators to CRLF and removes additional trailing line terminators
3672 at the end of a file. A bug in the canonicalisation process meant that
3673 some characters, such as form-feed, were incorrectly treated as whitespace
3674 and removed. This is contrary to the specification (RFC5485). This fix
3675 could mean that detached text data signed with an earlier version of
3676 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3677 signed with a fixed OpenSSL may fail to verify with an earlier version of
3678 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3679 and use the "-binary" flag (for the "cms" command line application) or set
3680 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3684 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3686 * Constructed ASN.1 types with a recursive definition could exceed the stack
3688 Constructed ASN.1 types with a recursive definition (such as can be found
3689 in PKCS7) could eventually exceed the stack given malicious input with
3690 excessive recursion. This could result in a Denial Of Service attack. There
3691 are no such structures used within SSL/TLS that come from untrusted sources
3692 so this is considered safe.
3694 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3700 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3702 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3703 effectively reduced to only comparing the least significant bit of each
3704 byte. This allows an attacker to forge messages that would be considered as
3705 authenticated in an amount of tries lower than that guaranteed by the
3706 security claims of the scheme. The module can only be compiled by the
3707 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3709 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3715 * Add a build target 'build_all_generated', to build all generated files
3716 and only that. This can be used to prepare everything that requires
3717 things like perl for a system that lacks perl and then move everything
3718 to that system and do the rest of the build there.
3722 * Backport SSL_OP_NO_RENGOTIATION
3724 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3725 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3726 changes this is no longer possible in 1.1.0. Therefore, the new
3727 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3728 1.1.0 to provide equivalent functionality.
3730 Note that if an application built against 1.1.0h headers (or above) is run
3731 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3732 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3736 * Removed the OS390-Unix config target. It relied on a script that doesn't
3741 * rsaz_1024_mul_avx2 overflow bug on x86_64
3743 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3744 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3745 Analysis suggests that attacks against RSA and DSA as a result of this
3746 defect would be very difficult to perform and are not believed likely.
3747 Attacks against DH1024 are considered just feasible, because most of the
3748 work necessary to deduce information about a private key may be performed
3749 offline. The amount of resources required for such an attack would be
3750 significant. However, for an attack on TLS to be meaningful, the server
3751 would have to share the DH1024 private key among multiple clients, which is
3752 no longer an option since CVE-2016-0701.
3754 This only affects processors that support the AVX2 but not ADX extensions
3755 like Intel Haswell (4th generation).
3757 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3758 was originally found via the OSS-Fuzz project.
3763 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3765 * bn_sqrx8x_internal carry bug on x86_64
3767 There is a carry propagating bug in the x86_64 Montgomery squaring
3768 procedure. No EC algorithms are affected. Analysis suggests that attacks
3769 against RSA and DSA as a result of this defect would be very difficult to
3770 perform and are not believed likely. Attacks against DH are considered just
3771 feasible (although very difficult) because most of the work necessary to
3772 deduce information about a private key may be performed offline. The amount
3773 of resources required for such an attack would be very significant and
3774 likely only accessible to a limited number of attackers. An attacker would
3775 additionally need online access to an unpatched system using the target
3776 private key in a scenario with persistent DH parameters and a private
3777 key that is shared between multiple clients.
3779 This only affects processors that support the BMI1, BMI2 and ADX extensions
3780 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3782 This issue was reported to OpenSSL by the OSS-Fuzz project.
3787 * Malformed X.509 IPAddressFamily could cause OOB read
3789 If an X.509 certificate has a malformed IPAddressFamily extension,
3790 OpenSSL could do a one-byte buffer overread. The most likely result
3791 would be an erroneous display of the certificate in text format.
3793 This issue was reported to OpenSSL by the OSS-Fuzz project.
3798 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3800 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3801 platform rather than 'mingw'.
3805 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3806 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3807 which is the minimum version we support.
3811 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3813 * Encrypt-Then-Mac renegotiation crash
3815 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3816 negotiated where it was not in the original handshake (or vice-versa) then
3817 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3818 and servers are affected.
3820 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3825 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3827 * Truncated packet could crash via OOB read
3829 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3830 cipher is being used, then a truncated packet can cause that host to
3831 perform an out-of-bounds read, usually resulting in a crash.
3833 This issue was reported to OpenSSL by Robert Święcki of Google.
3838 * Bad (EC)DHE parameters cause a client crash
3840 If a malicious server supplies bad parameters for a DHE or ECDHE key
3841 exchange then this can result in the client attempting to dereference a
3842 NULL pointer leading to a client crash. This could be exploited in a Denial
3845 This issue was reported to OpenSSL by Guido Vranken.
3850 * BN_mod_exp may produce incorrect results on x86_64
3852 There is a carry propagating bug in the x86_64 Montgomery squaring
3853 procedure. No EC algorithms are affected. Analysis suggests that attacks
3854 against RSA and DSA as a result of this defect would be very difficult to
3855 perform and are not believed likely. Attacks against DH are considered just
3856 feasible (although very difficult) because most of the work necessary to
3857 deduce information about a private key may be performed offline. The amount
3858 of resources required for such an attack would be very significant and
3859 likely only accessible to a limited number of attackers. An attacker would
3860 additionally need online access to an unpatched system using the target
3861 private key in a scenario with persistent DH parameters and a private
3862 key that is shared between multiple clients. For example this can occur by
3863 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3864 similar to CVE-2015-3193 but must be treated as a separate problem.
3866 This issue was reported to OpenSSL by the OSS-Fuzz project.
3871 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3873 * ChaCha20/Poly1305 heap-buffer-overflow
3875 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3876 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3877 crash. This issue is not considered to be exploitable beyond a DoS.
3879 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3884 * CMS Null dereference
3886 Applications parsing invalid CMS structures can crash with a NULL pointer
3887 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3888 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3889 structure callback if an attempt is made to free certain invalid encodings.
3890 Only CHOICE structures using a callback which do not handle NULL value are
3893 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3898 * Montgomery multiplication may produce incorrect results
3900 There is a carry propagating bug in the Broadwell-specific Montgomery
3901 multiplication procedure that handles input lengths divisible by, but
3902 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3903 and DH private keys are impossible. This is because the subroutine in
3904 question is not used in operations with the private key itself and an input
3905 of the attacker's direct choice. Otherwise the bug can manifest itself as
3906 transient authentication and key negotiation failures or reproducible
3907 erroneous outcome of public-key operations with specially crafted input.
3908 Among EC algorithms only Brainpool P-512 curves are affected and one
3909 presumably can attack ECDH key negotiation. Impact was not analyzed in
3910 detail, because pre-requisites for attack are considered unlikely. Namely
3911 multiple clients have to choose the curve in question and the server has to
3912 share the private key among them, neither of which is default behaviour.
3913 Even then only clients that chose the curve will be affected.
3915 This issue was publicly reported as transient failures and was not
3916 initially recognized as a security issue. Thanks to Richard Morgan for
3917 providing reproducible case.
3922 * Removed automatic addition of RPATH in shared libraries and executables,
3923 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3927 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3929 * Fix Use After Free for large message sizes
3931 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3932 message larger than approx 16k is received then the underlying buffer to
3933 store the incoming message is reallocated and moved. Unfortunately a
3934 dangling pointer to the old location is left which results in an attempt to
3935 write to the previously freed location. This is likely to result in a
3936 crash, however it could potentially lead to execution of arbitrary code.
3938 This issue only affects OpenSSL 1.1.0a.
3940 This issue was reported to OpenSSL by Robert Święcki.
3945 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3947 * OCSP Status Request extension unbounded memory growth
3949 A malicious client can send an excessively large OCSP Status Request
3950 extension. If that client continually requests renegotiation, sending a
3951 large OCSP Status Request extension each time, then there will be unbounded
3952 memory growth on the server. This will eventually lead to a Denial Of
3953 Service attack through memory exhaustion. Servers with a default
3954 configuration are vulnerable even if they do not support OCSP. Builds using
3955 the "no-ocsp" build time option are not affected.
3957 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3962 * SSL_peek() hang on empty record
3964 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3965 sends an empty record. This could be exploited by a malicious peer in a
3966 Denial Of Service attack.
3968 This issue was reported to OpenSSL by Alex Gaynor.
3973 * Excessive allocation of memory in tls_get_message_header() and
3974 dtls1_preprocess_fragment()
3976 A (D)TLS message includes 3 bytes for its length in the header for the
3977 message. This would allow for messages up to 16Mb in length. Messages of
3978 this length are excessive and OpenSSL includes a check to ensure that a
3979 peer is sending reasonably sized messages in order to avoid too much memory
3980 being consumed to service a connection. A flaw in the logic of version
3981 1.1.0 means that memory for the message is allocated too early, prior to
3982 the excessive message length check. Due to way memory is allocated in
3983 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3984 to service a connection. This could lead to a Denial of Service through
3985 memory exhaustion. However, the excessive message length check still takes
3986 place, and this would cause the connection to immediately fail. Assuming
3987 that the application calls SSL_free() on the failed connection in a timely
3988 manner then the 21Mb of allocated memory will then be immediately freed
3989 again. Therefore, the excessive memory allocation will be transitory in
3990 nature. This then means that there is only a security impact if:
3992 1) The application does not call SSL_free() in a timely manner in the event
3993 that the connection fails
3995 2) The application is working in a constrained environment where there is
3996 very little free memory
3998 3) The attacker initiates multiple connection attempts such that there are
3999 multiple connections in a state where memory has been allocated for the
4000 connection; SSL_free() has not yet been called; and there is insufficient
4001 memory to service the multiple requests.
4003 Except in the instance of (1) above any Denial Of Service is likely to be
4004 transitory because as soon as the connection fails the memory is
4005 subsequently freed again in the SSL_free() call. However there is an
4006 increased risk during this period of application crashes due to the lack of
4007 memory - which would then mean a more serious Denial of Service.
4009 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4010 (CVE-2016-6307 and CVE-2016-6308)
4014 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4015 had to be removed. Primary reason is that vendor assembler can't
4016 assemble our modules with -KPIC flag. As result it, assembly
4017 support, was not even available as option. But its lack means
4018 lack of side-channel resistant code, which is incompatible with
4019 security by todays standards. Fortunately gcc is readily available
4020 prepackaged option, which we firmly point at...
4024 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4026 * Windows command-line tool supports UTF-8 opt-in option for arguments
4027 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4028 (to any value) allows Windows user to access PKCS#12 file generated
4029 with Windows CryptoAPI and protected with non-ASCII password, as well
4030 as files generated under UTF-8 locale on Linux also protected with
4035 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4036 have been disabled by default and removed from DEFAULT, just like RC4.
4037 See the RC4 item below to re-enable both.
4041 * The method for finding the storage location for the Windows RAND seed file
4042 has changed. First we check %RANDFILE%. If that is not set then we check
4043 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4044 all else fails we fall back to C:\.
4048 * The EVP_EncryptUpdate() function has had its return type changed from void
4049 to int. A return of 0 indicates and error while a return of 1 indicates
4054 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4055 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4056 off the constant time implementation for RSA, DSA and DH have been made
4057 no-ops and deprecated.
4061 * Windows RAND implementation was simplified to only get entropy by
4062 calling CryptGenRandom(). Various other RAND-related tickets
4065 *Joseph Wylie Yandle, Rich Salz*
4067 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4068 and `OPENSSL_LH_`, respectively. The old names are available
4069 with API compatibility. They new names are now completely documented.
4073 * Unify TYPE_up_ref(obj) methods signature.
4074 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4075 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4076 int (instead of void) like all others TYPE_up_ref() methods.
4077 So now these methods also check the return value of CRYPTO_atomic_add(),
4078 and the validity of object reference counter.
4080 *fdasilvayy@gmail.com*
4082 * With Windows Visual Studio builds, the .pdb files are installed
4083 alongside the installed libraries and executables. For a static
4084 library installation, ossl_static.pdb is the associate compiler
4085 generated .pdb file to be used when linking programs.
4089 * Remove openssl.spec. Packaging files belong with the packagers.
4093 * Automatic Darwin/OSX configuration has had a refresh, it will now
4094 recognise x86_64 architectures automatically. You can still decide
4095 to build for a different bitness with the environment variable
4096 KERNEL_BITS (can be 32 or 64), for example:
4098 KERNEL_BITS=32 ./config
4102 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4103 256 bit AES and HMAC with SHA256.
4107 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4111 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4115 * To enable users to have their own config files and build file templates,
4116 Configure looks in the directory indicated by the environment variable
4117 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4118 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4119 name and is used as is.
4123 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4124 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4125 X509_CERT_FILE_CTX was removed.
4129 * "shared" builds are now the default. To create only static libraries use
4130 the "no-shared" Configure option.
4134 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4135 All of these option have not worked for some while and are fundamental
4140 * Make various cleanup routines no-ops and mark them as deprecated. Most
4141 global cleanup functions are no longer required because they are handled
4142 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4143 Explicitly de-initing can cause problems (e.g. where a library that uses
4144 OpenSSL de-inits, but an application is still using it). The affected
4145 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4146 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4147 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4148 COMP_zlib_cleanup().
4152 * --strict-warnings no longer enables runtime debugging options
4153 such as REF_DEBUG. Instead, debug options are automatically
4154 enabled with '--debug' builds.
4156 *Andy Polyakov, Emilia Käsper*
4158 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4159 have been moved out of the public header files. New functions for managing
4160 these have been added.
4164 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4165 objects have been moved out of the public header files. New
4166 functions for managing these have been added.
4170 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4171 have been moved out of the public header files. New functions for managing
4172 these have been added.
4176 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4177 moved out of the public header files. New functions for managing these
4182 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4186 * Removed the mk1mf build scripts.
4190 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4191 it is always safe to #include a header now.
4195 * Removed the aged BC-32 config and all its supporting scripts
4199 * Removed support for Ultrix, Netware, and OS/2.
4203 * Add support for HKDF.
4205 *Alessandro Ghedini*
4207 * Add support for blake2b and blake2s
4211 * Added support for "pipelining". Ciphers that have the
4212 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4213 encryptions/decryptions simultaneously. There are currently no built-in
4214 ciphers with this property but the expectation is that engines will be able
4215 to offer it to significantly improve throughput. Support has been extended
4216 into libssl so that multiple records for a single connection can be
4217 processed in one go (for >=TLS 1.1).
4221 * Added the AFALG engine. This is an async capable engine which is able to
4222 offload work to the Linux kernel. In this initial version it only supports
4223 AES128-CBC. The kernel must be version 4.1.0 or greater.
4227 * OpenSSL now uses a new threading API. It is no longer necessary to
4228 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4229 are two supported threading models: pthreads and windows threads. It is
4230 also possible to configure OpenSSL at compile time for "no-threads". The
4231 old threading API should no longer be used. The functions have been
4232 replaced with "no-op" compatibility macros.
4234 *Alessandro Ghedini, Matt Caswell*
4236 * Modify behavior of ALPN to invoke callback after SNI/servername
4237 callback, such that updates to the SSL_CTX affect ALPN.
4241 * Add SSL_CIPHER queries for authentication and key-exchange.
4245 * Changes to the DEFAULT cipherlist:
4246 - Prefer (EC)DHE handshakes over plain RSA.
4247 - Prefer AEAD ciphers over legacy ciphers.
4248 - Prefer ECDSA over RSA when both certificates are available.
4249 - Prefer TLSv1.2 ciphers/PRF.
4250 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4255 * Change the ECC default curve list to be this, in order: x25519,
4256 secp256r1, secp521r1, secp384r1.
4260 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4261 disabled by default. They can be re-enabled using the
4262 enable-weak-ssl-ciphers option to Configure.
4266 * If the server has ALPN configured, but supports no protocols that the
4267 client advertises, send a fatal "no_application_protocol" alert.
4268 This behaviour is SHALL in RFC 7301, though it isn't universally
4269 implemented by other servers.
4273 * Add X25519 support.
4274 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4275 for public and private key encoding using the format documented in
4276 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4277 key generation and key derivation.
4279 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4284 * Deprecate SRP_VBASE_get_by_user.
4285 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4286 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4287 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4288 seed, even if the seed is configured.
4290 Users should use SRP_VBASE_get1_by_user instead. Note that in
4291 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4292 also that even though configuring the SRP seed attempts to hide
4293 invalid usernames by continuing the handshake with fake
4294 credentials, this behaviour is not constant time and no strong
4295 guarantees are made that the handshake is indistinguishable from
4296 that of a valid user.
4300 * Configuration change; it's now possible to build dynamic engines
4301 without having to build shared libraries and vice versa. This
4302 only applies to the engines in `engines/`, those in `crypto/engine/`
4303 will always be built into libcrypto (i.e. "static").
4305 Building dynamic engines is enabled by default; to disable, use
4306 the configuration option "disable-dynamic-engine".
4308 The only requirements for building dynamic engines are the
4309 presence of the DSO module and building with position independent
4310 code, so they will also automatically be disabled if configuring
4311 with "disable-dso" or "disable-pic".
4313 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4314 are also taken away from openssl/opensslconf.h, as they are
4319 * Configuration change; if there is a known flag to compile
4320 position independent code, it will always be applied on the
4321 libcrypto and libssl object files, and never on the application
4322 object files. This means other libraries that use routines from
4323 libcrypto / libssl can be made into shared libraries regardless
4324 of how OpenSSL was configured.
4326 If this isn't desirable, the configuration options "disable-pic"
4327 or "no-pic" can be used to disable the use of PIC. This will
4328 also disable building shared libraries and dynamic engines.
4332 * Removed JPAKE code. It was experimental and has no wide use.
4336 * The INSTALL_PREFIX Makefile variable has been renamed to
4337 DESTDIR. That makes for less confusion on what this variable
4338 is for. Also, the configuration option --install_prefix is
4343 * Heartbeat for TLS has been removed and is disabled by default
4344 for DTLS; configure with enable-heartbeats. Code that uses the
4345 old #define's might need to be updated.
4347 *Emilia Käsper, Rich Salz*
4349 * Rename REF_CHECK to REF_DEBUG.
4353 * New "unified" build system
4355 The "unified" build system is aimed to be a common system for all
4356 platforms we support. With it comes new support for VMS.
4358 This system builds supports building in a different directory tree
4359 than the source tree. It produces one Makefile (for unix family
4360 or lookalikes), or one descrip.mms (for VMS).
4362 The source of information to make the Makefile / descrip.mms is
4363 small files called 'build.info', holding the necessary
4364 information for each directory with source to compile, and a
4365 template in Configurations, like unix-Makefile.tmpl or
4368 With this change, the library names were also renamed on Windows
4369 and on VMS. They now have names that are closer to the standard
4370 on Unix, and include the major version number, and in certain
4371 cases, the architecture they are built for. See "Notes on shared
4372 libraries" in INSTALL.
4374 We rely heavily on the perl module Text::Template.
4378 * Added support for auto-initialisation and de-initialisation of the library.
4379 OpenSSL no longer requires explicit init or deinit routines to be called,
4380 except in certain circumstances. See the OPENSSL_init_crypto() and
4381 OPENSSL_init_ssl() man pages for further information.
4385 * The arguments to the DTLSv1_listen function have changed. Specifically the
4386 "peer" argument is now expected to be a BIO_ADDR object.
4388 * Rewrite of BIO networking library. The BIO library lacked consistent
4389 support of IPv6, and adding it required some more extensive
4390 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4391 which hold all types of addresses and chains of address information.
4392 It also introduces a new API, with functions like BIO_socket,
4393 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4394 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4395 have been adapted accordingly.
4399 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4404 * CRIME protection: disable compression by default, even if OpenSSL is
4405 compiled with zlib enabled. Applications can still enable compression
4406 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4407 using the SSL_CONF library to configure compression.
4411 * The signature of the session callback configured with
4412 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4413 was explicitly marked as `const unsigned char*` instead of
4418 * Always DPURIFY. Remove the use of uninitialized memory in the
4419 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4423 * Removed many obsolete configuration items, including
4424 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4425 MD2_CHAR, MD2_INT, MD2_LONG
4427 IDEA_SHORT, IDEA_LONG
4428 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4430 *Rich Salz, with advice from Andy Polyakov*
4432 * Many BN internals have been moved to an internal header file.
4434 *Rich Salz with help from Andy Polyakov*
4436 * Configuration and writing out the results from it has changed.
4437 Files such as Makefile include/openssl/opensslconf.h and are now
4438 produced through general templates, such as Makefile.in and
4439 crypto/opensslconf.h.in and some help from the perl module
4442 Also, the center of configuration information is no longer
4443 Makefile. Instead, Configure produces a perl module in
4444 configdata.pm which holds most of the config data (in the hash
4445 table %config), the target data that comes from the target
4446 configuration in one of the `Configurations/*.conf` files (in
4451 * To clarify their intended purposes, the Configure options
4452 --prefix and --openssldir change their semantics, and become more
4453 straightforward and less interdependent.
4455 --prefix shall be used exclusively to give the location INSTALLTOP
4456 where programs, scripts, libraries, include files and manuals are
4457 going to be installed. The default is now /usr/local.
4459 --openssldir shall be used exclusively to give the default
4460 location OPENSSLDIR where certificates, private keys, CRLs are
4461 managed. This is also where the default openssl.cnf gets
4463 If the directory given with this option is a relative path, the
4464 values of both the --prefix value and the --openssldir value will
4465 be combined to become OPENSSLDIR.
4466 The default for --openssldir is INSTALLTOP/ssl.
4468 Anyone who uses --openssldir to specify where OpenSSL is to be
4469 installed MUST change to use --prefix instead.
4473 * The GOST engine was out of date and therefore it has been removed. An up
4474 to date GOST engine is now being maintained in an external repository.
4475 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4476 support for GOST ciphersuites (these are only activated if a GOST engine
4481 * EGD is no longer supported by default; use enable-egd when
4484 *Ben Kaduk and Rich Salz*
4486 * The distribution now has Makefile.in files, which are used to
4487 create Makefile's when Configure is run. *Configure must be run
4488 before trying to build now.*
4492 * The return value for SSL_CIPHER_description() for error conditions
4497 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4499 Obtaining and performing DNSSEC validation of TLSA records is
4500 the application's responsibility. The application provides
4501 the TLSA records of its choice to OpenSSL, and these are then
4502 used to authenticate the peer.
4504 The TLSA records need not even come from DNS. They can, for
4505 example, be used to implement local end-entity certificate or
4506 trust-anchor "pinning", where the "pin" data takes the form
4507 of TLSA records, which can augment or replace verification
4508 based on the usual WebPKI public certification authorities.
4512 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4513 continues to support deprecated interfaces in default builds.
4514 However, applications are strongly advised to compile their
4515 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4516 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4517 or the 1.1.0 releases.
4519 In environments in which all applications have been ported to
4520 not use any deprecated interfaces OpenSSL's Configure script
4521 should be used with the --api=1.1.0 option to entirely remove
4522 support for the deprecated features from the library and
4523 unconditionally disable them in the installed headers.
4524 Essentially the same effect can be achieved with the "no-deprecated"
4525 argument to Configure, except that this will always restrict
4526 the build to just the latest API, rather than a fixed API
4529 As applications are ported to future revisions of the API,
4530 they should update their compile-time OPENSSL_API_COMPAT define
4531 accordingly, but in most cases should be able to continue to
4532 compile with later releases.
4534 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4535 0x10000000L and 0x00908000L, respectively. However those
4536 versions did not support the OPENSSL_API_COMPAT feature, and
4537 so applications are not typically tested for explicit support
4538 of just the undeprecated features of either release.
4542 * Add support for setting the minimum and maximum supported protocol.
4543 It can bet set via the SSL_set_min_proto_version() and
4544 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4545 MaxProtocol. It's recommended to use the new APIs to disable
4546 protocols instead of disabling individual protocols using
4547 SSL_set_options() or SSL_CONF's Protocol. This change also
4548 removes support for disabling TLS 1.2 in the OpenSSL TLS
4549 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4553 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4557 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4558 and integrates ECDSA and ECDH functionality into EC. Implementations can
4559 now redirect key generation and no longer need to convert to or from
4562 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4563 include the ec.h header file instead.
4567 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4568 ciphers who are no longer supported and drops support the ephemeral RSA key
4569 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4573 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4574 opaque. For HMAC_CTX, the following constructors and destructors
4577 HMAC_CTX *HMAC_CTX_new(void);
4578 void HMAC_CTX_free(HMAC_CTX *ctx);
4580 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4581 destroy such methods has been added. See EVP_MD_meth_new(3) and
4582 EVP_CIPHER_meth_new(3) for documentation.
4585 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4586 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4587 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4588 an already created structure.
4589 2) For consistency with the majority of our object creators and
4590 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4591 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4592 for deprecated builds.
4596 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4597 cryptographic operations to be performed asynchronously as long as an
4598 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4599 further details. Libssl has also had this capability integrated with the
4600 introduction of the new mode SSL_MODE_ASYNC and associated error
4601 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4602 pages. This work was developed in partnership with Intel Corp.
4606 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4607 always enabled now. If you want to disable the support you should
4608 exclude it using the list of supported ciphers. This also means that the
4609 "-no_ecdhe" option has been removed from s_server.
4613 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4614 SSL_{CTX_}set1_curves() which can set a list.
4618 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4619 curve you want to support using SSL_{CTX_}set1_curves().
4623 * State machine rewrite. The state machine code has been significantly
4624 refactored in order to remove much duplication of code and solve issues
4625 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4626 further details). This change does have some associated API changes.
4627 Notably the SSL_state() function has been removed and replaced by
4628 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4629 SSL_set_state() has been removed altogether. The previous handshake states
4630 defined in ssl.h and ssl3.h have also been removed.
4634 * All instances of the string "ssleay" in the public API were replaced
4635 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4636 Some error codes related to internal RSA_eay API's were renamed.
4640 * The demo files in crypto/threads were moved to demo/threads.
4644 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4647 *Matt Caswell, Rich Salz*
4649 * New ASN.1 embed macro.
4651 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4652 structure is not allocated: it is part of the parent. That is instead of
4660 This reduces memory fragmentation and make it impossible to accidentally
4661 set a mandatory field to NULL.
4663 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4664 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4665 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4670 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4674 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4675 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4676 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4677 DES and RC4 ciphersuites.
4681 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4682 This changes the decoding behaviour for some invalid messages,
4683 though the change is mostly in the more lenient direction, and
4684 legacy behaviour is preserved as much as possible.
4688 * Fix no-stdio build.
4689 *David Woodhouse <David.Woodhouse@intel.com> and also*
4690 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4692 * New testing framework
4693 The testing framework has been largely rewritten and is now using
4694 perl and the perl modules Test::Harness and an extended variant of
4695 Test::More called OpenSSL::Test to do its work. All test scripts in
4696 test/ have been rewritten into test recipes, and all direct calls to
4697 executables in test/Makefile have become individual recipes using the
4698 simplified testing OpenSSL::Test::Simple.
4700 For documentation on our testing modules, do:
4702 perldoc test/testlib/OpenSSL/Test/Simple.pm
4703 perldoc test/testlib/OpenSSL/Test.pm
4707 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4708 are used; the latter aborts on memory leaks (usually checked on exit).
4709 Some undocumented "set malloc, etc., hooks" functions were removed
4710 and others were changed. All are now documented.
4714 * In DSA_generate_parameters_ex, if the provided seed is too short,
4717 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4719 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4720 from RFC4279, RFC4785, RFC5487, RFC5489.
4722 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4723 original RSA_PSK patch.
4727 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4728 era flag was never set throughout the codebase (only read). Also removed
4729 SSL3_FLAGS_POP_BUFFER which was only used if
4730 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4734 * Changed the default name options in the "ca", "crl", "req" and "x509"
4735 to be "oneline" instead of "compat".
4739 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4740 not aware of clients that still exhibit this bug, and the workaround
4741 hasn't been working properly for a while.
4745 * The return type of BIO_number_read() and BIO_number_written() as well as
4746 the corresponding num_read and num_write members in the BIO structure has
4747 changed from unsigned long to uint64_t. On platforms where an unsigned
4748 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4753 * Given the pervasive nature of TLS extensions it is inadvisable to run
4754 OpenSSL without support for them. It also means that maintaining
4755 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4756 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4760 * Removed support for the two export grade static DH ciphersuites
4761 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4762 were newly added (along with a number of other static DH ciphersuites) to
4763 1.0.2. However the two export ones have *never* worked since they were
4764 introduced. It seems strange in any case to be adding new export
4765 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4769 * Version negotiation has been rewritten. In particular SSLv23_method(),
4770 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4771 and turned into macros which simply call the new preferred function names
4772 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4773 should use the new names instead. Also as part of this change the ssl23.h
4774 header file has been removed.
4778 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4779 code and the associated standard is no longer considered fit-for-purpose.
4783 * RT2547 was closed. When generating a private key, try to make the
4784 output file readable only by the owner. This behavior change might
4785 be noticeable when interacting with other software.
4787 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4792 * Added HTTP GET support to the ocsp command.
4796 * Changed default digest for the dgst and enc commands from MD5 to
4801 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4805 * Added support for TLS extended master secret from
4806 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4807 initial patch which was a great help during development.
4811 * All libssl internal structures have been removed from the public header
4812 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4813 now redundant). Users should not attempt to access internal structures
4814 directly. Instead they should use the provided API functions.
4818 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4819 Access to deprecated functions can be re-enabled by running config with
4820 "enable-deprecated". In addition applications wishing to use deprecated
4821 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4822 will, by default, disable some transitive includes that previously existed
4823 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4827 * Added support for OCB mode. OpenSSL has been granted a patent license
4828 compatible with the OpenSSL license for use of OCB. Details are available
4829 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4830 for OCB can be removed by calling config with no-ocb.
4834 * SSLv2 support has been removed. It still supports receiving an SSLv2
4835 compatible client hello.
4839 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4840 done while fixing the error code for the key-too-small case.
4842 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4844 * CA.sh has been removed; use CA.pl instead.
4848 * Removed old DES API.
4852 * Remove various unsupported platforms:
4858 Sinix/ReliantUNIX RM400
4863 16-bit platforms such as WIN16
4867 * Clean up OPENSSL_NO_xxx #define's
4868 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4869 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4870 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4871 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4872 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4873 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4874 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4875 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4876 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4877 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4881 * Cleaned up dead code
4882 Remove all but one '#ifdef undef' which is to be looked at.
4886 * Clean up calling of xxx_free routines.
4887 Just like free(), fix most of the xxx_free routines to accept
4888 NULL. Remove the non-null checks from callers. Save much code.
4892 * Add secure heap for storage of private keys (when possible).
4893 Add BIO_s_secmem(), CBIGNUM, etc.
4894 Contributed by Akamai Technologies under our Corporate CLA.
4898 * Experimental support for a new, fast, unbiased prime candidate generator,
4899 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4901 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4903 * New output format NSS in the sess_id command line tool. This allows
4904 exporting the session id and the master key in NSS keylog format.
4906 *Martin Kaiser <martin@kaiser.cx>*
4908 * Harmonize version and its documentation. -f flag is used to display
4911 *mancha <mancha1@zoho.com>*
4913 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4914 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4916 *mancha <mancha1@zoho.com>*
4918 * Fix some double frees. These are not thought to be exploitable.
4920 *mancha <mancha1@zoho.com>*
4922 * A missing bounds check in the handling of the TLS heartbeat extension
4923 can be used to reveal up to 64k of memory to a connected client or
4926 Thanks for Neel Mehta of Google Security for discovering this bug and to
4927 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4928 preparing the fix ([CVE-2014-0160])
4930 *Adam Langley, Bodo Moeller*
4932 * Fix for the attack described in the paper "Recovering OpenSSL
4933 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4934 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4935 <http://eprint.iacr.org/2014/140>
4937 Thanks to Yuval Yarom and Naomi Benger for discovering this
4938 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4940 *Yuval Yarom and Naomi Benger*
4942 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4943 this fixes a limitation in previous versions of OpenSSL.
4947 * Experimental encrypt-then-mac support.
4949 Experimental support for encrypt then mac from
4950 draft-gutmann-tls-encrypt-then-mac-02.txt
4952 To enable it set the appropriate extension number (0x42 for the test
4953 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4955 For non-compliant peers (i.e. just about everything) this should have no
4958 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4962 * Add EVP support for key wrapping algorithms, to avoid problems with
4963 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4964 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4965 algorithms and include tests cases.
4969 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4974 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4975 MGF1 digest and OAEP label.
4979 * Make openssl verify return errors.
4981 *Chris Palmer <palmer@google.com> and Ben Laurie*
4983 * New function ASN1_TIME_diff to calculate the difference between two
4984 ASN1_TIME structures or one structure and the current time.
4988 * Update fips_test_suite to support multiple command line options. New
4989 test to induce all self test errors in sequence and check expected
4994 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4995 sign or verify all in one operation.
4999 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5000 test programs and fips_test_suite. Includes functionality to parse
5001 the minimal script output of fipsalgest.pl directly.
5005 * Add authorisation parameter to FIPS_module_mode_set().
5009 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5013 * Use separate DRBG fields for internal and external flags. New function
5014 FIPS_drbg_health_check() to perform on demand health checking. Add
5015 generation tests to fips_test_suite with reduced health check interval to
5016 demonstrate periodic health checking. Add "nodh" option to
5017 fips_test_suite to skip very slow DH test.
5021 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5026 * More extensive health check for DRBG checking many more failure modes.
5027 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5028 combination: call this in fips_test_suite.
5032 * Add support for canonical generation of DSA parameter 'g'. See
5035 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5036 POST to handle HMAC cases.
5040 * Add functions FIPS_module_version() and FIPS_module_version_text()
5041 to return numerical and string versions of the FIPS module number.
5045 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5046 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5047 outside the validated module in the FIPS capable OpenSSL.
5051 * Minor change to DRBG entropy callback semantics. In some cases
5052 there is no multiple of the block length between min_len and
5053 max_len. Allow the callback to return more than max_len bytes
5054 of entropy but discard any extra: it is the callback's responsibility
5055 to ensure that the extra data discarded does not impact the
5056 requested amount of entropy.
5060 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5061 information in FIPS186-3, SP800-57 and SP800-131A.
5065 * CCM support via EVP. Interface is very similar to GCM case except we
5066 must supply all data in one chunk (i.e. no update, final) and the
5067 message length must be supplied if AAD is used. Add algorithm test
5072 * Initial version of POST overhaul. Add POST callback to allow the status
5073 of POST to be monitored and/or failures induced. Modify fips_test_suite
5074 to use callback. Always run all selftests even if one fails.
5078 * XTS support including algorithm test driver in the fips_gcmtest program.
5079 Note: this does increase the maximum key length from 32 to 64 bytes but
5080 there should be no binary compatibility issues as existing applications
5081 will never use XTS mode.
5085 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5086 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5087 performs algorithm blocking for unapproved PRNG types. Also do not
5088 set PRNG type in FIPS_mode_set(): leave this to the application.
5089 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5090 the standard OpenSSL PRNG: set additional data to a date time vector.
5094 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5095 This shouldn't present any incompatibility problems because applications
5096 shouldn't be using these directly and any that are will need to rethink
5097 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5101 * Extensive self tests and health checking required by SP800-90 DRBG.
5102 Remove strength parameter from FIPS_drbg_instantiate and always
5103 instantiate at maximum supported strength.
5107 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5111 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5115 * New function DH_compute_key_padded() to compute a DH key and pad with
5116 leading zeroes if needed: this complies with SP800-56A et al.
5120 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5121 anything, incomplete, subject to change and largely untested at present.
5125 * Modify fipscanisteronly build option to only build the necessary object
5126 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5130 * Add experimental option FIPSSYMS to give all symbols in
5131 fipscanister.o and FIPS or fips prefix. This will avoid
5132 conflicts with future versions of OpenSSL. Add perl script
5133 util/fipsas.pl to preprocess assembly language source files
5134 and rename any affected symbols.
5138 * Add selftest checks and algorithm block of non-fips algorithms in
5139 FIPS mode. Remove DES2 from selftests.
5143 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5144 return internal method without any ENGINE dependencies. Add new
5145 tiny fips sign and verify functions.
5149 * New build option no-ec2m to disable characteristic 2 code.
5153 * New build option "fipscanisteronly". This only builds fipscanister.o
5154 and (currently) associated fips utilities. Uses the file Makefile.fips
5155 instead of Makefile.org as the prototype.
5159 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5160 Update fips_gcmtest to use IV generator.
5164 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5165 setting output buffer to NULL. The `*Final` function must be
5166 called although it will not retrieve any additional data. The tag
5167 can be set or retrieved with a ctrl. The IV length is by default 12
5168 bytes (96 bits) but can be set to an alternative value. If the IV
5169 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5174 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5175 underlying do_cipher function handles all cipher semantics itself
5176 including padding and finalisation. This is useful if (for example)
5177 an ENGINE cipher handles block padding itself. The behaviour of
5178 do_cipher is subtly changed if this flag is set: the return value
5179 is the number of characters written to the output buffer (zero is
5180 no longer an error code) or a negative error code. Also if the
5181 input buffer is NULL and length 0 finalisation should be performed.
5185 * If a candidate issuer certificate is already part of the constructed
5186 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5190 * Improve forward-security support: add functions
5192 void SSL_CTX_set_not_resumable_session_callback(
5193 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5194 void SSL_set_not_resumable_session_callback(
5195 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5197 for use by SSL/TLS servers; the callback function will be called whenever a
5198 new session is created, and gets to decide whether the session may be
5199 cached to make it resumable (return 0) or not (return 1). (As by the
5200 SSL/TLS protocol specifications, the session_id sent by the server will be
5201 empty to indicate that the session is not resumable; also, the server will
5202 not generate RFC 4507 (RFC 5077) session tickets.)
5204 A simple reasonable callback implementation is to return is_forward_secure.
5205 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5206 by the SSL/TLS server library, indicating whether it can provide forward
5209 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5211 * New -verify_name option in command line utilities to set verification
5216 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5217 Add CMAC pkey methods.
5221 * Experimental renegotiation in s_server -www mode. If the client
5222 browses /reneg connection is renegotiated. If /renegcert it is
5223 renegotiated requesting a certificate.
5227 * Add an "external" session cache for debugging purposes to s_server. This
5228 should help trace issues which normally are only apparent in deployed
5229 multi-process servers.
5233 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5234 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5235 BIO_set_cipher() and some obscure PEM functions were changed so they
5236 can now return an error. The RAND changes required a change to the
5237 RAND_METHOD structure.
5241 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5242 a gcc attribute to warn if the result of a function is ignored. This
5243 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5244 whose return value is often ignored.
5248 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5249 These allow SCTs (signed certificate timestamps) to be requested and
5250 validated when establishing a connection.
5252 *Rob Percival <robpercival@google.com>*
5257 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5259 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5260 used even when parsing explicit parameters, when loading a encoded key
5261 or calling `EC_GROUP_new_from_ecpkparameters()`/
5262 `EC_GROUP_new_from_ecparameters()`.
5263 This prevents bypass of security hardening and performance gains,
5264 especially for curves with specialized EC_METHODs.
5265 By default, if a key encoded with explicit parameters is loaded and later
5266 encoded, the output is still encoded with explicit parameters, even if
5267 internally a "named" EC_GROUP is used for computation.
5271 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5272 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5273 NULL. After this change, only the cofactor parameter can be NULL. It also
5274 does some minimal sanity checks on the passed order.
5279 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5280 An attack is simple, if the first CMS_recipientInfo is valid but the
5281 second CMS_recipientInfo is chosen ciphertext. If the second
5282 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5283 encryption key will be replaced by garbage, and the message cannot be
5284 decoded, but if the RSA decryption fails, the correct encryption key is
5285 used and the recipient will not notice the attack.
5286 As a work around for this potential attack the length of the decrypted
5287 key must be equal to the cipher default key length, in case the
5288 certificate is not given and all recipientInfo are tried out.
5289 The old behaviour can be re-enabled in the CMS code by setting the
5290 CMS_DEBUG_DECRYPT flag.
5295 * Document issue with installation paths in diverse Windows builds
5297 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5298 binaries and run-time config file.
5303 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5305 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5306 This changes the size when using the `genpkey` command when no size is given.
5307 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5308 generation commands to use 2048 bits by default.
5312 * Add FIPS support for Android Arm 64-bit
5314 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5315 Module in Version 2.0.10. For some reason, the corresponding target
5316 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5317 built with FIPS support on Android Arm 64-bit. This omission has been
5320 *Matthias St. Pierre*
5322 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5324 * 0-byte record padding oracle
5326 If an application encounters a fatal protocol error and then calls
5327 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5328 then OpenSSL can respond differently to the calling application if a 0 byte
5329 record is received with invalid padding compared to if a 0 byte record is
5330 received with an invalid MAC. If the application then behaves differently
5331 based on that in a way that is detectable to the remote peer, then this
5332 amounts to a padding oracle that could be used to decrypt data.
5334 In order for this to be exploitable "non-stitched" ciphersuites must be in
5335 use. Stitched ciphersuites are optimised implementations of certain
5336 commonly used ciphersuites. Also the application must call SSL_shutdown()
5337 twice even if a protocol error has occurred (applications should not do
5338 this but some do anyway).
5340 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5341 Aviram, with additional investigation by Steven Collison and Andrew
5342 Hourselt. It was reported to OpenSSL on 10th December 2018.
5347 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5351 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5353 * Microarchitecture timing vulnerability in ECC scalar multiplication
5355 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5356 shown to be vulnerable to a microarchitecture timing side channel attack.
5357 An attacker with sufficient access to mount local timing attacks during
5358 ECDSA signature generation could recover the private key.
5360 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5361 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5367 * Timing vulnerability in DSA signature generation
5369 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5370 timing side channel attack. An attacker could use variations in the signing
5371 algorithm to recover the private key.
5373 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5378 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5379 Module, accidentally introduced while backporting security fixes from the
5380 development branch and hindering the use of ECC in FIPS mode.
5384 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5386 * Client DoS due to large DH parameter
5388 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5389 malicious server can send a very large prime value to the client. This will
5390 cause the client to spend an unreasonably long period of time generating a
5391 key for this prime resulting in a hang until the client has finished. This
5392 could be exploited in a Denial Of Service attack.
5394 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5399 * Cache timing vulnerability in RSA Key Generation
5401 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5402 a cache timing side channel attack. An attacker with sufficient access to
5403 mount cache timing attacks during the RSA key generation process could
5404 recover the private key.
5406 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5407 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5412 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5413 parameter is no longer accepted, as it leads to a corrupt table. NULL
5414 pem_str is reserved for alias entries only.
5418 * Revert blinding in ECDSA sign and instead make problematic addition
5419 length-invariant. Switch even to fixed-length Montgomery multiplication.
5423 * Change generating and checking of primes so that the error rate of not
5424 being prime depends on the intended use based on the size of the input.
5425 For larger primes this will result in more rounds of Miller-Rabin.
5426 The maximal error rate for primes with more than 1080 bits is lowered
5429 *Kurt Roeckx, Annie Yousar*
5431 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5435 * Add blinding to ECDSA and DSA signatures to protect against side channel
5436 attacks discovered by Keegan Ryan (NCC Group).
5440 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5441 now allow empty (zero character) pass phrases.
5445 * Certificate time validation (X509_cmp_time) enforces stricter
5446 compliance with RFC 5280. Fractional seconds and timezone offsets
5447 are no longer allowed.
5451 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5453 * Constructed ASN.1 types with a recursive definition could exceed the stack
5455 Constructed ASN.1 types with a recursive definition (such as can be found
5456 in PKCS7) could eventually exceed the stack given malicious input with
5457 excessive recursion. This could result in a Denial Of Service attack. There
5458 are no such structures used within SSL/TLS that come from untrusted sources
5459 so this is considered safe.
5461 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5467 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5469 * Read/write after SSL object in error state
5471 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5472 mechanism. The intent was that if a fatal error occurred during a handshake
5473 then OpenSSL would move into the error state and would immediately fail if
5474 you attempted to continue the handshake. This works as designed for the
5475 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5476 SSL_connect()), however due to a bug it does not work correctly if
5477 SSL_read() or SSL_write() is called directly. In that scenario, if the
5478 handshake fails then a fatal error will be returned in the initial function
5479 call. If SSL_read()/SSL_write() is subsequently called by the application
5480 for the same SSL object then it will succeed and the data is passed without
5481 being decrypted/encrypted directly from the SSL/TLS record layer.
5483 In order to exploit this issue an application bug would have to be present
5484 that resulted in a call to SSL_read()/SSL_write() being issued after having
5485 already received a fatal error.
5487 This issue was reported to OpenSSL by David Benjamin (Google).
5492 * rsaz_1024_mul_avx2 overflow bug on x86_64
5494 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5495 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5496 Analysis suggests that attacks against RSA and DSA as a result of this
5497 defect would be very difficult to perform and are not believed likely.
5498 Attacks against DH1024 are considered just feasible, because most of the
5499 work necessary to deduce information about a private key may be performed
5500 offline. The amount of resources required for such an attack would be
5501 significant. However, for an attack on TLS to be meaningful, the server
5502 would have to share the DH1024 private key among multiple clients, which is
5503 no longer an option since CVE-2016-0701.
5505 This only affects processors that support the AVX2 but not ADX extensions
5506 like Intel Haswell (4th generation).
5508 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5509 was originally found via the OSS-Fuzz project.
5514 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5516 * bn_sqrx8x_internal carry bug on x86_64
5518 There is a carry propagating bug in the x86_64 Montgomery squaring
5519 procedure. No EC algorithms are affected. Analysis suggests that attacks
5520 against RSA and DSA as a result of this defect would be very difficult to
5521 perform and are not believed likely. Attacks against DH are considered just
5522 feasible (although very difficult) because most of the work necessary to
5523 deduce information about a private key may be performed offline. The amount
5524 of resources required for such an attack would be very significant and
5525 likely only accessible to a limited number of attackers. An attacker would
5526 additionally need online access to an unpatched system using the target
5527 private key in a scenario with persistent DH parameters and a private
5528 key that is shared between multiple clients.
5530 This only affects processors that support the BMI1, BMI2 and ADX extensions
5531 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5533 This issue was reported to OpenSSL by the OSS-Fuzz project.
5538 * Malformed X.509 IPAddressFamily could cause OOB read
5540 If an X.509 certificate has a malformed IPAddressFamily extension,
5541 OpenSSL could do a one-byte buffer overread. The most likely result
5542 would be an erroneous display of the certificate in text format.
5544 This issue was reported to OpenSSL by the OSS-Fuzz project.
5548 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5550 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5551 platform rather than 'mingw'.
5555 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5557 * Truncated packet could crash via OOB read
5559 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5560 cipher is being used, then a truncated packet can cause that host to
5561 perform an out-of-bounds read, usually resulting in a crash.
5563 This issue was reported to OpenSSL by Robert Święcki of Google.
5568 * BN_mod_exp may produce incorrect results on x86_64
5570 There is a carry propagating bug in the x86_64 Montgomery squaring
5571 procedure. No EC algorithms are affected. Analysis suggests that attacks
5572 against RSA and DSA as a result of this defect would be very difficult to
5573 perform and are not believed likely. Attacks against DH are considered just
5574 feasible (although very difficult) because most of the work necessary to
5575 deduce information about a private key may be performed offline. The amount
5576 of resources required for such an attack would be very significant and
5577 likely only accessible to a limited number of attackers. An attacker would
5578 additionally need online access to an unpatched system using the target
5579 private key in a scenario with persistent DH parameters and a private
5580 key that is shared between multiple clients. For example this can occur by
5581 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5582 similar to CVE-2015-3193 but must be treated as a separate problem.
5584 This issue was reported to OpenSSL by the OSS-Fuzz project.
5589 * Montgomery multiplication may produce incorrect results
5591 There is a carry propagating bug in the Broadwell-specific Montgomery
5592 multiplication procedure that handles input lengths divisible by, but
5593 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5594 and DH private keys are impossible. This is because the subroutine in
5595 question is not used in operations with the private key itself and an input
5596 of the attacker's direct choice. Otherwise the bug can manifest itself as
5597 transient authentication and key negotiation failures or reproducible
5598 erroneous outcome of public-key operations with specially crafted input.
5599 Among EC algorithms only Brainpool P-512 curves are affected and one
5600 presumably can attack ECDH key negotiation. Impact was not analyzed in
5601 detail, because pre-requisites for attack are considered unlikely. Namely
5602 multiple clients have to choose the curve in question and the server has to
5603 share the private key among them, neither of which is default behaviour.
5604 Even then only clients that chose the curve will be affected.
5606 This issue was publicly reported as transient failures and was not
5607 initially recognized as a security issue. Thanks to Richard Morgan for
5608 providing reproducible case.
5613 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5614 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5615 prevent issues where no progress is being made and the peer continually
5616 sends unrecognised record types, using up resources processing them.
5620 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5622 * Missing CRL sanity check
5624 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5625 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5626 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5628 This issue only affects the OpenSSL 1.0.2i
5633 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5635 * OCSP Status Request extension unbounded memory growth
5637 A malicious client can send an excessively large OCSP Status Request
5638 extension. If that client continually requests renegotiation, sending a
5639 large OCSP Status Request extension each time, then there will be unbounded
5640 memory growth on the server. This will eventually lead to a Denial Of
5641 Service attack through memory exhaustion. Servers with a default
5642 configuration are vulnerable even if they do not support OCSP. Builds using
5643 the "no-ocsp" build time option are not affected.
5645 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5650 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5653 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5659 * OOB write in MDC2_Update()
5661 An overflow can occur in MDC2_Update() either if called directly or
5662 through the EVP_DigestUpdate() function using MDC2. If an attacker
5663 is able to supply very large amounts of input data after a previous
5664 call to EVP_EncryptUpdate() with a partial block then a length check
5665 can overflow resulting in a heap corruption.
5667 The amount of data needed is comparable to SIZE_MAX which is impractical
5670 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5675 * Malformed SHA512 ticket DoS
5677 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5678 DoS attack where a malformed ticket will result in an OOB read which will
5681 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5682 a custom server callback and ticket lookup mechanism.
5684 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5689 * OOB write in BN_bn2dec()
5691 The function BN_bn2dec() does not check the return value of BN_div_word().
5692 This can cause an OOB write if an application uses this function with an
5693 overly large BIGNUM. This could be a problem if an overly large certificate
5694 or CRL is printed out from an untrusted source. TLS is not affected because
5695 record limits will reject an oversized certificate before it is parsed.
5697 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5702 * OOB read in TS_OBJ_print_bio()
5704 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5705 the total length the OID text representation would use and not the amount
5706 of data written. This will result in OOB reads when large OIDs are
5709 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5714 * Pointer arithmetic undefined behaviour
5716 Avoid some undefined pointer arithmetic
5718 A common idiom in the codebase is to check limits in the following manner:
5721 Where "p" points to some malloc'd data of SIZE bytes and
5724 "len" here could be from some externally supplied data (e.g. from a TLS
5727 The rules of C pointer arithmetic are such that "p + len" is only well
5728 defined where len <= SIZE. Therefore the above idiom is actually
5729 undefined behaviour.
5731 For example this could cause problems if some malloc implementation
5732 provides an address for "p" such that "p + len" actually overflows for
5733 values of len that are too big and therefore p + len < limit.
5735 This issue was reported to OpenSSL by Guido Vranken
5740 * Constant time flag not preserved in DSA signing
5742 Operations in the DSA signing algorithm should run in constant time in
5743 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5744 implementation means that a non-constant time codepath is followed for
5745 certain operations. This has been demonstrated through a cache-timing
5746 attack to be sufficient for an attacker to recover the private DSA key.
5748 This issue was reported by César Pereida (Aalto University), Billy Brumley
5749 (Tampere University of Technology), and Yuval Yarom (The University of
5750 Adelaide and NICTA).
5755 * DTLS buffered message DoS
5757 In a DTLS connection where handshake messages are delivered out-of-order
5758 those messages that OpenSSL is not yet ready to process will be buffered
5759 for later use. Under certain circumstances, a flaw in the logic means that
5760 those messages do not get removed from the buffer even though the handshake
5761 has been completed. An attacker could force up to approx. 15 messages to
5762 remain in the buffer when they are no longer required. These messages will
5763 be cleared when the DTLS connection is closed. The default maximum size for
5764 a message is 100k. Therefore, the attacker could force an additional 1500k
5765 to be consumed per connection. By opening many simultaneous connections an
5766 attacker could cause a DoS attack through memory exhaustion.
5768 This issue was reported to OpenSSL by Quan Luo.
5773 * DTLS replay protection DoS
5775 A flaw in the DTLS replay attack protection mechanism means that records
5776 that arrive for future epochs update the replay protection "window" before
5777 the MAC for the record has been validated. This could be exploited by an
5778 attacker by sending a record for the next epoch (which does not have to
5779 decrypt or have a valid MAC), with a very large sequence number. This means
5780 that all subsequent legitimate packets are dropped causing a denial of
5781 service for a specific DTLS connection.
5783 This issue was reported to OpenSSL by the OCAP audit team.
5788 * Certificate message OOB reads
5790 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5791 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5792 theoretical DoS risk but this has not been observed in practice on common
5795 The messages affected are client certificate, client certificate request
5796 and server certificate. As a result the attack can only be performed
5797 against a client or a server which enables client authentication.
5799 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5804 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5806 * Prevent padding oracle in AES-NI CBC MAC check
5808 A MITM attacker can use a padding oracle attack to decrypt traffic
5809 when the connection uses an AES CBC cipher and the server support
5812 This issue was introduced as part of the fix for Lucky 13 padding
5813 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5814 constant time by making sure that always the same bytes are read and
5815 compared against either the MAC or padding bytes. But it no longer
5816 checked that there was enough data to have both the MAC and padding
5819 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5823 * Fix EVP_EncodeUpdate overflow
5825 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5826 Base64 encoding of binary data. If an attacker is able to supply very large
5827 amounts of input data then a length check can overflow resulting in a heap
5830 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5831 the `PEM_write_bio*` family of functions. These are mainly used within the
5832 OpenSSL command line applications, so any application which processes data
5833 from an untrusted source and outputs it as a PEM file should be considered
5834 vulnerable to this issue. User applications that call these APIs directly
5835 with large amounts of untrusted data may also be vulnerable.
5837 This issue was reported by Guido Vranken.
5842 * Fix EVP_EncryptUpdate overflow
5844 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5845 is able to supply very large amounts of input data after a previous call to
5846 EVP_EncryptUpdate() with a partial block then a length check can overflow
5847 resulting in a heap corruption. Following an analysis of all OpenSSL
5848 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5849 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5850 the first called function after an EVP_EncryptInit(), and therefore that
5851 specific call must be safe. The second form is where the length passed to
5852 EVP_EncryptUpdate() can be seen from the code to be some small value and
5853 therefore there is no possibility of an overflow. Since all instances are
5854 one of these two forms, it is believed that there can be no overflows in
5855 internal code due to this problem. It should be noted that
5856 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5857 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5858 of these calls have also been analysed too and it is believed there are no
5859 instances in internal usage where an overflow could occur.
5861 This issue was reported by Guido Vranken.
5866 * Prevent ASN.1 BIO excessive memory allocation
5868 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5869 a short invalid encoding can cause allocation of large amounts of memory
5870 potentially consuming excessive resources or exhausting memory.
5872 Any application parsing untrusted data through d2i BIO functions is
5873 affected. The memory based functions such as d2i_X509() are *not* affected.
5874 Since the memory based functions are used by the TLS library, TLS
5875 applications are not affected.
5877 This issue was reported by Brian Carpenter.
5884 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5885 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5886 in arbitrary stack data being returned in the buffer.
5888 This issue was reported by Guido Vranken.
5893 * Modify behavior of ALPN to invoke callback after SNI/servername
5894 callback, such that updates to the SSL_CTX affect ALPN.
5898 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5903 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5904 methods are enabled and ssl2 is disabled the methods return NULL.
5908 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5910 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5911 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5912 provide any "EXPORT" or "LOW" strength ciphers.
5916 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5917 is by default disabled at build-time. Builds that are not configured with
5918 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5919 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5920 will need to explicitly call either of:
5922 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5924 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5926 as appropriate. Even if either of those is used, or the application
5927 explicitly uses the version-specific SSLv2_method() or its client and
5928 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5929 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5930 ciphers, and SSLv2 56-bit DES are no longer available.
5935 * Fix a double-free in DSA code
5937 A double free bug was discovered when OpenSSL parses malformed DSA private
5938 keys and could lead to a DoS attack or memory corruption for applications
5939 that receive DSA private keys from untrusted sources. This scenario is
5942 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5948 * Disable SRP fake user seed to address a server memory leak.
5950 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5952 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5953 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5954 was changed to ignore the "fake user" SRP seed, even if the seed
5957 Users should use SRP_VBASE_get1_by_user instead. Note that in
5958 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5959 also that even though configuring the SRP seed attempts to hide
5960 invalid usernames by continuing the handshake with fake
5961 credentials, this behaviour is not constant time and no strong
5962 guarantees are made that the handshake is indistinguishable from
5963 that of a valid user.
5968 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5970 In the BN_hex2bn function the number of hex digits is calculated using an
5971 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5972 large values of `i` this can result in `bn_expand` not allocating any
5973 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5974 field as NULL leading to a subsequent NULL ptr deref. For very large values
5975 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5976 In this case memory is allocated to the internal BIGNUM data field, but it
5977 is insufficiently sized leading to heap corruption. A similar issue exists
5978 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5979 is ever called by user applications with very large untrusted hex/dec data.
5980 This is anticipated to be a rare occurrence.
5982 All OpenSSL internal usage of these functions use data that is not expected
5983 to be untrusted, e.g. config file data or application command line
5984 arguments. If user developed applications generate config file data based
5985 on untrusted data then it is possible that this could also lead to security
5986 consequences. This is also anticipated to be rare.
5988 This issue was reported to OpenSSL by Guido Vranken.
5993 * Fix memory issues in `BIO_*printf` functions
5995 The internal `fmtstr` function used in processing a "%s" format string in
5996 the `BIO_*printf` functions could overflow while calculating the length of a
5997 string and cause an OOB read when printing very long strings.
5999 Additionally the internal `doapr_outch` function can attempt to write to an
6000 OOB memory location (at an offset from the NULL pointer) in the event of a
6001 memory allocation failure. In 1.0.2 and below this could be caused where
6002 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6003 could be in processing a very long "%s" format string. Memory leaks can
6006 The first issue may mask the second issue dependent on compiler behaviour.
6007 These problems could enable attacks where large amounts of untrusted data
6008 is passed to the `BIO_*printf` functions. If applications use these functions
6009 in this way then they could be vulnerable. OpenSSL itself uses these
6010 functions when printing out human-readable dumps of ASN.1 data. Therefore
6011 applications that print this data could be vulnerable if the data is from
6012 untrusted sources. OpenSSL command line applications could also be
6013 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6014 as command line arguments.
6016 Libssl is not considered directly vulnerable. Additionally certificates etc
6017 received via remote connections via libssl are also unlikely to be able to
6018 trigger these issues because of message size limits enforced within libssl.
6020 This issue was reported to OpenSSL Guido Vranken.
6025 * Side channel attack on modular exponentiation
6027 A side-channel attack was found which makes use of cache-bank conflicts on
6028 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6029 of RSA keys. The ability to exploit this issue is limited as it relies on
6030 an attacker who has control of code in a thread running on the same
6031 hyper-threaded core as the victim thread which is performing decryptions.
6033 This issue was reported to OpenSSL by Yuval Yarom, The University of
6034 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6035 Nadia Heninger, University of Pennsylvania with more information at
6036 <http://cachebleed.info>.
6041 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6042 if no keysize is specified with default_bits. This fixes an
6043 omission in an earlier change that changed all RSA/DSA key generation
6044 commands to use 2048 bits by default.
6048 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6050 * DH small subgroups
6052 Historically OpenSSL only ever generated DH parameters based on "safe"
6053 primes. More recently (in version 1.0.2) support was provided for
6054 generating X9.42 style parameter files such as those required for RFC 5114
6055 support. The primes used in such files may not be "safe". Where an
6056 application is using DH configured with parameters based on primes that are
6057 not "safe" then an attacker could use this fact to find a peer's private
6058 DH exponent. This attack requires that the attacker complete multiple
6059 handshakes in which the peer uses the same private DH exponent. For example
6060 this could be used to discover a TLS server's private DH exponent if it's
6061 reusing the private DH exponent or it's using a static DH ciphersuite.
6063 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6064 TLS. It is not on by default. If the option is not set then the server
6065 reuses the same private DH exponent for the life of the server process and
6066 would be vulnerable to this attack. It is believed that many popular
6067 applications do set this option and would therefore not be at risk.
6069 The fix for this issue adds an additional check where a "q" parameter is
6070 available (as is the case in X9.42 based parameters). This detects the
6071 only known attack, and is the only possible defense for static DH
6072 ciphersuites. This could have some performance impact.
6074 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6075 default and cannot be disabled. This could have some performance impact.
6077 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6082 * SSLv2 doesn't block disabled ciphers
6084 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6085 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6086 been disabled, provided that the SSLv2 protocol was not also disabled via
6089 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6090 and Sebastian Schinzel.
6095 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6097 * BN_mod_exp may produce incorrect results on x86_64
6099 There is a carry propagating bug in the x86_64 Montgomery squaring
6100 procedure. No EC algorithms are affected. Analysis suggests that attacks
6101 against RSA and DSA as a result of this defect would be very difficult to
6102 perform and are not believed likely. Attacks against DH are considered just
6103 feasible (although very difficult) because most of the work necessary to
6104 deduce information about a private key may be performed offline. The amount
6105 of resources required for such an attack would be very significant and
6106 likely only accessible to a limited number of attackers. An attacker would
6107 additionally need online access to an unpatched system using the target
6108 private key in a scenario with persistent DH parameters and a private
6109 key that is shared between multiple clients. For example this can occur by
6110 default in OpenSSL DHE based SSL/TLS ciphersuites.
6112 This issue was reported to OpenSSL by Hanno Böck.
6117 * Certificate verify crash with missing PSS parameter
6119 The signature verification routines will crash with a NULL pointer
6120 dereference if presented with an ASN.1 signature using the RSA PSS
6121 algorithm and absent mask generation function parameter. Since these
6122 routines are used to verify certificate signature algorithms this can be
6123 used to crash any certificate verification operation and exploited in a
6124 DoS attack. Any application which performs certificate verification is
6125 vulnerable including OpenSSL clients and servers which enable client
6128 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6133 * X509_ATTRIBUTE memory leak
6135 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6136 memory. This structure is used by the PKCS#7 and CMS routines so any
6137 application which reads PKCS#7 or CMS data from untrusted sources is
6138 affected. SSL/TLS is not affected.
6140 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6146 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6147 This changes the decoding behaviour for some invalid messages,
6148 though the change is mostly in the more lenient direction, and
6149 legacy behaviour is preserved as much as possible.
6153 * In DSA_generate_parameters_ex, if the provided seed is too short,
6156 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6158 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6160 * Alternate chains certificate forgery
6162 During certificate verification, OpenSSL will attempt to find an
6163 alternative certificate chain if the first attempt to build such a chain
6164 fails. An error in the implementation of this logic can mean that an
6165 attacker could cause certain checks on untrusted certificates to be
6166 bypassed, such as the CA flag, enabling them to use a valid leaf
6167 certificate to act as a CA and "issue" an invalid certificate.
6169 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6174 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6176 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6177 incompatibility in the handling of HMAC. The previous ABI has now been
6182 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6184 * Malformed ECParameters causes infinite loop
6186 When processing an ECParameters structure OpenSSL enters an infinite loop
6187 if the curve specified is over a specially malformed binary polynomial
6190 This can be used to perform denial of service against any
6191 system which processes public keys, certificate requests or
6192 certificates. This includes TLS clients and TLS servers with
6193 client authentication enabled.
6195 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6200 * Exploitable out-of-bounds read in X509_cmp_time
6202 X509_cmp_time does not properly check the length of the ASN1_TIME
6203 string and can read a few bytes out of bounds. In addition,
6204 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6207 An attacker can use this to craft malformed certificates and CRLs of
6208 various sizes and potentially cause a segmentation fault, resulting in
6209 a DoS on applications that verify certificates or CRLs. TLS clients
6210 that verify CRLs are affected. TLS clients and servers with client
6211 authentication enabled may be affected if they use custom verification
6214 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6215 independently by Hanno Böck.
6220 * PKCS7 crash with missing EnvelopedContent
6222 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6223 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6224 with missing content and trigger a NULL pointer dereference on parsing.
6226 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6227 structures from untrusted sources are affected. OpenSSL clients and
6228 servers are not affected.
6230 This issue was reported to OpenSSL by Michal Zalewski (Google).
6235 * CMS verify infinite loop with unknown hash function
6237 When verifying a signedData message the CMS code can enter an infinite loop
6238 if presented with an unknown hash function OID. This can be used to perform
6239 denial of service against any system which verifies signedData messages using
6241 This issue was reported to OpenSSL by Johannes Bauer.
6246 * Race condition handling NewSessionTicket
6248 If a NewSessionTicket is received by a multi-threaded client when attempting to
6249 reuse a previous ticket then a race condition can occur potentially leading to
6250 a double free of the ticket data.
6255 * Only support 256-bit or stronger elliptic curves with the
6256 'ecdh_auto' setting (server) or by default (client). Of supported
6257 curves, prefer P-256 (both).
6261 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6263 * ClientHello sigalgs DoS fix
6265 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6266 invalid signature algorithms extension a NULL pointer dereference will
6267 occur. This can be exploited in a DoS attack against the server.
6269 This issue was was reported to OpenSSL by David Ramos of Stanford
6273 *Stephen Henson and Matt Caswell*
6275 * Multiblock corrupted pointer fix
6277 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6278 feature only applies on 64 bit x86 architecture platforms that support AES
6279 NI instructions. A defect in the implementation of "multiblock" can cause
6280 OpenSSL's internal write buffer to become incorrectly set to NULL when
6281 using non-blocking IO. Typically, when the user application is using a
6282 socket BIO for writing, this will only result in a failed connection.
6283 However if some other BIO is used then it is likely that a segmentation
6284 fault will be triggered, thus enabling a potential DoS attack.
6286 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6291 * Segmentation fault in DTLSv1_listen fix
6293 The DTLSv1_listen function is intended to be stateless and processes the
6294 initial ClientHello from many peers. It is common for user code to loop
6295 over the call to DTLSv1_listen until a valid ClientHello is received with
6296 an associated cookie. A defect in the implementation of DTLSv1_listen means
6297 that state is preserved in the SSL object from one invocation to the next
6298 that can lead to a segmentation fault. Errors processing the initial
6299 ClientHello can trigger this scenario. An example of such an error could be
6300 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6303 This issue was reported to OpenSSL by Per Allansson.
6308 * Segmentation fault in ASN1_TYPE_cmp fix
6310 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6311 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6312 certificate signature algorithm consistency this can be used to crash any
6313 certificate verification operation and exploited in a DoS attack. Any
6314 application which performs certificate verification is vulnerable including
6315 OpenSSL clients and servers which enable client authentication.
6320 * Segmentation fault for invalid PSS parameters fix
6322 The signature verification routines will crash with a NULL pointer
6323 dereference if presented with an ASN.1 signature using the RSA PSS
6324 algorithm and invalid parameters. Since these routines are used to verify
6325 certificate signature algorithms this can be used to crash any
6326 certificate verification operation and exploited in a DoS attack. Any
6327 application which performs certificate verification is vulnerable including
6328 OpenSSL clients and servers which enable client authentication.
6330 This issue was was reported to OpenSSL by Brian Carpenter.
6335 * ASN.1 structure reuse memory corruption fix
6337 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6338 memory corruption via an invalid write. Such reuse is and has been
6339 strongly discouraged and is believed to be rare.
6341 Applications that parse structures containing CHOICE or ANY DEFINED BY
6342 components may be affected. Certificate parsing (d2i_X509 and related
6343 functions) are however not affected. OpenSSL clients and servers are
6349 * PKCS7 NULL pointer dereferences fix
6351 The PKCS#7 parsing code does not handle missing outer ContentInfo
6352 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6353 missing content and trigger a NULL pointer dereference on parsing.
6355 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6356 otherwise parse PKCS#7 structures from untrusted sources are
6357 affected. OpenSSL clients and servers are not affected.
6359 This issue was reported to OpenSSL by Michal Zalewski (Google).
6364 * DoS via reachable assert in SSLv2 servers fix
6366 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6367 servers that both support SSLv2 and enable export cipher suites by sending
6368 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6370 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6371 (OpenSSL development team).
6376 * Empty CKE with client auth and DHE fix
6378 If client auth is used then a server can seg fault in the event of a DHE
6379 ciphersuite being selected and a zero length ClientKeyExchange message
6380 being sent by the client. This could be exploited in a DoS attack.
6385 * Handshake with unseeded PRNG fix
6387 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6388 with an unseeded PRNG. The conditions are:
6389 - The client is on a platform where the PRNG has not been seeded
6390 automatically, and the user has not seeded manually
6391 - A protocol specific client method version has been used (i.e. not
6392 SSL_client_methodv23)
6393 - A ciphersuite is used that does not require additional random data from
6394 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6396 If the handshake succeeds then the client random that has been used will
6397 have been generated from a PRNG with insufficient entropy and therefore the
6398 output may be predictable.
6400 For example using the following command with an unseeded openssl will
6401 succeed on an unpatched platform:
6403 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6408 * Use After Free following d2i_ECPrivatekey error fix
6410 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6411 could cause a use after free condition. This, in turn, could cause a double
6412 free in several private key parsing functions (such as d2i_PrivateKey
6413 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6414 for applications that receive EC private keys from untrusted
6415 sources. This scenario is considered rare.
6417 This issue was discovered by the BoringSSL project and fixed in their
6423 * X509_to_X509_REQ NULL pointer deref fix
6425 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6426 the certificate key is invalid. This function is rarely used in practice.
6428 This issue was discovered by Brian Carpenter.
6433 * Removed the export ciphers from the DEFAULT ciphers
6437 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6439 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6440 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6441 So far those who have to target multiple platforms would compromise
6442 and argue that binary targeting say ARMv5 would still execute on
6443 ARMv8. "Universal" build resolves this compromise by providing
6444 near-optimal performance even on newer platforms.
6448 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6449 (other platforms pending).
6451 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6453 * Add support for the SignedCertificateTimestampList certificate and
6454 OCSP response extensions from RFC6962.
6458 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6459 for corner cases. (Certain input points at infinity could lead to
6460 bogus results, with non-infinity inputs mapped to infinity too.)
6464 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6465 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6466 common cases are optimized and there still is room for further
6467 improvements. Vector Permutation AES for Altivec is also added.
6471 * Add support for little-endian ppc64 Linux target.
6473 *Marcelo Cerri (IBM)*
6475 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6476 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6477 are optimized and there still is room for further improvements.
6478 Both 32- and 64-bit modes are supported.
6480 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6482 * Improved ARMv7 NEON support.
6486 * Support for SPARC Architecture 2011 crypto extensions, first
6487 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6488 SHA256/512, MD5, GHASH and modular exponentiation.
6490 *Andy Polyakov, David Miller*
6492 * Accelerated modular exponentiation for Intel processors, a.k.a.
6495 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6497 * Support for new and upcoming Intel processors, including AVX2,
6498 BMI and SHA ISA extensions. This includes additional "stitched"
6499 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6502 This work was sponsored by Intel Corp.
6506 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6507 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6508 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6512 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6513 this fixes a limitation in previous versions of OpenSSL.
6517 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6518 MGF1 digest and OAEP label.
6522 * Add EVP support for key wrapping algorithms, to avoid problems with
6523 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6524 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6525 algorithms and include tests cases.
6529 * Add functions to allocate and set the fields of an ECDSA_METHOD
6532 *Douglas E. Engert, Steve Henson*
6534 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6535 difference in days and seconds between two tm or ASN1_TIME structures.
6539 * Add -rev test option to s_server to just reverse order of characters
6540 received by client and send back to server. Also prints an abbreviated
6541 summary of the connection parameters.
6545 * New option -brief for s_client and s_server to print out a brief summary
6546 of connection parameters.
6550 * Add callbacks for arbitrary TLS extensions.
6552 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6554 * New option -crl_download in several openssl utilities to download CRLs
6555 from CRLDP extension in certificates.
6559 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6563 * New function X509_CRL_diff to generate a delta CRL from the difference
6564 of two full CRLs. Add support to "crl" utility.
6568 * New functions to set lookup_crls function and to retrieve
6569 X509_STORE from X509_STORE_CTX.
6573 * Print out deprecated issuer and subject unique ID fields in
6578 * Extend OCSP I/O functions so they can be used for simple general purpose
6579 HTTP as well as OCSP. New wrapper function which can be used to download
6580 CRLs using the OCSP API.
6584 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6588 * `SSL_CONF*` functions. These provide a common framework for application
6589 configuration using configuration files or command lines.
6593 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6594 message callback and prints the results. Needs compile time option
6595 "enable-ssl-trace". New options to s_client and s_server to enable
6600 * New ctrl and macro to retrieve supported points extensions.
6601 Print out extension in s_server and s_client.
6605 * New functions to retrieve certificate signature and signature
6610 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6615 * New Suite B modes for TLS code. These use and enforce the requirements
6616 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6617 only use Suite B curves. The Suite B modes can be set by using the
6618 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6622 * New chain verification flags for Suite B levels of security. Check
6623 algorithms are acceptable when flags are set in X509_verify_cert.
6627 * Make tls1_check_chain return a set of flags indicating checks passed
6628 by a certificate chain. Add additional tests to handle client
6629 certificates: checks for matching certificate type and issuer name
6634 * If an attempt is made to use a signature algorithm not in the peer
6635 preference list abort the handshake. If client has no suitable
6636 signature algorithms in response to a certificate request do not
6637 use the certificate.
6641 * If server EC tmp key is not in client preference list abort handshake.
6645 * Add support for certificate stores in CERT structure. This makes it
6646 possible to have different stores per SSL structure or one store in
6647 the parent SSL_CTX. Include distinct stores for certificate chain
6648 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6649 to build and store a certificate chain in CERT structure: returning
6650 an error if the chain cannot be built: this will allow applications
6651 to test if a chain is correctly configured.
6653 Note: if the CERT based stores are not set then the parent SSL_CTX
6654 store is used to retain compatibility with existing behaviour.
6658 * New function ssl_set_client_disabled to set a ciphersuite disabled
6659 mask based on the current session, check mask when sending client
6660 hello and checking the requested ciphersuite.
6664 * New ctrls to retrieve and set certificate types in a certificate
6665 request message. Print out received values in s_client. If certificate
6666 types is not set with custom values set sensible values based on
6667 supported signature algorithms.
6671 * Support for distinct client and server supported signature algorithms.
6675 * Add certificate callback. If set this is called whenever a certificate
6676 is required by client or server. An application can decide which
6677 certificate chain to present based on arbitrary criteria: for example
6678 supported signature algorithms. Add very simple example to s_server.
6679 This fixes many of the problems and restrictions of the existing client
6680 certificate callback: for example you can now clear an existing
6681 certificate and specify the whole chain.
6685 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6686 the certificate can be used for (if anything). Set valid_flags field
6687 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6688 to have similar checks in it.
6690 Add new "cert_flags" field to CERT structure and include a "strict mode".
6691 This enforces some TLS certificate requirements (such as only permitting
6692 certificate signature algorithms contained in the supported algorithms
6693 extension) which some implementations ignore: this option should be used
6694 with caution as it could cause interoperability issues.
6698 * Update and tidy signature algorithm extension processing. Work out
6699 shared signature algorithms based on preferences and peer algorithms
6700 and print them out in s_client and s_server. Abort handshake if no
6701 shared signature algorithms.
6705 * Add new functions to allow customised supported signature algorithms
6706 for SSL and SSL_CTX structures. Add options to s_client and s_server
6711 * New function SSL_certs_clear() to delete all references to certificates
6712 from an SSL structure. Before this once a certificate had been added
6713 it couldn't be removed.
6717 * Integrate hostname, email address and IP address checking with certificate
6718 verification. New verify options supporting checking in openssl utility.
6722 * Fixes and wildcard matching support to hostname and email checking
6723 functions. Add manual page.
6725 *Florian Weimer (Red Hat Product Security Team)*
6727 * New functions to check a hostname email or IP address against a
6728 certificate. Add options x509 utility to print results of checks against
6733 * Fix OCSP checking.
6735 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6737 * Initial experimental support for explicitly trusted non-root CAs.
6738 OpenSSL still tries to build a complete chain to a root but if an
6739 intermediate CA has a trust setting included that is used. The first
6740 setting is used: whether to trust (e.g., -addtrust option to the x509
6745 * Add -trusted_first option which attempts to find certificates in the
6746 trusted store even if an untrusted chain is also supplied.
6750 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6751 platform support for Linux and Android.
6755 * Support for linux-x32, ILP32 environment in x86_64 framework.
6759 * Experimental multi-implementation support for FIPS capable OpenSSL.
6760 When in FIPS mode the approved implementations are used as normal,
6761 when not in FIPS mode the internal unapproved versions are used instead.
6762 This means that the FIPS capable OpenSSL isn't forced to use the
6763 (often lower performance) FIPS implementations outside FIPS mode.
6767 * Transparently support X9.42 DH parameters when calling
6768 PEM_read_bio_DHparameters. This means existing applications can handle
6769 the new parameter format automatically.
6773 * Initial experimental support for X9.42 DH parameter format: mainly
6774 to support use of 'q' parameter for RFC5114 parameters.
6778 * Add DH parameters from RFC5114 including test data to dhtest.
6782 * Support for automatic EC temporary key parameter selection. If enabled
6783 the most preferred EC parameters are automatically used instead of
6784 hardcoded fixed parameters. Now a server just has to call:
6785 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6786 support ECDH and use the most appropriate parameters.
6790 * Enhance and tidy EC curve and point format TLS extension code. Use
6791 static structures instead of allocation if default values are used.
6792 New ctrls to set curves we wish to support and to retrieve shared curves.
6793 Print out shared curves in s_server. New options to s_server and s_client
6794 to set list of supported curves.
6798 * New ctrls to retrieve supported signature algorithms and
6799 supported curve values as an array of NIDs. Extend openssl utility
6800 to print out received values.
6804 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6805 between NIDs and the more common NIST names such as "P-256". Enhance
6806 ecparam utility and ECC method to recognise the NIST names for curves.
6810 * Enhance SSL/TLS certificate chain handling to support different
6811 chains for each certificate instead of one chain in the parent SSL_CTX.
6815 * Support for fixed DH ciphersuite client authentication: where both
6816 server and client use DH certificates with common parameters.
6820 * Support for fixed DH ciphersuites: those requiring DH server
6825 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6827 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6828 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6829 X509_CINF_get_signature were reverted post internal team review.
6834 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6836 * OCSP Status Request extension unbounded memory growth
6838 A malicious client can send an excessively large OCSP Status Request
6839 extension. If that client continually requests renegotiation, sending a
6840 large OCSP Status Request extension each time, then there will be unbounded
6841 memory growth on the server. This will eventually lead to a Denial Of
6842 Service attack through memory exhaustion. Servers with a default
6843 configuration are vulnerable even if they do not support OCSP. Builds using
6844 the "no-ocsp" build time option are not affected.
6846 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6851 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6854 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6860 * OOB write in MDC2_Update()
6862 An overflow can occur in MDC2_Update() either if called directly or
6863 through the EVP_DigestUpdate() function using MDC2. If an attacker
6864 is able to supply very large amounts of input data after a previous
6865 call to EVP_EncryptUpdate() with a partial block then a length check
6866 can overflow resulting in a heap corruption.
6868 The amount of data needed is comparable to SIZE_MAX which is impractical
6871 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6876 * Malformed SHA512 ticket DoS
6878 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6879 DoS attack where a malformed ticket will result in an OOB read which will
6882 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6883 a custom server callback and ticket lookup mechanism.
6885 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6890 * OOB write in BN_bn2dec()
6892 The function BN_bn2dec() does not check the return value of BN_div_word().
6893 This can cause an OOB write if an application uses this function with an
6894 overly large BIGNUM. This could be a problem if an overly large certificate
6895 or CRL is printed out from an untrusted source. TLS is not affected because
6896 record limits will reject an oversized certificate before it is parsed.
6898 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6903 * OOB read in TS_OBJ_print_bio()
6905 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6906 the total length the OID text representation would use and not the amount
6907 of data written. This will result in OOB reads when large OIDs are
6910 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6915 * Pointer arithmetic undefined behaviour
6917 Avoid some undefined pointer arithmetic
6919 A common idiom in the codebase is to check limits in the following manner:
6922 Where "p" points to some malloc'd data of SIZE bytes and
6925 "len" here could be from some externally supplied data (e.g. from a TLS
6928 The rules of C pointer arithmetic are such that "p + len" is only well
6929 defined where len <= SIZE. Therefore, the above idiom is actually
6930 undefined behaviour.
6932 For example this could cause problems if some malloc implementation
6933 provides an address for "p" such that "p + len" actually overflows for
6934 values of len that are too big and therefore p + len < limit.
6936 This issue was reported to OpenSSL by Guido Vranken
6941 * Constant time flag not preserved in DSA signing
6943 Operations in the DSA signing algorithm should run in constant time in
6944 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6945 implementation means that a non-constant time codepath is followed for
6946 certain operations. This has been demonstrated through a cache-timing
6947 attack to be sufficient for an attacker to recover the private DSA key.
6949 This issue was reported by César Pereida (Aalto University), Billy Brumley
6950 (Tampere University of Technology), and Yuval Yarom (The University of
6951 Adelaide and NICTA).
6956 * DTLS buffered message DoS
6958 In a DTLS connection where handshake messages are delivered out-of-order
6959 those messages that OpenSSL is not yet ready to process will be buffered
6960 for later use. Under certain circumstances, a flaw in the logic means that
6961 those messages do not get removed from the buffer even though the handshake
6962 has been completed. An attacker could force up to approx. 15 messages to
6963 remain in the buffer when they are no longer required. These messages will
6964 be cleared when the DTLS connection is closed. The default maximum size for
6965 a message is 100k. Therefore, the attacker could force an additional 1500k
6966 to be consumed per connection. By opening many simultaneous connections an
6967 attacker could cause a DoS attack through memory exhaustion.
6969 This issue was reported to OpenSSL by Quan Luo.
6974 * DTLS replay protection DoS
6976 A flaw in the DTLS replay attack protection mechanism means that records
6977 that arrive for future epochs update the replay protection "window" before
6978 the MAC for the record has been validated. This could be exploited by an
6979 attacker by sending a record for the next epoch (which does not have to
6980 decrypt or have a valid MAC), with a very large sequence number. This means
6981 that all subsequent legitimate packets are dropped causing a denial of
6982 service for a specific DTLS connection.
6984 This issue was reported to OpenSSL by the OCAP audit team.
6989 * Certificate message OOB reads
6991 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6992 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6993 theoretical DoS risk but this has not been observed in practice on common
6996 The messages affected are client certificate, client certificate request
6997 and server certificate. As a result the attack can only be performed
6998 against a client or a server which enables client authentication.
7000 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7005 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7007 * Prevent padding oracle in AES-NI CBC MAC check
7009 A MITM attacker can use a padding oracle attack to decrypt traffic
7010 when the connection uses an AES CBC cipher and the server support
7013 This issue was introduced as part of the fix for Lucky 13 padding
7014 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7015 constant time by making sure that always the same bytes are read and
7016 compared against either the MAC or padding bytes. But it no longer
7017 checked that there was enough data to have both the MAC and padding
7020 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7025 * Fix EVP_EncodeUpdate overflow
7027 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7028 Base64 encoding of binary data. If an attacker is able to supply very large
7029 amounts of input data then a length check can overflow resulting in a heap
7032 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7033 the `PEM_write_bio*` family of functions. These are mainly used within the
7034 OpenSSL command line applications, so any application which processes data
7035 from an untrusted source and outputs it as a PEM file should be considered
7036 vulnerable to this issue. User applications that call these APIs directly
7037 with large amounts of untrusted data may also be vulnerable.
7039 This issue was reported by Guido Vranken.
7044 * Fix EVP_EncryptUpdate overflow
7046 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7047 is able to supply very large amounts of input data after a previous call to
7048 EVP_EncryptUpdate() with a partial block then a length check can overflow
7049 resulting in a heap corruption. Following an analysis of all OpenSSL
7050 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7051 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7052 the first called function after an EVP_EncryptInit(), and therefore that
7053 specific call must be safe. The second form is where the length passed to
7054 EVP_EncryptUpdate() can be seen from the code to be some small value and
7055 therefore there is no possibility of an overflow. Since all instances are
7056 one of these two forms, it is believed that there can be no overflows in
7057 internal code due to this problem. It should be noted that
7058 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7059 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7060 of these calls have also been analysed too and it is believed there are no
7061 instances in internal usage where an overflow could occur.
7063 This issue was reported by Guido Vranken.
7068 * Prevent ASN.1 BIO excessive memory allocation
7070 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7071 a short invalid encoding can casuse allocation of large amounts of memory
7072 potentially consuming excessive resources or exhausting memory.
7074 Any application parsing untrusted data through d2i BIO functions is
7075 affected. The memory based functions such as d2i_X509() are *not* affected.
7076 Since the memory based functions are used by the TLS library, TLS
7077 applications are not affected.
7079 This issue was reported by Brian Carpenter.
7086 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7087 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7088 in arbitrary stack data being returned in the buffer.
7090 This issue was reported by Guido Vranken.
7095 * Modify behavior of ALPN to invoke callback after SNI/servername
7096 callback, such that updates to the SSL_CTX affect ALPN.
7100 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7105 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7106 methods are enabled and ssl2 is disabled the methods return NULL.
7110 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7112 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7113 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7114 provide any "EXPORT" or "LOW" strength ciphers.
7118 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7119 is by default disabled at build-time. Builds that are not configured with
7120 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7121 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7122 will need to explicitly call either of:
7124 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7126 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7128 as appropriate. Even if either of those is used, or the application
7129 explicitly uses the version-specific SSLv2_method() or its client and
7130 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7131 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7132 ciphers, and SSLv2 56-bit DES are no longer available.
7137 * Fix a double-free in DSA code
7139 A double free bug was discovered when OpenSSL parses malformed DSA private
7140 keys and could lead to a DoS attack or memory corruption for applications
7141 that receive DSA private keys from untrusted sources. This scenario is
7144 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7150 * Disable SRP fake user seed to address a server memory leak.
7152 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7154 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7155 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7156 was changed to ignore the "fake user" SRP seed, even if the seed
7159 Users should use SRP_VBASE_get1_by_user instead. Note that in
7160 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7161 also that even though configuring the SRP seed attempts to hide
7162 invalid usernames by continuing the handshake with fake
7163 credentials, this behaviour is not constant time and no strong
7164 guarantees are made that the handshake is indistinguishable from
7165 that of a valid user.
7170 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7172 In the BN_hex2bn function the number of hex digits is calculated using an
7173 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7174 large values of `i` this can result in `bn_expand` not allocating any
7175 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7176 field as NULL leading to a subsequent NULL ptr deref. For very large values
7177 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7178 In this case memory is allocated to the internal BIGNUM data field, but it
7179 is insufficiently sized leading to heap corruption. A similar issue exists
7180 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7181 is ever called by user applications with very large untrusted hex/dec data.
7182 This is anticipated to be a rare occurrence.
7184 All OpenSSL internal usage of these functions use data that is not expected
7185 to be untrusted, e.g. config file data or application command line
7186 arguments. If user developed applications generate config file data based
7187 on untrusted data then it is possible that this could also lead to security
7188 consequences. This is also anticipated to be rare.
7190 This issue was reported to OpenSSL by Guido Vranken.
7195 * Fix memory issues in `BIO_*printf` functions
7197 The internal `fmtstr` function used in processing a "%s" format string in
7198 the `BIO_*printf` functions could overflow while calculating the length of a
7199 string and cause an OOB read when printing very long strings.
7201 Additionally the internal `doapr_outch` function can attempt to write to an
7202 OOB memory location (at an offset from the NULL pointer) in the event of a
7203 memory allocation failure. In 1.0.2 and below this could be caused where
7204 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7205 could be in processing a very long "%s" format string. Memory leaks can
7208 The first issue may mask the second issue dependent on compiler behaviour.
7209 These problems could enable attacks where large amounts of untrusted data
7210 is passed to the `BIO_*printf` functions. If applications use these functions
7211 in this way then they could be vulnerable. OpenSSL itself uses these
7212 functions when printing out human-readable dumps of ASN.1 data. Therefore
7213 applications that print this data could be vulnerable if the data is from
7214 untrusted sources. OpenSSL command line applications could also be
7215 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7216 as command line arguments.
7218 Libssl is not considered directly vulnerable. Additionally certificates etc
7219 received via remote connections via libssl are also unlikely to be able to
7220 trigger these issues because of message size limits enforced within libssl.
7222 This issue was reported to OpenSSL Guido Vranken.
7227 * Side channel attack on modular exponentiation
7229 A side-channel attack was found which makes use of cache-bank conflicts on
7230 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7231 of RSA keys. The ability to exploit this issue is limited as it relies on
7232 an attacker who has control of code in a thread running on the same
7233 hyper-threaded core as the victim thread which is performing decryptions.
7235 This issue was reported to OpenSSL by Yuval Yarom, The University of
7236 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7237 Nadia Heninger, University of Pennsylvania with more information at
7238 <http://cachebleed.info>.
7243 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7244 if no keysize is specified with default_bits. This fixes an
7245 omission in an earlier change that changed all RSA/DSA key generation
7246 commands to use 2048 bits by default.
7250 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7252 * Protection for DH small subgroup attacks
7254 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7255 switched on by default and cannot be disabled. This could have some
7260 * SSLv2 doesn't block disabled ciphers
7262 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7263 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7264 been disabled, provided that the SSLv2 protocol was not also disabled via
7267 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7268 and Sebastian Schinzel.
7273 * Reject DH handshakes with parameters shorter than 1024 bits.
7277 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7279 * Certificate verify crash with missing PSS parameter
7281 The signature verification routines will crash with a NULL pointer
7282 dereference if presented with an ASN.1 signature using the RSA PSS
7283 algorithm and absent mask generation function parameter. Since these
7284 routines are used to verify certificate signature algorithms this can be
7285 used to crash any certificate verification operation and exploited in a
7286 DoS attack. Any application which performs certificate verification is
7287 vulnerable including OpenSSL clients and servers which enable client
7290 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7295 * X509_ATTRIBUTE memory leak
7297 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7298 memory. This structure is used by the PKCS#7 and CMS routines so any
7299 application which reads PKCS#7 or CMS data from untrusted sources is
7300 affected. SSL/TLS is not affected.
7302 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7308 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7309 This changes the decoding behaviour for some invalid messages,
7310 though the change is mostly in the more lenient direction, and
7311 legacy behaviour is preserved as much as possible.
7315 * In DSA_generate_parameters_ex, if the provided seed is too short,
7316 use a random seed, as already documented.
7318 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7320 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7322 * Alternate chains certificate forgery
7324 During certificate verfification, OpenSSL will attempt to find an
7325 alternative certificate chain if the first attempt to build such a chain
7326 fails. An error in the implementation of this logic can mean that an
7327 attacker could cause certain checks on untrusted certificates to be
7328 bypassed, such as the CA flag, enabling them to use a valid leaf
7329 certificate to act as a CA and "issue" an invalid certificate.
7331 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7337 * Race condition handling PSK identify hint
7339 If PSK identity hints are received by a multi-threaded client then
7340 the values are wrongly updated in the parent SSL_CTX structure. This can
7341 result in a race condition potentially leading to a double free of the
7347 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7349 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7350 incompatibility in the handling of HMAC. The previous ABI has now been
7353 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7355 * Malformed ECParameters causes infinite loop
7357 When processing an ECParameters structure OpenSSL enters an infinite loop
7358 if the curve specified is over a specially malformed binary polynomial
7361 This can be used to perform denial of service against any
7362 system which processes public keys, certificate requests or
7363 certificates. This includes TLS clients and TLS servers with
7364 client authentication enabled.
7366 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7371 * Exploitable out-of-bounds read in X509_cmp_time
7373 X509_cmp_time does not properly check the length of the ASN1_TIME
7374 string and can read a few bytes out of bounds. In addition,
7375 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7378 An attacker can use this to craft malformed certificates and CRLs of
7379 various sizes and potentially cause a segmentation fault, resulting in
7380 a DoS on applications that verify certificates or CRLs. TLS clients
7381 that verify CRLs are affected. TLS clients and servers with client
7382 authentication enabled may be affected if they use custom verification
7385 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7386 independently by Hanno Böck.
7391 * PKCS7 crash with missing EnvelopedContent
7393 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7394 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7395 with missing content and trigger a NULL pointer dereference on parsing.
7397 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7398 structures from untrusted sources are affected. OpenSSL clients and
7399 servers are not affected.
7401 This issue was reported to OpenSSL by Michal Zalewski (Google).
7406 * CMS verify infinite loop with unknown hash function
7408 When verifying a signedData message the CMS code can enter an infinite loop
7409 if presented with an unknown hash function OID. This can be used to perform
7410 denial of service against any system which verifies signedData messages using
7412 This issue was reported to OpenSSL by Johannes Bauer.
7417 * Race condition handling NewSessionTicket
7419 If a NewSessionTicket is received by a multi-threaded client when attempting to
7420 reuse a previous ticket then a race condition can occur potentially leading to
7421 a double free of the ticket data.
7426 * Reject DH handshakes with parameters shorter than 768 bits.
7428 *Kurt Roeckx and Emilia Kasper*
7430 * dhparam: generate 2048-bit parameters by default.
7432 *Kurt Roeckx and Emilia Kasper*
7434 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7436 * Segmentation fault in ASN1_TYPE_cmp fix
7438 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7439 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7440 certificate signature algorithm consistency this can be used to crash any
7441 certificate verification operation and exploited in a DoS attack. Any
7442 application which performs certificate verification is vulnerable including
7443 OpenSSL clients and servers which enable client authentication.
7448 * ASN.1 structure reuse memory corruption fix
7450 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7451 memory corruption via an invalid write. Such reuse is and has been
7452 strongly discouraged and is believed to be rare.
7454 Applications that parse structures containing CHOICE or ANY DEFINED BY
7455 components may be affected. Certificate parsing (d2i_X509 and related
7456 functions) are however not affected. OpenSSL clients and servers are
7462 * PKCS7 NULL pointer dereferences fix
7464 The PKCS#7 parsing code does not handle missing outer ContentInfo
7465 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7466 missing content and trigger a NULL pointer dereference on parsing.
7468 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7469 otherwise parse PKCS#7 structures from untrusted sources are
7470 affected. OpenSSL clients and servers are not affected.
7472 This issue was reported to OpenSSL by Michal Zalewski (Google).
7477 * DoS via reachable assert in SSLv2 servers fix
7479 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7480 servers that both support SSLv2 and enable export cipher suites by sending
7481 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7483 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7484 (OpenSSL development team).
7489 * Use After Free following d2i_ECPrivatekey error fix
7491 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7492 could cause a use after free condition. This, in turn, could cause a double
7493 free in several private key parsing functions (such as d2i_PrivateKey
7494 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7495 for applications that receive EC private keys from untrusted
7496 sources. This scenario is considered rare.
7498 This issue was discovered by the BoringSSL project and fixed in their
7504 * X509_to_X509_REQ NULL pointer deref fix
7506 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7507 the certificate key is invalid. This function is rarely used in practice.
7509 This issue was discovered by Brian Carpenter.
7514 * Removed the export ciphers from the DEFAULT ciphers
7518 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7520 * Build fixes for the Windows and OpenVMS platforms
7522 *Matt Caswell and Richard Levitte*
7524 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7526 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7527 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7528 dereference. This could lead to a Denial Of Service attack. Thanks to
7529 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7534 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7535 dtls1_buffer_record function under certain conditions. In particular this
7536 could occur if an attacker sent repeated DTLS records with the same
7537 sequence number but for the next epoch. The memory leak could be exploited
7538 by an attacker in a Denial of Service attack through memory exhaustion.
7539 Thanks to Chris Mueller for reporting this issue.
7544 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7545 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7546 method would be set to NULL which could later result in a NULL pointer
7547 dereference. Thanks to Frank Schmirler for reporting this issue.
7552 * Abort handshake if server key exchange message is omitted for ephemeral
7555 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7556 reporting this issue.
7561 * Remove non-export ephemeral RSA code on client and server. This code
7562 violated the TLS standard by allowing the use of temporary RSA keys in
7563 non-export ciphersuites and could be used by a server to effectively
7564 downgrade the RSA key length used to a value smaller than the server
7565 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7566 INRIA or reporting this issue.
7571 * Fixed issue where DH client certificates are accepted without verification.
7572 An OpenSSL server will accept a DH certificate for client authentication
7573 without the certificate verify message. This effectively allows a client to
7574 authenticate without the use of a private key. This only affects servers
7575 which trust a client certificate authority which issues certificates
7576 containing DH keys: these are extremely rare and hardly ever encountered.
7577 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7583 * Ensure that the session ID context of an SSL is updated when its
7584 SSL_CTX is updated via SSL_set_SSL_CTX.
7586 The session ID context is typically set from the parent SSL_CTX,
7587 and can vary with the CTX.
7591 * Fix various certificate fingerprint issues.
7593 By using non-DER or invalid encodings outside the signed portion of a
7594 certificate the fingerprint can be changed without breaking the signature.
7595 Although no details of the signed portion of the certificate can be changed
7596 this can cause problems with some applications: e.g. those using the
7597 certificate fingerprint for blacklists.
7599 1. Reject signatures with non zero unused bits.
7601 If the BIT STRING containing the signature has non zero unused bits reject
7602 the signature. All current signature algorithms require zero unused bits.
7604 2. Check certificate algorithm consistency.
7606 Check the AlgorithmIdentifier inside TBS matches the one in the
7607 certificate signature. NB: this will result in signature failure
7608 errors for some broken certificates.
7610 Thanks to Konrad Kraszewski from Google for reporting this issue.
7612 3. Check DSA/ECDSA signatures use DER.
7614 Re-encode DSA/ECDSA signatures and compare with the original received
7615 signature. Return an error if there is a mismatch.
7617 This will reject various cases including garbage after signature
7618 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7619 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7620 (negative or with leading zeroes).
7622 Further analysis was conducted and fixes were developed by Stephen Henson
7623 of the OpenSSL core team.
7629 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7630 results on some platforms, including x86_64. This bug occurs at random
7631 with a very low probability, and is not known to be exploitable in any
7632 way, though its exact impact is difficult to determine. Thanks to Pieter
7633 Wuille (Blockstream) who reported this issue and also suggested an initial
7634 fix. Further analysis was conducted by the OpenSSL development team and
7635 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7636 the OpenSSL core team.
7641 * Do not resume sessions on the server if the negotiated protocol
7642 version does not match the session's version. Resuming with a different
7643 version, while not strictly forbidden by the RFC, is of questionable
7644 sanity and breaks all known clients.
7646 *David Benjamin, Emilia Käsper*
7648 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7649 early CCS messages during renegotiation. (Note that because
7650 renegotiation is encrypted, this early CCS was not exploitable.)
7654 * Tighten client-side session ticket handling during renegotiation:
7655 ensure that the client only accepts a session ticket if the server sends
7656 the extension anew in the ServerHello. Previously, a TLS client would
7657 reuse the old extension state and thus accept a session ticket if one was
7658 announced in the initial ServerHello.
7660 Similarly, ensure that the client requires a session ticket if one
7661 was advertised in the ServerHello. Previously, a TLS client would
7662 ignore a missing NewSessionTicket message.
7666 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7670 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7671 sends a carefully crafted handshake message, to cause OpenSSL to fail
7672 to free up to 64k of memory causing a memory leak. This could be
7673 exploited in a Denial Of Service attack. This issue affects OpenSSL
7674 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7675 whether SRTP is used or configured. Implementations of OpenSSL that
7676 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7678 The fix was developed by the OpenSSL team.
7683 * Session Ticket Memory Leak.
7685 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7686 integrity of that ticket is first verified. In the event of a session
7687 ticket integrity check failing, OpenSSL will fail to free memory
7688 causing a memory leak. By sending a large number of invalid session
7689 tickets an attacker could exploit this issue in a Denial Of Service
7695 * Build option no-ssl3 is incomplete.
7697 When OpenSSL is configured with "no-ssl3" as a build option, servers
7698 could accept and complete an SSL 3.0 handshake, and clients could be
7699 configured to send them.
7702 *Akamai and the OpenSSL team*
7704 * Add support for TLS_FALLBACK_SCSV.
7705 Client applications doing fallback retries should call
7706 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7709 *Adam Langley, Bodo Moeller*
7711 * Add additional DigestInfo checks.
7713 Re-encode DigestInto in DER and check against the original when
7714 verifying RSA signature: this will reject any improperly encoded
7715 DigestInfo structures.
7717 Note: this is a precautionary measure and no attacks are currently known.
7721 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7723 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7724 SRP code can be overrun an internal buffer. Add sanity check that
7725 g, A, B < N to SRP code.
7727 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7728 Group for discovering this issue.
7733 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7734 TLS 1.0 instead of higher protocol versions when the ClientHello message
7735 is badly fragmented. This allows a man-in-the-middle attacker to force a
7736 downgrade to TLS 1.0 even if both the server and the client support a
7737 higher protocol version, by modifying the client's TLS records.
7739 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7740 researching this issue.
7745 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7746 to a denial of service attack. A malicious server can crash the client
7747 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7748 ciphersuite and sending carefully crafted handshake messages.
7750 Thanks to Felix Gröbert (Google) for discovering and researching this
7756 * By sending carefully crafted DTLS packets an attacker could cause openssl
7757 to leak memory. This can be exploited through a Denial of Service attack.
7758 Thanks to Adam Langley for discovering and researching this issue.
7763 * An attacker can force openssl to consume large amounts of memory whilst
7764 processing DTLS handshake messages. This can be exploited through a
7765 Denial of Service attack.
7766 Thanks to Adam Langley for discovering and researching this issue.
7771 * An attacker can force an error condition which causes openssl to crash
7772 whilst processing DTLS packets due to memory being freed twice. This
7773 can be exploited through a Denial of Service attack.
7774 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7780 * If a multithreaded client connects to a malicious server using a resumed
7781 session and the server sends an ec point format extension it could write
7782 up to 255 bytes to freed memory.
7784 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7790 * A malicious server can crash an OpenSSL client with a null pointer
7791 dereference (read) by specifying an SRP ciphersuite even though it was not
7792 properly negotiated with the client. This can be exploited through a
7793 Denial of Service attack.
7795 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7796 discovering and researching this issue.
7801 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7802 X509_name_oneline, X509_name_print_ex et al. to leak some information
7803 from the stack. Applications may be affected if they echo pretty printing
7804 output to the attacker.
7806 Thanks to Ivan Fratric (Google) for discovering this issue.
7809 *Emilia Käsper, and Steve Henson*
7811 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7812 for corner cases. (Certain input points at infinity could lead to
7813 bogus results, with non-infinity inputs mapped to infinity too.)
7817 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7819 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7820 handshake can force the use of weak keying material in OpenSSL
7821 SSL/TLS clients and servers.
7823 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7824 researching this issue. ([CVE-2014-0224])
7826 *KIKUCHI Masashi, Steve Henson*
7828 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7829 OpenSSL DTLS client the code can be made to recurse eventually crashing
7832 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7835 *Imre Rad, Steve Henson*
7837 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7838 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7839 client or server. This is potentially exploitable to run arbitrary
7840 code on a vulnerable client or server.
7842 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7844 *Jüri Aedla, Steve Henson*
7846 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7847 are subject to a denial of service attack.
7849 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7850 this issue. ([CVE-2014-3470])
7852 *Felix Gröbert, Ivan Fratric, Steve Henson*
7854 * Harmonize version and its documentation. -f flag is used to display
7857 *mancha <mancha1@zoho.com>*
7859 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7860 in i2d_ECPrivateKey.
7862 *mancha <mancha1@zoho.com>*
7864 * Fix some double frees. These are not thought to be exploitable.
7866 *mancha <mancha1@zoho.com>*
7868 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7870 * A missing bounds check in the handling of the TLS heartbeat extension
7871 can be used to reveal up to 64k of memory to a connected client or
7874 Thanks for Neel Mehta of Google Security for discovering this bug and to
7875 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7876 preparing the fix ([CVE-2014-0160])
7878 *Adam Langley, Bodo Moeller*
7880 * Fix for the attack described in the paper "Recovering OpenSSL
7881 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7882 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7883 <http://eprint.iacr.org/2014/140>
7885 Thanks to Yuval Yarom and Naomi Benger for discovering this
7886 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7888 *Yuval Yarom and Naomi Benger*
7890 * TLS pad extension: draft-agl-tls-padding-03
7892 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7893 TLS client Hello record length value would otherwise be > 255 and
7894 less that 512 pad with a dummy extension containing zeroes so it
7895 is at least 512 bytes long.
7897 *Adam Langley, Steve Henson*
7899 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7901 * Fix for TLS record tampering bug. A carefully crafted invalid
7902 handshake could crash OpenSSL with a NULL pointer exception.
7903 Thanks to Anton Johansson for reporting this issues.
7906 * Keep original DTLS digest and encryption contexts in retransmission
7907 structures so we can use the previous session parameters if they need
7908 to be resent. ([CVE-2013-6450])
7912 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7913 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7914 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7915 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7916 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7917 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7919 *Rob Stradling, Adam Langley*
7921 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7923 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7924 supporting platforms or when small records were transferred.
7926 *Andy Polyakov, Steve Henson*
7928 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7930 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7932 This addresses the flaw in CBC record processing discovered by
7933 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7934 at: <http://www.isg.rhul.ac.uk/tls/>
7936 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7937 Security Group at Royal Holloway, University of London
7938 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7939 Emilia Käsper for the initial patch.
7942 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7944 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7945 ciphersuites which can be exploited in a denial of service attack.
7946 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7947 and detecting this bug and to Wolfgang Ettlinger
7948 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7953 * Return an error when checking OCSP signatures when key is NULL.
7954 This fixes a DoS attack. ([CVE-2013-0166])
7958 * Make openssl verify return errors.
7960 *Chris Palmer <palmer@google.com> and Ben Laurie*
7962 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7963 the right response is stapled. Also change SSL_get_certificate()
7964 so it returns the certificate actually sent.
7965 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7967 *Rob Stradling <rob.stradling@comodo.com>*
7969 * Fix possible deadlock when decoding public keys.
7973 * Don't use TLS 1.0 record version number in initial client hello
7978 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7980 * Sanity check record length before skipping explicit IV in TLS
7981 1.2, 1.1 and DTLS to fix DoS attack.
7983 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7984 fuzzing as a service testing platform.
7989 * Initialise tkeylen properly when encrypting CMS messages.
7990 Thanks to Solar Designer of Openwall for reporting this issue.
7994 * In FIPS mode don't try to use composite ciphers as they are not
7999 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8001 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8002 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8003 mean any application compiled against OpenSSL 1.0.0 headers setting
8004 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8005 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8006 0x10000000L Any application which was previously compiled against
8007 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8008 will need to be recompiled as a result. Letting be results in
8009 inability to disable specifically TLS 1.1 and in client context,
8010 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8014 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8015 disable just protocol X, but all protocols above X *if* there are
8016 protocols *below* X still enabled. In more practical terms it means
8017 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8018 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8019 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8024 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8026 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8027 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8028 in CRYPTO_realloc_clean.
8030 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8031 issue and to Adam Langley <agl@chromium.org> for fixing it.
8034 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8036 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8040 * Workarounds for some broken servers that "hang" if a client hello
8041 record length exceeds 255 bytes.
8043 1. Do not use record version number > TLS 1.0 in initial client
8044 hello: some (but not all) hanging servers will now work.
8045 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8046 the number of ciphers sent in the client hello. This should be
8047 set to an even number, such as 50, for example by passing:
8048 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8049 Most broken servers should now work.
8050 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8051 TLS 1.2 client support entirely.
8055 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8059 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8061 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8062 STRING form instead of a DigestInfo.
8066 * The format used for MDC2 RSA signatures is inconsistent between EVP
8067 and the RSA_sign/RSA_verify functions. This was made more apparent when
8068 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8069 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8070 the correct format in RSA_verify so both forms transparently work.
8074 * Some servers which support TLS 1.0 can choke if we initially indicate
8075 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8076 encrypted premaster secret. As a workaround use the maximum permitted
8077 client version in client hello, this should keep such servers happy
8078 and still work with previous versions of OpenSSL.
8082 * Add support for TLS/DTLS heartbeats.
8084 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8086 * Add support for SCTP.
8088 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8090 * Improved PRNG seeding for VOS.
8092 *Paul Green <Paul.Green@stratus.com>*
8094 * Extensive assembler packs updates, most notably:
8096 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8097 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8098 - x86_64: bit-sliced AES implementation;
8099 - ARM: NEON support, contemporary platforms optimizations;
8100 - s390x: z196 support;
8101 - `*`: GHASH and GF(2^m) multiplication implementations;
8105 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8106 (removal of unnecessary code)
8108 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8110 * Add TLS key material exporter from RFC 5705.
8114 * Add DTLS-SRTP negotiation from RFC 5764.
8118 * Add Next Protocol Negotiation,
8119 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8120 disabled with a no-npn flag to config or Configure. Code donated
8123 *Adam Langley <agl@google.com> and Ben Laurie*
8125 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8126 NIST-P256, NIST-P521, with constant-time single point multiplication on
8127 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8128 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8129 Code made available under Apache License version 2.0.
8131 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8132 line to include this in your build of OpenSSL, and run "make depend" (or
8133 "make update"). This enables the following EC_METHODs:
8135 EC_GFp_nistp224_method()
8136 EC_GFp_nistp256_method()
8137 EC_GFp_nistp521_method()
8139 EC_GROUP_new_by_curve_name() will automatically use these (while
8140 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8143 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8145 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8146 all platforms. Move ssize_t definition from e_os.h to the public
8147 header file e_os2.h as it now appears in public header file cms.h
8151 * New -sigopt option to the ca, req and x509 utilities. Additional
8152 signature parameters can be passed using this option and in
8157 * Add RSA PSS signing function. This will generate and set the
8158 appropriate AlgorithmIdentifiers for PSS based on those in the
8159 corresponding EVP_MD_CTX structure. No application support yet.
8163 * Support for companion algorithm specific ASN1 signing routines.
8164 New function ASN1_item_sign_ctx() signs a pre-initialised
8165 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8166 the appropriate parameters.
8170 * Add new algorithm specific ASN1 verification initialisation function
8171 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8172 handling will be the same no matter what EVP_PKEY_METHOD is used.
8173 Add a PSS handler to support verification of PSS signatures: checked
8174 against a number of sample certificates.
8178 * Add signature printing for PSS. Add PSS OIDs.
8180 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8182 * Add algorithm specific signature printing. An individual ASN1 method
8183 can now print out signatures instead of the standard hex dump.
8185 More complex signatures (e.g. PSS) can print out more meaningful
8186 information. Include DSA version that prints out the signature
8191 * Password based recipient info support for CMS library: implementing
8196 * Split password based encryption into PBES2 and PBKDF2 functions. This
8197 neatly separates the code into cipher and PBE sections and is required
8198 for some algorithms that split PBES2 into separate pieces (such as
8199 password based CMS).
8203 * Session-handling fixes:
8204 - Fix handling of connections that are resuming with a session ID,
8205 but also support Session Tickets.
8206 - Fix a bug that suppressed issuing of a new ticket if the client
8207 presented a ticket with an expired session.
8208 - Try to set the ticket lifetime hint to something reasonable.
8209 - Make tickets shorter by excluding irrelevant information.
8210 - On the client side, don't ignore renewed tickets.
8212 *Adam Langley, Bodo Moeller (Google)*
8214 * Fix PSK session representation.
8218 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8220 This work was sponsored by Intel.
8224 * Add GCM support to TLS library. Some custom code is needed to split
8225 the IV between the fixed (from PRF) and explicit (from TLS record)
8226 portions. This adds all GCM ciphersuites supported by RFC5288 and
8227 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8228 add a special AESGCM string for GCM only.
8232 * Expand range of ctrls for AES GCM. Permit setting invocation
8233 field on decrypt and retrieval of invocation field only on encrypt.
8237 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8238 As required by RFC5289 these ciphersuites cannot be used if for
8239 versions of TLS earlier than 1.2.
8243 * For FIPS capable OpenSSL interpret a NULL default public key method
8244 as unset and return the appropriate default but do *not* set the default.
8245 This means we can return the appropriate method in applications that
8246 switch between FIPS and non-FIPS modes.
8250 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8251 ENGINE is used then we cannot handle that in the FIPS module so we
8252 keep original code iff non-FIPS operations are allowed.
8256 * Add -attime option to openssl utilities.
8258 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8260 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8264 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8265 FIPS EC methods unconditionally for now.
8269 * New build option no-ec2m to disable characteristic 2 code.
8273 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8274 all cases can be covered as some introduce binary incompatibilities.
8278 * Redirect RSA operations to FIPS module including keygen,
8279 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8283 * Add similar low-level API blocking to ciphers.
8287 * low-level digest APIs are not approved in FIPS mode: any attempt
8288 to use these will cause a fatal error. Applications that *really* want
8289 to use them can use the `private_*` version instead.
8293 * Redirect cipher operations to FIPS module for FIPS builds.
8297 * Redirect digest operations to FIPS module for FIPS builds.
8301 * Update build system to add "fips" flag which will link in fipscanister.o
8302 for static and shared library builds embedding a signature if needed.
8306 * Output TLS supported curves in preference order instead of numerical
8307 order. This is currently hardcoded for the highest order curves first.
8308 This should be configurable so applications can judge speed vs strength.
8312 * Add TLS v1.2 server support for client authentication.
8316 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8321 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8322 FIPS modules versions.
8326 * Add TLS v1.2 client side support for client authentication. Keep cache
8327 of handshake records longer as we don't know the hash algorithm to use
8328 until after the certificate request message is received.
8332 * Initial TLS v1.2 client support. Add a default signature algorithms
8333 extension including all the algorithms we support. Parse new signature
8334 format in client key exchange. Relax some ECC signing restrictions for
8335 TLS v1.2 as indicated in RFC5246.
8339 * Add server support for TLS v1.2 signature algorithms extension. Switch
8340 to new signature format when needed using client digest preference.
8341 All server ciphersuites should now work correctly in TLS v1.2. No client
8342 support yet and no support for client certificates.
8346 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8347 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8348 ciphersuites. At present only RSA key exchange ciphersuites work with
8349 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8350 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8351 and version checking.
8355 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8356 with this defined it will not be affected by any changes to ssl internal
8357 structures. Add several utility functions to allow openssl application
8358 to work with OPENSSL_NO_SSL_INTERN defined.
8362 * A long standing patch to add support for SRP from EdelWeb (Peter
8363 Sylvester and Christophe Renou) was integrated.
8364 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8365 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8368 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8372 * Permit abbreviated handshakes when renegotiating using the function
8373 SSL_renegotiate_abbreviated().
8375 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8377 * Add call to ENGINE_register_all_complete() to
8378 ENGINE_load_builtin_engines(), so some implementations get used
8379 automatically instead of needing explicit application support.
8383 * Add support for TLS key exporter as described in RFC5705.
8385 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8387 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8388 a few changes are required:
8390 Add SSL_OP_NO_TLSv1_1 flag.
8391 Add TLSv1_1 methods.
8392 Update version checking logic to handle version 1.1.
8393 Add explicit IV handling (ported from DTLS code).
8394 Add command line options to s_client/s_server.
8401 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8403 * X509_ATTRIBUTE memory leak
8405 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8406 memory. This structure is used by the PKCS#7 and CMS routines so any
8407 application which reads PKCS#7 or CMS data from untrusted sources is
8408 affected. SSL/TLS is not affected.
8410 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8416 * Race condition handling PSK identify hint
8418 If PSK identity hints are received by a multi-threaded client then
8419 the values are wrongly updated in the parent SSL_CTX structure. This can
8420 result in a race condition potentially leading to a double free of the
8426 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8428 * Malformed ECParameters causes infinite loop
8430 When processing an ECParameters structure OpenSSL enters an infinite loop
8431 if the curve specified is over a specially malformed binary polynomial
8434 This can be used to perform denial of service against any
8435 system which processes public keys, certificate requests or
8436 certificates. This includes TLS clients and TLS servers with
8437 client authentication enabled.
8439 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8444 * Exploitable out-of-bounds read in X509_cmp_time
8446 X509_cmp_time does not properly check the length of the ASN1_TIME
8447 string and can read a few bytes out of bounds. In addition,
8448 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8451 An attacker can use this to craft malformed certificates and CRLs of
8452 various sizes and potentially cause a segmentation fault, resulting in
8453 a DoS on applications that verify certificates or CRLs. TLS clients
8454 that verify CRLs are affected. TLS clients and servers with client
8455 authentication enabled may be affected if they use custom verification
8458 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8459 independently by Hanno Böck.
8464 * PKCS7 crash with missing EnvelopedContent
8466 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8467 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8468 with missing content and trigger a NULL pointer dereference on parsing.
8470 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8471 structures from untrusted sources are affected. OpenSSL clients and
8472 servers are not affected.
8474 This issue was reported to OpenSSL by Michal Zalewski (Google).
8479 * CMS verify infinite loop with unknown hash function
8481 When verifying a signedData message the CMS code can enter an infinite loop
8482 if presented with an unknown hash function OID. This can be used to perform
8483 denial of service against any system which verifies signedData messages using
8485 This issue was reported to OpenSSL by Johannes Bauer.
8490 * Race condition handling NewSessionTicket
8492 If a NewSessionTicket is received by a multi-threaded client when attempting to
8493 reuse a previous ticket then a race condition can occur potentially leading to
8494 a double free of the ticket data.
8499 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8501 * Segmentation fault in ASN1_TYPE_cmp fix
8503 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8504 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8505 certificate signature algorithm consistency this can be used to crash any
8506 certificate verification operation and exploited in a DoS attack. Any
8507 application which performs certificate verification is vulnerable including
8508 OpenSSL clients and servers which enable client authentication.
8513 * ASN.1 structure reuse memory corruption fix
8515 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8516 memory corruption via an invalid write. Such reuse is and has been
8517 strongly discouraged and is believed to be rare.
8519 Applications that parse structures containing CHOICE or ANY DEFINED BY
8520 components may be affected. Certificate parsing (d2i_X509 and related
8521 functions) are however not affected. OpenSSL clients and servers are
8527 * PKCS7 NULL pointer dereferences fix
8529 The PKCS#7 parsing code does not handle missing outer ContentInfo
8530 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8531 missing content and trigger a NULL pointer dereference on parsing.
8533 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8534 otherwise parse PKCS#7 structures from untrusted sources are
8535 affected. OpenSSL clients and servers are not affected.
8537 This issue was reported to OpenSSL by Michal Zalewski (Google).
8542 * DoS via reachable assert in SSLv2 servers fix
8544 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8545 servers that both support SSLv2 and enable export cipher suites by sending
8546 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8548 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8549 (OpenSSL development team).
8554 * Use After Free following d2i_ECPrivatekey error fix
8556 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8557 could cause a use after free condition. This, in turn, could cause a double
8558 free in several private key parsing functions (such as d2i_PrivateKey
8559 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8560 for applications that receive EC private keys from untrusted
8561 sources. This scenario is considered rare.
8563 This issue was discovered by the BoringSSL project and fixed in their
8569 * X509_to_X509_REQ NULL pointer deref fix
8571 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8572 the certificate key is invalid. This function is rarely used in practice.
8574 This issue was discovered by Brian Carpenter.
8579 * Removed the export ciphers from the DEFAULT ciphers
8583 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8585 * Build fixes for the Windows and OpenVMS platforms
8587 *Matt Caswell and Richard Levitte*
8589 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8591 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8592 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8593 dereference. This could lead to a Denial Of Service attack. Thanks to
8594 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8599 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8600 dtls1_buffer_record function under certain conditions. In particular this
8601 could occur if an attacker sent repeated DTLS records with the same
8602 sequence number but for the next epoch. The memory leak could be exploited
8603 by an attacker in a Denial of Service attack through memory exhaustion.
8604 Thanks to Chris Mueller for reporting this issue.
8609 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8610 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8611 method would be set to NULL which could later result in a NULL pointer
8612 dereference. Thanks to Frank Schmirler for reporting this issue.
8617 * Abort handshake if server key exchange message is omitted for ephemeral
8620 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8621 reporting this issue.
8626 * Remove non-export ephemeral RSA code on client and server. This code
8627 violated the TLS standard by allowing the use of temporary RSA keys in
8628 non-export ciphersuites and could be used by a server to effectively
8629 downgrade the RSA key length used to a value smaller than the server
8630 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8631 INRIA or reporting this issue.
8636 * Fixed issue where DH client certificates are accepted without verification.
8637 An OpenSSL server will accept a DH certificate for client authentication
8638 without the certificate verify message. This effectively allows a client to
8639 authenticate without the use of a private key. This only affects servers
8640 which trust a client certificate authority which issues certificates
8641 containing DH keys: these are extremely rare and hardly ever encountered.
8642 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8648 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8649 results on some platforms, including x86_64. This bug occurs at random
8650 with a very low probability, and is not known to be exploitable in any
8651 way, though its exact impact is difficult to determine. Thanks to Pieter
8652 Wuille (Blockstream) who reported this issue and also suggested an initial
8653 fix. Further analysis was conducted by the OpenSSL development team and
8654 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8655 the OpenSSL core team.
8660 * Fix various certificate fingerprint issues.
8662 By using non-DER or invalid encodings outside the signed portion of a
8663 certificate the fingerprint can be changed without breaking the signature.
8664 Although no details of the signed portion of the certificate can be changed
8665 this can cause problems with some applications: e.g. those using the
8666 certificate fingerprint for blacklists.
8668 1. Reject signatures with non zero unused bits.
8670 If the BIT STRING containing the signature has non zero unused bits reject
8671 the signature. All current signature algorithms require zero unused bits.
8673 2. Check certificate algorithm consistency.
8675 Check the AlgorithmIdentifier inside TBS matches the one in the
8676 certificate signature. NB: this will result in signature failure
8677 errors for some broken certificates.
8679 Thanks to Konrad Kraszewski from Google for reporting this issue.
8681 3. Check DSA/ECDSA signatures use DER.
8683 Reencode DSA/ECDSA signatures and compare with the original received
8684 signature. Return an error if there is a mismatch.
8686 This will reject various cases including garbage after signature
8687 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8688 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8689 (negative or with leading zeroes).
8691 Further analysis was conducted and fixes were developed by Stephen Henson
8692 of the OpenSSL core team.
8698 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8700 * Session Ticket Memory Leak.
8702 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8703 integrity of that ticket is first verified. In the event of a session
8704 ticket integrity check failing, OpenSSL will fail to free memory
8705 causing a memory leak. By sending a large number of invalid session
8706 tickets an attacker could exploit this issue in a Denial Of Service
8712 * Build option no-ssl3 is incomplete.
8714 When OpenSSL is configured with "no-ssl3" as a build option, servers
8715 could accept and complete an SSL 3.0 handshake, and clients could be
8716 configured to send them.
8719 *Akamai and the OpenSSL team*
8721 * Add support for TLS_FALLBACK_SCSV.
8722 Client applications doing fallback retries should call
8723 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8726 *Adam Langley, Bodo Moeller*
8728 * Add additional DigestInfo checks.
8730 Reencode DigestInto in DER and check against the original when
8731 verifying RSA signature: this will reject any improperly encoded
8732 DigestInfo structures.
8734 Note: this is a precautionary measure and no attacks are currently known.
8738 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8740 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8741 to a denial of service attack. A malicious server can crash the client
8742 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8743 ciphersuite and sending carefully crafted handshake messages.
8745 Thanks to Felix Gröbert (Google) for discovering and researching this
8751 * By sending carefully crafted DTLS packets an attacker could cause openssl
8752 to leak memory. This can be exploited through a Denial of Service attack.
8753 Thanks to Adam Langley for discovering and researching this issue.
8758 * An attacker can force openssl to consume large amounts of memory whilst
8759 processing DTLS handshake messages. This can be exploited through a
8760 Denial of Service attack.
8761 Thanks to Adam Langley for discovering and researching this issue.
8766 * An attacker can force an error condition which causes openssl to crash
8767 whilst processing DTLS packets due to memory being freed twice. This
8768 can be exploited through a Denial of Service attack.
8769 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8775 * If a multithreaded client connects to a malicious server using a resumed
8776 session and the server sends an ec point format extension it could write
8777 up to 255 bytes to freed memory.
8779 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8785 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8786 X509_name_oneline, X509_name_print_ex et al. to leak some information
8787 from the stack. Applications may be affected if they echo pretty printing
8788 output to the attacker.
8790 Thanks to Ivan Fratric (Google) for discovering this issue.
8793 *Emilia Käsper, and Steve Henson*
8795 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8796 for corner cases. (Certain input points at infinity could lead to
8797 bogus results, with non-infinity inputs mapped to infinity too.)
8801 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8803 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8804 handshake can force the use of weak keying material in OpenSSL
8805 SSL/TLS clients and servers.
8807 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8808 researching this issue. ([CVE-2014-0224])
8810 *KIKUCHI Masashi, Steve Henson*
8812 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8813 OpenSSL DTLS client the code can be made to recurse eventually crashing
8816 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8819 *Imre Rad, Steve Henson*
8821 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8822 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8823 client or server. This is potentially exploitable to run arbitrary
8824 code on a vulnerable client or server.
8826 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8828 *Jüri Aedla, Steve Henson*
8830 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8831 are subject to a denial of service attack.
8833 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8834 this issue. ([CVE-2014-3470])
8836 *Felix Gröbert, Ivan Fratric, Steve Henson*
8838 * Harmonize version and its documentation. -f flag is used to display
8841 *mancha <mancha1@zoho.com>*
8843 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8844 in i2d_ECPrivateKey.
8846 *mancha <mancha1@zoho.com>*
8848 * Fix some double frees. These are not thought to be exploitable.
8850 *mancha <mancha1@zoho.com>*
8852 * Fix for the attack described in the paper "Recovering OpenSSL
8853 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8854 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8855 <http://eprint.iacr.org/2014/140>
8857 Thanks to Yuval Yarom and Naomi Benger for discovering this
8858 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8860 *Yuval Yarom and Naomi Benger*
8862 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8864 * Keep original DTLS digest and encryption contexts in retransmission
8865 structures so we can use the previous session parameters if they need
8866 to be resent. ([CVE-2013-6450])
8870 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8871 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8872 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8873 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8874 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8875 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8877 *Rob Stradling, Adam Langley*
8879 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8881 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8883 This addresses the flaw in CBC record processing discovered by
8884 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8885 at: <http://www.isg.rhul.ac.uk/tls/>
8887 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8888 Security Group at Royal Holloway, University of London
8889 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8890 Emilia Käsper for the initial patch.
8893 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8895 * Return an error when checking OCSP signatures when key is NULL.
8896 This fixes a DoS attack. ([CVE-2013-0166])
8900 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8901 the right response is stapled. Also change SSL_get_certificate()
8902 so it returns the certificate actually sent.
8903 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8904 (This is a backport)
8906 *Rob Stradling <rob.stradling@comodo.com>*
8908 * Fix possible deadlock when decoding public keys.
8912 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8914 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8917 * Sanity check record length before skipping explicit IV in DTLS
8920 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8921 fuzzing as a service testing platform.
8926 * Initialise tkeylen properly when encrypting CMS messages.
8927 Thanks to Solar Designer of Openwall for reporting this issue.
8931 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8933 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8934 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8935 in CRYPTO_realloc_clean.
8937 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8938 issue and to Adam Langley <agl@chromium.org> for fixing it.
8941 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8943 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8945 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8946 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8947 content decryption and always return the same error. Note: this attack
8948 needs on average 2^20 messages so it only affects automated senders. The
8949 old behaviour can be re-enabled in the CMS code by setting the
8950 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8951 an MMA defence is not necessary.
8952 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8953 this issue. ([CVE-2012-0884])
8957 * Fix CVE-2011-4619: make sure we really are receiving a
8958 client hello before rejecting multiple SGC restarts. Thanks to
8959 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8963 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8965 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8966 Thanks to Antonio Martin, Enterprise Secure Access Research and
8967 Development, Cisco Systems, Inc. for discovering this bug and
8968 preparing a fix. ([CVE-2012-0050])
8972 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8974 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8975 of the Vaudenay padding oracle attack on CBC mode encryption
8976 which enables an efficient plaintext recovery attack against
8977 the OpenSSL implementation of DTLS. Their attack exploits timing
8978 differences arising during decryption processing. A research
8979 paper describing this attack can be found at:
8980 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8981 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8982 Security Group at Royal Holloway, University of London
8983 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8984 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8985 for preparing the fix. ([CVE-2011-4108])
8987 *Robin Seggelmann, Michael Tuexen*
8989 * Clear bytes used for block padding of SSL 3.0 records.
8992 *Adam Langley (Google)*
8994 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8995 Kadianakis <desnacked@gmail.com> for discovering this issue and
8996 Adam Langley for preparing the fix. ([CVE-2011-4619])
8998 *Adam Langley (Google)*
9000 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9002 *Andrey Kulikov <amdeich@gmail.com>*
9004 * Prevent malformed RFC3779 data triggering an assertion failure.
9005 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9006 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9008 *Rob Austein <sra@hactrn.net>*
9010 * Improved PRNG seeding for VOS.
9012 *Paul Green <Paul.Green@stratus.com>*
9014 * Fix ssl_ciph.c set-up race.
9016 *Adam Langley (Google)*
9018 * Fix spurious failures in ecdsatest.c.
9020 *Emilia Käsper (Google)*
9022 * Fix the BIO_f_buffer() implementation (which was mixing different
9023 interpretations of the `..._len` fields).
9025 *Adam Langley (Google)*
9027 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9028 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9029 threads won't reuse the same blinding coefficients.
9031 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9032 lock to call BN_BLINDING_invert_ex, and avoids one use of
9033 BN_BLINDING_update for each BN_BLINDING structure (previously,
9034 the last update always remained unused).
9036 *Emilia Käsper (Google)*
9038 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9040 *Bob Buckholz (Google)*
9042 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9044 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9045 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9047 *Kaspar Brand <ossl@velox.ch>*
9049 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9050 for multi-threaded use of ECDH. ([CVE-2011-3210])
9052 *Adam Langley (Google)*
9054 * Fix x509_name_ex_d2i memory leak on bad inputs.
9058 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9059 signature public key algorithm by using OID xref utilities instead.
9060 Before this you could only use some ECC ciphersuites with SHA1 only.
9064 * Add protection against ECDSA timing attacks as mentioned in the paper
9065 by Billy Bob Brumley and Nicola Tuveri, see:
9066 <http://eprint.iacr.org/2011/232.pdf>
9068 *Billy Bob Brumley and Nicola Tuveri*
9070 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9072 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9074 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9076 * Fix bug in string printing code: if *any* escaping is enabled we must
9077 escape the escape character (backslash) or the resulting string is
9082 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9084 * Disable code workaround for ancient and obsolete Netscape browsers
9085 and servers: an attacker can use it in a ciphersuite downgrade attack.
9086 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9090 * Fixed J-PAKE implementation error, originally discovered by
9091 Sebastien Martini, further info and confirmation from Stefan
9092 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9096 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9098 * Fix extension code to avoid race conditions which can result in a buffer
9099 overrun vulnerability: resumed sessions must not be modified as they can
9100 be shared by multiple threads. CVE-2010-3864
9104 * Fix WIN32 build system to correctly link an ENGINE directory into
9109 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9111 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9114 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9116 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9118 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9119 context. The operation can be customised via the ctrl mechanism in
9120 case ENGINEs want to include additional functionality.
9124 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9128 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9129 output hashes compatible with older versions of OpenSSL.
9131 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9133 * Fix compression algorithm handling: if resuming a session use the
9134 compression algorithm of the resumed session instead of determining
9135 it from client hello again. Don't allow server to change algorithm.
9139 * Add load_crls() function to commands tidying load_certs() too. Add option
9140 to verify utility to allow additional CRLs to be included.
9144 * Update OCSP request code to permit adding custom headers to the request:
9145 some responders need this.
9149 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9152 *Julia Lawall <julia@diku.dk>*
9154 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9155 needlessly dereferenced structures, used obsolete functions and
9156 didn't handle all updated verify codes correctly.
9160 * Disable MD2 in the default configuration.
9164 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9165 indicate the initial BIO being pushed or popped. This makes it possible
9166 to determine whether the BIO is the one explicitly called or as a result
9167 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9168 it handles reference counts correctly and doesn't zero out the I/O bio
9169 when it is not being explicitly popped. WARNING: applications which
9170 included workarounds for the old buggy behaviour will need to be modified
9171 or they could free up already freed BIOs.
9175 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9176 renaming to all platforms (within the 0.9.8 branch, this was
9177 done conditionally on Netware platforms to avoid a name clash).
9179 *Guenter <lists@gknw.net>*
9181 * Add ECDHE and PSK support to DTLS.
9183 *Michael Tuexen <tuexen@fh-muenster.de>*
9185 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9190 * Add "missing" function EVP_MD_flags() (without this the only way to
9191 retrieve a digest flags is by accessing the structure directly. Update
9192 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9193 or cipher is registered as in the "from" argument. Print out all
9194 registered digests in the dgst usage message instead of manually
9195 attempting to work them out.
9199 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9200 this allows the use of compression and extensions. Change default cipher
9201 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9202 by default unless an application cipher string requests it.
9206 * Alter match criteria in PKCS12_parse(). It used to try to use local
9207 key ids to find matching certificates and keys but some PKCS#12 files
9208 don't follow the (somewhat unwritten) rules and this strategy fails.
9209 Now just gather all certificates together and the first private key
9210 then look for the first certificate that matches the key.
9214 * Support use of registered digest and cipher names for dgst and cipher
9215 commands instead of having to add each one as a special case. So now
9222 openssl dgst -sha256 foo
9224 and this works for ENGINE based algorithms too.
9228 * Update Gost ENGINE to support parameter files.
9230 *Victor B. Wagner <vitus@cryptocom.ru>*
9232 * Support GeneralizedTime in ca utility.
9234 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9236 * Enhance the hash format used for certificate directory links. The new
9237 form uses the canonical encoding (meaning equivalent names will work
9238 even if they aren't identical) and uses SHA1 instead of MD5. This form
9239 is incompatible with the older format and as a result c_rehash should
9240 be used to rebuild symbolic links.
9244 * Make PKCS#8 the default write format for private keys, replacing the
9245 traditional format. This form is standardised, more secure and doesn't
9246 include an implicit MD5 dependency.
9250 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9251 committed to OpenSSL should pass this lot as a minimum.
9255 * Add session ticket override functionality for use by EAP-FAST.
9257 *Jouni Malinen <j@w1.fi>*
9259 * Modify HMAC functions to return a value. Since these can be implemented
9260 in an ENGINE errors can occur.
9264 * Type-checked OBJ_bsearch_ex.
9268 * Type-checked OBJ_bsearch. Also some constification necessitated
9269 by type-checking. Still to come: TXT_DB, bsearch(?),
9270 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9275 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9276 seconds to a tm structure directly, instead of going through OS
9277 specific date routines. This avoids any issues with OS routines such
9278 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9279 and X509_time_adj_ex() to cover the extended range. The existing
9280 X509_time_adj() is still usable and will no longer have any date issues.
9284 * Delta CRL support. New use deltas option which will attempt to locate
9285 and search any appropriate delta CRLs available.
9287 This work was sponsored by Google.
9291 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9292 code and add additional score elements. Validate alternate CRL paths
9293 as part of the CRL checking and indicate a new error "CRL path validation
9294 error" in this case. Applications wanting additional details can use
9295 the verify callback and check the new "parent" field. If this is not
9296 NULL CRL path validation is taking place. Existing applications won't
9297 see this because it requires extended CRL support which is off by
9300 This work was sponsored by Google.
9304 * Support for freshest CRL extension.
9306 This work was sponsored by Google.
9310 * Initial indirect CRL support. Currently only supported in the CRLs
9311 passed directly and not via lookup. Process certificate issuer
9312 CRL entry extension and lookup CRL entries by bother issuer name
9313 and serial number. Check and process CRL issuer entry in IDP extension.
9315 This work was sponsored by Google.
9319 * Add support for distinct certificate and CRL paths. The CRL issuer
9320 certificate is validated separately in this case. Only enabled if
9321 an extended CRL support flag is set: this flag will enable additional
9322 CRL functionality in future.
9324 This work was sponsored by Google.
9328 * Add support for policy mappings extension.
9330 This work was sponsored by Google.
9334 * Fixes to pathlength constraint, self issued certificate handling,
9335 policy processing to align with RFC3280 and PKITS tests.
9337 This work was sponsored by Google.
9341 * Support for name constraints certificate extension. DN, email, DNS
9342 and URI types are currently supported.
9344 This work was sponsored by Google.
9348 * To cater for systems that provide a pointer-based thread ID rather
9349 than numeric, deprecate the current numeric thread ID mechanism and
9350 replace it with a structure and associated callback type. This
9351 mechanism allows a numeric "hash" to be extracted from a thread ID in
9352 either case, and on platforms where pointers are larger than 'long',
9353 mixing is done to help ensure the numeric 'hash' is usable even if it
9354 can't be guaranteed unique. The default mechanism is to use "&errno"
9355 as a pointer-based thread ID to distinguish between threads.
9357 Applications that want to provide their own thread IDs should now use
9358 CRYPTO_THREADID_set_callback() to register a callback that will call
9359 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9361 Note that ERR_remove_state() is now deprecated, because it is tied
9362 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9363 to free the current thread's error state should be replaced by
9364 ERR_remove_thread_state(NULL).
9366 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9367 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9368 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9369 application was previously providing a numeric thread callback that
9370 was inappropriate for distinguishing threads, then uniqueness might
9371 have been obtained with &errno that happened immediately in the
9372 intermediate development versions of OpenSSL; this is no longer the
9373 case, the numeric thread callback will now override the automatic use
9376 *Geoff Thorpe, with help from Bodo Moeller*
9378 * Initial support for different CRL issuing certificates. This covers a
9379 simple case where the self issued certificates in the chain exist and
9380 the real CRL issuer is higher in the existing chain.
9382 This work was sponsored by Google.
9386 * Removed effectively defunct crypto/store from the build.
9390 * Revamp of STACK to provide stronger type-checking. Still to come:
9391 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9392 ASN1_STRING, CONF_VALUE.
9396 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9397 RAM on SSL connections. This option can save about 34k per idle SSL.
9401 * Revamp of LHASH to provide stronger type-checking. Still to come:
9402 STACK, TXT_DB, bsearch, qsort.
9406 * Initial support for Cryptographic Message Syntax (aka CMS) based
9407 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9408 support for data, signedData, compressedData, digestedData and
9409 encryptedData, envelopedData types included. Scripts to check against
9410 RFC4134 examples draft and interop and consistency checks of many
9411 content types and variants.
9415 * Add options to enc utility to support use of zlib compression BIO.
9419 * Extend mk1mf to support importing of options and assembly language
9420 files from Configure script, currently only included in VC-WIN32.
9421 The assembly language rules can now optionally generate the source
9422 files from the associated perl scripts.
9426 * Implement remaining functionality needed to support GOST ciphersuites.
9427 Interop testing has been performed using CryptoPro implementations.
9429 *Victor B. Wagner <vitus@cryptocom.ru>*
9431 * s390x assembler pack.
9435 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9440 * Implement Opaque PRF Input TLS extension as specified in
9441 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9442 official specification yet and no extension type assignment by
9443 IANA exists, this extension (for now) will have to be explicitly
9444 enabled when building OpenSSL by providing the extension number
9445 to use. For example, specify an option
9447 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9449 to the "config" or "Configure" script to enable the extension,
9450 assuming extension number 0x9527 (which is a completely arbitrary
9451 and unofficial assignment based on the MD5 hash of the Internet
9452 Draft). Note that by doing so, you potentially lose
9453 interoperability with other TLS implementations since these might
9454 be using the same extension number for other purposes.
9456 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9457 opaque PRF input value to use in the handshake. This will create
9458 an internal copy of the length-'len' string at 'src', and will
9459 return non-zero for success.
9461 To get more control and flexibility, provide a callback function
9464 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9465 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9469 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9472 Callback function 'cb' will be called in handshakes, and is
9473 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9474 Argument 'arg' is for application purposes (the value as given to
9475 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9476 be provided to the callback function). The callback function
9477 has to return non-zero to report success: usually 1 to use opaque
9478 PRF input just if possible, or 2 to enforce use of the opaque PRF
9479 input. In the latter case, the library will abort the handshake
9480 if opaque PRF input is not successfully negotiated.
9482 Arguments 'peerinput' and 'len' given to the callback function
9483 will always be NULL and 0 in the case of a client. A server will
9484 see the client's opaque PRF input through these variables if
9485 available (NULL and 0 otherwise). Note that if the server
9486 provides an opaque PRF input, the length must be the same as the
9487 length of the client's opaque PRF input.
9489 Note that the callback function will only be called when creating
9490 a new session (session resumption can resume whatever was
9491 previously negotiated), and will not be called in SSL 2.0
9492 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9493 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9494 for applications that need to enforce opaque PRF input.
9498 * Update ssl code to support digests other than SHA1+MD5 for handshake
9501 *Victor B. Wagner <vitus@cryptocom.ru>*
9503 * Add RFC4507 support to OpenSSL. This includes the corrections in
9504 RFC4507bis. The encrypted ticket format is an encrypted encoded
9505 SSL_SESSION structure, that way new session features are automatically
9508 If a client application caches session in an SSL_SESSION structure
9509 support is transparent because tickets are now stored in the encoded
9512 The SSL_CTX structure automatically generates keys for ticket
9513 protection in servers so again support should be possible
9514 with no application modification.
9516 If a client or server wishes to disable RFC4507 support then the option
9517 SSL_OP_NO_TICKET can be set.
9519 Add a TLS extension debugging callback to allow the contents of any client
9520 or server extensions to be examined.
9522 This work was sponsored by Google.
9526 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9527 OpenSSL should now compile cleanly on gcc 4.2
9529 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9531 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9532 support including streaming MAC support: this is required for GOST
9533 ciphersuite support.
9535 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9537 * Add option -stream to use PKCS#7 streaming in smime utility. New
9538 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9539 to output in BER and PEM format.
9543 * Experimental support for use of HMAC via EVP_PKEY interface. This
9544 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9545 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9546 ENGINE support for HMAC keys which are unextractable. New -mac and
9547 -macopt options to dgst utility.
9551 * New option -sigopt to dgst utility. Update dgst to use
9552 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9553 alternative signing parameters such as X9.31 or PSS in the dgst
9558 * Change ssl_cipher_apply_rule(), the internal function that does
9559 the work each time a ciphersuite string requests enabling
9560 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9561 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9562 the order of disabled ciphersuites such that those ciphersuites
9563 that most recently went from enabled to disabled not only stay
9564 in order with respect to each other, but also have higher priority
9565 than other disabled ciphersuites the next time ciphersuites are
9568 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9569 the same ciphersuites as with "HIGH" alone, but in a specific
9570 order where the PSK ciphersuites come first (since they are the
9571 most recently disabled ciphersuites when "HIGH" is parsed).
9573 Also, change ssl_create_cipher_list() (using this new
9574 functionality) such that between otherwise identical
9575 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9580 * Change ssl_create_cipher_list() so that it automatically
9581 arranges the ciphersuites in reasonable order before starting
9582 to process the rule string. Thus, the definition for "DEFAULT"
9583 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9584 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9585 This makes it much easier to arrive at a reasonable default order
9586 in applications for which anonymous ciphers are OK (meaning
9587 that you can't actually use DEFAULT).
9589 *Bodo Moeller; suggested by Victor Duchovni*
9591 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9592 processing) into multiple integers instead of setting
9593 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9594 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9595 (These masks as well as the individual bit definitions are hidden
9596 away into the non-exported interface ssl/ssl_locl.h, so this
9597 change to the definition of the SSL_CIPHER structure shouldn't
9598 affect applications.) This give us more bits for each of these
9599 categories, so there is no longer a need to coagulate AES128 and
9600 AES256 into a single algorithm bit, and to coagulate Camellia128
9601 and Camellia256 into a single algorithm bit, which has led to all
9604 Thus, among other things, the kludge introduced in 0.9.7m and
9605 0.9.8e for masking out AES256 independently of AES128 or masking
9606 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9608 With the change, we also introduce new ciphersuite aliases that
9609 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9614 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9615 Use the leftmost N bytes of the signature input if the input is
9616 larger than the prime q (with N being the size in bytes of q).
9620 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9621 it yet and it is largely untested.
9625 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9629 * Initial incomplete changes to avoid need for function casts in OpenSSL
9630 some compilers (gcc 4.2 and later) reject their use. Safestack is
9631 reimplemented. Update ASN1 to avoid use of legacy functions.
9635 * Win32/64 targets are linked with Winsock2.
9639 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9640 to external functions. This can be used to increase CRL handling
9641 efficiency especially when CRLs are very large by (for example) storing
9642 the CRL revoked certificates in a database.
9646 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9647 new CRLs added to a directory can be used. New command line option
9648 -verify_return_error to s_client and s_server. This causes real errors
9649 to be returned by the verify callback instead of carrying on no matter
9650 what. This reflects the way a "real world" verify callback would behave.
9654 * GOST engine, supporting several GOST algorithms and public key formats.
9655 Kindly donated by Cryptocom.
9659 * Partial support for Issuing Distribution Point CRL extension. CRLs
9660 partitioned by DP are handled but no indirect CRL or reason partitioning
9661 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9662 selected via a scoring technique which handles IDP and AKID in CRLs.
9666 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9667 will ultimately be used for all verify operations: this will remove the
9668 X509_STORE dependency on certificate verification and allow alternative
9669 lookup methods. X509_STORE based implementations of these two callbacks.
9673 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9674 Modify get_crl() to find a valid (unexpired) CRL if possible.
9678 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9679 this would be called X509_CRL_cmp() but that name is already used by
9680 a function that just compares CRL issuer names. Cache several CRL
9681 extensions in X509_CRL structure and cache CRLDP in X509.
9685 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9686 this maps equivalent X509_NAME structures into a consistent structure.
9687 Name comparison can then be performed rapidly using memcmp().
9691 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9696 * Allow digests to supply their own micalg string for S/MIME type using
9697 the ctrl EVP_MD_CTRL_MICALG.
9701 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9702 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9703 ctrl. It can then customise the structure before and/or after signing
9708 * New function OBJ_add_sigid() to allow application defined signature OIDs
9709 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9710 to free up any added signature OIDs.
9714 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9715 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9716 digest and cipher tables. New options added to openssl utility:
9717 list-message-digest-algorithms and list-cipher-algorithms.
9721 * Change the array representation of binary polynomials: the list
9722 of degrees of non-zero coefficients is now terminated with -1.
9723 Previously it was terminated with 0, which was also part of the
9724 value; thus, the array representation was not applicable to
9725 polynomials where t^0 has coefficient zero. This change makes
9726 the array representation useful in a more general context.
9730 * Various modifications and fixes to SSL/TLS cipher string
9731 handling. For ECC, the code now distinguishes between fixed ECDH
9732 with RSA certificates on the one hand and with ECDSA certificates
9733 on the other hand, since these are separate ciphersuites. The
9734 unused code for Fortezza ciphersuites has been removed.
9736 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9737 (not "ECDHE"). For consistency with the code for DH
9738 certificates, use of ECDH certificates is now considered ECDH
9739 authentication, not RSA or ECDSA authentication (the latter is
9740 merely the CA's signing algorithm and not actively used in the
9743 The temporary ciphersuite alias "ECCdraft" is no longer
9744 available, and ECC ciphersuites are no longer excluded from "ALL"
9745 and "DEFAULT". The following aliases now exist for RFC 4492
9746 ciphersuites, most of these by analogy with the DH case:
9748 kECDHr - ECDH cert, signed with RSA
9749 kECDHe - ECDH cert, signed with ECDSA
9750 kECDH - ECDH cert (signed with either RSA or ECDSA)
9751 kEECDH - ephemeral ECDH
9752 ECDH - ECDH cert or ephemeral ECDH
9758 AECDH - anonymous ECDH
9759 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9763 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9764 Use correct micalg parameters depending on digest(s) in signed message.
9768 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9769 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9773 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9774 an engine to register a method. Add ENGINE lookups for methods and
9775 functional reference processing.
9779 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9780 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9785 * New -resign option to smime utility. This adds one or more signers
9786 to an existing PKCS#7 signedData structure. Also -md option to use an
9787 alternative message digest algorithm for signing.
9791 * Tidy up PKCS#7 routines and add new functions to make it easier to
9792 create PKCS7 structures containing multiple signers. Update smime
9793 application to support multiple signers.
9797 * New -macalg option to pkcs12 utility to allow setting of an alternative
9802 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9803 Reorganize PBE internals to lookup from a static table using NIDs,
9804 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9805 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9806 PRF which will be automatically used with PBES2.
9810 * Replace the algorithm specific calls to generate keys in "req" with the
9815 * Update PKCS#7 enveloped data routines to use new API. This is now
9816 supported by any public key method supporting the encrypt operation. A
9817 ctrl is added to allow the public key algorithm to examine or modify
9818 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9823 * Add a ctrl to asn1 method to allow a public key algorithm to express
9824 a default digest type to use. In most cases this will be SHA1 but some
9825 algorithms (such as GOST) need to specify an alternative digest. The
9826 return value indicates how strong the preference is 1 means optional and
9827 2 is mandatory (that is it is the only supported type). Modify
9828 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9829 use the default md. Update openssl utilities to use the default digest
9830 type for signing if it is not explicitly indicated.
9834 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9835 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9836 signing method from the key type. This effectively removes the link
9837 between digests and public key types.
9841 * Add an OID cross reference table and utility functions. Its purpose is to
9842 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9843 rsaEncryption. This will allow some of the algorithm specific hackery
9844 needed to use the correct OID to be removed.
9848 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9849 structures for PKCS7_sign(). They are now set up by the relevant public
9854 * Add provisional EC pkey method with support for ECDSA and ECDH.
9858 * Add support for key derivation (agreement) in the API, DH method and
9863 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9864 public and private key formats. As a side effect these add additional
9865 command line functionality not previously available: DSA signatures can be
9866 generated and verified using pkeyutl and DH key support and generation in
9873 *Oliver Tappe <zooey@hirschkaefer.de>*
9875 * New make target "install_html_docs" installs HTML renditions of the
9878 *Oliver Tappe <zooey@hirschkaefer.de>*
9880 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9881 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9882 support key and parameter generation and add initial key generation
9883 functionality for RSA.
9887 * Add functions for main EVP_PKEY_method operations. The undocumented
9888 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9889 `EVP_PKEY_{encrypt,decrypt}_old`.
9893 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9894 key API, doesn't do much yet.
9898 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9899 public key algorithms. New option to openssl utility:
9900 "list-public-key-algorithms" to print out info.
9904 * Implement the Supported Elliptic Curves Extension for
9905 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9909 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9910 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9914 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9915 utilities such as rsa, dsa, dsaparam etc except they process any key
9920 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9921 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9922 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9927 * Initial support for pluggable public key ASN1.
9928 De-spaghettify the public key ASN1 handling. Move public and private
9929 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9930 algorithm specific handling to a single module within the relevant
9931 algorithm directory. Add functions to allow (near) opaque processing
9932 of public and private key structures.
9936 * Implement the Supported Point Formats Extension for
9937 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9941 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9942 for the psk identity [hint] and the psk callback functions to the
9943 SSL_SESSION, SSL and SSL_CTX structure.
9946 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9950 SSL_CTX_use_psk_identity_hint
9951 SSL_get_psk_identity_hint
9952 SSL_get_psk_identity
9953 SSL_use_psk_identity_hint
9955 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9957 * Add RFC 3161 compliant time stamp request creation, response generation
9958 and response verification functionality.
9960 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9962 * Add initial support for TLS extensions, specifically for the server_name
9963 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9964 have new members for a hostname. The SSL data structure has an
9965 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9966 stored in that context to allow for session resumption, even after the
9967 SSL has been switched to a new SSL_CTX in reaction to a client's
9968 server_name extension.
9970 New functions (subject to change):
9972 SSL_get_servername()
9973 SSL_get_servername_type()
9976 New CTRL codes and macros (subject to change):
9978 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9979 - SSL_CTX_set_tlsext_servername_callback()
9980 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9981 - SSL_CTX_set_tlsext_servername_arg()
9982 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9984 openssl s_client has a new '-servername ...' option.
9986 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9987 '-key2 ...', '-servername_fatal' (subject to change). This allows
9988 testing the HostName extension for a specific single hostname ('-cert'
9989 and '-key' remain fallbacks for handshakes without HostName
9990 negotiation). If the unrecognized_name alert has to be sent, this by
9991 default is a warning; it becomes fatal with the '-servername_fatal'
9994 *Peter Sylvester, Remy Allais, Christophe Renou*
9996 * Whirlpool hash implementation is added.
10000 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10001 bn(64,32). Because of instruction set limitations it doesn't have
10002 any negative impact on performance. This was done mostly in order
10003 to make it possible to share assembler modules, such as bn_mul_mont
10004 implementations, between 32- and 64-bit builds without hassle.
10008 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10009 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10014 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10015 dedicated Montgomery multiplication procedure, is introduced.
10016 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10017 "64-bit" performance on certain 32-bit targets.
10021 * New option SSL_OP_NO_COMP to disable use of compression selectively
10022 in SSL structures. New SSL ctrl to set maximum send fragment size.
10023 Save memory by setting the I/O buffer sizes dynamically instead of
10024 using the maximum available value.
10028 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10029 in addition to the text details.
10033 * Very, very preliminary EXPERIMENTAL support for printing of general
10034 ASN1 structures. This currently produces rather ugly output and doesn't
10035 handle several customised structures at all.
10039 * Integrated support for PVK file format and some related formats such
10040 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10041 these in the 'rsa' and 'dsa' utilities.
10045 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10049 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10050 place for the (very old) "NETSCAPE" format certificates which are now
10051 handled using new ASN1 code equivalents.
10055 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10056 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10057 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10061 * Modify CRL distribution points extension code to print out previously
10062 unsupported fields. Enhance extension setting code to allow setting of
10067 * Add print and set support for Issuing Distribution Point CRL extension.
10071 * Change 'Configure' script to enable Camellia by default.
10078 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10080 * When rejecting SSL/TLS records due to an incorrect version number, never
10081 update s->server with a new major version number. As of
10082 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10083 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10084 the previous behavior could result in a read attempt at NULL when
10085 receiving specific incorrect SSL/TLS records once record payload
10086 protection is active. ([CVE-2010-0740])
10088 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10090 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10091 could be crashed if the relevant tables were not present (e.g. chrooted).
10093 *Tomas Hoger <thoger@redhat.com>*
10095 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10097 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10099 *Martin Olsson, Neel Mehta*
10101 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10102 accommodate for stack sorting, always a write lock!).
10106 * On some versions of WIN32 Heap32Next is very slow. This can cause
10107 excessive delays in the RAND_poll(): over a minute. As a workaround
10108 include a time check in the inner Heap32Next loop too.
10112 * The code that handled flushing of data in SSL/TLS originally used the
10113 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10114 the problem outlined in PR#1949. The fix suggested there however can
10115 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10116 of Apache). So instead simplify the code to flush unconditionally.
10117 This should be fine since flushing with no data to flush is a no op.
10121 * Handle TLS versions 2.0 and later properly and correctly use the
10122 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10123 off ancient servers have a habit of sticking around for a while...
10127 * Modify compression code so it frees up structures without using the
10128 ex_data callbacks. This works around a problem where some applications
10129 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10130 restarting) then use compression (e.g. SSL with compression) later.
10131 This results in significant per-connection memory leaks and
10132 has caused some security issues including CVE-2008-1678 and
10137 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10138 change when encrypting or decrypting.
10142 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10143 connect and renegotiate with servers which do not support RI.
10144 Until RI is more widely deployed this option is enabled by default.
10148 * Add "missing" ssl ctrls to clear options and mode.
10152 * If client attempts to renegotiate and doesn't support RI respond with
10153 a no_renegotiation alert as required by RFC5746. Some renegotiating
10154 TLS clients will continue a connection gracefully when they receive
10155 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10156 waiting for a server hello which it will never receive. Now we treat a
10157 received no_renegotiation alert as a fatal error. This is because
10158 applications requesting a renegotiation might well expect it to succeed
10159 and would have no code in place to handle the server denying it so the
10160 only safe thing to do is to terminate the connection.
10164 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10165 peer supports secure renegotiation and 0 otherwise. Print out peer
10166 renegotiation support in s_client/s_server.
10170 * Replace the highly broken and deprecated SPKAC certification method with
10171 the updated NID creation version. This should correctly handle UTF8.
10175 * Implement RFC5746. Re-enable renegotiation but require the extension
10176 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10177 turns out to be a bad idea. It has been replaced by
10178 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10179 SSL_CTX_set_options(). This is really not recommended unless you
10180 know what you are doing.
10182 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10184 * Fixes to stateless session resumption handling. Use initial_ctx when
10185 issuing and attempting to decrypt tickets in case it has changed during
10186 servername handling. Use a non-zero length session ID when attempting
10187 stateless session resumption: this makes it possible to determine if
10188 a resumption has occurred immediately after receiving server hello
10189 (several places in OpenSSL subtly assume this) instead of later in
10194 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10195 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10196 fixes for a few places where the return code is not checked
10199 *Julia Lawall <julia@diku.dk>*
10201 * Add --strict-warnings option to Configure script to include devteam
10202 warnings in other configurations.
10206 * Add support for --libdir option and LIBDIR variable in makefiles. This
10207 makes it possible to install openssl libraries in locations which
10208 have names other than "lib", for example "/usr/lib64" which some
10211 *Steve Henson, based on patch from Jeremy Utley*
10213 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10214 X690 8.9.12 and can produce some misleading textual output of OIDs.
10216 *Steve Henson, reported by Dan Kaminsky*
10218 * Delete MD2 from algorithm tables. This follows the recommendation in
10219 several standards that it is not used in new applications due to
10220 several cryptographic weaknesses. For binary compatibility reasons
10221 the MD2 API is still compiled in by default.
10225 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10230 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10231 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10234 *Guenter <lists@gknw.net>*
10236 * Fix the server certificate chain building code to use X509_verify_cert(),
10237 it used to have an ad-hoc builder which was unable to cope with anything
10238 other than a simple chain.
10240 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10242 * Don't check self signed certificate signatures in X509_verify_cert()
10243 by default (a flag can override this): it just wastes time without
10244 adding any security. As a useful side effect self signed root CAs
10245 with non-FIPS digests are now usable in FIPS mode.
10249 * In dtls1_process_out_of_seq_message() the check if the current message
10250 is already buffered was missing. For every new message was memory
10251 allocated, allowing an attacker to perform an denial of service attack
10252 with sending out of seq handshake messages until there is no memory
10253 left. Additionally every future message was buffered, even if the
10254 sequence number made no sense and would be part of another handshake.
10255 So only messages with sequence numbers less than 10 in advance will be
10256 buffered. ([CVE-2009-1378])
10258 *Robin Seggelmann, discovered by Daniel Mentz*
10260 * Records are buffered if they arrive with a future epoch to be
10261 processed after finishing the corresponding handshake. There is
10262 currently no limitation to this buffer allowing an attacker to perform
10263 a DOS attack with sending records with future epochs until there is no
10264 memory left. This patch adds the pqueue_size() function to determine
10265 the size of a buffer and limits the record buffer to 100 entries.
10268 *Robin Seggelmann, discovered by Daniel Mentz*
10270 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10271 parent structure is freed. ([CVE-2009-1379])
10275 * Handle non-blocking I/O properly in SSL_shutdown() call.
10277 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10279 * Add `2.5.4.*` OIDs
10281 *Ilya O. <vrghost@gmail.com>*
10283 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10285 * Disable renegotiation completely - this fixes a severe security
10286 problem ([CVE-2009-3555]) at the cost of breaking all
10287 renegotiation. Renegotiation can be re-enabled by setting
10288 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10289 run-time. This is really not recommended unless you know what
10294 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10296 * Don't set val to NULL when freeing up structures, it is freed up by
10297 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10298 zeroing past the valid field. ([CVE-2009-0789])
10300 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10302 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10303 checked correctly. This would allow some invalid signed attributes to
10304 appear to verify correctly. ([CVE-2009-0591])
10306 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10308 * Reject UniversalString and BMPString types with invalid lengths. This
10309 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10310 a legal length. ([CVE-2009-0590])
10314 * Set S/MIME signing as the default purpose rather than setting it
10315 unconditionally. This allows applications to override it at the store
10320 * Permit restricted recursion of ASN1 strings. This is needed in practice
10321 to handle some structures.
10325 * Improve efficiency of mem_gets: don't search whole buffer each time
10328 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10330 * New -hex option for openssl rand.
10334 * Print out UTF8String and NumericString when parsing ASN1.
10338 * Support NumericString type for name components.
10342 * Allow CC in the environment to override the automatically chosen
10343 compiler. Note that nothing is done to ensure flags work with the
10348 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10350 * Properly check EVP_VerifyFinal() and similar return values
10353 *Ben Laurie, Bodo Moeller, Google Security Team*
10355 * Enable TLS extensions by default.
10359 * Allow the CHIL engine to be loaded, whether the application is
10360 multithreaded or not. (This does not release the developer from the
10361 obligation to set up the dynamic locking callbacks.)
10363 *Sander Temme <sander@temme.net>*
10365 * Use correct exit code if there is an error in dgst command.
10367 *Steve Henson; problem pointed out by Roland Dirlewanger*
10369 * Tweak Configure so that you need to say "experimental-jpake" to enable
10370 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10374 * Add experimental JPAKE support, including demo authentication in
10375 s_client and s_server.
10379 * Set the comparison function in v3_addr_canonize().
10381 *Rob Austein <sra@hactrn.net>*
10383 * Add support for XMPP STARTTLS in s_client.
10385 *Philip Paeps <philip@freebsd.org>*
10387 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10388 to ensure that even with this option, only ciphersuites in the
10389 server's preference list will be accepted. (Note that the option
10390 applies only when resuming a session, so the earlier behavior was
10391 just about the algorithm choice for symmetric cryptography.)
10395 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10397 * Fix NULL pointer dereference if a DTLS server received
10398 ChangeCipherSpec as first record ([CVE-2009-1386]).
10402 * Fix a state transition in s3_srvr.c and d1_srvr.c
10403 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10405 *Nagendra Modadugu*
10407 * The fix in 0.9.8c that supposedly got rid of unsafe
10408 double-checked locking was incomplete for RSA blinding,
10409 addressing just one layer of what turns out to have been
10410 doubly unsafe triple-checked locking.
10412 So now fix this for real by retiring the MONT_HELPER macro
10413 in crypto/rsa/rsa_eay.c.
10415 *Bodo Moeller; problem pointed out by Marius Schilder*
10417 * Various precautionary measures:
10419 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10421 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10422 (NB: This would require knowledge of the secret session ticket key
10423 to exploit, in which case you'd be SOL either way.)
10425 - Change bn_nist.c so that it will properly handle input BIGNUMs
10426 outside the expected range.
10428 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10431 *Neel Mehta, Bodo Moeller*
10433 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10434 the load fails. Useful for distros.
10436 *Ben Laurie and the FreeBSD team*
10438 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10442 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10446 * Expand ENGINE to support engine supplied SSL client certificate functions.
10448 This work was sponsored by Logica.
10452 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10453 keystores. Support for SSL/TLS client authentication too.
10454 Not compiled unless enable-capieng specified to Configure.
10456 This work was sponsored by Logica.
10460 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10461 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10462 attribute creation routines such as certificate requests and PKCS#12
10467 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10469 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10470 handshake which could lead to a client crash as found using the
10471 Codenomicon TLS test suite ([CVE-2008-1672])
10473 *Steve Henson, Mark Cox*
10475 * Fix double free in TLS server name extensions which could lead to
10476 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10480 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10482 Clear the error queue to ensure that error entries left from
10483 older function calls do not interfere with the correct operation.
10485 *Lutz Jaenicke, Erik de Castro Lopo*
10487 * Remove root CA certificates of commercial CAs:
10489 The OpenSSL project does not recommend any specific CA and does not
10490 have any policy with respect to including or excluding any CA.
10491 Therefore, it does not make any sense to ship an arbitrary selection
10492 of root CA certificates with the OpenSSL software.
10496 * RSA OAEP patches to fix two separate invalid memory reads.
10497 The first one involves inputs when 'lzero' is greater than
10498 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10499 before the beginning of from). The second one involves inputs where
10500 the 'db' section contains nothing but zeroes (there is a one-byte
10501 invalid read after the end of 'db').
10503 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10505 * Partial backport from 0.9.9-dev:
10507 Introduce bn_mul_mont (dedicated Montgomery multiplication
10508 procedure) as a candidate for BIGNUM assembler implementation.
10509 While 0.9.9-dev uses assembler for various architectures, only
10510 x86_64 is available by default here in the 0.9.8 branch, and
10511 32-bit x86 is available through a compile-time setting.
10513 To try the 32-bit x86 assembler implementation, use Configure
10514 option "enable-montasm" (which exists only for this backport).
10516 As "enable-montasm" for 32-bit x86 disclaims code stability
10517 anyway, in this constellation we activate additional code
10518 backported from 0.9.9-dev for further performance improvements,
10519 namely BN_from_montgomery_word. (To enable this otherwise,
10520 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10522 *Andy Polyakov (backport partially by Bodo Moeller)*
10524 * Add TLS session ticket callback. This allows an application to set
10525 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10526 values. This is useful for key rollover for example where several key
10527 sets may exist with different names.
10531 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10532 This was broken until now in 0.9.8 releases, such that the only way
10533 a registered ENGINE could be used (assuming it initialises
10534 successfully on the host) was to explicitly set it as the default
10535 for the relevant algorithms. This is in contradiction with 0.9.7
10536 behaviour and the documentation. With this fix, when an ENGINE is
10537 registered into a given algorithm's table of implementations, the
10538 'uptodate' flag is reset so that auto-discovery will be used next
10539 time a new context for that algorithm attempts to select an
10542 *Ian Lister (tweaked by Geoff Thorpe)*
10544 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10545 implementation in the following ways:
10547 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10550 Lack of BER streaming support means one pass streaming processing is
10551 only supported if data is detached: setting the streaming flag is
10552 ignored for embedded content.
10554 CMS support is disabled by default and must be explicitly enabled
10555 with the enable-cms configuration option.
10559 * Update the GMP engine glue to do direct copies between BIGNUM and
10560 mpz_t when openssl and GMP use the same limb size. Otherwise the
10561 existing "conversion via a text string export" trick is still used.
10563 *Paul Sheer <paulsheer@gmail.com>*
10565 * Zlib compression BIO. This is a filter BIO which compressed and
10566 uncompresses any data passed through it.
10570 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10571 RFC3394 compatible AES key wrapping.
10575 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10576 sets string data without copying. X509_ALGOR_set0() and
10577 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10578 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10579 from an X509_ATTRIBUTE structure optionally checking it occurs only
10580 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10585 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10586 to get the expected BN_FLG_CONSTTIME behavior.
10588 *Bodo Moeller (Google)*
10592 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10593 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10594 - added some more tests to do_tests.pl
10595 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10596 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10597 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10598 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10599 - various changes to netware.pl to enable gcc-cross builds on Win32
10601 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10602 - various changes to fix missing prototype warnings
10603 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10604 - added AES, WHIRLPOOL and CPUID assembler code to build files
10605 - added missing AES assembler make rules to mk1mf.pl
10606 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10608 *Guenter Knauf <eflash@gmx.net>*
10610 * Implement certificate status request TLS extension defined in RFC3546.
10611 A client can set the appropriate parameters and receive the encoded
10612 OCSP response via a callback. A server can query the supplied parameters
10613 and set the encoded OCSP response in the callback. Add simplified examples
10614 to s_client and s_server.
10618 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10620 * Fix various bugs:
10621 + Binary incompatibility of ssl_ctx_st structure
10622 + DTLS interoperation with non-compliant servers
10623 + Don't call get_session_cb() without proposed session
10624 + Fix ia64 assembler code
10626 *Andy Polyakov, Steve Henson*
10628 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10630 * DTLS Handshake overhaul. There were longstanding issues with
10631 OpenSSL DTLS implementation, which were making it impossible for
10632 RFC 4347 compliant client to communicate with OpenSSL server.
10633 Unfortunately just fixing these incompatibilities would "cut off"
10634 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10635 server keeps tolerating non RFC compliant syntax. The opposite is
10636 not true, 0.9.8f client can not communicate with earlier server.
10637 This update even addresses CVE-2007-4995.
10641 * Changes to avoid need for function casts in OpenSSL: some compilers
10642 (gcc 4.2 and later) reject their use.
10643 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10646 * Add RFC4507 support to OpenSSL. This includes the corrections in
10647 RFC4507bis. The encrypted ticket format is an encrypted encoded
10648 SSL_SESSION structure, that way new session features are automatically
10651 If a client application caches session in an SSL_SESSION structure
10652 support is transparent because tickets are now stored in the encoded
10655 The SSL_CTX structure automatically generates keys for ticket
10656 protection in servers so again support should be possible
10657 with no application modification.
10659 If a client or server wishes to disable RFC4507 support then the option
10660 SSL_OP_NO_TICKET can be set.
10662 Add a TLS extension debugging callback to allow the contents of any client
10663 or server extensions to be examined.
10665 This work was sponsored by Google.
10669 * Add initial support for TLS extensions, specifically for the server_name
10670 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10671 have new members for a hostname. The SSL data structure has an
10672 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10673 stored in that context to allow for session resumption, even after the
10674 SSL has been switched to a new SSL_CTX in reaction to a client's
10675 server_name extension.
10677 New functions (subject to change):
10679 SSL_get_servername()
10680 SSL_get_servername_type()
10683 New CTRL codes and macros (subject to change):
10685 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10686 - SSL_CTX_set_tlsext_servername_callback()
10687 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10688 - SSL_CTX_set_tlsext_servername_arg()
10689 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10691 openssl s_client has a new '-servername ...' option.
10693 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10694 '-key2 ...', '-servername_fatal' (subject to change). This allows
10695 testing the HostName extension for a specific single hostname ('-cert'
10696 and '-key' remain fallbacks for handshakes without HostName
10697 negotiation). If the unrecognized_name alert has to be sent, this by
10698 default is a warning; it becomes fatal with the '-servername_fatal'
10701 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10703 * Add AES and SSE2 assembly language support to VC++ build.
10707 * Mitigate attack on final subtraction in Montgomery reduction.
10711 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10712 (which previously caused an internal error).
10716 * Squeeze another 10% out of IGE mode when in != out.
10720 * AES IGE mode speedup.
10722 *Dean Gaudet (Google)*
10724 * Add the Korean symmetric 128-bit cipher SEED (see
10725 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10726 add SEED ciphersuites from RFC 4162:
10728 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10729 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10730 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10731 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10733 To minimize changes between patchlevels in the OpenSSL 0.9.8
10734 series, SEED remains excluded from compilation unless OpenSSL
10735 is configured with 'enable-seed'.
10737 *KISA, Bodo Moeller*
10739 * Mitigate branch prediction attacks, which can be practical if a
10740 single processor is shared, allowing a spy process to extract
10741 information. For detailed background information, see
10742 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10743 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10744 and Necessary Software Countermeasures"). The core of the change
10745 are new versions BN_div_no_branch() and
10746 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10747 respectively, which are slower, but avoid the security-relevant
10748 conditional branches. These are automatically called by BN_div()
10749 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10750 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10751 remove a conditional branch.
10753 BN_FLG_CONSTTIME is the new name for the previous
10754 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10755 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10756 in the exponent causes BN_mod_exp_mont() to use the alternative
10757 implementation in BN_mod_exp_mont_consttime().) The old name
10758 remains as a deprecated alias.
10760 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10761 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10762 constant-time implementations for more than just exponentiation.
10763 Here too the old name is kept as a deprecated alias.
10765 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10766 the BN_BLINDING structure gets an independent copy of the
10767 modulus. This means that the previous `BIGNUM *m` argument to
10768 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10769 essentially becomes `const BIGNUM *m`, although we can't actually
10770 change this in the header file before 0.9.9. It allows
10771 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10772 enable BN_FLG_CONSTTIME.
10774 *Matthew D Wood (Intel Corp)*
10776 * In the SSL/TLS server implementation, be strict about session ID
10777 context matching (which matters if an application uses a single
10778 external cache for different purposes). Previously,
10779 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10780 set. This did ensure strict client verification, but meant that,
10781 with applications using a single external cache for quite
10782 different requirements, clients could circumvent ciphersuite
10783 restrictions for a given session ID context by starting a session
10784 in a different context.
10788 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10789 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10790 authentication-only ciphersuites.
10794 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10795 not complete and could lead to a possible single byte overflow
10796 ([CVE-2007-5135]) [Ben Laurie]
10798 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10800 * Since AES128 and AES256 (and similarly Camellia128 and
10801 Camellia256) share a single mask bit in the logic of
10802 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10803 kludge to work properly if AES128 is available and AES256 isn't
10804 (or if Camellia128 is available and Camellia256 isn't).
10808 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10809 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10810 When a point or a seed is encoded in a BIT STRING, we need to
10811 prevent the removal of trailing zero bits to get the proper DER
10812 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10813 of a NamedBitList, for which trailing 0 bits need to be removed.)
10817 * Have SSL/TLS server implementation tolerate "mismatched" record
10818 protocol version while receiving ClientHello even if the
10819 ClientHello is fragmented. (The server can't insist on the
10820 particular protocol version it has chosen before the ServerHello
10821 message has informed the client about his choice.)
10825 * Add RFC 3779 support.
10827 *Rob Austein for ARIN, Ben Laurie*
10829 * Load error codes if they are not already present instead of using a
10830 static variable. This allows them to be cleanly unloaded and reloaded.
10831 Improve header file function name parsing.
10835 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10836 or CAPABILITY handshake as required by RFCs.
10838 *Goetz Babin-Ebell*
10840 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10842 * Introduce limits to prevent malicious keys being able to
10843 cause a denial of service. ([CVE-2006-2940])
10845 *Steve Henson, Bodo Moeller*
10847 * Fix ASN.1 parsing of certain invalid structures that can result
10848 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10850 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10851 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10853 * Fix SSL client code which could crash if connecting to a
10854 malicious SSLv2 server. ([CVE-2006-4343])
10856 *Tavis Ormandy and Will Drewry, Google Security Team*
10858 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10859 match only those. Before that, "AES256-SHA" would be interpreted
10860 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10861 the same strength classification in 0.9.7h) as we currently only
10862 have a single AES bit in the ciphersuite description bitmap.
10863 That change, however, also applied to ciphersuite strings such as
10864 "RC4-MD5" that intentionally matched multiple ciphersuites --
10865 namely, SSL 2.0 ciphersuites in addition to the more common ones
10866 from SSL 3.0/TLS 1.0.
10868 So we change the selection algorithm again: Naming an explicit
10869 ciphersuite selects this one ciphersuite, and any other similar
10870 ciphersuite (same bitmap) from *other* protocol versions.
10871 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10872 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10874 Since SSL 2.0 does not have any ciphersuites for which the
10875 128/256 bit distinction would be relevant, this works for now.
10876 The proper fix will be to use different bits for AES128 and
10877 AES256, which would have avoided the problems from the beginning;
10878 however, bits are scarce, so we can only do this in a new release
10879 (not just a patchlevel) when we can change the SSL_CIPHER
10880 definition to split the single 'unsigned long mask' bitmap into
10881 multiple values to extend the available space.
10885 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10887 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10888 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10890 * Add AES IGE and biIGE modes.
10894 * Change the Unix randomness entropy gathering to use poll() when
10895 possible instead of select(), since the latter has some
10896 undesirable limitations.
10898 *Darryl Miles via Richard Levitte and Bodo Moeller*
10900 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10901 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10902 cannot be implicitly activated as part of, e.g., the "AES" alias.
10903 However, please upgrade to OpenSSL 0.9.9[-dev] for
10904 non-experimental use of the ECC ciphersuites to get TLS extension
10905 support, which is required for curve and point format negotiation
10906 to avoid potential handshake problems.
10910 * Disable rogue ciphersuites:
10912 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10913 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10914 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10916 The latter two were purportedly from
10917 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10920 Also deactivate the remaining ciphersuites from
10921 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10922 unofficial, and the ID has long expired.
10926 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10927 dual-core machines) and other potential thread-safety issues.
10931 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10932 versions), which is now available for royalty-free use
10933 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10934 Also, add Camellia TLS ciphersuites from RFC 4132.
10936 To minimize changes between patchlevels in the OpenSSL 0.9.8
10937 series, Camellia remains excluded from compilation unless OpenSSL
10938 is configured with 'enable-camellia'.
10942 * Disable the padding bug check when compression is in use. The padding
10943 bug check assumes the first packet is of even length, this is not
10944 necessarily true if compression is enabled and can result in false
10945 positives causing handshake failure. The actual bug test is ancient
10946 code so it is hoped that implementations will either have fixed it by
10947 now or any which still have the bug do not support compression.
10951 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10953 * When applying a cipher rule check to see if string match is an explicit
10954 cipher suite and only match that one cipher suite if it is.
10958 * Link in manifests for VC++ if needed.
10960 *Austin Ziegler <halostatue@gmail.com>*
10962 * Update support for ECC-based TLS ciphersuites according to
10963 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10964 TLS extensions, which are supported starting with the 0.9.9
10965 branch, not in the OpenSSL 0.9.8 branch).
10969 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10970 opaque EVP_CIPHER_CTX handling.
10974 * Fixes and enhancements to zlib compression code. We now only use
10975 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10976 to conform with the standards mentioned here:
10977 <http://www.zlib.net/DLL_FAQ.txt>
10978 Static zlib linking now works on Windows and the new --with-zlib-include
10979 --with-zlib-lib options to Configure can be used to supply the location
10980 of the headers and library. Gracefully handle case where zlib library
10985 * Several fixes and enhancements to the OID generation code. The old code
10986 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10987 handle numbers larger than ULONG_MAX, truncated printing and had a
10988 non standard OBJ_obj2txt() behaviour.
10992 * Add support for building of engines under engine/ as shared libraries
10993 under VC++ build system.
10997 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10998 Hopefully, we will not see any false combination of paths any more.
11002 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11004 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11005 (part of SSL_OP_ALL). This option used to disable the
11006 countermeasure against man-in-the-middle protocol-version
11007 rollback in the SSL 2.0 server implementation, which is a bad
11008 idea. ([CVE-2005-2969])
11010 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11011 for Information Security, National Institute of Advanced Industrial
11012 Science and Technology [AIST], Japan)*
11014 * Add two function to clear and return the verify parameter flags.
11018 * Keep cipherlists sorted in the source instead of sorting them at
11019 runtime, thus removing the need for a lock.
11023 * Avoid some small subgroup attacks in Diffie-Hellman.
11025 *Nick Mathewson and Ben Laurie*
11027 * Add functions for well-known primes.
11031 * Extended Windows CE support.
11033 *Satoshi Nakamura and Andy Polyakov*
11035 * Initialize SSL_METHOD structures at compile time instead of during
11036 runtime, thus removing the need for a lock.
11040 * Make PKCS7_decrypt() work even if no certificate is supplied by
11041 attempting to decrypt each encrypted key in turn. Add support to
11046 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11048 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11051 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11055 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11056 key into the same file any more.
11060 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11064 * Add -utf8 command line and config file option to 'ca'.
11066 *Stefan <stf@udoma.org*
11068 * Removed the macro des_crypt(), as it seems to conflict with some
11069 libraries. Use DES_crypt().
11073 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11074 involves renaming the source and generated shared-libs for
11075 both. The engines will accept the corrected or legacy ids
11076 ('ncipher' and '4758_cca' respectively) when binding. NB,
11077 this only applies when building 'shared'.
11079 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11081 * Add attribute functions to EVP_PKEY structure. Modify
11082 PKCS12_create() to recognize a CSP name attribute and
11083 use it. Make -CSP option work again in pkcs12 utility.
11087 * Add new functionality to the bn blinding code:
11088 - automatic re-creation of the BN_BLINDING parameters after
11089 a fixed number of uses (currently 32)
11090 - add new function for parameter creation
11091 - introduce flags to control the update behaviour of the
11092 BN_BLINDING parameters
11093 - hide BN_BLINDING structure
11094 Add a second BN_BLINDING slot to the RSA structure to improve
11095 performance when a single RSA object is shared among several
11100 * Add support for DTLS.
11102 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11104 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11105 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11109 * Remove buggy and incomplete DH cert support from
11110 ssl/ssl_rsa.c and ssl/s3_both.c
11114 * Use SHA-1 instead of MD5 as the default digest algorithm for
11115 the `apps/openssl` commands.
11119 * Compile clean with "-Wall -Wmissing-prototypes
11120 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11121 DEBUG_SAFESTACK must also be set.
11125 * Change ./Configure so that certain algorithms can be disabled by default.
11126 The new counterpiece to "no-xxx" is "enable-xxx".
11128 The patented RC5 and MDC2 algorithms will now be disabled unless
11129 "enable-rc5" and "enable-mdc2", respectively, are specified.
11131 (IDEA remains enabled despite being patented. This is because IDEA
11132 is frequently required for interoperability, and there is no license
11133 fee for non-commercial use. As before, "no-idea" can be used to
11134 avoid this algorithm.)
11138 * Add processing of proxy certificates (see RFC 3820). This work was
11139 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11140 EGEE (Enabling Grids for E-science in Europe).
11144 * RC4 performance overhaul on modern architectures/implementations, such
11145 as Intel P4, IA-64 and AMD64.
11149 * New utility extract-section.pl. This can be used specify an alternative
11150 section number in a pod file instead of having to treat each file as
11151 a separate case in Makefile. This can be done by adding two lines to the
11154 =for comment openssl_section:XXX
11156 The blank line is mandatory.
11160 * New arguments -certform, -keyform and -pass for s_client and s_server
11161 to allow alternative format key and certificate files and passphrase
11166 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11167 update associated structures and add various utility functions.
11169 Add new policy related verify parameters, include policy checking in
11170 standard verify code. Enhance 'smime' application with extra parameters
11171 to support policy checking and print out.
11175 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11176 Nehemiah processors. These extensions support AES encryption in hardware
11177 as well as RNG (though RNG support is currently disabled).
11179 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11181 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11185 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11187 *Andy Polyakov and a number of other people*
11189 * Improved PowerPC platform support. Most notably BIGNUM assembler
11190 implementation contributed by IBM.
11192 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11194 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11195 exponent rather than 'unsigned long'. There is a corresponding change to
11196 the new 'rsa_keygen' element of the RSA_METHOD structure.
11198 *Jelte Jansen, Geoff Thorpe*
11200 * Functionality for creating the initial serial number file is now
11201 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11203 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11204 number file to 1, which is bound to cause problems. To avoid
11205 the problems while respecting compatibility between different 0.9.7
11206 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11207 CA.pl for serial number initialization. With the new release 0.9.8,
11208 we can fix the problem directly in the 'ca' utility.)
11212 * Reduced header interdependencies by declaring more opaque objects in
11213 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11214 give fewer recursive includes, which could break lazy source code - so
11215 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11216 developers should define this symbol when building and using openssl to
11217 ensure they track the recommended behaviour, interfaces, [etc], but
11218 backwards-compatible behaviour prevails when this isn't defined.
11222 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11226 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11227 This will generate a random key of the appropriate length based on the
11228 cipher context. The EVP_CIPHER can provide its own random key generation
11229 routine to support keys of a specific form. This is used in the des and
11230 3des routines to generate a key of the correct parity. Update S/MIME
11231 code to use new functions and hence generate correct parity DES keys.
11232 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11233 valid (weak or incorrect parity).
11237 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11238 as looking them up. This is useful when the verified structure may contain
11239 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11240 present unless the new PKCS7_NO_CRL flag is asserted.
11244 * Extend ASN1 oid configuration module. It now additionally accepts the
11247 shortName = some long name, 1.2.3.4
11251 * Reimplemented the BN_CTX implementation. There is now no more static
11252 limitation on the number of variables it can handle nor the depth of the
11253 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11254 information can now expand as required, and rather than having a single
11255 static array of bignums, BN_CTX now uses a linked-list of such arrays
11256 allowing it to expand on demand whilst maintaining the usefulness of
11257 BN_CTX's "bundling".
11261 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11262 to allow all RSA operations to function using a single BN_CTX.
11266 * Preliminary support for certificate policy evaluation and checking. This
11267 is initially intended to pass the tests outlined in "Conformance Testing
11268 of Relying Party Client Certificate Path Processing Logic" v1.07.
11272 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11273 remained unused and not that useful. A variety of other little bignum
11274 tweaks and fixes have also been made continuing on from the audit (see
11279 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11280 associated ASN1, EVP and SSL functions and old ASN1 macros.
11284 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11285 and this should never fail. So the return value from the use of
11286 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11287 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11291 * BN_CTX_get() should return zero-valued bignums, providing the same
11292 initialised value as BN_new().
11294 *Geoff Thorpe, suggested by Ulf Möller*
11296 * Support for inhibitAnyPolicy certificate extension.
11300 * An audit of the BIGNUM code is underway, for which debugging code is
11301 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11302 is considered valid when processing BIGNUMs, and causes execution to
11303 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11304 further steps are taken to deliberately pollute unused data in BIGNUM
11305 structures to try and expose faulty code further on. For now, openssl will
11306 (in its default mode of operation) continue to tolerate the inconsistent
11307 forms that it has tolerated in the past, but authors and packagers should
11308 consider trying openssl and their own applications when compiled with
11309 these debugging symbols defined. It will help highlight potential bugs in
11310 their own code, and will improve the test coverage for OpenSSL itself. At
11311 some point, these tighter rules will become openssl's default to improve
11312 maintainability, though the assert()s and other overheads will remain only
11313 in debugging configurations. See bn.h for more details.
11315 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11317 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11318 that can only be obtained through BN_CTX_new() (which implicitly
11319 initialises it). The presence of this function only made it possible
11320 to overwrite an existing structure (and cause memory leaks).
11324 * Because of the callback-based approach for implementing LHASH as a
11325 template type, lh_insert() adds opaque objects to hash-tables and
11326 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11327 to clean up those corresponding objects before destroying the hash table
11328 (and losing the object pointers). So some over-zealous constifications in
11329 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11330 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11331 prototyped to have "const" restrictions on the object pointers they are
11332 given (and so aren't required to cast them away any more).
11336 * The tmdiff.h API was so ugly and minimal that our own timing utility
11337 (speed) prefers to use its own implementation. The two implementations
11338 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11339 its object type properly exposed (MS_TM) instead of casting to/from
11340 `char *`. This may still change yet if someone realises MS_TM and
11342 aren't necessarily the greatest nomenclatures - but this is what was used
11343 internally to the implementation so I've used that for now.
11347 * Ensure that deprecated functions do not get compiled when
11348 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11349 the self-tests were still using deprecated key-generation functions so
11350 these have been updated also.
11354 * Reorganise PKCS#7 code to separate the digest location functionality
11355 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11356 New function PKCS7_set_digest() to set the digest type for PKCS#7
11357 digestedData type. Add additional code to correctly generate the
11358 digestedData type and add support for this type in PKCS7 initialization
11363 * New function PKCS7_set0_type_other() this initializes a PKCS7
11364 structure of type "other".
11368 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11369 sure the loop does correctly stop and breaking ("division by zero")
11370 modulus operations are not performed. The (pre-generated) prime
11371 table crypto/bn/bn_prime.h was already correct, but it could not be
11372 re-generated on some platforms because of the "division by zero"
11373 situation in the script.
11375 *Ralf S. Engelschall*
11377 * Update support for ECC-based TLS ciphersuites according to
11378 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11379 SHA-1 now is only used for "small" curves (where the
11380 representation of a field element takes up to 24 bytes); for
11381 larger curves, the field element resulting from ECDH is directly
11382 used as premaster secret.
11384 *Douglas Stebila (Sun Microsystems Laboratories)*
11386 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11387 curve secp160r1 to the tests.
11389 *Douglas Stebila (Sun Microsystems Laboratories)*
11391 * Add the possibility to load symbols globally with DSO.
11393 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11395 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11396 control of the error stack.
11400 * Add support for STORE in ENGINE.
11404 * Add the STORE type. The intention is to provide a common interface
11405 to certificate and key stores, be they simple file-based stores, or
11406 HSM-type store, or LDAP stores, or...
11407 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11411 * Add a generic structure called OPENSSL_ITEM. This can be used to
11412 pass a list of arguments to any function as well as provide a way
11413 for a function to pass data back to the caller.
11417 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11418 works like BUF_strdup() but can be used to duplicate a portion of
11419 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11424 * Add the function sk_find_ex() which works like sk_find(), but will
11425 return an index to an element even if an exact match couldn't be
11426 found. The index is guaranteed to point at the element where the
11427 searched-for key would be inserted to preserve sorting order.
11431 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11432 takes an extra flags argument for optional functionality. Currently,
11433 the following flags are defined:
11435 OBJ_BSEARCH_VALUE_ON_NOMATCH
11436 This one gets OBJ_bsearch_ex() to return a pointer to the first
11437 element where the comparing function returns a negative or zero
11440 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11441 This one gets OBJ_bsearch_ex() to return a pointer to the first
11442 element where the comparing function returns zero. This is useful
11443 if there are more than one element where the comparing function
11448 * Make it possible to create self-signed certificates with 'openssl ca'
11449 in such a way that the self-signed certificate becomes part of the
11450 CA database and uses the same mechanisms for serial number generation
11451 as all other certificate signing. The new flag '-selfsign' enables
11452 this functionality. Adapt CA.sh and CA.pl.in.
11456 * Add functionality to check the public key of a certificate request
11457 against a given private. This is useful to check that a certificate
11458 request can be signed by that key (self-signing).
11462 * Make it possible to have multiple active certificates with the same
11463 subject in the CA index file. This is done only if the keyword
11464 'unique_subject' is set to 'no' in the main CA section (default
11465 if 'CA_default') of the configuration file. The value is saved
11466 with the database itself in a separate index attribute file,
11467 named like the index file with '.attr' appended to the name.
11471 * Generate multi-valued AVAs using '+' notation in config files for
11476 * Support for nameConstraints certificate extension.
11480 * Support for policyConstraints certificate extension.
11484 * Support for policyMappings certificate extension.
11488 * Make sure the default DSA_METHOD implementation only uses its
11489 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11490 and change its own handlers to be NULL so as to remove unnecessary
11491 indirection. This lets alternative implementations fallback to the
11492 default implementation more easily.
11496 * Support for directoryName in GeneralName related extensions
11501 * Make it possible to link applications using Makefile.shared.
11502 Make that possible even when linking against static libraries!
11506 * Support for single pass processing for S/MIME signing. This now
11507 means that S/MIME signing can be done from a pipe, in addition
11508 cleartext signing (multipart/signed type) is effectively streaming
11509 and the signed data does not need to be all held in memory.
11511 This is done with a new flag PKCS7_STREAM. When this flag is set
11512 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11513 is done after the data is output (and digests calculated) in
11514 SMIME_write_PKCS7().
11518 * Add full support for -rpath/-R, both in shared libraries and
11519 applications, at least on the platforms where it's known how
11524 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11525 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11526 will now compute a table of multiples of the generator that
11527 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11528 faster (notably in the case of a single point multiplication,
11529 scalar * generator).
11531 *Nils Larsch, Bodo Moeller*
11533 * IPv6 support for certificate extensions. The various extensions
11534 which use the IP:a.b.c.d can now take IPv6 addresses using the
11535 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11540 * Added an ENGINE that implements RSA by performing private key
11541 exponentiations with the GMP library. The conversions to and from
11542 GMP's mpz_t format aren't optimised nor are any montgomery forms
11543 cached, and on x86 it appears OpenSSL's own performance has caught up.
11544 However there are likely to be other architectures where GMP could
11545 provide a boost. This ENGINE is not built in by default, but it can be
11546 specified at Configure time and should be accompanied by the necessary
11547 linker additions, eg;
11548 ./config -DOPENSSL_USE_GMP -lgmp
11552 * "openssl engine" will not display ENGINE/DSO load failure errors when
11553 testing availability of engines with "-t" - the old behaviour is
11554 produced by increasing the feature's verbosity with "-tt".
11558 * ECDSA routines: under certain error conditions uninitialized BN objects
11559 could be freed. Solution: make sure initialization is performed early
11560 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11565 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11566 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11567 software implementations. For DSA and DH, parameter generation can
11568 also be overridden by providing the appropriate method callbacks.
11572 * Change the "progress" mechanism used in key-generation and
11573 primality testing to functions that take a new BN_GENCB pointer in
11574 place of callback/argument pairs. The new API functions have `_ex`
11575 postfixes and the older functions are reimplemented as wrappers for
11576 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11577 declarations of the old functions to help (graceful) attempts to
11578 migrate to the new functions. Also, the new key-generation API
11579 functions operate on a caller-supplied key-structure and return
11580 success/failure rather than returning a key or NULL - this is to
11581 help make "keygen" another member function of RSA_METHOD etc.
11583 Example for using the new callback interface:
11585 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11586 void *my_arg = ...;
11589 BN_GENCB_set(&my_cb, my_callback, my_arg);
11591 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11592 /* For the meaning of a, b in calls to my_callback(), see the
11593 * documentation of the function that calls the callback.
11594 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11595 * my_callback should return 1 if it wants BN_is_prime_ex()
11596 * to continue, or 0 to stop.
11601 * Change the ZLIB compression method to be stateful, and make it
11602 available to TLS with the number defined in
11603 draft-ietf-tls-compression-04.txt.
11607 * Add the ASN.1 structures and functions for CertificatePair, which
11608 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11610 CertificatePair ::= SEQUENCE {
11611 forward [0] Certificate OPTIONAL,
11612 reverse [1] Certificate OPTIONAL,
11613 -- at least one of the pair shall be present -- }
11615 Also implement the PEM functions to read and write certificate
11616 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11618 This needed to be defined, mostly for the sake of the LDAP
11619 attribute crossCertificatePair, but may prove useful elsewhere as
11624 * Make it possible to inhibit symlinking of shared libraries in
11625 Makefile.shared, for Cygwin's sake.
11629 * Extend the BIGNUM API by creating a function
11630 void BN_set_negative(BIGNUM *a, int neg);
11631 and a macro that behave like
11632 int BN_is_negative(const BIGNUM *a);
11634 to avoid the need to access 'a->neg' directly in applications.
11638 * Implement fast modular reduction for pseudo-Mersenne primes
11639 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11640 EC_GROUP_new_curve_GFp() will now automatically use this
11643 *Nils Larsch <nla@trustcenter.de>*
11645 * Add new lock type (CRYPTO_LOCK_BN).
11649 * Change the ENGINE framework to automatically load engines
11650 dynamically from specific directories unless they could be
11651 found to already be built in or loaded. Move all the
11652 current engines except for the cryptodev one to a new
11653 directory engines/.
11654 The engines in engines/ are built as shared libraries if
11655 the "shared" options was given to ./Configure or ./config.
11656 Otherwise, they are inserted in libcrypto.a.
11657 /usr/local/ssl/engines is the default directory for dynamic
11658 engines, but that can be overridden at configure time through
11659 the usual use of --prefix and/or --openssldir, and at run
11660 time with the environment variable OPENSSL_ENGINES.
11662 *Geoff Thorpe and Richard Levitte*
11664 * Add Makefile.shared, a helper makefile to build shared
11665 libraries. Adapt Makefile.org.
11669 * Add version info to Win32 DLLs.
11671 *Peter 'Luna' Runestig" <peter@runestig.com>*
11673 * Add new 'medium level' PKCS#12 API. Certificates and keys
11674 can be added using this API to created arbitrary PKCS#12
11675 files while avoiding the low-level API.
11677 New options to PKCS12_create(), key or cert can be NULL and
11678 will then be omitted from the output file. The encryption
11679 algorithm NIDs can be set to -1 for no encryption, the mac
11680 iteration count can be set to 0 to omit the mac.
11682 Enhance pkcs12 utility by making the -nokeys and -nocerts
11683 options work when creating a PKCS#12 file. New option -nomac
11684 to omit the mac, NONE can be set for an encryption algorithm.
11685 New code is modified to use the enhanced PKCS12_create()
11686 instead of the low-level API.
11690 * Extend ASN1 encoder to support indefinite length constructed
11691 encoding. This can output sequences tags and octet strings in
11692 this form. Modify pk7_asn1.c to support indefinite length
11693 encoding. This is experimental and needs additional code to
11694 be useful, such as an ASN1 bio and some enhanced streaming
11697 Extend template encode functionality so that tagging is passed
11698 down to the template encoder.
11702 * Let 'openssl req' fail if an argument to '-newkey' is not
11703 recognized instead of using RSA as a default.
11707 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11708 As these are not official, they are not included in "ALL";
11709 the "ECCdraft" ciphersuite group alias can be used to select them.
11711 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11713 * Add ECDH engine support.
11715 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11717 * Add ECDH in new directory crypto/ecdh/.
11719 *Douglas Stebila (Sun Microsystems Laboratories)*
11721 * Let BN_rand_range() abort with an error after 100 iterations
11722 without success (which indicates a broken PRNG).
11726 * Change BN_mod_sqrt() so that it verifies that the input value
11727 is really the square of the return value. (Previously,
11728 BN_mod_sqrt would show GIGO behaviour.)
11732 * Add named elliptic curves over binary fields from X9.62, SECG,
11733 and WAP/WTLS; add OIDs that were still missing.
11735 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11737 * Extend the EC library for elliptic curves over binary fields
11738 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11741 EC_GF2m_simple_method
11745 EC_GROUP_new_curve_GF2m
11746 EC_GROUP_set_curve_GF2m
11747 EC_GROUP_get_curve_GF2m
11748 EC_POINT_set_affine_coordinates_GF2m
11749 EC_POINT_get_affine_coordinates_GF2m
11750 EC_POINT_set_compressed_coordinates_GF2m
11752 Point compression for binary fields is disabled by default for
11753 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11756 As binary polynomials are represented as BIGNUMs, various members
11757 of the EC_GROUP and EC_POINT data structures can be shared
11758 between the implementations for prime fields and binary fields;
11759 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11760 are essentially identical to their `..._GFp` counterparts.
11761 (For simplicity, the `..._GFp` prefix has been dropped from
11762 various internal method names.)
11764 An internal 'field_div' method (similar to 'field_mul' and
11765 'field_sqr') has been added; this is used only for binary fields.
11767 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11769 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11770 through methods ('mul', 'precompute_mult').
11772 The generic implementations (now internally called 'ec_wNAF_mul'
11773 and 'ec_wNAF_precomputed_mult') remain the default if these
11774 methods are undefined.
11776 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11778 * New function EC_GROUP_get_degree, which is defined through
11779 EC_METHOD. For curves over prime fields, this returns the bit
11780 length of the modulus.
11782 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11784 * New functions EC_GROUP_dup, EC_POINT_dup.
11785 (These simply call ..._new and ..._copy).
11787 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11789 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11790 Polynomials are represented as BIGNUMs (where the sign bit is not
11791 used) in the following functions [macros]:
11794 BN_GF2m_sub [= BN_GF2m_add]
11795 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11796 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11797 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11799 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11800 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11801 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11802 BN_GF2m_cmp [= BN_ucmp]
11804 (Note that only the 'mod' functions are actually for fields GF(2^m).
11805 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11807 For some functions, an the irreducible polynomial defining a
11808 field can be given as an 'unsigned int[]' with strictly
11809 decreasing elements giving the indices of those bits that are set;
11810 i.e., p[] represents the polynomial
11811 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11813 p[0] > p[1] > ... > p[k] = 0.
11814 This applies to the following functions:
11817 BN_GF2m_mod_mul_arr
11818 BN_GF2m_mod_sqr_arr
11819 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11820 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11821 BN_GF2m_mod_exp_arr
11822 BN_GF2m_mod_sqrt_arr
11823 BN_GF2m_mod_solve_quad_arr
11827 Conversion can be performed by the following functions:
11832 bntest.c has additional tests for binary polynomial arithmetic.
11834 Two implementations for BN_GF2m_mod_div() are available.
11835 The default algorithm simply uses BN_GF2m_mod_inv() and
11836 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11837 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11838 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11840 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11842 * Add new error code 'ERR_R_DISABLED' that can be used when some
11843 functionality is disabled at compile-time.
11845 *Douglas Stebila <douglas.stebila@sun.com>*
11847 * Change default behaviour of 'openssl asn1parse' so that more
11848 information is visible when viewing, e.g., a certificate:
11850 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11851 mode the content of non-printable OCTET STRINGs is output in a
11852 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11853 avoid the appearance of a printable string.
11855 *Nils Larsch <nla@trustcenter.de>*
11857 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11859 EC_GROUP_set_asn1_flag()
11860 EC_GROUP_get_asn1_flag()
11861 EC_GROUP_set_point_conversion_form()
11862 EC_GROUP_get_point_conversion_form()
11863 These control ASN1 encoding details:
11864 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11865 has been set to OPENSSL_EC_NAMED_CURVE.
11866 - Points are encoded in uncompressed form by default; options for
11867 asn1_for are as for point2oct, namely
11868 POINT_CONVERSION_COMPRESSED
11869 POINT_CONVERSION_UNCOMPRESSED
11870 POINT_CONVERSION_HYBRID
11872 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11874 EC_GROUP_set_seed()
11875 EC_GROUP_get0_seed()
11876 EC_GROUP_get_seed_len()
11877 This is used only for ASN1 purposes (so far).
11879 *Nils Larsch <nla@trustcenter.de>*
11881 * Add 'field_type' member to EC_METHOD, which holds the NID
11882 of the appropriate field type OID. The new function
11883 EC_METHOD_get_field_type() returns this value.
11885 *Nils Larsch <nla@trustcenter.de>*
11888 EC_POINT_point2bn()
11889 EC_POINT_bn2point()
11890 EC_POINT_point2hex()
11891 EC_POINT_hex2point()
11892 providing useful interfaces to EC_POINT_point2oct() and
11893 EC_POINT_oct2point().
11895 *Nils Larsch <nla@trustcenter.de>*
11897 * Change internals of the EC library so that the functions
11898 EC_GROUP_set_generator()
11899 EC_GROUP_get_generator()
11900 EC_GROUP_get_order()
11901 EC_GROUP_get_cofactor()
11902 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11903 to methods, which would lead to unnecessary code duplication when
11904 adding different types of curves.
11906 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11908 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11909 arithmetic, and such that modified wNAFs are generated
11910 (which avoid length expansion in many cases).
11914 * Add a function EC_GROUP_check_discriminant() (defined via
11915 EC_METHOD) that verifies that the curve discriminant is non-zero.
11917 Add a function EC_GROUP_check() that makes some sanity tests
11918 on a EC_GROUP, its generator and order. This includes
11919 EC_GROUP_check_discriminant().
11921 *Nils Larsch <nla@trustcenter.de>*
11923 * Add ECDSA in new directory crypto/ecdsa/.
11925 Add applications 'openssl ecparam' and 'openssl ecdsa'
11926 (these are based on 'openssl dsaparam' and 'openssl dsa').
11928 ECDSA support is also included in various other files across the
11929 library. Most notably,
11930 - 'openssl req' now has a '-newkey ecdsa:file' option;
11931 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11932 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11933 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11934 them suitable for ECDSA where domain parameters must be
11935 extracted before the specific public key;
11936 - ECDSA engine support has been added.
11938 *Nils Larsch <nla@trustcenter.de>*
11940 * Include some named elliptic curves, and add OIDs from X9.62,
11941 SECG, and WAP/WTLS. Each curve can be obtained from the new
11943 EC_GROUP_new_by_curve_name(),
11944 and the list of available named curves can be obtained with
11945 EC_get_builtin_curves().
11946 Also add a 'curve_name' member to EC_GROUP objects, which can be
11948 EC_GROUP_set_curve_name()
11949 EC_GROUP_get_curve_name()
11951 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11953 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11954 was actually never needed) and in BN_mul(). The removal in BN_mul()
11955 required a small change in bn_mul_part_recursive() and the addition
11956 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11957 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11958 bn_sub_words() and bn_add_words() except they take arrays with
11963 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11965 * Cleanse PEM buffers before freeing them since they may contain
11968 *Benjamin Bennett <ben@psc.edu>*
11970 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11971 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11972 authentication-only ciphersuites.
11976 * Since AES128 and AES256 share a single mask bit in the logic of
11977 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11978 kludge to work properly if AES128 is available and AES256 isn't.
11982 * Expand security boundary to match 1.1.1 module.
11986 * Remove redundant features: hash file source, editing of test vectors
11987 modify fipsld to use external fips_premain.c signature.
11991 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11992 run algorithm test programs.
11996 * Make algorithm test programs more tolerant of whitespace.
12000 * Have SSL/TLS server implementation tolerate "mismatched" record
12001 protocol version while receiving ClientHello even if the
12002 ClientHello is fragmented. (The server can't insist on the
12003 particular protocol version it has chosen before the ServerHello
12004 message has informed the client about his choice.)
12008 * Load error codes if they are not already present instead of using a
12009 static variable. This allows them to be cleanly unloaded and reloaded.
12013 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12015 * Introduce limits to prevent malicious keys being able to
12016 cause a denial of service. ([CVE-2006-2940])
12018 *Steve Henson, Bodo Moeller*
12020 * Fix ASN.1 parsing of certain invalid structures that can result
12021 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12023 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12024 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12026 * Fix SSL client code which could crash if connecting to a
12027 malicious SSLv2 server. ([CVE-2006-4343])
12029 *Tavis Ormandy and Will Drewry, Google Security Team*
12031 * Change ciphersuite string processing so that an explicit
12032 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12033 will no longer include "AES128-SHA"), and any other similar
12034 ciphersuite (same bitmap) from *other* protocol versions (so that
12035 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12036 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12037 changes from 0.9.8b and 0.9.8d.
12041 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12043 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12044 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12046 * Change the Unix randomness entropy gathering to use poll() when
12047 possible instead of select(), since the latter has some
12048 undesirable limitations.
12050 *Darryl Miles via Richard Levitte and Bodo Moeller*
12052 * Disable rogue ciphersuites:
12054 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12055 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12056 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12058 The latter two were purportedly from
12059 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12062 Also deactivate the remaining ciphersuites from
12063 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12064 unofficial, and the ID has long expired.
12068 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12069 dual-core machines) and other potential thread-safety issues.
12073 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12075 * Adapt fipsld and the build system to link against the validated FIPS
12076 module in FIPS mode.
12080 * Fixes for VC++ 2005 build under Windows.
12084 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12085 from a Windows bash shell such as MSYS. It is autodetected from the
12086 "config" script when run from a VC++ environment. Modify standard VC++
12087 build to use fipscanister.o from the GNU make build.
12091 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12093 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12094 The value now differs depending on if you build for FIPS or not.
12095 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12096 safely run with a non-FIPSed libcrypto, as it may crash because of
12097 the difference induced by this change.
12101 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12103 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12104 (part of SSL_OP_ALL). This option used to disable the
12105 countermeasure against man-in-the-middle protocol-version
12106 rollback in the SSL 2.0 server implementation, which is a bad
12107 idea. ([CVE-2005-2969])
12109 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12110 for Information Security, National Institute of Advanced Industrial
12111 Science and Technology [AIST, Japan)]*
12113 * Minimal support for X9.31 signatures and PSS padding modes. This is
12114 mainly for FIPS compliance and not fully integrated at this stage.
12118 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12119 the exponentiation using a fixed-length exponent. (Otherwise,
12120 the information leaked through timing could expose the secret key
12121 after many signatures; cf. Bleichenbacher's attack on DSA with
12126 * Make a new fixed-window mod_exp implementation the default for
12127 RSA, DSA, and DH private-key operations so that the sequence of
12128 squares and multiplies and the memory access pattern are
12129 independent of the particular secret key. This will mitigate
12130 cache-timing and potential related attacks.
12132 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12133 and this is automatically used by BN_mod_exp_mont() if the new flag
12134 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12135 will use this BN flag for private exponents unless the flag
12136 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12137 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12139 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12141 * Change the client implementation for SSLv23_method() and
12142 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12143 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12144 (Previously, the SSL 2.0 backwards compatible Client Hello
12145 message format would be used even with SSL_OP_NO_SSLv2.)
12149 * Add support for smime-type MIME parameter in S/MIME messages which some
12154 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12155 a threadsafe manner. Modify rsa code to use new function and add calls
12156 to dsa and dh code (which had race conditions before).
12160 * Include the fixed error library code in the C error file definitions
12161 instead of fixing them up at runtime. This keeps the error code
12162 structures constant.
12166 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12168 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12171 * Fixes for newer kerberos headers. NB: the casts are needed because
12172 the 'length' field is signed on one version and unsigned on another
12173 with no (?) obvious way to tell the difference, without these VC++
12174 complains. Also the "definition" of FAR (blank) is no longer included
12175 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12176 some needed definitions.
12180 * Undo Cygwin change.
12184 * Added support for proxy certificates according to RFC 3820.
12185 Because they may be a security thread to unaware applications,
12186 they must be explicitly allowed in run-time. See
12187 docs/HOWTO/proxy_certificates.txt for further information.
12191 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12193 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12194 server and client random values. Previously
12195 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12196 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12198 This change has negligible security impact because:
12200 1. Server and client random values still have 24 bytes of pseudo random
12203 2. Server and client random values are sent in the clear in the initial
12206 3. The master secret is derived using the premaster secret (48 bytes in
12207 size for static RSA ciphersuites) as well as client server and random
12210 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12213 *Stephen Henson, reported by UK NISCC*
12215 * Use Windows randomness collection on Cygwin.
12219 * Fix hang in EGD/PRNGD query when communication socket is closed
12220 prematurely by EGD/PRNGD.
12222 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12224 * Prompt for pass phrases when appropriate for PKCS12 input format.
12228 * Back-port of selected performance improvements from development
12229 branch, as well as improved support for PowerPC platforms.
12233 * Add lots of checks for memory allocation failure, error codes to indicate
12234 failure and freeing up memory if a failure occurs.
12236 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12238 * Add new -passin argument to dgst.
12242 * Perform some character comparisons of different types in X509_NAME_cmp:
12243 this is needed for some certificates that re-encode DNs into UTF8Strings
12244 (in violation of RFC3280) and can't or won't issue name rollover
12249 * Make an explicit check during certificate validation to see that
12250 the CA setting in each certificate on the chain is correct. As a
12251 side effect always do the following basic checks on extensions,
12252 not just when there's an associated purpose to the check:
12254 - if there is an unhandled critical extension (unless the user
12255 has chosen to ignore this fault)
12256 - if the path length has been exceeded (if one is set at all)
12257 - that certain extensions fit the associated purpose (if one has
12262 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12264 * Avoid a race condition when CRLs are checked in a multi threaded
12265 environment. This would happen due to the reordering of the revoked
12266 entries during signature checking and serial number lookup. Now the
12267 encoding is cached and the serial number sort performed under a lock.
12268 Add new STACK function sk_is_sorted().
12272 * Add Delta CRL to the extension code.
12276 * Various fixes to s3_pkt.c so alerts are sent properly.
12278 *David Holmes <d.holmes@f5.com>*
12280 * Reduce the chances of duplicate issuer name and serial numbers (in
12281 violation of RFC3280) using the OpenSSL certificate creation utilities.
12282 This is done by creating a random 64 bit value for the initial serial
12283 number when a serial number file is created or when a self signed
12284 certificate is created using 'openssl req -x509'. The initial serial
12285 number file is created using 'openssl x509 -next_serial' in CA.pl
12286 rather than being initialized to 1.
12290 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12292 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12293 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12295 *Joe Orton, Steve Henson*
12297 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12300 *Joe Orton, Steve Henson*
12302 * Make it possible to have multiple active certificates with the same
12303 subject in the CA index file. This is done only if the keyword
12304 'unique_subject' is set to 'no' in the main CA section (default
12305 if 'CA_default') of the configuration file. The value is saved
12306 with the database itself in a separate index attribute file,
12307 named like the index file with '.attr' appended to the name.
12311 * X509 verify fixes. Disable broken certificate workarounds when
12312 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12313 keyUsage extension present. Don't accept CRLs with unhandled critical
12314 extensions: since verify currently doesn't process CRL extensions this
12315 rejects a CRL with *any* critical extensions. Add new verify error codes
12320 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12321 A clarification of RFC2560 will require the use of OCTET STRINGs and
12322 some implementations cannot handle the current raw format. Since OpenSSL
12323 copies and compares OCSP nonces as opaque blobs without any attempt at
12324 parsing them this should not create any compatibility issues.
12328 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12329 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12330 this HMAC (and other) operations are several times slower than OpenSSL
12335 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12337 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12339 * Use the correct content when signing type "other".
12343 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12345 * Fix various bugs revealed by running the NISCC test suite:
12347 Stop out of bounds reads in the ASN1 code when presented with
12348 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12350 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12352 If verify callback ignores invalid public key errors don't try to check
12353 certificate signature with the NULL public key.
12357 * New -ignore_err option in ocsp application to stop the server
12358 exiting on the first error in a request.
12362 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12363 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12368 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12369 extra data after the compression methods not only for TLS 1.0
12370 but also for SSL 3.0 (as required by the specification).
12372 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12374 * Change X509_certificate_type() to mark the key as exported/exportable
12375 when it's 512 *bits* long, not 512 bytes.
12379 * Change AES_cbc_encrypt() so it outputs exact multiple of
12380 blocks during encryption.
12384 * Various fixes to base64 BIO and non blocking I/O. On write
12385 flushes were not handled properly if the BIO retried. On read
12386 data was not being buffered properly and had various logic bugs.
12387 This also affects blocking I/O when the data being decoded is a
12392 * Various S/MIME bugfixes and compatibility changes:
12393 output correct application/pkcs7 MIME type if
12394 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12395 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12396 of files as .eml work). Correctly handle very long lines in MIME
12401 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12403 * Countermeasure against the Klima-Pokorny-Rosa extension of
12404 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12405 a protocol version number mismatch like a decryption error
12406 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12410 * Turn on RSA blinding by default in the default implementation
12411 to avoid a timing attack. Applications that don't want it can call
12412 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12413 They would be ill-advised to do so in most cases.
12415 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12417 * Change RSA blinding code so that it works when the PRNG is not
12418 seeded (in this case, the secret RSA exponent is abused as
12419 an unpredictable seed -- if it is not unpredictable, there
12420 is no point in blinding anyway). Make RSA blinding thread-safe
12421 by remembering the creator's thread ID in rsa->blinding and
12422 having all other threads use local one-time blinding factors
12423 (this requires more computation than sharing rsa->blinding, but
12424 avoids excessive locking; and if an RSA object is not shared
12425 between threads, blinding will still be very fast).
12429 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12430 ENGINE as defaults for all supported algorithms irrespective of
12431 the 'flags' parameter. 'flags' is now honoured, so applications
12432 should make sure they are passing it correctly.
12436 * Target "mingw" now allows native Windows code to be generated in
12437 the Cygwin environment as well as with the MinGW compiler.
12441 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12443 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12444 via timing by performing a MAC computation even if incorrect
12445 block cipher padding has been found. This is a countermeasure
12446 against active attacks where the attacker has to distinguish
12447 between bad padding and a MAC verification error. ([CVE-2003-0078])
12449 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12450 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12451 Martin Vuagnoux (EPFL, Ilion)*
12453 * Make the no-err option work as intended. The intention with no-err
12454 is not to have the whole error stack handling routines removed from
12455 libcrypto, it's only intended to remove all the function name and
12456 reason texts, thereby removing some of the footprint that may not
12457 be interesting if those errors aren't displayed anyway.
12459 NOTE: it's still possible for any application or module to have its
12460 own set of error texts inserted. The routines are there, just not
12461 used by default when no-err is given.
12465 * Add support for FreeBSD on IA64.
12467 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12469 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12470 Kerberos function mit_des_cbc_cksum(). Before this change,
12471 the value returned by DES_cbc_cksum() was like the one from
12472 mit_des_cbc_cksum(), except the bytes were swapped.
12474 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12476 * Allow an application to disable the automatic SSL chain building.
12477 Before this a rather primitive chain build was always performed in
12478 ssl3_output_cert_chain(): an application had no way to send the
12479 correct chain if the automatic operation produced an incorrect result.
12481 Now the chain builder is disabled if either:
12483 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12485 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12487 The reasoning behind this is that an application would not want the
12488 auto chain building to take place if extra chain certificates are
12489 present and it might also want a means of sending no additional
12490 certificates (for example the chain has two certificates and the
12495 * Add the possibility to build without the ENGINE framework.
12497 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12499 * Under Win32 gmtime() can return NULL: check return value in
12500 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12504 * DSA routines: under certain error conditions uninitialized BN objects
12505 could be freed. Solution: make sure initialization is performed early
12506 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12507 Nils Larsch <nla@trustcenter.de> via PR#459)
12511 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12512 checked on reconnect on the client side, therefore session resumption
12513 could still fail with a "ssl session id is different" error. This
12514 behaviour is masked when SSL_OP_ALL is used due to
12515 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12516 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12517 followup to PR #377.
12521 * IA-32 assembler support enhancements: unified ELF targets, support
12522 for SCO/Caldera platforms, fix for Cygwin shared build.
12526 * Add support for FreeBSD on sparc64. As a consequence, support for
12527 FreeBSD on non-x86 processors is separate from x86 processors on
12528 the config script, much like the NetBSD support.
12530 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12532 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12534 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12537 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12538 code (06) was taken as the first octet of the session ID and the last
12539 octet was ignored consequently. As a result SSLv2 client side session
12540 caching could not have worked due to the session ID mismatch between
12542 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12547 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12548 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12553 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12554 seems that in spite of existing for more than a year, many application
12555 author have done nothing to provide the necessary callbacks, which
12556 means that this particular engine will not work properly anywhere.
12557 This is a very unfortunate situation which forces us, in the name
12558 of usability, to give the hw_ncipher.c a static lock, which is part
12560 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12561 appear in 0.9.8 or later. We EXPECT application authors to have
12562 dealt properly with this when 0.9.8 is released (unless we actually
12563 make such changes in the libcrypto locking code that changes will
12564 have to be made anyway).
12568 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12569 octets have been read, EOF or an error occurs. Without this change
12570 some truncated ASN1 structures will not produce an error.
12574 * Disable Heimdal support, since it hasn't been fully implemented.
12575 Still give the possibility to force the use of Heimdal, but with
12576 warnings and a request that patches get sent to openssl-dev.
12580 * Add the VC-CE target, introduce the WINCE sysname, and add
12581 INSTALL.WCE and appropriate conditionals to make it build.
12583 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12585 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12586 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12587 edit numbers of the version.
12589 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12591 * Introduce safe string copy and catenation functions
12592 (BUF_strlcpy() and BUF_strlcat()).
12594 *Ben Laurie (CHATS) and Richard Levitte*
12596 * Avoid using fixed-size buffers for one-line DNs.
12598 *Ben Laurie (CHATS)*
12600 * Add BUF_MEM_grow_clean() to avoid information leakage when
12601 resizing buffers containing secrets, and use where appropriate.
12603 *Ben Laurie (CHATS)*
12605 * Avoid using fixed size buffers for configuration file location.
12607 *Ben Laurie (CHATS)*
12609 * Avoid filename truncation for various CA files.
12611 *Ben Laurie (CHATS)*
12613 * Use sizeof in preference to magic numbers.
12615 *Ben Laurie (CHATS)*
12617 * Avoid filename truncation in cert requests.
12619 *Ben Laurie (CHATS)*
12621 * Add assertions to check for (supposedly impossible) buffer
12624 *Ben Laurie (CHATS)*
12626 * Don't cache truncated DNS entries in the local cache (this could
12627 potentially lead to a spoofing attack).
12629 *Ben Laurie (CHATS)*
12631 * Fix various buffers to be large enough for hex/decimal
12632 representations in a platform independent manner.
12634 *Ben Laurie (CHATS)*
12636 * Add CRYPTO_realloc_clean() to avoid information leakage when
12637 resizing buffers containing secrets, and use where appropriate.
12639 *Ben Laurie (CHATS)*
12641 * Add BIO_indent() to avoid much slightly worrying code to do
12644 *Ben Laurie (CHATS)*
12646 * Convert sprintf()/BIO_puts() to BIO_printf().
12648 *Ben Laurie (CHATS)*
12650 * buffer_gets() could terminate with the buffer only half
12653 *Ben Laurie (CHATS)*
12655 * Add assertions to prevent user-supplied crypto functions from
12656 overflowing internal buffers by having large block sizes, etc.
12658 *Ben Laurie (CHATS)*
12660 * New OPENSSL_assert() macro (similar to assert(), but enabled
12663 *Ben Laurie (CHATS)*
12665 * Eliminate unused copy of key in RC4.
12667 *Ben Laurie (CHATS)*
12669 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12671 *Ben Laurie (CHATS)*
12673 * Fix off-by-one error in EGD path.
12675 *Ben Laurie (CHATS)*
12677 * If RANDFILE path is too long, ignore instead of truncating.
12679 *Ben Laurie (CHATS)*
12681 * Eliminate unused and incorrectly sized X.509 structure
12684 *Ben Laurie (CHATS)*
12686 * Eliminate unused and dangerous function knumber().
12688 *Ben Laurie (CHATS)*
12690 * Eliminate unused and dangerous structure, KSSL_ERR.
12692 *Ben Laurie (CHATS)*
12694 * Protect against overlong session ID context length in an encoded
12695 session object. Since these are local, this does not appear to be
12698 *Ben Laurie (CHATS)*
12700 * Change from security patch (see 0.9.6e below) that did not affect
12701 the 0.9.6 release series:
12703 Remote buffer overflow in SSL3 protocol - an attacker could
12704 supply an oversized master key in Kerberos-enabled versions.
12707 *Ben Laurie (CHATS)*
12709 * Change the SSL kerb5 codes to match RFC 2712.
12713 * Make -nameopt work fully for req and add -reqopt switch.
12715 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12717 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12719 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12721 * Make sure tests can be performed even if the corresponding algorithms
12722 have been removed entirely. This was also the last step to make
12723 OpenSSL compilable with DJGPP under all reasonable conditions.
12725 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12727 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12728 to allow version independent disabling of normally unselected ciphers,
12729 which may be activated as a side-effect of selecting a single cipher.
12731 (E.g., cipher list string "RSA" enables ciphersuites that are left
12732 out of "ALL" because they do not provide symmetric encryption.
12733 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12735 *Lutz Jaenicke, Bodo Moeller*
12737 * Add appropriate support for separate platform-dependent build
12738 directories. The recommended way to make a platform-dependent
12739 build directory is the following (tested on Linux), maybe with
12742 # Place yourself outside of the OpenSSL source tree. In
12743 # this example, the environment variable OPENSSL_SOURCE
12744 # is assumed to contain the absolute OpenSSL source directory.
12745 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12746 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12747 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12748 mkdir -p `dirname $F`
12749 ln -s $OPENSSL_SOURCE/$F $F
12752 To be absolutely sure not to disturb the source tree, a "make clean"
12753 is a good thing. If it isn't successful, don't worry about it,
12754 it probably means the source directory is very clean.
12758 * Make sure any ENGINE control commands make local copies of string
12759 pointers passed to them whenever necessary. Otherwise it is possible
12760 the caller may have overwritten (or deallocated) the original string
12761 data when a later ENGINE operation tries to use the stored values.
12763 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12765 * Improve diagnostics in file reading and command-line digests.
12767 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12769 * Add AES modes CFB and OFB to the object database. Correct an
12770 error in AES-CFB decryption.
12774 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12775 allows existing EVP_CIPHER_CTX structures to be reused after
12776 calling `EVP_*Final()`. This behaviour is used by encryption
12777 BIOs and some applications. This has the side effect that
12778 applications must explicitly clean up cipher contexts with
12779 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12783 * Check the values of dna and dnb in bn_mul_recursive before calling
12784 bn_mul_comba (a non zero value means the a or b arrays do not contain
12785 n2 elements) and fallback to bn_mul_normal if either is not zero.
12789 * Fix escaping of non-ASCII characters when using the -subj option
12790 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12794 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12795 form for "surname", serialNumber has no short form.
12796 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12797 therefore remove "mail" short name for "internet 7".
12798 The OID for unique identifiers in X509 certificates is
12799 x500UniqueIdentifier, not uniqueIdentifier.
12800 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12804 * Add an "init" command to the ENGINE config module and auto initialize
12805 ENGINEs. Without any "init" command the ENGINE will be initialized
12806 after all ctrl commands have been executed on it. If init=1 the
12807 ENGINE is initialized at that point (ctrls before that point are run
12808 on the uninitialized ENGINE and after on the initialized one). If
12809 init=0 then the ENGINE will not be initialized at all.
12813 * Fix the 'app_verify_callback' interface so that the user-defined
12814 argument is actually passed to the callback: In the
12815 SSL_CTX_set_cert_verify_callback() prototype, the callback
12816 declaration has been changed from
12819 int (*cb)(X509_STORE_CTX *,void *);
12820 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12821 i=s->ctx->app_verify_callback(&ctx)
12822 has been changed into
12823 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12825 To update applications using SSL_CTX_set_cert_verify_callback(),
12826 a dummy argument can be added to their callback functions.
12828 *D. K. Smetters <smetters@parc.xerox.com>*
12830 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12832 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12834 * Add and OPENSSL_LOAD_CONF define which will cause
12835 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12836 This allows older applications to transparently support certain
12837 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12838 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12839 load the config file and OPENSSL_add_all_algorithms_conf() which will
12840 always load it have also been added.
12844 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12845 Adjust NIDs and EVP layer.
12847 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12849 * Config modules support in openssl utility.
12851 Most commands now load modules from the config file,
12852 though in a few (such as version) this isn't done
12853 because it couldn't be used for anything.
12855 In the case of ca and req the config file used is
12856 the same as the utility itself: that is the -config
12857 command line option can be used to specify an
12862 * Move default behaviour from OPENSSL_config(). If appname is NULL
12863 use "openssl_conf" if filename is NULL use default openssl config file.
12867 * Add an argument to OPENSSL_config() to allow the use of an alternative
12868 config section name. Add a new flag to tolerate a missing config file
12869 and move code to CONF_modules_load_file().
12873 * Support for crypto accelerator cards from Accelerated Encryption
12874 Processing, www.aep.ie. (Use engine 'aep')
12875 The support was copied from 0.9.6c [engine] and adapted/corrected
12876 to work with the new engine framework.
12878 *AEP Inc. and Richard Levitte*
12880 * Support for SureWare crypto accelerator cards from Baltimore
12881 Technologies. (Use engine 'sureware')
12882 The support was copied from 0.9.6c [engine] and adapted
12883 to work with the new engine framework.
12887 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12888 make the newer ENGINE framework commands for the CHIL engine work.
12890 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12892 * Make it possible to produce shared libraries on ReliantUNIX.
12894 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12896 * Add the configuration target debug-linux-ppro.
12897 Make 'openssl rsa' use the general key loading routines
12898 implemented in `apps.c`, and make those routines able to
12899 handle the key format FORMAT_NETSCAPE and the variant
12902 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12904 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12906 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12908 * Add -keyform to rsautl, and document -engine.
12910 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12912 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12913 BIO_R_NO_SUCH_FILE error code rather than the generic
12914 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12918 * Add new functions
12919 ERR_peek_last_error
12920 ERR_peek_last_error_line
12921 ERR_peek_last_error_line_data.
12922 These are similar to
12924 ERR_peek_error_line
12925 ERR_peek_error_line_data,
12926 but report on the latest error recorded rather than the first one
12927 still in the error queue.
12929 *Ben Laurie, Bodo Moeller*
12931 * default_algorithms option in ENGINE config module. This allows things
12933 default_algorithms = ALL
12934 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12938 * Preliminary ENGINE config module.
12942 * New experimental application configuration code.
12946 * Change the AES code to follow the same name structure as all other
12947 symmetric ciphers, and behave the same way. Move everything to
12948 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12950 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12952 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12954 *Ben Laurie and Theo de Raadt*
12956 * Add option to output public keys in req command.
12958 *Massimiliano Pala madwolf@openca.org*
12960 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12961 (up to about 10% better than before for P-192 and P-224).
12965 * New functions/macros
12967 SSL_CTX_set_msg_callback(ctx, cb)
12968 SSL_CTX_set_msg_callback_arg(ctx, arg)
12969 SSL_set_msg_callback(ssl, cb)
12970 SSL_set_msg_callback_arg(ssl, arg)
12972 to request calling a callback function
12974 void cb(int write_p, int version, int content_type,
12975 const void *buf, size_t len, SSL *ssl, void *arg)
12977 whenever a protocol message has been completely received
12978 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12979 protocol version according to which the SSL library interprets
12980 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12981 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12982 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12983 specification (change_cipher_spec(20), alert(21), handshake(22)).
12984 'buf' and 'len' point to the actual message, 'ssl' to the
12985 SSL object, and 'arg' is the application-defined value set by
12986 SSL[_CTX]_set_msg_callback_arg().
12988 'openssl s_client' and 'openssl s_server' have new '-msg' options
12989 to enable a callback that displays all protocol messages.
12993 * Change the shared library support so shared libraries are built as
12994 soon as the corresponding static library is finished, and thereby get
12995 openssl and the test programs linked against the shared library.
12996 This still only happens when the keyword "shard" has been given to
12997 the configuration scripts.
12999 NOTE: shared library support is still an experimental thing, and
13000 backward binary compatibility is still not guaranteed.
13002 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13004 * Add support for Subject Information Access extension.
13006 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13008 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13009 additional bytes when new memory had to be allocated, not just
13010 when reusing an existing buffer.
13014 * New command line and configuration option 'utf8' for the req command.
13015 This allows field values to be specified as UTF8 strings.
13019 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13020 runs for the former and machine-readable output for the latter.
13024 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13025 of the e-mail address in the DN (i.e., it will go into a certificate
13026 extension only). The new configuration file option 'email_in_dn = no'
13027 has the same effect.
13029 *Massimiliano Pala madwolf@openca.org*
13031 * Change all functions with names starting with `des_` to be starting
13032 with `DES_` instead. Add wrappers that are compatible with libdes,
13033 but are named `_ossl_old_des_*`. Finally, add macros that map the
13034 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13035 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13036 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13039 Since we provide two compatibility mappings, the user needs to
13040 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13041 compatibility is desired. The default (i.e., when that macro
13042 isn't defined) is OpenSSL 0.9.6c compatibility.
13044 There are also macros that enable and disable the support of old
13045 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13046 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13047 are defined, the default will apply: to support the old des routines.
13049 In either case, one must include openssl/des.h to get the correct
13050 definitions. Do not try to just include openssl/des_old.h, that
13053 NOTE: This is a major break of an old API into a new one. Software
13054 authors are encouraged to switch to the `DES_` style functions. Some
13055 time in the future, des_old.h and the libdes compatibility functions
13056 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13057 default), and then completely removed.
13061 * Test for certificates which contain unsupported critical extensions.
13062 If such a certificate is found during a verify operation it is
13063 rejected by default: this behaviour can be overridden by either
13064 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13065 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13066 X509_supported_extension() has also been added which returns 1 if a
13067 particular extension is supported.
13071 * Modify the behaviour of EVP cipher functions in similar way to digests
13072 to retain compatibility with existing code.
13076 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13077 compatibility with existing code. In particular the 'ctx' parameter does
13078 not have to be to be initialized before the call to EVP_DigestInit() and
13079 it is tidied up after a call to EVP_DigestFinal(). New function
13080 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13081 EVP_MD_CTX_copy() changed to not require the destination to be
13082 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13083 requires the destination to be valid.
13085 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13086 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13090 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13091 so that complete 'Handshake' protocol structures are kept in memory
13092 instead of overwriting 'msg_type' and 'length' with 'body' data.
13096 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13098 *Massimo Santin via Richard Levitte*
13100 * Major restructuring to the underlying ENGINE code. This includes
13101 reduction of linker bloat, separation of pure "ENGINE" manipulation
13102 (initialisation, etc) from functionality dealing with implementations
13103 of specific crypto interfaces. This change also introduces integrated
13104 support for symmetric ciphers and digest implementations - so ENGINEs
13105 can now accelerate these by providing EVP_CIPHER and EVP_MD
13106 implementations of their own. This is detailed in
13107 [crypto/engine/README.md](crypto/engine/README.md)
13108 as it couldn't be adequately described here. However, there are a few
13109 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13110 were changed in the original introduction of ENGINE code have now
13111 reverted back - the hooking from this code to ENGINE is now a good
13112 deal more passive and at run-time, operations deal directly with
13113 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13114 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13115 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13116 they were not being used by the framework as there is no concept of a
13117 BIGNUM_METHOD and they could not be generalised to the new
13118 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13119 ENGINE_cpy() has been removed as it cannot be consistently defined in
13124 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13128 * Change mkdef.pl to sort symbols that get the same entry number,
13129 and make sure the automatically generated functions `ERR_load_*`
13130 become part of libeay.num as well.
13134 * New function SSL_renegotiate_pending(). This returns true once
13135 renegotiation has been requested (either SSL_renegotiate() call
13136 or HelloRequest/ClientHello received from the peer) and becomes
13137 false once a handshake has been completed.
13138 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13139 sends a HelloRequest, but does not ensure that a handshake takes
13140 place. SSL_renegotiate_pending() is useful for checking if the
13141 client has followed the request.)
13145 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13146 By default, clients may request session resumption even during
13147 renegotiation (if session ID contexts permit); with this option,
13148 session resumption is possible only in the first handshake.
13150 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13151 more bits available for options that should not be part of
13152 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13156 * Add some demos for certificate and certificate request creation.
13160 * Make maximum certificate chain size accepted from the peer application
13161 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13162 "Douglas E. Engert" <deengert@anl.gov>.
13166 * Add support for shared libraries for Unixware-7
13167 (Boyd Lynn Gerber <gerberb@zenez.com>).
13171 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13172 be done prior to destruction. Use this to unload error strings from
13173 ENGINEs that load their own error strings. NB: This adds two new API
13174 functions to "get" and "set" this destroy handler in an ENGINE.
13178 * Alter all existing ENGINE implementations (except "openssl" and
13179 "openbsd") to dynamically instantiate their own error strings. This
13180 makes them more flexible to be built both as statically-linked ENGINEs
13181 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13182 Also, add stub code to each that makes building them as self-contained
13183 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13187 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13188 implementations into applications that are completely implemented in
13189 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13190 commands that can be used to configure what shared-library to load and
13191 to control aspects of the way it is handled. Also, made an update to
13192 the [README-Engine.md](README-Engine.md) file
13193 that brings its information up-to-date and
13194 provides some information and instructions on the "dynamic" ENGINE
13195 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13199 * Make it possible to unload ranges of ERR strings with a new
13200 "ERR_unload_strings" function.
13204 * Add a copy() function to EVP_MD.
13208 * Make EVP_MD routines take a context pointer instead of just the
13209 md_data void pointer.
13213 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13214 that the digest can only process a single chunk of data
13215 (typically because it is provided by a piece of
13216 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13217 is only going to provide a single chunk of data, and hence the
13218 framework needn't accumulate the data for oneshot drivers.
13222 * As with "ERR", make it possible to replace the underlying "ex_data"
13223 functions. This change also alters the storage and management of global
13224 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13225 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13226 index counters. The API functions that use this state have been changed
13227 to take a "class_index" rather than pointers to the class's local STACK
13228 and counter, and there is now an API function to dynamically create new
13229 classes. This centralisation allows us to (a) plug a lot of the
13230 thread-safety problems that existed, and (b) makes it possible to clean
13231 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13232 such data would previously have always leaked in application code and
13233 workarounds were in place to make the memory debugging turn a blind eye
13234 to it. Application code that doesn't use this new function will still
13235 leak as before, but their memory debugging output will announce it now
13236 rather than letting it slide.
13238 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13239 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13240 has a return value to indicate success or failure.
13244 * Make it possible to replace the underlying "ERR" functions such that the
13245 global state (2 LHASH tables and 2 locks) is only used by the "default"
13246 implementation. This change also adds two functions to "get" and "set"
13247 the implementation prior to it being automatically set the first time
13248 any other ERR function takes place. Ie. an application can call "get",
13249 pass the return value to a module it has just loaded, and that module
13250 can call its own "set" function using that value. This means the
13251 module's "ERR" operations will use (and modify) the error state in the
13252 application and not in its own statically linked copy of OpenSSL code.
13256 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13257 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13258 the operation, and provides a more encapsulated way for external code
13259 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13260 to use these functions rather than manually incrementing the counts.
13262 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13266 * Add EVP test program.
13270 * Add symmetric cipher support to ENGINE. Expect the API to change!
13274 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13275 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13276 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13277 These allow a CRL to be built without having to access X509_CRL fields
13278 directly. Modify 'ca' application to use new functions.
13282 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13283 bug workarounds. Rollback attack detection is a security feature.
13284 The problem will only arise on OpenSSL servers when TLSv1 is not
13285 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13286 Software authors not wanting to support TLSv1 will have special reasons
13287 for their choice and can explicitly enable this option.
13289 *Bodo Moeller, Lutz Jaenicke*
13291 * Rationalise EVP so it can be extended: don't include a union of
13292 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13293 (similar to those existing for EVP_CIPHER_CTX).
13298 EVP_MD_CTX_init(&md); /* new function call */
13299 EVP_DigestInit(&md, EVP_sha1());
13300 EVP_DigestUpdate(&md, in, len);
13301 EVP_DigestFinal(&md, out, NULL);
13302 EVP_MD_CTX_cleanup(&md); /* new function call */
13306 * Make DES key schedule conform to the usual scheme, as well as
13307 correcting its structure. This means that calls to DES functions
13308 now have to pass a pointer to a des_key_schedule instead of a
13309 plain des_key_schedule (which was actually always a pointer
13312 des_key_schedule ks;
13314 des_set_key_checked(..., &ks);
13315 des_ncbc_encrypt(..., &ks, ...);
13317 (Note that a later change renames 'des_...' into 'DES_...'.)
13321 * Initial reduction of linker bloat: the use of some functions, such as
13322 PEM causes large amounts of unused functions to be linked in due to
13323 poor organisation. For example pem_all.c contains every PEM function
13324 which has a knock on effect of linking in large amounts of (unused)
13325 ASN1 code. Grouping together similar functions and splitting unrelated
13326 functions prevents this.
13330 * Cleanup of EVP macros.
13334 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13335 correct `_ecb suffix`.
13339 * Add initial OCSP responder support to ocsp application. The
13340 revocation information is handled using the text based index
13341 use by the ca application. The responder can either handle
13342 requests generated internally, supplied in files (for example
13343 via a CGI script) or using an internal minimal server.
13347 * Add configuration choices to get zlib compression for TLS.
13351 * Changes to Kerberos SSL for RFC 2712 compliance:
13352 1. Implemented real KerberosWrapper, instead of just using
13353 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13354 2. Implemented optional authenticator field of KerberosWrapper.
13356 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13357 and authenticator structs; see crypto/krb5/.
13359 Generalized Kerberos calls to support multiple Kerberos libraries.
13360 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13361 via Richard Levitte*
13363 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13364 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13365 values for each of the key sizes rather than having just
13366 parameters (and 'speed' generating keys each time).
13370 * Speed up EVP routines.
13373 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13374 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13375 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13376 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13378 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13379 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13380 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13383 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13385 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13389 * Added the OS2-EMX target.
13391 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13393 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13394 New functions to support `NCONF` routines in extension code.
13395 New function `CONF_set_nconf()`
13396 to allow functions which take an `NCONF` to also handle the old `LHASH`
13397 structure: this means that the old `CONF` compatible routines can be
13398 retained (in particular w.rt. extensions) without having to duplicate the
13399 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13403 * Enhance the general user interface with mechanisms for inner control
13404 and with possibilities to have yes/no kind of prompts.
13408 * Change all calls to low-level digest routines in the library and
13409 applications to use EVP. Add missing calls to HMAC_cleanup() and
13410 don't assume HMAC_CTX can be copied using memcpy().
13412 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13414 * Add the possibility to control engines through control names but with
13415 arbitrary arguments instead of just a string.
13416 Change the key loaders to take a UI_METHOD instead of a callback
13417 function pointer. NOTE: this breaks binary compatibility with earlier
13418 versions of OpenSSL [engine].
13419 Adapt the nCipher code for these new conditions and add a card insertion
13424 * Enhance the general user interface with mechanisms to better support
13425 dialog box interfaces, application-defined prompts, the possibility
13426 to use defaults (for example default passwords from somewhere else)
13427 and interrupts/cancellations.
13431 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13432 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13436 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13437 tidy up some unnecessarily weird code in 'sk_new()').
13439 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13441 * Change the key loading routines for ENGINEs to use the same kind
13442 callback (pem_password_cb) as all other routines that need this
13447 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13448 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13449 than this minimum value is recommended.
13453 * New random seeder for OpenVMS, using the system process statistics
13454 that are easily reachable.
13458 * Windows apparently can't transparently handle global
13459 variables defined in DLLs. Initialisations such as:
13461 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13463 won't compile. This is used by the any applications that need to
13464 declare their own ASN1 modules. This was fixed by adding the option
13465 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13466 needed for static libraries under Win32.
13470 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13471 setting of purpose and trust fields. New X509_STORE trust and
13472 purpose functions and tidy up setting in other SSL functions.
13476 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13477 structure. These are inherited by X509_STORE_CTX when it is
13478 initialised. This allows various defaults to be set in the
13479 X509_STORE structure (such as flags for CRL checking and custom
13480 purpose or trust settings) for functions which only use X509_STORE_CTX
13481 internally such as S/MIME.
13483 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13484 trust settings if they are not set in X509_STORE. This allows X509_STORE
13485 purposes and trust (in S/MIME for example) to override any set by default.
13487 Add command line options for CRL checking to smime, s_client and s_server
13492 * Initial CRL based revocation checking. If the CRL checking flag(s)
13493 are set then the CRL is looked up in the X509_STORE structure and
13494 its validity and signature checked, then if the certificate is found
13495 in the CRL the verify fails with a revoked error.
13497 Various new CRL related callbacks added to X509_STORE_CTX structure.
13499 Command line options added to 'verify' application to support this.
13501 This needs some additional work, such as being able to handle multiple
13502 CRLs with different times, extension based lookup (rather than just
13503 by subject name) and ultimately more complete V2 CRL extension
13508 * Add a general user interface API (crypto/ui/). This is designed
13509 to replace things like des_read_password and friends (backward
13510 compatibility functions using this new API are provided).
13511 The purpose is to remove prompting functions from the DES code
13512 section as well as provide for prompting through dialog boxes in
13513 a window system and the like.
13517 * Add "ex_data" support to ENGINE so implementations can add state at a
13518 per-structure level rather than having to store it globally.
13522 * Make it possible for ENGINE structures to be copied when retrieved by
13523 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13524 This causes the "original" ENGINE structure to act like a template,
13525 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13526 operational state can be localised to each ENGINE structure, despite the
13527 fact they all share the same "methods". New ENGINE structures returned in
13528 this case have no functional references and the return value is the single
13529 structural reference. This matches the single structural reference returned
13530 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13535 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13536 needs to match any other type at all we need to manually clear the
13541 * Changes to the "openssl engine" utility to include;
13542 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13543 about an ENGINE's available control commands.
13544 - executing control commands from command line arguments using the
13545 '-pre' and '-post' switches. '-post' is only used if '-t' is
13546 specified and the ENGINE is successfully initialised. The syntax for
13547 the individual commands are colon-separated, for example;
13548 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13552 * New dynamic control command support for ENGINEs. ENGINEs can now
13553 declare their own commands (numbers), names (strings), descriptions,
13554 and input types for run-time discovery by calling applications. A
13555 subset of these commands are implicitly classed as "executable"
13556 depending on their input type, and only these can be invoked through
13557 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13558 can be based on user input, config files, etc). The distinction is
13559 that "executable" commands cannot return anything other than a boolean
13560 result and can only support numeric or string input, whereas some
13561 discoverable commands may only be for direct use through
13562 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13563 pointers, or other custom uses. The "executable" commands are to
13564 support parameterisations of ENGINE behaviour that can be
13565 unambiguously defined by ENGINEs and used consistently across any
13566 OpenSSL-based application. Commands have been added to all the
13567 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13568 control over shared-library paths without source code alterations.
13572 * Changed all ENGINE implementations to dynamically allocate their
13573 ENGINEs rather than declaring them statically. Apart from this being
13574 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13575 this also allows the implementations to compile without using the
13576 internal engine_int.h header.
13580 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13581 'const' value. Any code that should be able to modify a RAND_METHOD
13582 should already have non-const pointers to it (ie. they should only
13583 modify their own ones).
13587 * Made a variety of little tweaks to the ENGINE code.
13588 - "atalla" and "ubsec" string definitions were moved from header files
13589 to C code. "nuron" string definitions were placed in variables
13590 rather than hard-coded - allowing parameterisation of these values
13591 later on via ctrl() commands.
13592 - Removed unused "#if 0"'d code.
13593 - Fixed engine list iteration code so it uses ENGINE_free() to release
13594 structural references.
13595 - Constified the RAND_METHOD element of ENGINE structures.
13596 - Constified various get/set functions as appropriate and added
13597 missing functions (including a catch-all ENGINE_cpy that duplicates
13598 all ENGINE values onto a new ENGINE except reference counts/state).
13599 - Removed NULL parameter checks in get/set functions. Setting a method
13600 or function to NULL is a way of cancelling out a previously set
13601 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13602 and doesn't justify the extra error symbols and code.
13603 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13604 flags from engine_int.h to engine.h.
13605 - Changed prototypes for ENGINE handler functions (init(), finish(),
13606 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13610 * Implement binary inversion algorithm for BN_mod_inverse in addition
13611 to the algorithm using long division. The binary algorithm can be
13612 used only if the modulus is odd. On 32-bit systems, it is faster
13613 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13614 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13615 up to 450 bits. In 64-bit environments, the binary algorithm
13616 appears to be advantageous for much longer moduli; here we use it
13617 for moduli up to 2048 bits.
13621 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13622 could not support the combine flag in choice fields.
13626 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13627 extensions from a certificate request to the certificate.
13631 * Allow multiple 'certopt' and 'nameopt' options to be separated
13632 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13633 file: this allows the display of the certificate about to be
13634 signed to be customised, to allow certain fields to be included
13635 or excluded and extension details. The old system didn't display
13636 multicharacter strings properly, omitted fields not in the policy
13637 and couldn't display additional details such as extensions.
13641 * Function EC_POINTs_mul for multiple scalar multiplication
13642 of an arbitrary number of elliptic curve points
13643 \sum scalars[i]*points[i],
13644 optionally including the generator defined for the EC_GROUP:
13645 scalar*generator + \sum scalars[i]*points[i].
13647 EC_POINT_mul is a simple wrapper function for the typical case
13648 that the point list has just one item (besides the optional
13653 * First EC_METHODs for curves over GF(p):
13655 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13656 operations and provides various method functions that can also
13657 operate with faster implementations of modular arithmetic.
13659 EC_GFp_mont_method() reuses most functions that are part of
13660 EC_GFp_simple_method, but uses Montgomery arithmetic.
13662 *Bodo Moeller; point addition and point doubling
13663 implementation directly derived from source code provided by
13664 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13666 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13667 crypto/ec/ec_lib.c):
13669 Curves are EC_GROUP objects (with an optional group generator)
13670 based on EC_METHODs that are built into the library.
13672 Points are EC_POINT objects based on EC_GROUP objects.
13674 Most of the framework would be able to handle curves over arbitrary
13675 finite fields, but as there are no obvious types for fields other
13676 than GF(p), some functions are limited to that for now.
13680 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13681 that the file contains a complete HTTP response.
13685 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13686 change the def and num file printf format specifier from "%-40sXXX"
13687 to "%-39s XXX". The latter will always guarantee a space after the
13688 field while the former will cause them to run together if the field
13689 is 40 of more characters long.
13693 * Constify the cipher and digest 'method' functions and structures
13694 and modify related functions to take constant EVP_MD and EVP_CIPHER
13699 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13700 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13704 * Modify `EVP_Digest*()` routines so they now return values. Although the
13705 internal software routines can never fail additional hardware versions
13710 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13712 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13713 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13716 ERR_R_NESTED_ASN1_ERROR
13718 ERR_R_MISSING_ASN1_EOS
13719 were 4 .. 9, conflicting with
13720 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13722 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13723 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13725 Add new error code 'ERR_R_INTERNAL_ERROR'.
13729 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13734 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13735 sets the subject name for a new request or supersedes the
13736 subject name in a given request. Formats that can be parsed are
13737 'CN=Some Name, OU=myOU, C=IT'
13739 'CN=Some Name/OU=myOU/C=IT'.
13741 Add options '-batch' and '-verbose' to 'openssl req'.
13743 *Massimiliano Pala <madwolf@hackmasters.net>*
13745 * Introduce the possibility to access global variables through
13746 functions on platform were that's the best way to handle exporting
13747 global variables in shared libraries. To enable this functionality,
13748 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13749 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13750 is normally done by Configure or something similar).
13752 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13753 in the source file (foo.c) like this:
13755 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13756 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13758 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13759 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13761 OPENSSL_DECLARE_GLOBAL(int,foo);
13762 #define foo OPENSSL_GLOBAL_REF(foo)
13763 OPENSSL_DECLARE_GLOBAL(double,bar);
13764 #define bar OPENSSL_GLOBAL_REF(bar)
13766 The #defines are very important, and therefore so is including the
13767 header file everywhere where the defined globals are used.
13769 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13770 of ASN.1 items, but that structure is a bit different.
13772 The largest change is in util/mkdef.pl which has been enhanced with
13773 better and easier to understand logic to choose which symbols should
13774 go into the Windows .def files as well as a number of fixes and code
13775 cleanup (among others, algorithm keywords are now sorted
13776 lexicographically to avoid constant rewrites).
13780 * In BN_div() keep a copy of the sign of 'num' before writing the
13781 result to 'rm' because if rm==num the value will be overwritten
13782 and produce the wrong result if 'num' is negative: this caused
13783 problems with BN_mod() and BN_nnmod().
13787 * Function OCSP_request_verify(). This checks the signature on an
13788 OCSP request and verifies the signer certificate. The signer
13789 certificate is just checked for a generic purpose and OCSP request
13794 * Add OCSP_check_validity() function to check the validity of OCSP
13795 responses. OCSP responses are prepared in real time and may only
13796 be a few seconds old. Simply checking that the current time lies
13797 between thisUpdate and nextUpdate max reject otherwise valid responses
13798 caused by either OCSP responder or client clock inaccuracy. Instead
13799 we allow thisUpdate and nextUpdate to fall within a certain period of
13800 the current time. The age of the response can also optionally be
13801 checked. Two new options -validity_period and -status_age added to
13806 * If signature or public key algorithm is unrecognized print out its
13807 OID rather that just UNKNOWN.
13811 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13812 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13813 ID to be generated from the issuer certificate alone which can then be
13814 passed to OCSP_id_issuer_cmp().
13818 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13819 ASN1 modules to export functions returning ASN1_ITEM pointers
13820 instead of the ASN1_ITEM structures themselves. This adds several
13821 new macros which allow the underlying ASN1 function/structure to
13822 be accessed transparently. As a result code should not use ASN1_ITEM
13823 references directly (such as &X509_it) but instead use the relevant
13824 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13825 use of the new ASN1 code on platforms where exporting structures
13826 is problematical (for example in shared libraries) but exporting
13827 functions returning pointers to structures is not.
13831 * Add support for overriding the generation of SSL/TLS session IDs.
13832 These callbacks can be registered either in an SSL_CTX or per SSL.
13833 The purpose of this is to allow applications to control, if they wish,
13834 the arbitrary values chosen for use as session IDs, particularly as it
13835 can be useful for session caching in multiple-server environments. A
13836 command-line switch for testing this (and any client code that wishes
13837 to use such a feature) has been added to "s_server".
13839 *Geoff Thorpe, Lutz Jaenicke*
13841 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13842 of the form `#if defined(...) || defined(...) || ...` and
13843 `#if !defined(...) && !defined(...) && ...`. This also avoids
13844 the growing number of special cases it was previously handling.
13848 * Make all configuration macros available for application by making
13849 sure they are available in opensslconf.h, by giving them names starting
13850 with `OPENSSL_` to avoid conflicts with other packages and by making
13851 sure e_os2.h will cover all platform-specific cases together with
13853 Additionally, it is now possible to define configuration/platform-
13854 specific names (called "system identities"). In the C code, these
13855 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13856 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13857 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13862 * New option -set_serial to 'req' and 'x509' this allows the serial
13863 number to use to be specified on the command line. Previously self
13864 signed certificates were hard coded with serial number 0 and the
13865 CA options of 'x509' had to use a serial number in a file which was
13870 * New options to 'ca' utility to support V2 CRL entry extensions.
13871 Currently CRL reason, invalidity date and hold instruction are
13872 supported. Add new CRL extensions to V3 code and some new objects.
13876 * New function EVP_CIPHER_CTX_set_padding() this is used to
13877 disable standard block padding (aka PKCS#5 padding) in the EVP
13878 API, which was previously mandatory. This means that the data is
13879 not padded in any way and so the total length much be a multiple
13880 of the block size, otherwise an error occurs.
13884 * Initial (incomplete) OCSP SSL support.
13888 * New function OCSP_parse_url(). This splits up a URL into its host,
13889 port and path components: primarily to parse OCSP URLs. New -url
13890 option to ocsp utility.
13894 * New nonce behavior. The return value of OCSP_check_nonce() now
13895 reflects the various checks performed. Applications can decide
13896 whether to tolerate certain situations such as an absent nonce
13897 in a response when one was present in a request: the ocsp application
13898 just prints out a warning. New function OCSP_add1_basic_nonce()
13899 this is to allow responders to include a nonce in a response even if
13900 the request is nonce-less.
13904 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13905 skipped when using openssl x509 multiple times on a single input file,
13906 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13910 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13911 set string type: to handle setting ASN1_TIME structures. Fix ca
13912 utility to correctly initialize revocation date of CRLs.
13916 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13917 the clients preferred ciphersuites and rather use its own preferences.
13918 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13919 Internet Explorer by ensuring unchanged hash method during stepup.
13920 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13924 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13925 to aes and add a new 'exist' option to print out symbols that don't
13930 * Additional options to ocsp utility to allow flags to be set and
13931 additional certificates supplied.
13935 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13936 OCSP client a number of certificate to only verify the response
13941 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13942 handle the new API. Currently only ECB, CBC modes supported. Add new
13945 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13946 Encryption Standard (AES) Ciphersuites for Transport Layer
13947 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13948 not enabled by default and were not part of the "ALL" ciphersuite
13949 alias because they were not yet official; they could be
13950 explicitly requested by specifying the "AESdraft" ciphersuite
13951 group alias. In the final release of OpenSSL 0.9.7, the group
13952 alias is called "AES" and is part of "ALL".)
13954 *Ben Laurie, Steve Henson, Bodo Moeller*
13956 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13957 request to response.
13961 * Functions for OCSP responders. OCSP_request_onereq_count(),
13962 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13963 extract information from a certificate request. OCSP_response_create()
13964 creates a response and optionally adds a basic response structure.
13965 OCSP_basic_add1_status() adds a complete single response to a basic
13966 response and returns the OCSP_SINGLERESP structure just added (to allow
13967 extensions to be included for example). OCSP_basic_add1_cert() adds a
13968 certificate to a basic response and OCSP_basic_sign() signs a basic
13969 response with various flags. New helper functions ASN1_TIME_check()
13970 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13971 (converts ASN1_TIME to GeneralizedTime).
13975 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13976 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13977 structure from a certificate. X509_pubkey_digest() digests the public_key
13978 contents: this is used in various key identifiers.
13982 * Make sk_sort() tolerate a NULL argument.
13984 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13986 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13987 passed by the function are trusted implicitly. If any of them signed the
13988 response then it is assumed to be valid and is not verified.
13992 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13993 to data. This was previously part of the PKCS7 ASN1 code. This
13994 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13995 *Steve Henson, reported by Kenneth R. Robinette
13996 <support@securenetterm.com>*
13998 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13999 routines: without these tracing memory leaks is very painful.
14000 Fix leaks in PKCS12 and PKCS7 routines.
14004 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14005 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14006 effectively meant GeneralizedTime would never be used. Now it
14007 is initialised to -1 but X509_time_adj() now has to check the value
14008 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14009 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14010 *Steve Henson, reported by Kenneth R. Robinette
14011 <support@securenetterm.com>*
14013 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14014 result in a zero length in the ASN1_INTEGER structure which was
14015 not consistent with the structure when d2i_ASN1_INTEGER() was used
14016 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14017 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14018 where it did not print out a minus for negative ASN1_INTEGER.
14022 * Add summary printout to ocsp utility. The various functions which
14023 convert status values to strings have been renamed to:
14024 OCSP_response_status_str(), OCSP_cert_status_str() and
14025 OCSP_crl_reason_str() and are no longer static. New options
14026 to verify nonce values and to disable verification. OCSP response
14027 printout format cleaned up.
14031 * Add additional OCSP certificate checks. These are those specified
14032 in RFC2560. This consists of two separate checks: the CA of the
14033 certificate being checked must either be the OCSP signer certificate
14034 or the issuer of the OCSP signer certificate. In the latter case the
14035 OCSP signer certificate must contain the OCSP signing extended key
14036 usage. This check is performed by attempting to match the OCSP
14037 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14038 in the OCSP_CERTID structures of the response.
14042 * Initial OCSP certificate verification added to OCSP_basic_verify()
14043 and related routines. This uses the standard OpenSSL certificate
14044 verify routines to perform initial checks (just CA validity) and
14045 to obtain the certificate chain. Then additional checks will be
14046 performed on the chain. Currently the root CA is checked to see
14047 if it is explicitly trusted for OCSP signing. This is used to set
14048 a root CA as a global signing root: that is any certificate that
14049 chains to that CA is an acceptable OCSP signing certificate.
14053 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14054 extensions from a separate configuration file.
14055 As when reading extensions from the main configuration file,
14056 the '-extensions ...' option may be used for specifying the
14059 *Massimiliano Pala <madwolf@comune.modena.it>*
14061 * New OCSP utility. Allows OCSP requests to be generated or
14062 read. The request can be sent to a responder and the output
14063 parsed, outputted or printed in text form. Not complete yet:
14064 still needs to check the OCSP response validity.
14068 * New subcommands for 'openssl ca':
14069 `openssl ca -status <serial>` prints the status of the cert with
14070 the given serial number (according to the index file).
14071 `openssl ca -updatedb` updates the expiry status of certificates
14074 *Massimiliano Pala <madwolf@comune.modena.it>*
14076 * New '-newreq-nodes' command option to CA.pl. This is like
14077 '-newreq', but calls 'openssl req' with the '-nodes' option
14078 so that the resulting key is not encrypted.
14080 *Damien Miller <djm@mindrot.org>*
14082 * New configuration for the GNU Hurd.
14084 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14086 * Initial code to implement OCSP basic response verify. This
14087 is currently incomplete. Currently just finds the signer's
14088 certificate and verifies the signature on the response.
14092 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14093 value of OPENSSLDIR. This is available via the new '-d' option
14094 to 'openssl version', and is also included in 'openssl version -a'.
14098 * Allowing defining memory allocation callbacks that will be given
14099 file name and line number information in additional arguments
14100 (a `const char*` and an int). The basic functionality remains, as
14101 well as the original possibility to just replace malloc(),
14102 realloc() and free() by functions that do not know about these
14103 additional arguments. To register and find out the current
14104 settings for extended allocation functions, the following
14105 functions are provided:
14107 CRYPTO_set_mem_ex_functions
14108 CRYPTO_set_locked_mem_ex_functions
14109 CRYPTO_get_mem_ex_functions
14110 CRYPTO_get_locked_mem_ex_functions
14112 These work the same way as CRYPTO_set_mem_functions and friends.
14113 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14114 extended allocation function is enabled.
14115 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14116 a conventional allocation function is enabled.
14118 *Richard Levitte, Bodo Moeller*
14120 * Finish off removing the remaining LHASH function pointer casts.
14121 There should no longer be any prototype-casting required when using
14122 the LHASH abstraction, and any casts that remain are "bugs". See
14123 the callback types and macros at the head of lhash.h for details
14124 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14128 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14129 If /dev/[u]random devices are not available or do not return enough
14130 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14132 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14133 /etc/entropy will be queried once each in this sequence, querying stops
14134 when enough entropy was collected without querying more sockets.
14138 * Change the Unix RAND_poll() variant to be able to poll several
14139 random devices, as specified by DEVRANDOM, until a sufficient amount
14140 of data has been collected. We spend at most 10 ms on each file
14141 (select timeout) and read in non-blocking mode. DEVRANDOM now
14142 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14143 (previously it was just the string "/dev/urandom"), so on typical
14144 platforms the 10 ms delay will never occur.
14145 Also separate out the Unix variant to its own file, rand_unix.c.
14146 For VMS, there's a currently-empty rand_vms.c.
14150 * Move OCSP client related routines to ocsp_cl.c. These
14151 provide utility functions which an application needing
14152 to issue a request to an OCSP responder and analyse the
14153 response will typically need: as opposed to those which an
14154 OCSP responder itself would need which will be added later.
14156 OCSP_request_sign() signs an OCSP request with an API similar
14157 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14158 response. OCSP_response_get1_basic() extracts basic response
14159 from response. OCSP_resp_find_status(): finds and extracts status
14160 information from an OCSP_CERTID structure (which will be created
14161 when the request structure is built). These are built from lower
14162 level functions which work on OCSP_SINGLERESP structures but
14163 won't normally be used unless the application wishes to examine
14164 extensions in the OCSP response for example.
14166 Replace nonce routines with a pair of functions.
14167 OCSP_request_add1_nonce() adds a nonce value and optionally
14168 generates a random value. OCSP_check_nonce() checks the
14169 validity of the nonce in an OCSP response.
14173 * Change function OCSP_request_add() to OCSP_request_add0_id().
14174 This doesn't copy the supplied OCSP_CERTID and avoids the
14175 need to free up the newly created id. Change return type
14176 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14177 This can then be used to add extensions to the request.
14178 Deleted OCSP_request_new(), since most of its functionality
14179 is now in OCSP_REQUEST_new() (and the case insensitive name
14180 clash) apart from the ability to set the request name which
14181 will be added elsewhere.
14185 * Update OCSP API. Remove obsolete extensions argument from
14186 various functions. Extensions are now handled using the new
14187 OCSP extension code. New simple OCSP HTTP function which
14188 can be used to send requests and parse the response.
14192 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14193 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14194 uses the special reorder version of SET OF to sort the attributes
14195 and reorder them to match the encoded order. This resolves a long
14196 standing problem: a verify on a PKCS7 structure just after signing
14197 it used to fail because the attribute order did not match the
14198 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14199 it uses the received order. This is necessary to tolerate some broken
14200 software that does not order SET OF. This is handled by encoding
14201 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14202 to produce the required SET OF.
14206 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14207 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14208 files to get correct declarations of the ASN.1 item variables.
14212 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14213 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14214 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14215 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14216 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14217 ASN1_ITEM and no wrapper functions.
14221 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14222 replace the old function pointer based I/O routines. Change most of
14223 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14227 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14228 lines, recognize more "algorithms" that can be deselected, and make
14229 it complain about algorithm deselection that isn't recognised.
14233 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14234 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14235 to use new functions. Add NO_ASN1_OLD which can be set to remove
14236 some old style ASN1 functions: this can be used to determine if old
14237 code will still work when these eventually go away.
14241 * New extension functions for OCSP structures, these follow the
14242 same conventions as certificates and CRLs.
14246 * New function X509V3_add1_i2d(). This automatically encodes and
14247 adds an extension. Its behaviour can be customised with various
14248 flags to append, replace or delete. Various wrappers added for
14249 certificates and CRLs.
14253 * Fix to avoid calling the underlying ASN1 print routine when
14254 an extension cannot be parsed. Correct a typo in the
14255 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14259 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14260 entries for variables.
14264 * Add functionality to `apps/openssl.c` for detecting locking
14265 problems: As the program is single-threaded, all we have
14266 to do is register a locking callback using an array for
14267 storing which locks are currently held by the program.
14271 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14272 SSL_get_ex_data_X509_STORE_idx(), which is used in
14273 ssl_verify_cert_chain() and thus can be called at any time
14274 during TLS/SSL handshakes so that thread-safety is essential.
14275 Unfortunately, the ex_data design is not at all suited
14276 for multi-threaded use, so it probably should be abolished.
14280 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14282 *Broadcom, tweaked and integrated by Geoff Thorpe*
14284 * Move common extension printing code to new function
14285 X509V3_print_extensions(). Reorganise OCSP print routines and
14286 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14290 * New function X509_signature_print() to remove duplication in some
14295 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14296 set (this was treated exactly the same as SET OF previously). This
14297 is used to reorder the STACK representing the structure to match the
14298 encoding. This will be used to get round a problem where a PKCS7
14299 structure which was signed could not be verified because the STACK
14300 order did not reflect the encoded order.
14304 * Reimplement the OCSP ASN1 module using the new code.
14308 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14309 for its ASN1 operations. The old style function pointers still exist
14310 for now but they will eventually go away.
14314 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14315 completely replaces the old ASN1 functionality with a table driven
14316 encoder and decoder which interprets an ASN1_ITEM structure describing
14317 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14318 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14319 has also been converted to the new form.
14323 * Change BN_mod_exp_recp so that negative moduli are tolerated
14324 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14325 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14326 for negative moduli.
14330 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14331 of not touching the result's sign bit.
14335 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14340 * Changed the LHASH code to use prototypes for callbacks, and created
14341 macros to declare and implement thin (optionally static) functions
14342 that provide type-safety and avoid function pointer casting for the
14343 type-specific callbacks.
14347 * Added Kerberos Cipher Suites to be used with TLS, as written in
14349 *Veers Staats <staatsvr@asc.hpc.mil>,
14350 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14352 * Reformat the FAQ so the different questions and answers can be divided
14353 in sections depending on the subject.
14357 * Have the zlib compression code load ZLIB.DLL dynamically under
14362 * New function BN_mod_sqrt for computing square roots modulo a prime
14363 (using the probabilistic Tonelli-Shanks algorithm unless
14364 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14365 be handled deterministically).
14367 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14369 * Make BN_mod_inverse faster by explicitly handling small quotients
14370 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14371 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14375 * New function BN_kronecker.
14379 * Fix BN_gcd so that it works on negative inputs; the result is
14380 positive unless both parameters are zero.
14381 Previously something reasonably close to an infinite loop was
14382 possible because numbers could be growing instead of shrinking
14383 in the implementation of Euclid's algorithm.
14387 * Fix BN_is_word() and BN_is_one() macros to take into account the
14388 sign of the number in question.
14390 Fix BN_is_word(a,w) to work correctly for w == 0.
14392 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14393 because its test if the absolute value of 'a' equals 'w'.
14394 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14395 it exists mostly for use in the implementations of BN_is_zero(),
14396 BN_is_one(), and BN_is_word().
14400 * New function BN_swap.
14404 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14405 the exponentiation functions are more likely to produce reasonable
14406 results on negative inputs.
14410 * Change BN_mod_mul so that the result is always non-negative.
14411 Previously, it could be negative if one of the factors was negative;
14412 I don't think anyone really wanted that behaviour.
14416 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14417 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14418 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14419 and add new functions:
14428 BN_mod_lshift1_quick
14430 BN_mod_lshift_quick
14432 These functions always generate non-negative results.
14434 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14435 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14437 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14438 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14439 be reduced modulo `m`.
14441 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14444 The following entry accidentally appeared in the CHANGES file
14445 distributed with OpenSSL 0.9.7. The modifications described in
14446 it do *not* apply to OpenSSL 0.9.7.
14448 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14449 was actually never needed) and in BN_mul(). The removal in BN_mul()
14450 required a small change in bn_mul_part_recursive() and the addition
14451 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14452 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14453 bn_sub_words() and bn_add_words() except they take arrays with
14459 * In 'openssl passwd', verify passwords read from the terminal
14460 unless the '-salt' option is used (which usually means that
14461 verification would just waste user's time since the resulting
14462 hash is going to be compared with some given password hash)
14463 or the new '-noverify' option is used.
14465 This is an incompatible change, but it does not affect
14466 non-interactive use of 'openssl passwd' (passwords on the command
14467 line, '-stdin' option, '-in ...' option) and thus should not
14468 cause any problems.
14472 * Remove all references to RSAref, since there's no more need for it.
14476 * Make DSO load along a path given through an environment variable
14477 (SHLIB_PATH) with shl_load().
14481 * Constify the ENGINE code as a result of BIGNUM constification.
14482 Also constify the RSA code and most things related to it. In a
14483 few places, most notable in the depth of the ASN.1 code, ugly
14484 casts back to non-const were required (to be solved at a later
14489 * Make it so the openssl application has all engines loaded by default.
14493 * Constify the BIGNUM routines a little more.
14497 * Add the following functions:
14499 ENGINE_load_cswift()
14501 ENGINE_load_atalla()
14502 ENGINE_load_nuron()
14503 ENGINE_load_builtin_engines()
14505 That way, an application can itself choose if external engines that
14506 are built-in in OpenSSL shall ever be used or not. The benefit is
14507 that applications won't have to be linked with libdl or other dso
14508 libraries unless it's really needed.
14510 Changed 'openssl engine' to load all engines on demand.
14511 Changed the engine header files to avoid the duplication of some
14512 declarations (they differed!).
14516 * 'openssl engine' can now list capabilities.
14520 * Better error reporting in 'openssl engine'.
14524 * Never call load_dh_param(NULL) in s_server.
14528 * Add engine application. It can currently list engines by name and
14529 identity, and test if they are actually available.
14533 * Improve RPM specification file by forcing symbolic linking and making
14534 sure the installed documentation is also owned by root.root.
14536 *Damien Miller <djm@mindrot.org>*
14538 * Give the OpenSSL applications more possibilities to make use of
14539 keys (public as well as private) handled by engines.
14543 * Add OCSP code that comes from CertCo.
14547 * Add VMS support for the Rijndael code.
14551 * Added untested support for Nuron crypto accelerator.
14555 * Add support for external cryptographic devices. This code was
14556 previously distributed separately as the "engine" branch.
14558 *Geoff Thorpe, Richard Levitte*
14560 * Rework the filename-translation in the DSO code. It is now possible to
14561 have far greater control over how a "name" is turned into a filename
14562 depending on the operating environment and any oddities about the
14563 different shared library filenames on each system.
14567 * Support threads on FreeBSD-elf in Configure.
14571 * Fix for SHA1 assembly problem with MASM: it produces
14572 warnings about corrupt line number information when assembling
14573 with debugging information. This is caused by the overlapping
14576 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14579 NCONF_get_number() has no error checking at all. As a replacement,
14580 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14581 promoted strongly. The old NCONF_get_number is kept around for
14582 binary backward compatibility.
14583 Make it possible for methods to load from something other than a BIO,
14584 by providing a function pointer that is given a name instead of a BIO.
14585 For example, this could be used to load configuration data from an
14590 * Fix for non blocking accept BIOs. Added new I/O special reason
14591 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14592 with non blocking I/O was not possible because no retry code was
14593 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14598 * Added the beginnings of Rijndael support.
14602 * Fix for bug in DirectoryString mask setting. Add support for
14603 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14604 to allow certificate printing to more controllable, additional
14605 'certopt' option to 'x509' to allow new printing options to be
14610 * Clean old EAY MD5 hack from e_os.h.
14614 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14616 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14617 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14619 *Joe Orton, Steve Henson*
14621 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14623 * Fix additional bug revealed by the NISCC test suite:
14625 Stop bug triggering large recursion when presented with
14626 certain ASN.1 tags ([CVE-2003-0851])
14630 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14632 * Fix various bugs revealed by running the NISCC test suite:
14634 Stop out of bounds reads in the ASN1 code when presented with
14635 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14637 If verify callback ignores invalid public key errors don't try to check
14638 certificate signature with the NULL public key.
14642 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14643 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14648 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14649 extra data after the compression methods not only for TLS 1.0
14650 but also for SSL 3.0 (as required by the specification).
14652 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14654 * Change X509_certificate_type() to mark the key as exported/exportable
14655 when it's 512 *bits* long, not 512 bytes.
14659 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14661 * Countermeasure against the Klima-Pokorny-Rosa extension of
14662 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14663 a protocol version number mismatch like a decryption error
14664 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14668 * Turn on RSA blinding by default in the default implementation
14669 to avoid a timing attack. Applications that don't want it can call
14670 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14671 They would be ill-advised to do so in most cases.
14673 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14675 * Change RSA blinding code so that it works when the PRNG is not
14676 seeded (in this case, the secret RSA exponent is abused as
14677 an unpredictable seed -- if it is not unpredictable, there
14678 is no point in blinding anyway). Make RSA blinding thread-safe
14679 by remembering the creator's thread ID in rsa->blinding and
14680 having all other threads use local one-time blinding factors
14681 (this requires more computation than sharing rsa->blinding, but
14682 avoids excessive locking; and if an RSA object is not shared
14683 between threads, blinding will still be very fast).
14687 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14689 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14690 via timing by performing a MAC computation even if incorrect
14691 block cipher padding has been found. This is a countermeasure
14692 against active attacks where the attacker has to distinguish
14693 between bad padding and a MAC verification error. ([CVE-2003-0078])
14695 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14696 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14697 Martin Vuagnoux (EPFL, Ilion)*
14699 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14701 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14702 memory from its contents. This is done with a counter that will
14703 place alternating values in each byte. This can be used to solve
14704 two issues: 1) the removal of calls to memset() by highly optimizing
14705 compilers, and 2) cleansing with other values than 0, since those can
14706 be read through on certain media, for example a swap space on disk.
14710 * Bugfix: client side session caching did not work with external caching,
14711 because the session->cipher setting was not restored when reloading
14712 from the external cache. This problem was masked, when
14713 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14714 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14718 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14719 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14721 *Zeev Lieber <zeev-l@yahoo.com>*
14723 * Undo an undocumented change introduced in 0.9.6e which caused
14724 repeated calls to OpenSSL_add_all_ciphers() and
14725 OpenSSL_add_all_digests() to be ignored, even after calling
14730 * Change the default configuration reader to deal with last line not
14731 being properly terminated.
14735 * Change X509_NAME_cmp() so it applies the special rules on handling
14736 DN values that are of type PrintableString, as well as RDNs of type
14737 emailAddress where the value has the type ia5String.
14739 *stefank@valicert.com via Richard Levitte*
14741 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14742 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14743 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14744 the bitwise-OR of the two for use by the majority of applications
14745 wanting this behaviour, and update the docs. The documented
14746 behaviour and actual behaviour were inconsistent and had been
14747 changing anyway, so this is more a bug-fix than a behavioural
14750 *Geoff Thorpe, diagnosed by Nadav Har'El*
14752 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14753 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14757 * Fix initialization code race conditions in
14758 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14759 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14760 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14761 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14762 ssl2_get_cipher_by_char(),
14763 ssl3_get_cipher_by_char().
14765 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14767 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14768 the cached sessions are flushed, as the remove_cb() might use ex_data
14769 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14770 (see [openssl.org #212]).
14772 *Geoff Thorpe, Lutz Jaenicke*
14774 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14775 length, instead of the encoding length to d2i_ASN1_OBJECT.
14779 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14781 * [In 0.9.6g-engine release:]
14782 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14784 *Lynn Gazis <lgazis@rainbow.com>*
14786 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14788 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14789 and get fix the header length calculation.
14790 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14791 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14793 * Use proper error handling instead of 'assertions' in buffer
14794 overflow checks added in 0.9.6e. This prevents DoS (the
14795 assertions could call abort()).
14797 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14799 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14801 * Add various sanity checks to asn1_get_length() to reject
14802 the ASN1 length bytes if they exceed sizeof(long), will appear
14803 negative or the content length exceeds the length of the
14806 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14808 * Fix cipher selection routines: ciphers without encryption had no flags
14809 for the cipher strength set and where therefore not handled correctly
14810 by the selection routines (PR #130).
14814 * Fix EVP_dsa_sha macro.
14819 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14820 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14821 that was added in OpenSSL 0.9.6d.
14823 As the countermeasure turned out to be incompatible with some
14824 broken SSL implementations, the new option is part of SSL_OP_ALL.
14825 SSL_OP_ALL is usually employed when compatibility with weird SSL
14826 implementations is desired (e.g. '-bugs' option to 's_client' and
14827 's_server'), so the new option is automatically set in many
14832 * Changes in security patch:
14834 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14835 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14836 Air Force Materiel Command, USAF, under agreement number
14839 * Add various sanity checks to asn1_get_length() to reject
14840 the ASN1 length bytes if they exceed sizeof(long), will appear
14841 negative or the content length exceeds the length of the
14842 supplied buffer. ([CVE-2002-0659])
14844 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14846 * Assertions for various potential buffer overflows, not known to
14847 happen in practice.
14849 *Ben Laurie (CHATS)*
14851 * Various temporary buffers to hold ASCII versions of integers were
14852 too small for 64 bit platforms. ([CVE-2002-0655])
14853 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14855 * Remote buffer overflow in SSL3 protocol - an attacker could
14856 supply an oversized session ID to a client. ([CVE-2002-0656])
14858 *Ben Laurie (CHATS)*
14860 * Remote buffer overflow in SSL2 protocol - an attacker could
14861 supply an oversized client master key. ([CVE-2002-0656])
14863 *Ben Laurie (CHATS)*
14865 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14867 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14868 encoded as NULL) with id-dsa-with-sha1.
14870 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14872 * Check various `X509_...()` return values in `apps/req.c`.
14874 *Nils Larsch <nla@trustcenter.de>*
14876 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14877 an end-of-file condition would erroneously be flagged, when the CRLF
14878 was just at the end of a processed block. The bug was discovered when
14879 processing data through a buffering memory BIO handing the data to a
14880 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14881 <ptsekov@syntrex.com> and Nedelcho Stanev.
14885 * Implement a countermeasure against a vulnerability recently found
14886 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14887 before application data chunks to avoid the use of known IVs
14888 with data potentially chosen by the attacker.
14892 * Fix length checks in ssl3_get_client_hello().
14896 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14897 to prevent ssl3_read_internal() from incorrectly assuming that
14898 ssl3_read_bytes() found application data while handshake
14899 processing was enabled when in fact s->s3->in_read_app_data was
14900 merely automatically cleared during the initial handshake.
14902 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14904 * Fix object definitions for Private and Enterprise: they were not
14905 recognized in their shortname (=lowercase) representation. Extend
14906 obj_dat.pl to issue an error when using undefined keywords instead
14907 of silently ignoring the problem (Svenning Sorensen
14908 <sss@sss.dnsalias.net>).
14912 * Fix DH_generate_parameters() so that it works for 'non-standard'
14913 generators, i.e. generators other than 2 and 5. (Previously, the
14914 code did not properly initialise the 'add' and 'rem' values to
14915 BN_generate_prime().)
14917 In the new general case, we do not insist that 'generator' is
14918 actually a primitive root: This requirement is rather pointless;
14919 a generator of the order-q subgroup is just as good, if not
14924 * Map new X509 verification errors to alerts. Discovered and submitted by
14925 Tom Wu <tom@arcot.com>.
14929 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14930 returning non-zero before the data has been completely received
14931 when using non-blocking I/O.
14933 *Bodo Moeller; problem pointed out by John Hughes*
14935 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14937 *Ben Laurie, Lutz Jaenicke*
14939 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14940 Yoram Zahavi <YoramZ@gilian.com>).
14944 * Add information about CygWin 1.3 and on, and preserve proper
14945 configuration for the versions before that.
14947 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14949 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14950 check whether we deal with a copy of a session and do not delete from
14951 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14952 <izhar@checkpoint.com>.
14956 * Do not store session data into the internal session cache, if it
14957 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14958 flag is set). Proposed by Aslam <aslam@funk.com>.
14962 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14967 * [In 0.9.6d-engine release:]
14968 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14970 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14972 * Add the configuration target linux-s390x.
14974 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14976 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14977 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14978 variable as an indication that a ClientHello message has been
14979 received. As the flag value will be lost between multiple
14980 invocations of ssl3_accept when using non-blocking I/O, the
14981 function may not be aware that a handshake has actually taken
14982 place, thus preventing a new session from being added to the
14985 To avoid this problem, we now set s->new_session to 2 instead of
14986 using a local variable.
14988 *Lutz Jaenicke, Bodo Moeller*
14990 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14991 if the SSL_R_LENGTH_MISMATCH error is detected.
14993 *Geoff Thorpe, Bodo Moeller*
14995 * New 'shared_ldflag' column in Configure platform table.
14999 * Fix EVP_CIPHER_mode macro.
15001 *"Dan S. Camper" <dan@bti.net>*
15003 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15004 type, we must throw them away by setting rr->length to 0.
15006 *D P Chang <dpc@qualys.com>*
15008 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15010 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15011 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15012 worked incorrectly for those cases where range = `10..._2` and
15013 `3*range` is two bits longer than range.)
15017 * Only add signing time to PKCS7 structures if it is not already
15022 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15023 OBJ_ld_ce should be OBJ_id_ce.
15024 Also some ip-pda OIDs in crypto/objects/objects.txt were
15025 incorrect (cf. RFC 3039).
15027 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15029 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15030 returns early because it has nothing to do.
15032 *Andy Schneider <andy.schneider@bjss.co.uk>*
15034 * [In 0.9.6c-engine release:]
15035 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15037 *Andy Schneider <andy.schneider@bjss.co.uk>*
15039 * [In 0.9.6c-engine release:]
15040 Add support for Cryptographic Appliance's keyserver technology.
15041 (Use engine 'keyclient')
15043 *Cryptographic Appliances and Geoff Thorpe*
15045 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15046 is called via tools/c89.sh because arguments have to be
15047 rearranged (all '-L' options must appear before the first object
15050 *Richard Shapiro <rshapiro@abinitio.com>*
15052 * [In 0.9.6c-engine release:]
15053 Add support for Broadcom crypto accelerator cards, backported
15056 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15058 * [In 0.9.6c-engine release:]
15059 Add support for SureWare crypto accelerator cards from
15060 Baltimore Technologies. (Use engine 'sureware')
15062 *Baltimore Technologies and Mark Cox*
15064 * [In 0.9.6c-engine release:]
15065 Add support for crypto accelerator cards from Accelerated
15066 Encryption Processing, www.aep.ie. (Use engine 'aep')
15068 *AEP Inc. and Mark Cox*
15070 * Add a configuration entry for gcc on UnixWare.
15072 *Gary Benson <gbenson@redhat.com>*
15074 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15075 messages are stored in a single piece (fixed-length part and
15076 variable-length part combined) and fix various bugs found on the way.
15080 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15081 instead. BIO_gethostbyname() does not know what timeouts are
15082 appropriate, so entries would stay in cache even when they have
15084 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15086 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15087 faced with a pathologically small ClientHello fragment that does
15088 not contain client_version: Instead of aborting with an error,
15089 simply choose the highest available protocol version (i.e.,
15090 TLS 1.0 unless it is disabled). In practice, ClientHello
15091 messages are never sent like this, but this change gives us
15092 strictly correct behaviour at least for TLS.
15096 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15097 never resets s->method to s->ctx->method when called from within
15098 one of the SSL handshake functions.
15100 *Bodo Moeller; problem pointed out by Niko Baric*
15102 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15103 (sent using the client's version number) if client_version is
15104 smaller than the protocol version in use. Also change
15105 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15106 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15107 the client will at least see that alert.
15111 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15116 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15117 client receives HelloRequest while in a handshake.
15119 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15121 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15122 should end in 'break', not 'goto end' which circumvents various
15123 cleanups done in state SSL_ST_OK. But session related stuff
15124 must be disabled for SSL_ST_OK in the case that we just sent a
15127 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15128 before just sending a HelloRequest.
15130 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15132 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15133 reveal whether illegal block cipher padding was found or a MAC
15134 verification error occurred. (Neither SSLerr() codes nor alerts
15135 are directly visible to potential attackers, but the information
15136 may leak via logfiles.)
15138 Similar changes are not required for the SSL 2.0 implementation
15139 because the number of padding bytes is sent in clear for SSL 2.0,
15140 and the extra bytes are just ignored. However ssl/s2_pkt.c
15141 failed to verify that the purported number of padding bytes is in
15146 * Add OpenUNIX-8 support including shared libraries
15147 (Boyd Lynn Gerber <gerberb@zenez.com>).
15151 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15152 'wristwatch attack' using huge encoding parameters (cf.
15153 James H. Manger's CRYPTO 2001 paper). Note that the
15154 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15155 encoding parameters and hence was not vulnerable.
15159 * BN_sqr() bug fix.
15161 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15163 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15164 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15165 followed by modular reduction.
15167 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15169 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15170 equivalent based on BN_pseudo_rand() instead of BN_rand().
15174 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15175 This function was broken, as the check for a new client hello message
15176 to handle SGC did not allow these large messages.
15177 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15181 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15185 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15186 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15190 * Rework the configuration and shared library support for Tru64 Unix.
15191 The configuration part makes use of modern compiler features and
15192 still retains old compiler behavior for those that run older versions
15193 of the OS. The shared library support part includes a variant that
15194 uses the RPATH feature, and is available through the special
15195 configuration target "alpha-cc-rpath", which will never be selected
15198 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15200 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15201 with the same message size as in ssl3_get_certificate_request().
15202 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15203 messages might inadvertently be reject as too long.
15205 *Petr Lampa <lampa@fee.vutbr.cz>*
15207 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15211 * Modified SSL library such that the verify_callback that has been set
15212 specifically for an SSL object with SSL_set_verify() is actually being
15213 used. Before the change, a verify_callback set with this function was
15214 ignored and the verify_callback() set in the SSL_CTX at the time of
15215 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15216 to allow the necessary settings.
15220 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15221 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15222 done automatically (in contradiction to the requirements of the C
15223 standard). This made problems when used from OpenSSH.
15227 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15228 dh->length and always used
15230 BN_rand_range(priv_key, dh->p).
15232 BN_rand_range() is not necessary for Diffie-Hellman, and this
15233 specific range makes Diffie-Hellman unnecessarily inefficient if
15234 dh->length (recommended exponent length) is much smaller than the
15235 length of dh->p. We could use BN_rand_range() if the order of
15236 the subgroup was stored in the DH structure, but we only have
15241 BN_rand(priv_key, l, ...)
15243 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15250 RSA_eay_public_encrypt
15251 RSA_eay_private_decrypt
15252 RSA_eay_private_encrypt (signing)
15253 RSA_eay_public_decrypt (signature verification)
15255 (default implementations for RSA_public_encrypt,
15256 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15257 always reject numbers >= n.
15261 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15262 to synchronize access to 'locking_thread'. This is necessary on
15263 systems where access to 'locking_thread' (an 'unsigned long'
15264 variable) is not atomic.
15268 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15269 *before* setting the 'crypto_lock_rand' flag. The previous code had
15270 a race condition if 0 is a valid thread ID.
15272 *Travis Vitek <vitek@roguewave.com>*
15274 * Add support for shared libraries under Irix.
15276 *Albert Chin-A-Young <china@thewrittenword.com>*
15278 * Add configuration option to build on Linux on both big-endian and
15279 little-endian MIPS.
15281 *Ralf Baechle <ralf@uni-koblenz.de>*
15283 * Add the possibility to create shared libraries on HP-UX.
15287 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15289 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15290 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15291 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15292 PRNG state recovery was possible based on the output of
15293 one PRNG request appropriately sized to gain knowledge on
15294 'md' followed by enough consecutive 1-byte PRNG requests
15295 to traverse all of 'state'.
15297 1. When updating 'md_local' (the current thread's copy of 'md')
15298 during PRNG output generation, hash all of the previous
15299 'md_local' value, not just the half used for PRNG output.
15301 2. Make the number of bytes from 'state' included into the hash
15302 independent from the number of PRNG bytes requested.
15304 The first measure alone would be sufficient to avoid
15305 Markku-Juhani's attack. (Actually it had never occurred
15306 to me that the half of 'md_local' used for chaining was the
15307 half from which PRNG output bytes were taken -- I had always
15308 assumed that the secret half would be used.) The second
15309 measure makes sure that additional data from 'state' is never
15310 mixed into 'md_local' in small portions; this heuristically
15311 further strengthens the PRNG.
15315 * Fix crypto/bn/asm/mips3.s.
15319 * When only the key is given to "enc", the IV is undefined. Print out
15320 an error message in this case.
15324 * Handle special case when X509_NAME is empty in X509 printing routines.
15328 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15329 positive and less than q.
15333 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15334 used: it isn't thread safe and the add_lock_callback should handle
15337 *Paul Rose <Paul.Rose@bridge.com>*
15339 * Verify that incoming data obeys the block size in
15340 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15346 *Ulf Möller, Bodo Möller*
15348 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15349 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15350 when fixing the server behaviour for backwards-compatible 'client
15351 hello' messages. (Note that the attack is impractical against
15352 SSL 3.0 and TLS 1.0 anyway because length and version checking
15353 means that the probability of guessing a valid ciphertext is
15354 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15357 Before 0.9.5, the countermeasure (hide the error by generating a
15358 random 'decryption result') did not work properly because
15359 ERR_clear_error() was missing, meaning that SSL_get_error() would
15360 detect the supposedly ignored error.
15362 Both problems are now fixed.
15366 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15367 (previously it was 1024).
15371 * Fix for compatibility mode trust settings: ignore trust settings
15372 unless some valid trust or reject settings are present.
15376 * Fix for blowfish EVP: its a variable length cipher.
15380 * Fix various bugs related to DSA S/MIME verification. Handle missing
15381 parameters in DSA public key structures and return an error in the
15382 DSA routines if parameters are absent.
15386 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15387 in the current directory if neither $RANDFILE nor $HOME was set.
15388 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15389 caused some confusion to Windows users who haven't defined $HOME.
15390 Thus RAND_file_name() is changed again: e_os.h can define a
15391 DEFAULT_HOME, which will be used if $HOME is not set.
15392 For Windows, we use "C:"; on other platforms, we still require
15393 environment variables.
15395 * Move 'if (!initialized) RAND_poll()' into regions protected by
15396 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15397 having multiple threads call RAND_poll() concurrently.
15401 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15402 combination of a flag and a thread ID variable.
15403 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15404 flag), *other* threads can enter ssleay_add_bytes without obeying
15405 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15406 that they do not hold after the first thread unsets add_do_not_lock).
15410 * Change bctest again: '-x' expressions are not available in all
15411 versions of 'test'.
15415 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15417 * Fix a couple of memory leaks in PKCS7_dataDecode()
15419 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15421 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15422 the default extension for executables, if any. Also, make the perl
15423 scripts that use symlink() to test if it really exists and use "cp"
15424 if it doesn't. All this made OpenSSL compilable and installable in
15429 * Fix for asn1_GetSequence() for indefinite length constructed data.
15430 If SEQUENCE is length is indefinite just set c->slen to the total
15431 amount of data available.
15433 *Steve Henson, reported by shige@FreeBSD.org*
15435 *This change does not apply to 0.9.7.*
15437 * Change bctest to avoid here-documents inside command substitution
15438 (workaround for FreeBSD /bin/sh bug).
15439 For compatibility with Ultrix, avoid shell functions (introduced
15440 in the bctest version that searches along $PATH).
15444 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15445 with des_encrypt() defined on some operating systems, like Solaris
15450 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15451 On the Importance of Eliminating Errors in Cryptographic
15452 Computations, J. Cryptology 14 (2001) 2, 101-119,
15453 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15457 * MIPS assembler BIGNUM division bug fix.
15461 * Disabled incorrect Alpha assembler code.
15465 * Fix PKCS#7 decode routines so they correctly update the length
15466 after reading an EOC for the EXPLICIT tag.
15470 *This change does not apply to 0.9.7.*
15472 * Fix bug in PKCS#12 key generation routines. This was triggered
15473 if a 3DES key was generated with a 0 initial byte. Include
15474 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15475 (but broken) behaviour.
15479 * Enhance bctest to search for a working bc along $PATH and print
15482 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15484 * Fix memory leaks in err.c: free err_data string if necessary;
15485 don't write to the wrong index in ERR_set_error_data.
15489 * Implement ssl23_peek (analogous to ssl23_read), which previously
15494 * Replace rdtsc with `_emit` statements for VC++ version 5.
15496 *Jeremy Cooper <jeremy@baymoo.org>*
15498 * Make it possible to reuse SSLv2 sessions.
15502 * In copy_email() check for >= 0 as a return value for
15503 X509_NAME_get_index_by_NID() since 0 is a valid index.
15505 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15507 * Avoid coredump with unsupported or invalid public keys by checking if
15508 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15509 PKCS7_verify() fails with non detached data.
15513 * Don't use getenv in library functions when run as setuid/setgid.
15514 New function OPENSSL_issetugid().
15518 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15519 due to incorrect handling of multi-threading:
15521 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15523 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15525 3. Count how many times MemCheck_off() has been called so that
15526 nested use can be treated correctly. This also avoids
15527 inband-signalling in the previous code (which relied on the
15528 assumption that thread ID 0 is impossible).
15532 * Add "-rand" option also to s_client and s_server.
15536 * Fix CPU detection on Irix 6.x.
15537 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15538 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15540 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15545 *This change does not apply to 0.9.7.*
15547 * Use the cached encoding of an X509_NAME structure rather than
15548 copying it. This is apparently the reason for the libsafe "errors"
15549 but the code is actually correct.
15553 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15554 Bleichenbacher's DSA attack.
15555 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15556 to be set and top=0 forces the highest bit to be set; top=-1 is new
15557 and leaves the highest bit random.
15559 *Ulf Moeller, Bodo Moeller*
15561 * In the `NCONF_...`-based implementations for `CONF_...` queries
15562 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15563 a temporary CONF structure with the data component set to NULL
15564 (which gives segmentation faults in lh_retrieve).
15565 Instead, use NULL for the CONF pointer in CONF_get_string and
15566 CONF_get_number (which may use environment variables) and directly
15567 return NULL from CONF_get_section.
15571 * Fix potential buffer overrun for EBCDIC.
15575 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15576 keyUsage if basicConstraints absent for a CA.
15580 * Make SMIME_write_PKCS7() write mail header values with a format that
15581 is more generally accepted (no spaces before the semicolon), since
15582 some programs can't parse those values properly otherwise. Also make
15583 sure BIO's that break lines after each write do not create invalid
15588 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15589 macros previously used would not encode an empty SEQUENCE OF
15590 and break the signature.
15594 *This change does not apply to 0.9.7.*
15596 * Zero the premaster secret after deriving the master secret in
15601 * Add some EVP_add_digest_alias registrations (as found in
15602 OpenSSL_add_all_digests()) to SSL_library_init()
15603 aka OpenSSL_add_ssl_algorithms(). This provides improved
15604 compatibility with peers using X.509 certificates
15605 with unconventional AlgorithmIdentifier OIDs.
15609 * Fix for Irix with NO_ASM.
15611 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15613 * ./config script fixes.
15615 *Ulf Moeller, Richard Levitte*
15617 * Fix 'openssl passwd -1'.
15621 * Change PKCS12_key_gen_asc() so it can cope with non null
15622 terminated strings whose length is passed in the passlen
15623 parameter, for example from PEM callbacks. This was done
15624 by adding an extra length parameter to asc2uni().
15626 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15628 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15629 call failed, free the DSA structure.
15633 * Fix to uni2asc() to cope with zero length Unicode strings.
15634 These are present in some PKCS#12 files.
15638 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15639 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15640 when writing a 32767 byte record.
15642 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15644 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15645 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15647 (RSA objects have a reference count access to which is protected
15648 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15649 so they are meant to be shared between threads.)
15650 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15651 "Reddie, Steven" <Steven.Reddie@ca.com>*
15653 * Fix a deadlock in CRYPTO_mem_leaks().
15657 * Use better test patterns in bntest.
15661 * rand_win.c fix for Borland C.
15665 * BN_rshift bugfix for n == 0.
15669 * Add a 'bctest' script that checks for some known 'bc' bugs
15670 so that 'make test' does not abort just because 'bc' is broken.
15674 * Store verify_result within SSL_SESSION also for client side to
15675 avoid potential security hole. (Re-used sessions on the client side
15676 always resulted in verify_result==X509_V_OK, not using the original
15677 result of the server certificate verification.)
15681 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15682 SSL3_RT_APPLICATION_DATA, return 0.
15683 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15688 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15689 releases, have been re-implemented by renaming the previous
15690 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15691 and ssl3_read_internal, respectively, and adding 'peek' parameters
15692 to them. The new ssl[23]_{read,peek} functions are calls to
15693 ssl[23]_read_internal with the 'peek' flag set appropriately.
15694 A 'peek' parameter has also been added to ssl3_read_bytes, which
15695 does the actual work for ssl3_read_internal.
15699 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15700 the method-specific "init()" handler. Also clean up ex_data after
15701 calling the method-specific "finish()" handler. Previously, this was
15702 happening the other way round.
15706 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15707 The previous value, 12, was not always sufficient for BN_mod_exp().
15711 * Make sure that shared libraries get the internal name engine with
15712 the full version number and not just 0. This should mark the
15713 shared libraries as not backward compatible. Of course, this should
15714 be changed again when we can guarantee backward binary compatibility.
15718 * Fix typo in get_cert_by_subject() in by_dir.c
15720 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15722 * Rework the system to generate shared libraries:
15724 - Make note of the expected extension for the shared libraries and
15725 if there is a need for symbolic links from for example libcrypto.so.0
15726 to libcrypto.so.0.9.7. There is extended info in Configure for
15729 - Make as few rebuilds of the shared libraries as possible.
15731 - Still avoid linking the OpenSSL programs with the shared libraries.
15733 - When installing, install the shared libraries separately from the
15738 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15740 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15741 and not in SSL_clear because the latter is also used by the
15742 accept/connect functions; previously, the settings made by
15743 SSL_set_read_ahead would be lost during the handshake.
15745 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15747 * Correct util/mkdef.pl to be selective about disabled algorithms.
15748 Previously, it would create entries for disabled algorithms no
15753 * Added several new manual pages for SSL_* function.
15757 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15759 * In ssl23_get_client_hello, generate an error message when faced
15760 with an initial SSL 3.0/TLS record that is too small to contain the
15761 first two bytes of the ClientHello message, i.e. client_version.
15762 (Note that this is a pathologic case that probably has never happened
15763 in real life.) The previous approach was to use the version number
15764 from the record header as a substitute; but our protocol choice
15765 should not depend on that one because it is not authenticated
15766 by the Finished messages.
15770 * More robust randomness gathering functions for Windows.
15772 *Jeffrey Altman <jaltman@columbia.edu>*
15774 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15775 not set then we don't setup the error code for issuer check errors
15776 to avoid possibly overwriting other errors which the callback does
15777 handle. If an application does set the flag then we assume it knows
15778 what it is doing and can handle the new informational codes
15783 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15784 a general "ANY" type, as such it should be able to decode anything
15785 including tagged types. However it didn't check the class so it would
15786 wrongly interpret tagged types in the same way as their universal
15787 counterpart and unknown types were just rejected. Changed so that the
15788 tagged and unknown types are handled in the same way as a SEQUENCE:
15789 that is the encoding is stored intact. There is also a new type
15790 "V_ASN1_OTHER" which is used when the class is not universal, in this
15791 case we have no idea what the actual type is so we just lump them all
15796 * On VMS, stdout may very well lead to a file that is written to
15797 in a record-oriented fashion. That means that every write() will
15798 write a separate record, which will be read separately by the
15799 programs trying to read from it. This can be very confusing.
15801 The solution is to put a BIO filter in the way that will buffer
15802 text until a linefeed is reached, and then write everything a
15803 line at a time, so every record written will be an actual line,
15804 not chunks of lines and not (usually doesn't happen, but I've
15805 seen it once) several lines in one record. BIO_f_linebuffer() is
15808 Currently, it's a VMS-only method, because that's where it has
15809 been tested well enough.
15813 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15814 it can return incorrect results.
15815 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15816 but it was in 0.9.6-beta[12].)
15820 * Disable the check for content being present when verifying detached
15821 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15822 include zero length content when signing messages.
15826 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15827 BIO_ctrl (for BIO pairs).
15831 * Add DSO method for VMS.
15835 * Bug fix: Montgomery multiplication could produce results with the
15840 * Add RPM specification openssl.spec and modify it to build three
15841 packages. The default package contains applications, application
15842 documentation and run-time libraries. The devel package contains
15843 include files, static libraries and function documentation. The
15844 doc package contains the contents of the doc directory. The original
15845 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15849 * Add a large number of documentation files for many SSL routines.
15851 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15853 * Add a configuration entry for Sony News 4.
15855 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15857 * Don't set the two most significant bits to one when generating a
15858 random number < q in the DSA library.
15862 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15863 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15864 the underlying transport is blocking) if a handshake took place.
15865 (The default behaviour is needed by applications such as s_client
15866 and s_server that use select() to determine when to use SSL_read;
15867 but for applications that know in advance when to expect data, it
15868 just makes things more complicated.)
15872 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15877 * Add a few more EBCDIC conditionals that make `req` and `x509`
15878 work better on such systems.
15880 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15882 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15883 Update PKCS12_parse() so it copies the friendlyName and the
15884 keyid to the certificates aux info.
15888 * Fix bug in PKCS7_verify() which caused an infinite loop
15889 if there was more than one signature.
15891 *Sven Uszpelkat <su@celocom.de>*
15893 * Major change in util/mkdef.pl to include extra information
15894 about each symbol, as well as presenting variables as well
15895 as functions. This change means that there's n more need
15896 to rebuild the .num files when some algorithms are excluded.
15900 * Allow the verify time to be set by an application,
15901 rather than always using the current time.
15905 * Phase 2 verify code reorganisation. The certificate
15906 verify code now looks up an issuer certificate by a
15907 number of criteria: subject name, authority key id
15908 and key usage. It also verifies self signed certificates
15909 by the same criteria. The main comparison function is
15910 X509_check_issued() which performs these checks.
15912 Lot of changes were necessary in order to support this
15913 without completely rewriting the lookup code.
15915 Authority and subject key identifier are now cached.
15917 The LHASH 'certs' is X509_STORE has now been replaced
15918 by a STACK_OF(X509_OBJECT). This is mainly because an
15919 LHASH can't store or retrieve multiple objects with
15920 the same hash value.
15922 As a result various functions (which were all internal
15923 use only) have changed to handle the new X509_STORE
15924 structure. This will break anything that messed round
15925 with X509_STORE internally.
15927 The functions X509_STORE_add_cert() now checks for an
15928 exact match, rather than just subject name.
15930 The X509_STORE API doesn't directly support the retrieval
15931 of multiple certificates matching a given criteria, however
15932 this can be worked round by performing a lookup first
15933 (which will fill the cache with candidate certificates)
15934 and then examining the cache for matches. This is probably
15935 the best we can do without throwing out X509_LOOKUP
15936 entirely (maybe later...).
15938 The X509_VERIFY_CTX structure has been enhanced considerably.
15940 All certificate lookup operations now go via a get_issuer()
15941 callback. Although this currently uses an X509_STORE it
15942 can be replaced by custom lookups. This is a simple way
15943 to bypass the X509_STORE hackery necessary to make this
15944 work and makes it possible to use more efficient techniques
15945 in future. A very simple version which uses a simple
15946 STACK for its trusted certificate store is also provided
15947 using X509_STORE_CTX_trusted_stack().
15949 The verify_cb() and verify() callbacks now have equivalents
15950 in the X509_STORE_CTX structure.
15952 X509_STORE_CTX also has a 'flags' field which can be used
15953 to customise the verify behaviour.
15957 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15958 excludes S/MIME capabilities.
15962 * When a certificate request is read in keep a copy of the
15963 original encoding of the signed data and use it when outputting
15964 again. Signatures then use the original encoding rather than
15965 a decoded, encoded version which may cause problems if the
15966 request is improperly encoded.
15970 * For consistency with other BIO_puts implementations, call
15971 buffer_write(b, ...) directly in buffer_puts instead of calling
15974 In BIO_puts, increment b->num_write as in BIO_write.
15976 *Peter.Sylvester@EdelWeb.fr*
15978 * Fix BN_mul_word for the case where the word is 0. (We have to use
15979 BN_zero, we may not return a BIGNUM with an array consisting of
15980 words set to zero.)
15984 * Avoid calling abort() from within the library when problems are
15985 detected, except if preprocessor symbols have been defined
15986 (such as REF_CHECK, BN_DEBUG etc.).
15990 * New openssl application 'rsautl'. This utility can be
15991 used for low-level RSA operations. DER public key
15992 BIO/fp routines also added.
15996 * New Configure entry and patches for compiling on QNX 4.
15998 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16000 * A demo state-machine implementation was sponsored by
16001 Nuron (<http://www.nuron.com/>) and is now available in
16002 demos/state_machine.
16006 * New options added to the 'dgst' utility for signature
16007 generation and verification.
16011 * Unrecognized PKCS#7 content types are now handled via a
16012 catch all ASN1_TYPE structure. This allows unsupported
16013 types to be stored as a "blob" and an application can
16014 encode and decode it manually.
16018 * Fix various signed/unsigned issues to make a_strex.c
16019 compile under VC++.
16021 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16023 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16024 length if passed a buffer. ASN1_INTEGER_to_BN failed
16025 if passed a NULL BN and its argument was negative.
16027 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16029 * Modification to PKCS#7 encoding routines to output definite
16030 length encoding. Since currently the whole structures are in
16031 memory there's not real point in using indefinite length
16032 constructed encoding. However if OpenSSL is compiled with
16033 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16037 * Added BIO_vprintf() and BIO_vsnprintf().
16041 * Added more prefixes to parse for in the strings written
16042 through a logging bio, to cover all the levels that are available
16043 through syslog. The prefixes are now:
16045 PANIC, EMERG, EMR => LOG_EMERG
16046 ALERT, ALR => LOG_ALERT
16047 CRIT, CRI => LOG_CRIT
16048 ERROR, ERR => LOG_ERR
16049 WARNING, WARN, WAR => LOG_WARNING
16050 NOTICE, NOTE, NOT => LOG_NOTICE
16051 INFO, INF => LOG_INFO
16052 DEBUG, DBG => LOG_DEBUG
16054 and as before, if none of those prefixes are present at the
16055 beginning of the string, LOG_ERR is chosen.
16057 On Win32, the `LOG_*` levels are mapped according to this:
16059 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16060 LOG_WARNING => EVENTLOG_WARNING_TYPE
16061 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16065 * Made it possible to reconfigure with just the configuration
16066 argument "reconf" or "reconfigure". The command line arguments
16067 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16068 and are retrieved from there when reconfiguring.
16074 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16076 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16080 * The obj_dat.pl script was messing up the sorting of object
16081 names. The reason was that it compared the quoted version
16082 of strings as a result "OCSP" > "OCSP Signing" because
16083 " > SPACE. Changed script to store unquoted versions of
16084 names and add quotes on output. It was also omitting some
16085 names from the lookup table if they were given a default
16086 value (that is if SN is missing it is given the same
16087 value as LN and vice versa), these are now added on the
16088 grounds that if an object has a name we should be able to
16089 look it up. Finally added warning output when duplicate
16090 short or long names are found.
16094 * Changes needed for Tandem NSK.
16096 *Scott Uroff <scott@xypro.com>*
16098 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16099 RSA_padding_check_SSLv23(), special padding was never detected
16100 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16101 version rollback attacks was not effective.
16103 In s23_clnt.c, don't use special rollback-attack detection padding
16104 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16105 client; similarly, in s23_srvr.c, don't do the rollback check if
16106 SSL 2.0 is the only protocol enabled in the server.
16110 * Make it possible to get hexdumps of unprintable data with 'openssl
16111 asn1parse'. By implication, the functions ASN1_parse_dump() and
16112 BIO_dump_indent() are added.
16116 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16117 these print out strings and name structures based on various
16118 flags including RFC2253 support and proper handling of
16119 multibyte characters. Added options to the 'x509' utility
16120 to allow the various flags to be set.
16124 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16125 Also change the functions X509_cmp_current_time() and
16126 X509_gmtime_adj() work with an ASN1_TIME structure,
16127 this will enable certificates using GeneralizedTime in validity
16128 dates to be checked.
16132 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16133 negative public key encodings) on by default,
16134 NO_NEG_PUBKEY_BUG can be set to disable it.
16138 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16139 content octets. An i2c_ASN1_OBJECT is unnecessary because
16140 the encoding can be trivially obtained from the structure.
16144 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16145 not read locks (`CRYPTO_r_[un]lock`).
16149 * A first attempt at creating official support for shared
16150 libraries through configuration. I've kept it so the
16151 default is static libraries only, and the OpenSSL programs
16152 are always statically linked for now, but there are
16153 preparations for dynamic linking in place.
16154 This has been tested on Linux and Tru64.
16158 * Randomness polling function for Win9x, as described in:
16159 Peter Gutmann, Software Generation of Practically Strong
16164 * Fix so PRNG is seeded in req if using an already existing
16169 * New options to smime application. -inform and -outform
16170 allow alternative formats for the S/MIME message including
16171 PEM and DER. The -content option allows the content to be
16172 specified separately. This should allow things like Netscape
16173 form signing output easier to verify.
16177 * Fix the ASN1 encoding of tags using the 'long form'.
16181 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16182 STRING types. These convert content octets to and from the
16183 underlying type. The actual tag and length octets are
16184 already assumed to have been read in and checked. These
16185 are needed because all other string types have virtually
16186 identical handling apart from the tag. By having versions
16187 of the ASN1 functions that just operate on content octets
16188 IMPLICIT tagging can be handled properly. It also allows
16189 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16190 and ASN1_INTEGER are identical apart from the tag.
16194 * Change the handling of OID objects as follows:
16196 - New object identifiers are inserted in objects.txt, following
16197 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16198 - objects.pl is used to process obj_mac.num and create a new
16200 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16203 This is currently kind of a hack, and the perl code in objects.pl
16204 isn't very elegant, but it works as I intended. The simplest way
16205 to check that it worked correctly is to look in obj_dat.h and
16206 check the array nid_objs and make sure the objects haven't moved
16207 around (this is important!). Additions are OK, as well as
16208 consistent name changes.
16212 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16216 * Addition of the command line parameter '-rand file' to 'openssl req'.
16217 The given file adds to whatever has already been seeded into the
16218 random pool through the RANDFILE configuration file option or
16219 environment variable, or the default random state file.
16223 * mkstack.pl now sorts each macro group into lexical order.
16224 Previously the output order depended on the order the files
16225 appeared in the directory, resulting in needless rewriting
16230 * Patches to make OpenSSL compile under Win32 again. Mostly
16231 work arounds for the VC++ problem that it treats func() as
16232 func(void). Also stripped out the parts of mkdef.pl that
16233 added extra typesafe functions: these no longer exist.
16237 * Reorganisation of the stack code. The macros are now all
16238 collected in safestack.h . Each macro is defined in terms of
16239 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16240 DEBUG_SAFESTACK is now handled in terms of function casts,
16241 this has the advantage of retaining type safety without the
16242 use of additional functions. If DEBUG_SAFESTACK is not defined
16243 then the non typesafe macros are used instead. Also modified the
16244 mkstack.pl script to handle the new form. Needs testing to see
16245 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16246 the default if no major problems. Similar behaviour for ASN1_SET_OF
16247 and PKCS12_STACK_OF.
16251 * When some versions of IIS use the 'NET' form of private key the
16252 key derivation algorithm is different. Normally MD5(password) is
16253 used as a 128 bit RC4 key. In the modified case
16254 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16255 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16256 as the old Netscape_RSA functions except they have an additional
16257 'sgckey' parameter which uses the modified algorithm. Also added
16258 an -sgckey command line option to the rsa utility. Thanks to
16259 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16260 algorithm to openssl-dev.
16264 * The evp_local.h macros were using 'c.##kname' which resulted in
16265 invalid expansion on some systems (SCO 5.0.5 for example).
16266 Corrected to 'c.kname'.
16268 *Phillip Porch <root@theporch.com>*
16270 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16271 a STACK of email addresses from a certificate or request, these look
16272 in the subject name and the subject alternative name extensions and
16273 omit any duplicate addresses.
16277 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16278 This makes DSA verification about 2 % faster.
16282 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16283 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16284 plus overhead for 1024 bit moduli).
16285 This makes exponentiations about 0.5 % faster for 1024 bit
16286 exponents (as measured by "openssl speed rsa2048").
16290 * Rename memory handling macros to avoid conflicts with other
16292 Malloc => OPENSSL_malloc
16293 Malloc_locked => OPENSSL_malloc_locked
16294 Realloc => OPENSSL_realloc
16295 Free => OPENSSL_free
16299 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16300 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16304 * CygWin32 support.
16306 *John Jarvie <jjarvie@newsguy.com>*
16308 * The type-safe stack code has been rejigged. It is now only compiled
16309 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16310 by default all type-specific stack functions are "#define"d back to
16311 standard stack functions. This results in more streamlined output
16312 but retains the type-safety checking possibilities of the original
16317 * The STACK code has been cleaned up, and certain type declarations
16318 that didn't make a lot of sense have been brought in line. This has
16319 also involved a cleanup of sorts in safestack.h to more correctly
16320 map type-safe stack functions onto their plain stack counterparts.
16321 This work has also resulted in a variety of "const"ifications of
16322 lots of the code, especially `_cmp` operations which should normally
16323 be prototyped with "const" parameters anyway.
16327 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16328 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16329 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16330 where all of 'md' is used each time the PRNG is used, but 'state'
16331 is used only indexed by a cyclic counter. As entropy may not be
16332 well distributed from the beginning, 'md' is important as a
16333 chaining variable. However, the output function chains only half
16334 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16335 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16336 in all of 'state' being rewritten, with the new values depending
16337 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16341 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16342 the handshake is continued after ssl_verify_cert_chain();
16343 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16344 can lead to 'unexplainable' connection aborts later.
16346 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16348 * Major EVP API cipher revision.
16349 Add hooks for extra EVP features. This allows various cipher
16350 parameters to be set in the EVP interface. Support added for variable
16351 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16352 setting of RC2 and RC5 parameters.
16354 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16357 Remove lots of duplicated code from the EVP library. For example *every*
16358 cipher init() function handles the 'iv' in the same way according to the
16359 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16360 for CFB and OFB modes they zero ctx->num.
16362 New functionality allows removal of S/MIME code RC2 hack.
16364 Most of the routines have the same form and so can be declared in terms
16367 By shifting this to the top level EVP_CipherInit() it can be removed from
16368 all individual ciphers. If the cipher wants to handle IVs or keys
16369 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16372 Change lots of functions like EVP_EncryptUpdate() to now return a
16373 value: although software versions of the algorithms cannot fail
16374 any installed hardware versions can.
16378 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16379 this option is set, tolerate broken clients that send the negotiated
16380 protocol version number instead of the requested protocol version
16385 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16386 i.e. non-zero for export ciphersuites, zero otherwise.
16387 Previous versions had this flag inverted, inconsistent with
16388 rsa_tmp_cb (..._TMP_RSA_CB).
16390 *Bodo Moeller; problem reported by Amit Chopra*
16392 * Add missing DSA library text string. Work around for some IIS
16393 key files with invalid SEQUENCE encoding.
16397 * Add a document (doc/standards.txt) that list all kinds of standards
16398 and so on that are implemented in OpenSSL.
16402 * Enhance c_rehash script. Old version would mishandle certificates
16403 with the same subject name hash and wouldn't handle CRLs at all.
16404 Added -fingerprint option to crl utility, to support new c_rehash
16409 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16413 * Fix for SSL server purpose checking. Server checking was
16414 rejecting certificates which had extended key usage present
16415 but no ssl client purpose.
16417 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16419 * Make PKCS#12 code work with no password. The PKCS#12 spec
16420 is a little unclear about how a blank password is handled.
16421 Since the password in encoded as a BMPString with terminating
16422 double NULL a zero length password would end up as just the
16423 double NULL. However no password at all is different and is
16424 handled differently in the PKCS#12 key generation code. NS
16425 treats a blank password as zero length. MSIE treats it as no
16426 password on export: but it will try both on import. We now do
16427 the same: PKCS12_parse() tries zero length and no password if
16428 the password is set to "" or NULL (NULL is now a valid password:
16429 it wasn't before) as does the pkcs12 application.
16433 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16434 perror when PEM_read_bio_X509_REQ fails, the error message must
16435 be obtained from the error queue.
16439 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16440 it in ERR_remove_state if appropriate, and change ERR_get_state
16441 accordingly to avoid race conditions (this is necessary because
16442 thread_hash is no longer constant once set).
16446 * Bugfix for linux-elf makefile.one.
16450 * RSA_get_default_method() will now cause a default
16451 RSA_METHOD to be chosen if one doesn't exist already.
16452 Previously this was only set during a call to RSA_new()
16453 or RSA_new_method(NULL) meaning it was possible for
16454 RSA_get_default_method() to return NULL.
16458 * Added native name translation to the existing DSO code
16459 that will convert (if the flag to do so is set) filenames
16460 that are sufficiently small and have no path information
16461 into a canonical native form. Eg. "blah" converted to
16462 "libblah.so" or "blah.dll" etc.
16466 * New function ERR_error_string_n(e, buf, len) which is like
16467 ERR_error_string(e, buf), but writes at most 'len' bytes
16468 including the 0 terminator. For ERR_error_string_n, 'buf'
16471 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16473 * CONF library reworked to become more general. A new CONF
16474 configuration file reader "class" is implemented as well as a
16475 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16476 old `CONF_*` functions are still there, but are reimplemented to
16477 work in terms of the new functions. Also, a set of functions
16478 to handle the internal storage of the configuration data is
16479 provided to make it easier to write new configuration file
16480 reader "classes" (I can definitely see something reading a
16481 configuration file in XML format, for example), called `_CONF_*`,
16482 or "the configuration storage API"...
16484 The new configuration file reading functions are:
16486 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16487 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16489 NCONF_default, NCONF_WIN32
16491 NCONF_dump_fp, NCONF_dump_bio
16493 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16494 NCONF_new creates a new CONF object. This works in the same way
16495 as other interfaces in OpenSSL, like the BIO interface.
16496 `NCONF_dump_*` dump the internal storage of the configuration file,
16497 which is useful for debugging. All other functions take the same
16498 arguments as the old `CONF_*` functions with the exception of the
16499 first that must be a `CONF *` instead of a `LHASH *`.
16501 To make it easier to use the new classes with the old `CONF_*` functions,
16502 the function CONF_set_default_method is provided.
16506 * Add '-tls1' option to 'openssl ciphers', which was already
16507 mentioned in the documentation but had not been implemented.
16508 (This option is not yet really useful because even the additional
16509 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16513 * Initial DSO code added into libcrypto for letting OpenSSL (and
16514 OpenSSL-based applications) load shared libraries and bind to
16515 them in a portable way.
16517 *Geoff Thorpe, with contributions from Richard Levitte*
16519 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16521 * Make sure _lrotl and _lrotr are only used with MSVC.
16523 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16524 (the default implementation of RAND_status).
16526 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16527 to '-clrext' (= clear extensions), as intended and documented.
16528 *Bodo Moeller; inconsistency pointed out by Michael Attili
16529 <attili@amaxo.com>*
16531 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16532 was larger than the MD block size.
16534 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16536 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16537 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16538 using the passed key: if the passed key was a private key the result
16539 of X509_print(), for example, would be to print out all the private key
16544 * des_quad_cksum() byte order bug fix.
16545 *Ulf Möller, using the problem description in krb4-0.9.7, where
16546 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16548 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16551 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16553 * For easily testing in shell scripts whether some command
16554 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16555 returns with exit code 0 iff no command of the given name is available.
16556 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16557 the output goes to stdout and nothing is printed to stderr.
16558 Additional arguments are always ignored.
16560 Since for each cipher there is a command of the same name,
16561 the 'no-cipher' compilation switches can be tested this way.
16563 ('openssl no-XXX' is not able to detect pseudo-commands such
16564 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16568 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16572 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16573 is set; it will be thrown away anyway because each handshake creates
16575 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16576 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16577 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16578 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16582 * New s_client option -ign_eof: EOF at stdin is ignored, and
16583 'Q' and 'R' lose their special meanings (quit/renegotiate).
16584 This is part of what -quiet does; unlike -quiet, -ign_eof
16585 does not suppress any output.
16589 * Add compatibility options to the purpose and trust code. The
16590 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16591 accepts a certificate or CA, this was the previous behaviour,
16592 with all the associated security issues.
16594 X509_TRUST_COMPAT is the old trust behaviour: only and
16595 automatically trust self signed roots in certificate store. A
16596 new trust setting X509_TRUST_DEFAULT is used to specify that
16597 a purpose has no associated trust setting and it should instead
16598 use the value in the default purpose.
16602 * Fix the PKCS#8 DSA private key code so it decodes keys again
16603 and fix a memory leak.
16607 * In util/mkerr.pl (which implements 'make errors'), preserve
16608 reason strings from the previous version of the .c file, as
16609 the default to have only downcase letters (and digits) in
16610 automatically generated reasons codes is not always appropriate.
16614 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16615 using strerror. Previously, ERR_reason_error_string() returned
16616 library names as reason strings for SYSerr; but SYSerr is a special
16617 case where small numbers are errno values, not library numbers.
16621 * Add '-dsaparam' option to 'openssl dhparam' application. This
16622 converts DSA parameters into DH parameters. (When creating parameters,
16623 DSA_generate_parameters is used.)
16627 * Include 'length' (recommended exponent length) in C code generated
16628 by 'openssl dhparam -C'.
16632 * The second argument to set_label in perlasm was already being used
16633 so couldn't be used as a "file scope" flag. Moved to third argument
16638 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16639 instead of RAND_bytes for encryption IVs and salts.
16643 * Include RAND_status() into RAND_METHOD instead of implementing
16644 it only for md_rand.c Otherwise replacing the PRNG by calling
16645 RAND_set_rand_method would be impossible.
16649 * Don't let DSA_generate_key() enter an infinite loop if the random
16650 number generation fails.
16654 * New 'rand' application for creating pseudo-random output.
16658 * Added configuration support for Linux/IA64
16660 *Rolf Haberrecker <rolf@suse.de>*
16662 * Assembler module support for Mingw32.
16666 * Shared library support for HPUX (in shlib/).
16668 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16670 * Shared library support for Solaris gcc.
16672 *Lutz Behnke <behnke@trustcenter.de>*
16674 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16676 * PKCS7_encrypt() was adding text MIME headers twice because they
16677 were added manually and by SMIME_crlf_copy().
16681 * In bntest.c don't call BN_rand with zero bits argument.
16683 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16685 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16686 case was implemented. This caused BN_div_recp() to fail occasionally.
16690 * Add an optional second argument to the set_label() in the perl
16691 assembly language builder. If this argument exists and is set
16692 to 1 it signals that the assembler should use a symbol whose
16693 scope is the entire file, not just the current function. This
16694 is needed with MASM which uses the format label:: for this scope.
16696 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16698 * Change the ASN1 types so they are typedefs by default. Before
16699 almost all types were #define'd to ASN1_STRING which was causing
16700 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16705 * Change names of new functions to the new get1/get0 naming
16706 convention: After 'get1', the caller owns a reference count
16707 and has to call `..._free`; 'get0' returns a pointer to some
16708 data structure without incrementing reference counters.
16709 (Some of the existing 'get' functions increment a reference
16710 counter, some don't.)
16711 Similarly, 'set1' and 'add1' functions increase reference
16712 counters or duplicate objects.
16716 * Allow for the possibility of temp RSA key generation failure:
16717 the code used to assume it always worked and crashed on failure.
16721 * Fix potential buffer overrun problem in BIO_printf().
16722 *Ulf Möller, using public domain code by Patrick Powell; problem
16723 pointed out by David Sacerdote <das33@cornell.edu>*
16725 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16726 RAND_egd() and RAND_status(). In the command line application,
16727 the EGD socket can be specified like a seed file using RANDFILE
16732 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16733 Some CAs (e.g. Verisign) distribute certificates in this form.
16737 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16738 list to exclude them. This means that no special compilation option
16739 is needed to use anonymous DH: it just needs to be included in the
16744 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16745 EVP_MD_type. The old functionality is available in a new macro called
16746 EVP_MD_md(). Change code that uses it and update docs.
16750 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16751 where the `void *` argument is replaced by a function pointer argument.
16752 Previously `void *` was abused to point to functions, which works on
16753 many platforms, but is not correct. As these functions are usually
16754 called by macros defined in OpenSSL header files, most source code
16755 should work without changes.
16759 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16760 sections with information on -D... compiler switches used for
16761 compiling the library so that applications can see them. To enable
16762 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16763 must be defined. E.g.,
16764 #define OPENSSL_ALGORITHM_DEFINES
16765 #include <openssl/opensslconf.h>
16766 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16768 *Richard Levitte, Ulf and Bodo Möller*
16770 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16775 * Change the 'other' type in certificate aux info to a STACK_OF
16776 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16777 the required ASN1 format: arbitrary types determined by an OID.
16781 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16782 argument to 'req'. This is not because the function is newer or
16783 better than others it just uses the work 'NEW' in the certificate
16784 request header lines. Some software needs this.
16788 * Reorganise password command line arguments: now passwords can be
16789 obtained from various sources. Delete the PEM_cb function and make
16790 it the default behaviour: i.e. if the callback is NULL and the
16791 usrdata argument is not NULL interpret it as a null terminated pass
16792 phrase. If usrdata and the callback are NULL then the pass phrase
16793 is prompted for as usual.
16797 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16798 the support is automatically enabled. The resulting binaries will
16799 autodetect the card and use it if present.
16801 *Ben Laurie and Compaq Inc.*
16803 * Work around for Netscape hang bug. This sends certificate request
16804 and server done in one record. Since this is perfectly legal in the
16805 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16806 the bugs/SSLv3 entry for more info.
16810 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16814 * Add -rand argument to smime and pkcs12 applications and read/write
16819 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16823 * Add command line password options to the remaining applications.
16827 * Bug fix for BN_div_recp() for numerators with an even number of
16832 * More tests in bntest.c, and changed test_bn output.
16836 * ./config recognizes MacOS X now.
16840 * Bug fix for BN_div() when the first words of num and divisor are
16841 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16845 * Add support for various broken PKCS#8 formats, and command line
16846 options to produce them.
16850 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16851 get temporary BIGNUMs from a BN_CTX.
16855 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16860 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16861 include a #define from the old name to the new. The original intent
16862 was that statically linked binaries could for example just call
16863 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16864 link with digests. This never worked because SSLeay_add_all_digests()
16865 and SSLeay_add_all_ciphers() were in the same source file so calling
16866 one would link with the other. They are now in separate source files.
16870 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16874 * Use a less unusual form of the Miller-Rabin primality test (it used
16875 a binary algorithm for exponentiation integrated into the Miller-Rabin
16876 loop, our standard modexp algorithms are faster).
16880 * Support for the EBCDIC character set completed.
16882 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16884 * Source code cleanups: use const where appropriate, eliminate casts,
16885 use `void *` instead of `char *` in lhash.
16889 * Bugfix: ssl3_send_server_key_exchange was not restartable
16890 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16891 this the server could overwrite ephemeral keys that the client
16896 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16897 using 50 iterations of the Rabin-Miller test.
16899 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16900 iterations of the Rabin-Miller test as required by the appendix
16901 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16902 As BN_is_prime_fasttest includes trial division, DSA parameter
16903 generation becomes much faster.
16905 This implies a change for the callback functions in DSA_is_prime
16906 and DSA_generate_parameters: The callback function is called once
16907 for each positive witness in the Rabin-Miller test, not just
16908 occasionally in the inner loop; and the parameters to the
16909 callback function now provide an iteration count for the outer
16910 loop rather than for the current invocation of the inner loop.
16911 DSA_generate_parameters additionally can call the callback
16912 function with an 'iteration count' of -1, meaning that a
16913 candidate has passed the trial division test (when q is generated
16914 from an application-provided seed, trial division is skipped).
16918 * New function BN_is_prime_fasttest that optionally does trial
16919 division before starting the Rabin-Miller test and has
16920 an additional BN_CTX * argument (whereas BN_is_prime always
16921 has to allocate at least one BN_CTX).
16922 'callback(1, -1, cb_arg)' is called when a number has passed the
16923 trial division stage.
16927 * Fix for bug in CRL encoding. The validity dates weren't being handled
16932 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16936 * New function BN_pseudo_rand().
16940 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16941 bignum version of BN_from_montgomery() with the working code from
16942 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16947 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16948 made it impossible to use the same SSL_SESSION data structure in
16949 SSL2 clients in multiple threads.
16953 * The return value of RAND_load_file() no longer counts bytes obtained
16954 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16955 to seed the PRNG (previously an explicit byte count was required).
16957 *Ulf Möller, Bodo Möller*
16959 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16960 used `char *` instead of `void *` and had casts all over the place.
16964 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16968 * Retain source code compatibility for BN_prime_checks macro:
16969 BN_is_prime(..., BN_prime_checks, ...) now uses
16970 BN_prime_checks_for_size to determine the appropriate number of
16971 Rabin-Miller iterations.
16975 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16976 DH_CHECK_P_NOT_SAFE_PRIME.
16977 (Check if this is true? OpenPGP calls them "strong".)
16981 * Merge the functionality of "dh" and "gendh" programs into a new program
16982 "dhparam". The old programs are retained for now but will handle DH keys
16983 (instead of parameters) in future.
16987 * Make the ciphers, s_server and s_client programs check the return values
16988 when a new cipher list is set.
16992 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16993 ciphers. Before when the 56bit ciphers were enabled the sorting was
16996 The syntax for the cipher sorting has been extended to support sorting by
16997 cipher-strength (using the strength_bits hard coded in the tables).
16998 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17000 Fix a bug in the cipher-command parser: when supplying a cipher command
17001 string with an "undefined" symbol (neither command nor alphanumeric
17002 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17003 an error is flagged.
17005 Due to the strength-sorting extension, the code of the
17006 ssl_create_cipher_list() function was completely rearranged. I hope that
17007 the readability was also increased :-)
17009 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17011 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17012 for the first serial number and places 2 in the serial number file. This
17013 avoids problems when the root CA is created with serial number zero and
17014 the first user certificate has the same issuer name and serial number
17019 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17020 the new code. Add documentation for this stuff.
17024 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17025 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17026 structures and behave in an analogous way to the X509v3 functions:
17027 they shouldn't be called directly but wrapper functions should be used
17030 So we also now have some wrapper functions that call the X509at functions
17031 when passed certificate requests. (TO DO: similar things can be done with
17032 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17033 things. Some of these need some d2i or i2d and print functionality
17034 because they handle more complex structures.)
17038 * Add missing #ifndefs that caused missing symbols when building libssl
17039 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17040 NO_RSA in `ssl/s2*.c`.
17042 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17044 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17045 has a return value which indicates the quality of the random data
17046 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17047 error queue. New function RAND_pseudo_bytes() generates output that is
17048 guaranteed to be unique but not unpredictable. RAND_add is like
17049 RAND_seed, but takes an extra argument for an entropy estimate
17050 (RAND_seed always assumes full entropy).
17054 * Do more iterations of Rabin-Miller probable prime test (specifically,
17055 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17056 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17057 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17058 false-positive rate of at most 2^-80 for random input.
17062 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17066 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17067 in the 0.9.5 release), this returns the chain
17068 from an X509_CTX structure with a dup of the stack and all
17069 the X509 reference counts upped: so the stack will exist
17070 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17073 Also make SSL_SESSION_print() print out the verify return
17078 * Add manpage for the pkcs12 command. Also change the default
17079 behaviour so MAC iteration counts are used unless the new
17080 -nomaciter option is used. This improves file security and
17081 only older versions of MSIE (4.0 for example) need it.
17085 * Honor the no-xxx Configure options when creating .DEF files.
17089 * Add PKCS#10 attributes to field table: challengePassword,
17090 unstructuredName and unstructuredAddress. These are taken from
17091 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17092 international characters are used.
17094 More changes to X509_ATTRIBUTE code: allow the setting of types
17095 based on strings. Remove the 'loc' parameter when adding
17096 attributes because these will be a SET OF encoding which is sorted
17101 * Initial changes to the 'req' utility to allow request generation
17102 automation. This will allow an application to just generate a template
17103 file containing all the field values and have req construct the
17106 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17107 used all over the place including certificate requests and PKCS#7
17108 structures. They are currently handled manually where necessary with
17109 some primitive wrappers for PKCS#7. The new functions behave in a
17110 manner analogous to the X509 extension functions: they allow
17111 attributes to be looked up by NID and added.
17113 Later something similar to the X509V3 code would be desirable to
17114 automatically handle the encoding, decoding and printing of the
17115 more complex types. The string types like challengePassword can
17116 be handled by the string table functions.
17118 Also modified the multi byte string table handling. Now there is
17119 a 'global mask' which masks out certain types. The table itself
17120 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17121 is useful when for example there is only one permissible type
17122 (as in countryName) and using the mask might result in no valid
17127 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17128 SSL_get_peer_finished to allow applications to obtain the latest
17129 Finished messages sent to the peer or expected from the peer,
17130 respectively. (SSL_get_peer_finished is usually the Finished message
17131 actually received from the peer, otherwise the protocol will be aborted.)
17133 As the Finished message are message digests of the complete handshake
17134 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17135 be used for external authentication procedures when the authentication
17136 provided by SSL/TLS is not desired or is not enough.
17140 * Enhanced support for Alpha Linux is added. Now ./config checks if
17141 the host supports BWX extension and if Compaq C is present on the
17142 $PATH. Just exploiting of the BWX extension results in 20-30%
17143 performance kick for some algorithms, e.g. DES and RC4 to mention
17144 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17149 * Add support for MS "fast SGC". This is arguably a violation of the
17150 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17151 weak crypto and after checking the certificate is SGC a second one
17152 with strong crypto. MS SGC stops the first handshake after receiving
17153 the server certificate message and sends a second client hello. Since
17154 a server will typically do all the time consuming operations before
17155 expecting any further messages from the client (server key exchange
17156 is the most expensive) there is little difference between the two.
17158 To get OpenSSL to support MS SGC we have to permit a second client
17159 hello message after we have sent server done. In addition we have to
17160 reset the MAC if we do get this second client hello.
17164 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17165 if a DER encoded private key is RSA or DSA traditional format. Changed
17166 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17167 format DER encoded private key. Newer code should use PKCS#8 format which
17168 has the key type encoded in the ASN1 structure. Added DER private key
17169 support to pkcs8 application.
17173 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17174 ciphersuites has been selected (as required by the SSL 3/TLS 1
17175 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17176 is set, we interpret this as a request to violate the specification
17177 (the worst that can happen is a handshake failure, and 'correct'
17178 behaviour would result in a handshake failure anyway).
17182 * In SSL_CTX_add_session, take into account that there might be multiple
17183 SSL_SESSION structures with the same session ID (e.g. when two threads
17184 concurrently obtain them from an external cache).
17185 The internal cache can handle only one SSL_SESSION with a given ID,
17186 so if there's a conflict, we now throw out the old one to achieve
17191 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17192 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17193 some routines that use cipher OIDs: some ciphers do not have OIDs
17194 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17199 * Simplify the trust setting structure and code. Now we just have
17200 two sequences of OIDs for trusted and rejected settings. These will
17201 typically have values the same as the extended key usage extension
17202 and any application specific purposes.
17204 The trust checking code now has a default behaviour: it will just
17205 check for an object with the same NID as the passed id. Functions can
17206 be provided to override either the default behaviour or the behaviour
17207 for a given id. SSL client, server and email already have functions
17208 in place for compatibility: they check the NID and also return "trusted"
17209 if the certificate is self signed.
17213 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17214 traditional format into an EVP_PKEY structure.
17218 * Add a password callback function PEM_cb() which either prompts for
17219 a password if usr_data is NULL or otherwise assumes it is a null
17220 terminated password. Allow passwords to be passed on command line
17221 environment or config files in a few more utilities.
17225 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17226 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17227 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17228 Update documentation.
17232 * Support for ASN1 "NULL" type. This could be handled before by using
17233 ASN1_TYPE but there wasn't any function that would try to read a NULL
17234 and produce an error if it couldn't. For compatibility we also have
17235 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17236 don't allocate anything because they don't need to.
17240 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17243 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17245 * Rebuild of the memory allocation routines used by OpenSSL code and
17246 possibly others as well. The purpose is to make an interface that
17247 provide hooks so anyone can build a separate set of allocation and
17248 deallocation routines to be used by OpenSSL, for example memory
17249 pool implementations, or something else, which was previously hard
17250 since Malloc(), Realloc() and Free() were defined as macros having
17251 the values malloc, realloc and free, respectively (except for Win32
17252 compilations). The same is provided for memory debugging code.
17253 OpenSSL already comes with functionality to find memory leaks, but
17254 this gives people a chance to debug other memory problems.
17256 With these changes, a new set of functions and macros have appeared:
17258 CRYPTO_set_mem_debug_functions() [F]
17259 CRYPTO_get_mem_debug_functions() [F]
17260 CRYPTO_dbg_set_options() [F]
17261 CRYPTO_dbg_get_options() [F]
17262 CRYPTO_malloc_debug_init() [M]
17264 The memory debug functions are NULL by default, unless the library
17265 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17266 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17267 gives the standard debugging functions that come with OpenSSL) or
17268 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17269 provided by the library user) must be used. When the standard
17270 debugging functions are used, CRYPTO_dbg_set_options can be used to
17271 request additional information:
17272 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17273 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17275 Also, things like CRYPTO_set_mem_functions will always give the
17276 expected result (the new set of functions is used for allocation
17277 and deallocation) at all times, regardless of platform and compiler
17280 To finish it up, some functions that were never use in any other
17281 way than through macros have a new API and new semantic:
17283 CRYPTO_dbg_malloc()
17284 CRYPTO_dbg_realloc()
17287 All macros of value have retained their old syntax.
17289 *Richard Levitte and Bodo Moeller*
17291 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17292 ordering of SMIMECapabilities wasn't in "strength order" and there
17293 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17298 * Some ASN1 types with illegal zero length encoding (INTEGER,
17299 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17301 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17303 * Merge in my S/MIME library for OpenSSL. This provides a simple
17304 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17305 functionality to handle multipart/signed properly) and a utility
17306 called 'smime' to call all this stuff. This is based on code I
17307 originally wrote for Celo who have kindly allowed it to be
17308 included in OpenSSL.
17312 * Add variants des_set_key_checked and des_set_key_unchecked of
17313 des_set_key (aka des_key_sched). Global variable des_check_key
17314 decides which of these is called by des_set_key; this way
17315 des_check_key behaves as it always did, but applications and
17316 the library itself, which was buggy for des_check_key == 1,
17317 have a cleaner way to pick the version they need.
17321 * New function PKCS12_newpass() which changes the password of a
17326 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17327 dynamic mix. In both cases the ids can be used as an index into the
17328 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17329 functions so they accept a list of the field values and the
17330 application doesn't need to directly manipulate the X509_TRUST
17335 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17340 * Modify the way the V3 extension code looks up extensions. This now
17341 works in a similar way to the object code: we have some "standard"
17342 extensions in a static table which is searched with OBJ_bsearch()
17343 and the application can add dynamic ones if needed. The file
17344 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17345 updated whenever a new extension is added to the core code and kept
17346 in ext_nid order. There is a simple program 'tabtest.c' which checks
17347 this. New extensions are not added too often so this file can readily
17348 be maintained manually.
17350 There are two big advantages in doing things this way. The extensions
17351 can be looked up immediately and no longer need to be "added" using
17352 X509V3_add_standard_extensions(): this function now does nothing.
17353 Side note: I get *lots* of email saying the extension code doesn't
17354 work because people forget to call this function.
17355 Also no dynamic allocation is done unless new extensions are added:
17356 so if we don't add custom extensions there is no need to call
17357 X509V3_EXT_cleanup().
17361 * Modify enc utility's salting as follows: make salting the default. Add a
17362 magic header, so unsalted files fail gracefully instead of just decrypting
17363 to garbage. This is because not salting is a big security hole, so people
17364 should be discouraged from doing it.
17368 * Fixes and enhancements to the 'x509' utility. It allowed a message
17369 digest to be passed on the command line but it only used this
17370 parameter when signing a certificate. Modified so all relevant
17371 operations are affected by the digest parameter including the
17372 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17373 DSA key was used because it didn't fix the digest.
17377 * Initial certificate chain verify code. Currently tests the untrusted
17378 certificates for consistency with the verify purpose (which is set
17379 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17381 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17382 this is because it will reject chains with invalid extensions whereas
17383 every previous version of OpenSSL and SSLeay made no checks at all.
17385 Trust code: checks the root CA for the relevant trust settings. Trust
17386 settings have an initial value consistent with the verify purpose: e.g.
17387 if the verify purpose is for SSL client use it expects the CA to be
17388 trusted for SSL client use. However the default value can be changed to
17389 permit custom trust settings: one example of this would be to only trust
17390 certificates from a specific "secure" set of CAs.
17392 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17393 which should be used for version portability: especially since the
17394 verify structure is likely to change more often now.
17396 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17397 to set them. If not set then assume SSL clients will verify SSL servers
17400 Two new options to the verify program: -untrusted allows a set of
17401 untrusted certificates to be passed in and -purpose which sets the
17402 intended purpose of the certificate. If a purpose is set then the
17403 new chain verify code is used to check extension consistency.
17407 * Support for the authority information access extension.
17411 * Modify RSA and DSA PEM read routines to transparently handle
17412 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17413 public keys in a format compatible with certificate
17414 SubjectPublicKeyInfo structures. Unfortunately there were already
17415 functions called *_PublicKey_* which used various odd formats so
17416 these are retained for compatibility: however the DSA variants were
17417 never in a public release so they have been deleted. Changed dsa/rsa
17418 utilities to handle the new format: note no releases ever handled public
17419 keys so we should be OK.
17421 The primary motivation for this change is to avoid the same fiasco
17422 that dogs private keys: there are several incompatible private key
17423 formats some of which are standard and some OpenSSL specific and
17424 require various evil hacks to allow partial transparent handling and
17425 even then it doesn't work with DER formats. Given the option anything
17426 other than PKCS#8 should be dumped: but the other formats have to
17427 stay in the name of compatibility.
17429 With public keys and the benefit of hindsight one standard format
17430 is used which works with EVP_PKEY, RSA or DSA structures: though
17431 it clearly returns an error if you try to read the wrong kind of key.
17433 Added a -pubkey option to the 'x509' utility to output the public key.
17434 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17435 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17436 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17437 that do the same as the `EVP_PKEY_assign_*()` except they up the
17438 reference count of the added key (they don't "swallow" the
17443 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17444 CRLs would fail if the file contained no certificates or no CRLs:
17445 added a new function to read in both types and return the number
17446 read: this means that if none are read it will be an error. The
17447 DER versions of the certificate and CRL reader would always fail
17448 because it isn't possible to mix certificates and CRLs in DER format
17449 without choking one or the other routine. Changed this to just read
17450 a certificate: this is the best we can do. Also modified the code
17451 in `apps/verify.c` to take notice of return codes: it was previously
17452 attempting to read in certificates from NULL pointers and ignoring
17453 any errors: this is one reason why the cert and CRL reader seemed
17454 to work. It doesn't check return codes from the default certificate
17455 routines: these may well fail if the certificates aren't installed.
17459 * Code to support otherName option in GeneralName.
17463 * First update to verify code. Change the verify utility
17464 so it warns if it is passed a self signed certificate:
17465 for consistency with the normal behaviour. X509_verify
17466 has been modified to it will now verify a self signed
17467 certificate if *exactly* the same certificate appears
17468 in the store: it was previously impossible to trust a
17469 single self signed certificate. This means that:
17470 openssl verify ss.pem
17471 now gives a warning about a self signed certificate but
17472 openssl verify -CAfile ss.pem ss.pem
17477 * For servers, store verify_result in SSL_SESSION data structure
17478 (and add it to external session representation).
17479 This is needed when client certificate verifications fails,
17480 but an application-provided verification callback (set by
17481 SSL_CTX_set_cert_verify_callback) allows accepting the session
17482 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17483 but returns 1): When the session is reused, we have to set
17484 ssl->verify_result to the appropriate error code to avoid
17487 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17489 * Fix a bug in the new PKCS#7 code: it didn't consider the
17490 case in PKCS7_dataInit() where the signed PKCS7 structure
17491 didn't contain any existing data because it was being created.
17493 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17495 * Add a salt to the key derivation routines in enc.c. This
17496 forms the first 8 bytes of the encrypted file. Also add a
17497 -S option to allow a salt to be input on the command line.
17501 * New function X509_cmp(). Oddly enough there wasn't a function
17502 to compare two certificates. We do this by working out the SHA1
17503 hash and comparing that. X509_cmp() will be needed by the trust
17508 * SSL_get1_session() is like SSL_get_session(), but increments
17509 the reference count in the SSL_SESSION returned.
17511 *Geoff Thorpe <geoff@eu.c2.net>*
17513 * Fix for 'req': it was adding a null to request attributes.
17514 Also change the X509_LOOKUP and X509_INFO code to handle
17515 certificate auxiliary information.
17519 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17524 * Add the possibility to add extra information to the memory leak
17525 detecting output, to form tracebacks, showing from where each
17526 allocation was originated: CRYPTO_push_info("constant string") adds
17527 the string plus current file name and line number to a per-thread
17528 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17529 is like calling CYRPTO_pop_info() until the stack is empty.
17530 Also updated memory leak detection code to be multi-thread-safe.
17534 * Add options -text and -noout to pkcs7 utility and delete the
17535 encryption options which never did anything. Update docs.
17539 * Add options to some of the utilities to allow the pass phrase
17540 to be included on either the command line (not recommended on
17541 OSes like Unix) or read from the environment. Update the
17542 manpages and fix a few bugs.
17546 * Add a few manpages for some of the openssl commands.
17550 * Fix the -revoke option in ca. It was freeing up memory twice,
17551 leaking and not finding already revoked certificates.
17555 * Extensive changes to support certificate auxiliary information.
17556 This involves the use of X509_CERT_AUX structure and X509_AUX
17557 functions. An X509_AUX function such as PEM_read_X509_AUX()
17558 can still read in a certificate file in the usual way but it
17559 will also read in any additional "auxiliary information". By
17560 doing things this way a fair degree of compatibility can be
17561 retained: existing certificates can have this information added
17562 using the new 'x509' options.
17564 Current auxiliary information includes an "alias" and some trust
17565 settings. The trust settings will ultimately be used in enhanced
17566 certificate chain verification routines: currently a certificate
17567 can only be trusted if it is self signed and then it is trusted
17572 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17573 The problem was that one of the replacement routines had not been working
17574 since SSLeay releases. For now the offending routine has been replaced
17575 with non-optimised assembler. Even so, this now gives around 95%
17576 performance improvement for 1024 bit RSA signs.
17580 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17581 handling. Most clients have the effective key size in bits equal to
17582 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17583 A few however don't do this and instead use the size of the decrypted key
17584 to determine the RC2 key length and the AlgorithmIdentifier to determine
17585 the effective key length. In this case the effective key length can still
17586 be 40 bits but the key length can be 168 bits for example. This is fixed
17587 by manually forcing an RC2 key into the EVP_PKEY structure because the
17588 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17589 the key length and effective key length are equal.
17593 * Add a bunch of functions that should simplify the creation of
17594 X509_NAME structures. Now you should be able to do:
17595 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17596 and have it automatically work out the correct field type and fill in
17597 the structures. The more adventurous can try:
17598 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17599 and it will (hopefully) work out the correct multibyte encoding.
17603 * Change the 'req' utility to use the new field handling and multibyte
17604 copy routines. Before the DN field creation was handled in an ad hoc
17605 way in req, ca, and x509 which was rather broken and didn't support
17606 BMPStrings or UTF8Strings. Since some software doesn't implement
17607 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17608 using the dirstring_type option. See the new comment in the default
17609 openssl.cnf for more info.
17613 * Make crypto/rand/md_rand.c more robust:
17614 - Assure unique random numbers after fork().
17615 - Make sure that concurrent threads access the global counter and
17616 md serializably so that we never lose entropy in them
17617 or use exactly the same state in multiple threads.
17618 Access to the large state is not always serializable because
17619 the additional locking could be a performance killer, and
17620 md should be large enough anyway.
17624 * New file `apps/app_rand.c` with commonly needed functionality
17625 for handling the random seed file.
17627 Use the random seed file in some applications that previously did not:
17629 dsaparam -genkey (which also ignored its '-rand' option),
17632 x509 (when signing).
17633 Except on systems with /dev/urandom, it is crucial to have a random
17634 seed file at least for key creation, DSA signing, and for DH exchanges;
17635 for RSA signatures we could do without one.
17637 gendh and gendsa (unlike genrsa) used to read only the first byte
17638 of each file listed in the '-rand' option. The function as previously
17639 found in genrsa is now in app_rand.c and is used by all programs
17640 that support '-rand'.
17644 * In RAND_write_file, use mode 0600 for creating files;
17645 don't just chmod when it may be too late.
17649 * Report an error from X509_STORE_load_locations
17650 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17654 * New function ASN1_mbstring_copy() this copies a string in either
17655 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17656 into an ASN1_STRING type. A mask of permissible types is passed
17657 and it chooses the "minimal" type to use or an error if not type
17662 * Add function equivalents to the various macros in asn1.h. The old
17663 macros are retained with an `M_` prefix. Code inside the library can
17664 use the `M_` macros. External code (including the openssl utility)
17665 should *NOT* in order to be "shared library friendly".
17669 * Add various functions that can check a certificate's extensions
17670 to see if it usable for various purposes such as SSL client,
17671 server or S/MIME and CAs of these types. This is currently
17672 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17673 verification. Also added a -purpose flag to x509 utility to
17674 print out all the purposes.
17678 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17683 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17684 for, obtain and decode and extension and obtain its critical flag.
17685 This allows all the necessary extension code to be handled in a
17686 single function call.
17690 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17691 platforms. See crypto/rc4/rc4_enc.c for further details.
17695 * New -noout option to asn1parse. This causes no output to be produced
17696 its main use is when combined with -strparse and -out to extract data
17697 from a file (which may not be in ASN.1 format).
17701 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17702 when producing the local key id.
17704 *Richard Levitte <levitte@stacken.kth.se>*
17706 * New option -dhparam in s_server. This allows a DH parameter file to be
17707 stated explicitly. If it is not stated then it tries the first server
17708 certificate file. The previous behaviour hard coded the filename
17713 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17714 a public key to be input or output. For example:
17715 openssl rsa -in key.pem -pubout -out pubkey.pem
17716 Also added necessary DSA public key functions to handle this.
17720 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17721 in the message. This was handled by allowing
17722 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17724 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17726 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17727 to the end of the strings whereas this didn't. This would cause problems
17728 if strings read with d2i_ASN1_bytes() were later modified.
17730 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17732 * Fix for base64 decode bug. When a base64 bio reads only one line of
17733 data and it contains EOF it will end up returning an error. This is
17734 caused by input 46 bytes long. The cause is due to the way base64
17735 BIOs find the start of base64 encoded data. They do this by trying a
17736 trial decode on each line until they find one that works. When they
17737 do a flag is set and it starts again knowing it can pass all the
17738 data directly through the decoder. Unfortunately it doesn't reset
17739 the context it uses. This means that if EOF is reached an attempt
17740 is made to pass two EOFs through the context and this causes the
17741 resulting error. This can also cause other problems as well. As is
17742 usual with these problems it takes *ages* to find and the fix is
17743 trivial: move one line.
17745 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17747 * Ugly workaround to get s_client and s_server working under Windows. The
17748 old code wouldn't work because it needed to select() on sockets and the
17749 tty (for keypresses and to see if data could be written). Win32 only
17750 supports select() on sockets so we select() with a 1s timeout on the
17751 sockets and then see if any characters are waiting to be read, if none
17752 are present then we retry, we also assume we can always write data to
17753 the tty. This isn't nice because the code then blocks until we've
17754 received a complete line of data and it is effectively polling the
17755 keyboard at 1s intervals: however it's quite a bit better than not
17756 working at all :-) A dedicated Windows application might handle this
17757 with an event loop for example.
17761 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17762 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17763 will be called when RSA_sign() and RSA_verify() are used. This is useful
17764 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17765 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17766 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17767 This necessitated the support of an extra signature type NID_md5_sha1
17768 for SSL signatures and modifications to the SSL library to use it instead
17769 of calling RSA_public_decrypt() and RSA_private_encrypt().
17773 * Add new -verify -CAfile and -CApath options to the crl program, these
17774 will lookup a CRL issuers certificate and verify the signature in a
17775 similar way to the verify program. Tidy up the crl program so it
17776 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17777 less strict. It will now permit CRL extensions even if it is not
17778 a V2 CRL: this will allow it to tolerate some broken CRLs.
17782 * Initialize all non-automatic variables each time one of the openssl
17783 sub-programs is started (this is necessary as they may be started
17784 multiple times from the "OpenSSL>" prompt).
17786 *Lennart Bang, Bodo Moeller*
17788 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17789 removing all other RSA functionality (this is what NO_RSA does). This
17790 is so (for example) those in the US can disable those operations covered
17791 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17796 * Non-copying interface to BIO pairs.
17797 (still largely untested)
17801 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17802 ASCII string. This was handled independently in various places before.
17806 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17807 UTF8 strings a character at a time.
17811 * Use client_version from client hello to select the protocol
17812 (s23_srvr.c) and for RSA client key exchange verification
17813 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17817 * Add various utility functions to handle SPKACs, these were previously
17818 handled by poking round in the structure internals. Added new function
17819 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17820 print, verify and generate SPKACs. Based on an original idea from
17821 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17825 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17829 * Allow the config file extension section to be overwritten on the
17830 command line. Based on an original idea from Massimiliano Pala
17831 <madwolf@comune.modena.it>. The new option is called -extensions
17832 and can be applied to ca, req and x509. Also -reqexts to override
17833 the request extensions in req and -crlexts to override the crl extensions
17838 * Add new feature to the SPKAC handling in ca. Now you can include
17839 the same field multiple times by preceding it by "XXXX." for example:
17842 this is the same syntax as used in the req config file.
17846 * Allow certificate extensions to be added to certificate requests. These
17847 are specified in a 'req_extensions' option of the req section of the
17848 config file. They can be printed out with the -text option to req but
17849 are otherwise ignored at present.
17853 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17854 data read consists of only the final block it would not decrypted because
17855 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17856 A misplaced 'break' also meant the decrypted final block might not be
17857 copied until the next read.
17861 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17862 a few extra parameters to the DH structure: these will be useful if
17863 for example we want the value of 'q' or implement X9.42 DH.
17867 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17868 provides hooks that allow the default DSA functions or functions on a
17869 "per key" basis to be replaced. This allows hardware acceleration and
17870 hardware key storage to be handled without major modification to the
17871 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17872 associated functions.
17876 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17877 as "read only": it can't be written to and the buffer it points to will
17878 not be freed. Reading from a read only BIO is much more efficient than
17879 a normal memory BIO. This was added because there are several times when
17880 an area of memory needs to be read from a BIO. The previous method was
17881 to create a memory BIO and write the data to it, this results in two
17882 copies of the data and an O(n^2) reading algorithm. There is a new
17883 function BIO_new_mem_buf() which creates a read only memory BIO from
17884 an area of memory. Also modified the PKCS#7 routines to use read only
17889 * Bugfix: ssl23_get_client_hello did not work properly when called in
17890 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17891 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17892 but a retry condition occurred while trying to read the rest.
17896 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17897 NID_pkcs7_encrypted by default: this was wrong since this should almost
17898 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17899 the encrypted data type: this is a more sensible place to put it and it
17900 allows the PKCS#12 code to be tidied up that duplicated this
17905 * Changed obj_dat.pl script so it takes its input and output files on
17906 the command line. This should avoid shell escape redirection problems
17911 * Initial support for certificate extension requests, these are included
17912 in things like Xenroll certificate requests. Included functions to allow
17913 extensions to be obtained and added.
17917 * -crlf option to s_client and s_server for sending newlines as
17918 CRLF (as required by many protocols).
17922 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17924 * Install libRSAglue.a when OpenSSL is built with RSAref.
17926 *Ralf S. Engelschall*
17928 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17930 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17932 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17937 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17938 DH parameters/keys (q is lost during that conversion, but the resulting
17939 DH parameters contain its length).
17941 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17942 much faster than DH_generate_parameters (which creates parameters
17943 where `p = 2*q + 1`), and also the smaller q makes DH computations
17944 much more efficient (160-bit exponentiation instead of 1024-bit
17945 exponentiation); so this provides a convenient way to support DHE
17946 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17947 utter importance to use
17948 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17950 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17951 when such DH parameters are used, because otherwise small subgroup
17952 attacks may become possible!
17956 * Avoid memory leak in i2d_DHparams.
17960 * Allow the -k option to be used more than once in the enc program:
17961 this allows the same encrypted message to be read by multiple recipients.
17965 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17966 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17967 it will always use the numerical form of the OID, even if it has a short
17972 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17973 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17974 otherwise bn_mod_exp was called. In the case of hardware keys for example
17975 no private key components need be present and it might store extra data
17976 in the RSA structure, which cannot be accessed from bn_mod_exp.
17977 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17978 private key operations.
17982 * Added support for SPARC Linux.
17986 * pem_password_cb function type incompatibly changed from
17987 typedef int pem_password_cb(char *buf, int size, int rwflag);
17989 ....(char *buf, int size, int rwflag, void *userdata);
17990 so that applications can pass data to their callbacks:
17991 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17992 additional void * argument, which is just handed through whenever
17993 the password callback is called.
17995 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17997 New function SSL_CTX_set_default_passwd_cb_userdata.
17999 Compatibility note: As many C implementations push function arguments
18000 onto the stack in reverse order, the new library version is likely to
18001 interoperate with programs that have been compiled with the old
18002 pem_password_cb definition (PEM_whatever takes some data that
18003 happens to be on the stack as its last argument, and the callback
18004 just ignores this garbage); but there is no guarantee whatsoever that
18007 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18008 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18009 problems not only on Windows, but also on some Unix platforms.
18010 To avoid problematic command lines, these definitions are now in an
18011 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18012 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18016 * MIPS III/IV assembler module is reimplemented.
18020 * More DES library cleanups: remove references to srand/rand and
18021 delete an unused file.
18025 * Add support for the free Netwide assembler (NASM) under Win32,
18026 since not many people have MASM (ml) and it can be hard to obtain.
18027 This is currently experimental but it seems to work OK and pass all
18028 the tests. Check out INSTALL.W32 for info.
18032 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18033 without temporary keys kept an extra copy of the server key,
18034 and connections with temporary keys did not free everything in case
18039 * New function RSA_check_key and new openssl rsa option -check
18040 for verifying the consistency of RSA keys.
18042 *Ulf Moeller, Bodo Moeller*
18044 * Various changes to make Win32 compile work:
18045 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18046 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18047 comparison" warnings.
18048 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18052 * Add a debugging option to PKCS#5 v2 key generation function: when
18053 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18054 derived keys are printed to stderr.
18058 * Copy the flags in ASN1_STRING_dup().
18060 *Roman E. Pavlov <pre@mo.msk.ru>*
18062 * The x509 application mishandled signing requests containing DSA
18063 keys when the signing key was also DSA and the parameters didn't match.
18065 It was supposed to omit the parameters when they matched the signing key:
18066 the verifying software was then supposed to automatically use the CA's
18067 parameters if they were absent from the end user certificate.
18069 Omitting parameters is no longer recommended. The test was also
18070 the wrong way round! This was probably due to unusual behaviour in
18071 EVP_cmp_parameters() which returns 1 if the parameters match.
18072 This meant that parameters were omitted when they *didn't* match and
18073 the certificate was useless. Certificates signed with 'ca' didn't have
18076 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18078 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18079 The interface is as follows:
18080 Applications can use
18081 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18082 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18083 "off" is now the default.
18084 The library internally uses
18085 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18086 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18087 to disable memory-checking temporarily.
18089 Some inconsistent states that previously were possible (and were
18090 even the default) are now avoided.
18092 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18093 with each memory chunk allocated; this is occasionally more helpful
18094 than just having a counter.
18096 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18098 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18103 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18104 which largely parallels "options", but is for changing API behaviour,
18105 whereas "options" are about protocol behaviour.
18106 Initial "mode" flags are:
18108 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18109 a single record has been written.
18110 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18111 retries use the same buffer location.
18112 (But all of the contents must be
18117 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18120 * Fix problems with no-hmac etc.
18122 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18124 * New functions RSA_get_default_method(), RSA_set_method() and
18125 RSA_get_method(). These allows replacement of RSA_METHODs without having
18126 to mess around with the internals of an RSA structure.
18130 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18131 Also really enable memory leak checks in openssl.c and in some
18134 *Chad C. Mulligan, Bodo Moeller*
18136 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18137 up the length of negative integers. This has now been simplified to just
18138 store the length when it is first determined and use it later, rather
18139 than trying to keep track of where data is copied and updating it to
18141 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18143 * Add a new function PKCS7_signatureVerify. This allows the verification
18144 of a PKCS#7 signature but with the signing certificate passed to the
18145 function itself. This contrasts with PKCS7_dataVerify which assumes the
18146 certificate is present in the PKCS#7 structure. This isn't always the
18147 case: certificates can be omitted from a PKCS#7 structure and be
18148 distributed by "out of band" means (such as a certificate database).
18152 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18153 function prototypes in pem.h, also change util/mkdef.pl to add the
18154 necessary function names.
18158 * mk1mf.pl (used by Windows builds) did not properly read the
18159 options set by Configure in the top level Makefile, and Configure
18160 was not even able to write more than one option correctly.
18161 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18165 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18166 file to be loaded from a BIO or FILE pointer. The BIO version will
18167 for example allow memory BIOs to contain config info.
18171 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18172 Whoever hopes to achieve shared-library compatibility across versions
18173 must use this, not the compile-time macro.
18174 (Exercise 0.9.4: Which is the minimum library version required by
18176 Note: All this applies only to multi-threaded programs, others don't
18181 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18182 through a BIO pair triggered the default case, i.e.
18183 SSLerr(...,SSL_R_UNKNOWN_STATE).
18187 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18188 can use the SSL library even if none of the specific BIOs is
18193 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18194 for the encoded length.
18196 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18198 * Add initial documentation of the X509V3 functions.
18202 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18203 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18204 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18205 secure PKCS#8 private key format with a high iteration count.
18209 * Fix determination of Perl interpreter: A perl or perl5
18210 *directory* in $PATH was also accepted as the interpreter.
18212 *Ralf S. Engelschall*
18214 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18215 wrong with it but it was very old and did things like calling
18216 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18217 unusual formatting.
18221 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18222 to use the new extension code.
18226 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18227 with macros. This should make it easier to change their form, add extra
18228 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18233 * Add to configuration table a new entry that can specify an alternative
18234 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18235 according to Mark Crispin <MRC@Panda.COM>.
18239 * DES CBC did not update the IV. Weird.
18243 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18244 Changing the behaviour of the former might break existing programs --
18245 where IV updating is needed, des_ncbc_encrypt can be used.
18248 * When bntest is run from "make test" it drives bc to check its
18249 calculations, as well as internally checking them. If an internal check
18250 fails, it needs to cause bc to give a non-zero result or make test carries
18251 on without noticing the failure. Fixed.
18255 * DES library cleanups.
18259 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18260 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18261 ciphers. NOTE: although the key derivation function has been verified
18262 against some published test vectors it has not been extensively tested
18263 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18268 * Instead of "mkdir -p", which is not fully portable, use new
18269 Perl script "util/mkdir-p.pl".
18273 * Rewrite the way password based encryption (PBE) is handled. It used to
18274 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18275 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18276 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18277 the 'parameter' field of the AlgorithmIdentifier is passed to the
18278 underlying key generation function so it must do its own ASN1 parsing.
18279 This has also changed the EVP_PBE_CipherInit() function which now has a
18280 'parameter' argument instead of literal salt and iteration count values
18281 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18285 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18286 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18287 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18288 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18289 value was just used as a "magic string" and not used directly its
18290 value doesn't matter.
18294 * Introduce some semblance of const correctness to BN. Shame C doesn't
18299 * "linux-sparc64" configuration (ultrapenguin).
18301 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18302 "linux-sparc" configuration.
18304 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18306 * config now generates no-xxx options for missing ciphers.
18310 * Support the EBCDIC character set (work in progress).
18311 File ebcdic.c not yet included because it has a different license.
18313 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18315 * Support BS2000/OSD-POSIX.
18317 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18319 * Make callbacks for key generation use `void *` instead of `char *`.
18323 * Make S/MIME samples compile (not yet tested).
18327 * Additional typesafe stacks.
18331 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18335 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18337 * New configuration variant "sco5-gcc".
18339 * Updated some demos.
18341 *Sean O Riordain, Wade Scholine*
18343 * Add missing BIO_free at exit of pkcs12 application.
18347 * Fix memory leak in conf.c.
18351 * Updates for Win32 to assembler version of MD5.
18355 * Set #! path to perl in `apps/der_chop` to where we found it
18356 instead of using a fixed path.
18360 * SHA library changes for irix64-mips4-cc.
18364 * Improvements for VMS support.
18368 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18370 * Bignum library bug fix. IRIX 6 passes "make test" now!
18371 This also avoids the problems with SC4.2 and unpatched SC5.
18373 *Andy Polyakov <appro@fy.chalmers.se>*
18375 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18376 These are required because of the typesafe stack would otherwise break
18377 existing code. If old code used a structure member which used to be STACK
18378 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18379 sk_num or sk_value it would produce an error because the num, data members
18380 are not present in STACK_OF. Now it just produces a warning. sk_set
18381 replaces the old method of assigning a value to sk_value
18382 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18383 that does this will no longer work (and should use sk_set instead) but
18384 this could be regarded as a "questionable" behaviour anyway.
18388 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18389 correctly handle encrypted S/MIME data.
18393 * Change type of various DES function arguments from des_cblock
18394 (which means, in function argument declarations, pointer to char)
18395 to des_cblock * (meaning pointer to array with 8 char elements),
18396 which allows the compiler to do more typechecking; it was like
18397 that back in SSLeay, but with lots of ugly casts.
18399 Introduce new type const_des_cblock.
18403 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18404 problems: find RecipientInfo structure that matches recipient certificate
18405 and initialise the ASN1 structures properly based on passed cipher.
18409 * Belatedly make the BN tests actually check the results.
18413 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18414 to and from BNs: it was completely broken. New compilation option
18415 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18416 key elements as negative integers.
18420 * Reorganize and speed up MD5.
18422 *Andy Polyakov <appro@fy.chalmers.se>*
18426 *Richard Levitte <richard@levitte.org>*
18428 * New option -out to asn1parse to allow the parsed structure to be
18429 output to a file. This is most useful when combined with the -strparse
18430 option to examine the output of things like OCTET STRINGS.
18434 * Make SSL library a little more fool-proof by not requiring any longer
18435 that `SSL_set_{accept,connect}_state` be called before
18436 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18437 in many applications because usually everything *appeared* to work as
18438 intended anyway -- now it really works as intended).
18442 * Move openssl.cnf out of lib/.
18446 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18447 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18448 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18450 *Ralf S. Engelschall*
18452 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18453 handle PKCS#7 enveloped data properly.
18455 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18457 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18458 copying pointers. The cert_st handling is changed by this in
18459 various ways (and thus what used to be known as ctx->default_cert
18460 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18461 any longer when s->cert does not give us what we need).
18462 ssl_cert_instantiate becomes obsolete by this change.
18463 As soon as we've got the new code right (possibly it already is?),
18464 we have solved a couple of bugs of the earlier code where s->cert
18465 was used as if it could not have been shared with other SSL structures.
18467 Note that using the SSL API in certain dirty ways now will result
18468 in different behaviour than observed with earlier library versions:
18469 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18470 does not influence s as it used to.
18472 In order to clean up things more thoroughly, inside SSL_SESSION
18473 we don't use CERT any longer, but a new structure SESS_CERT
18474 that holds per-session data (if available); currently, this is
18475 the peer's certificate chain and, for clients, the server's certificate
18476 and temporary key. CERT holds only those values that can have
18477 meaningful defaults in an SSL_CTX.
18481 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18482 from the internal representation. Various PKCS#7 fixes: remove some
18483 evil casts and set the enc_dig_alg field properly based on the signing
18488 * Allow PKCS#12 password to be set from the command line or the
18489 environment. Let 'ca' get its config file name from the environment
18490 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18495 * Allow certificate policies extension to use an IA5STRING for the
18496 organization field. This is contrary to the PKIX definition but
18497 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18502 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18503 without disallowing inline assembler and the like for non-pedantic builds.
18507 * Support Borland C++ builder.
18509 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18515 * SHA-1 cleanups and performance enhancements.
18517 *Andy Polyakov <appro@fy.chalmers.se>*
18519 * Sparc v8plus assembler for the bignum library.
18521 *Andy Polyakov <appro@fy.chalmers.se>*
18523 * Accept any -xxx and +xxx compiler options in Configure.
18527 * Update HPUX configuration.
18531 * Add missing `sk_<type>_unshift()` function to safestack.h
18533 *Ralf S. Engelschall*
18535 * New function SSL_CTX_use_certificate_chain_file that sets the
18536 "extra_cert"s in addition to the certificate. (This makes sense
18537 only for "PEM" format files, as chains as a whole are not
18542 * Support verify_depth from the SSL API.
18543 x509_vfy.c had what can be considered an off-by-one-error:
18544 Its depth (which was not part of the external interface)
18545 was actually counting the number of certificates in a chain;
18546 now it really counts the depth.
18550 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18551 instead of X509err, which often resulted in confusing error
18552 messages since the error codes are not globally unique
18553 (e.g. an alleged error in ssl3_accept when a certificate
18554 didn't match the private key).
18556 * New function SSL_CTX_set_session_id_context that allows to set a default
18557 value (so that you don't need SSL_set_session_id_context for each
18558 connection using the SSL_CTX).
18562 * OAEP decoding bug fix.
18566 * Support INSTALL_PREFIX for package builders, as proposed by
18571 * New Configure options "threads" and "no-threads". For systems
18572 where the proper compiler options are known (currently Solaris
18573 and Linux), "threads" is the default.
18577 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18581 * Install various scripts to $(OPENSSLDIR)/misc, not to
18582 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18583 such as /usr/local/bin.
18587 * "make linux-shared" to build shared libraries.
18589 *Niels Poppe <niels@netbox.org>*
18591 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18595 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18596 extension adding in x509 utility.
18600 * Remove NOPROTO sections and error code comments.
18604 * Partial rewrite of the DEF file generator to now parse the ANSI
18609 * New Configure options --prefix=DIR and --openssldir=DIR.
18613 * Complete rewrite of the error code script(s). It is all now handled
18614 by one script at the top level which handles error code gathering,
18615 header rewriting and C source file generation. It should be much better
18616 than the old method: it now uses a modified version of Ulf's parser to
18617 read the ANSI prototypes in all header files (thus the old K&R definitions
18618 aren't needed for error creation any more) and do a better job of
18619 translating function codes into names. The old 'ASN1 error code embedded
18620 in a comment' is no longer necessary and it doesn't use .err files which
18621 have now been deleted. Also the error code call doesn't have to appear all
18622 on one line (which resulted in some large lines...).
18626 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18630 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18631 0 (which usually indicates a closed connection), but continue reading.
18635 * Fix some race conditions.
18639 * Add support for CRL distribution points extension. Add Certificate
18640 Policies and CRL distribution points documentation.
18644 * Move the autogenerated header file parts to crypto/opensslconf.h.
18648 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18649 8 of keying material. Merlin has also confirmed interop with this fix
18650 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18652 *Merlin Hughes <merlin@baltimore.ie>*
18654 * Fix lots of warnings.
18656 *Richard Levitte <levitte@stacken.kth.se>*
18658 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18659 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18661 *Richard Levitte <levitte@stacken.kth.se>*
18663 * Fix problems with sizeof(long) == 8.
18665 *Andy Polyakov <appro@fy.chalmers.se>*
18667 * Change functions to ANSI C.
18671 * Fix typos in error codes.
18673 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18675 * Remove defunct assembler files from Configure.
18679 * SPARC v8 assembler BIGNUM implementation.
18681 *Andy Polyakov <appro@fy.chalmers.se>*
18683 * Support for Certificate Policies extension: both print and set.
18684 Various additions to support the r2i method this uses.
18688 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18689 return a const string when you are expecting an allocated buffer.
18693 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18694 types DirectoryString and DisplayText.
18698 * Add code to allow r2i extensions to access the configuration database,
18699 add an LHASH database driver and add several ctx helper functions.
18703 * Fix an evil bug in bn_expand2() which caused various BN functions to
18704 fail when they extended the size of a BIGNUM.
18708 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18709 support typesafe stack.
18713 * Fix typo in SSL_[gs]et_options().
18715 *Nils Frostberg <nils@medcom.se>*
18717 * Delete various functions and files that belonged to the (now obsolete)
18718 old X509V3 handling code.
18722 * New Configure option "rsaref".
18726 * Don't auto-generate pem.h.
18730 * Introduce type-safe ASN.1 SETs.
18734 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18736 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18738 * Introduce type-safe STACKs. This will almost certainly break lots of code
18739 that links with OpenSSL (well at least cause lots of warnings), but fear
18740 not: the conversion is trivial, and it eliminates loads of evil casts. A
18741 few STACKed things have been converted already. Feel free to convert more.
18742 In the fullness of time, I'll do away with the STACK type altogether.
18746 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18747 specified in `<certfile>` by updating the entry in the index.txt file.
18748 This way one no longer has to edit the index.txt file manually for
18749 revoking a certificate. The -revoke option does the gory details now.
18751 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18753 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18754 `-text` option at all and this way the `-noout -text` combination was
18755 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18757 *Ralf S. Engelschall*
18759 * Make sure a corresponding plain text error message exists for the
18760 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18761 verify callback function determined that a certificate was revoked.
18763 *Ralf S. Engelschall*
18765 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18766 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18767 all available ciphers including rc5, which was forgotten until now.
18768 In order to let the testing shell script know which algorithms
18769 are available, a new (up to now undocumented) command
18770 `openssl list-cipher-commands` is used.
18774 * Bugfix: s_client occasionally would sleep in select() when
18775 it should have checked SSL_pending() first.
18779 * New functions DSA_do_sign and DSA_do_verify to provide access to
18780 the raw DSA values prior to ASN.1 encoding.
18784 * Tweaks to Configure
18786 *Niels Poppe <niels@netbox.org>*
18788 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18793 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18797 * New config option to avoid instructions that are illegal on the 80386.
18798 The default code is faster, but requires at least a 486.
18802 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18803 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18804 same as SSL2_VERSION anyway.
18808 * New "-showcerts" option for s_client.
18812 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18813 application. Various cleanups and fixes.
18817 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18818 modify error routines to work internally. Add error codes and PBE init
18819 to library startup routines.
18823 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18824 packing functions to asn1 and evp. Changed function names and error
18825 codes along the way.
18829 * PKCS12 integration: and so it begins... First of several patches to
18830 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18831 objects to objects.h
18835 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18836 and display support for Thawte strong extranet extension.
18840 * Add LinuxPPC support.
18842 *Jeff Dubrule <igor@pobox.org>*
18844 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18845 bn_div_words in alpha.s.
18847 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18849 * Make sure the RSA OAEP test is skipped under -DRSAref because
18850 OAEP isn't supported when OpenSSL is built with RSAref.
18852 *Ulf Moeller <ulf@fitug.de>*
18854 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18855 so they no longer are missing under -DNOPROTO.
18857 *Soren S. Jorvang <soren@t.dk>*
18859 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18861 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18862 doesn't work when the session is reused. Coming soon!
18866 * Fix a security hole, that allows sessions to be reused in the wrong
18867 context thus bypassing client cert protection! All software that uses
18868 client certs and session caches in multiple contexts NEEDS PATCHING to
18869 allow session reuse! A fuller solution is in the works.
18871 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18873 * Some more source tree cleanups (removed obsolete files
18874 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18875 permission on "config" script to be executable) and a fix for the INSTALL
18878 *Ulf Moeller <ulf@fitug.de>*
18880 * Remove some legacy and erroneous uses of malloc, free instead of
18883 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18885 * Make rsa_oaep_test return non-zero on error.
18887 *Ulf Moeller <ulf@fitug.de>*
18889 * Add support for native Solaris shared libraries. Configure
18890 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18891 if someone would make that last step automatic.
18893 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18895 * ctx_size was not built with the right compiler during "make links". Fixed.
18899 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18900 except NULL ciphers". This means the default cipher list will no longer
18901 enable NULL ciphers. They need to be specifically enabled e.g. with
18902 the string "DEFAULT:eNULL".
18906 * Fix to RSA private encryption routines: if p < q then it would
18907 occasionally produce an invalid result. This will only happen with
18908 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18912 * Be less restrictive and allow also `perl util/perlpath.pl
18913 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18914 because this way one can also use an interpreter named `perl5` (which is
18915 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18916 installed as `perl`).
18918 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18920 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18922 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18924 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18925 advapi32.lib to Win32 build and change the pem test comparison
18926 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18927 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18928 and crypto/des/ede_cbcm_enc.c.
18932 * DES quad checksum was broken on big-endian architectures. Fixed.
18936 * Comment out two functions in bio.h that aren't implemented. Fix up the
18937 Win32 test batch file so it (might) work again. The Win32 test batch file
18938 is horrible: I feel ill....
18942 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18943 in e_os.h. Audit of header files to check ANSI and non ANSI
18944 sections: 10 functions were absent from non ANSI section and not exported
18945 from Windows DLLs. Fixed up libeay.num for new functions.
18949 * Make `openssl version` output lines consistent.
18951 *Ralf S. Engelschall*
18953 * Fix Win32 symbol export lists for BIO functions: Added
18954 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18955 to ms/libeay{16,32}.def.
18957 *Ralf S. Engelschall*
18959 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18960 fine under Unix and passes some trivial tests I've now added. But the
18961 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18962 added to make sure no one expects that this stuff really works in the
18963 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18964 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18967 *Ralf S. Engelschall*
18969 * Fix the generation of two part addresses in perl.
18971 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18973 * Add config entry for Linux on MIPS.
18975 *John Tobey <jtobey@channel1.com>*
18977 * Make links whenever Configure is run, unless we are on Windoze.
18981 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18982 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18987 * Add a useful kludge to allow package maintainers to specify compiler and
18988 other platforms details on the command line without having to patch the
18989 Configure script every time: One now can use
18990 `perl Configure <id>:<details>`,
18991 i.e. platform ids are allowed to have details appended
18992 to them (separated by colons). This is treated as there would be a static
18993 pre-configured entry in Configure's %table under key `<id>` with value
18994 `<details>` and `perl Configure <id>` is called. So, when you want to
18995 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18996 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18997 now, which overrides the FreeBSD-elf entry on-the-fly.
18999 *Ralf S. Engelschall*
19001 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19005 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19006 on the `perl Configure ...` command line. This way one can compile
19007 OpenSSL libraries with Position Independent Code (PIC) which is needed
19008 for linking it into DSOs.
19010 *Ralf S. Engelschall*
19012 * Remarkably, export ciphers were totally broken and no-one had noticed!
19017 * Cleaned up the LICENSE document: The official contact for any license
19018 questions now is the OpenSSL core team under openssl-core@openssl.org.
19019 And add a paragraph about the dual-license situation to make sure people
19020 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19021 to the OpenSSL toolkit.
19023 *Ralf S. Engelschall*
19025 * General source tree makefile cleanups: Made `making xxx in yyy...`
19026 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19027 Additionally cleaned up the `make links` target: Remove unnecessary
19028 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19029 to speed processing and no longer clutter the display with confusing
19030 stuff. Instead only the actually done links are displayed.
19032 *Ralf S. Engelschall*
19034 * Permit null encryption ciphersuites, used for authentication only. It used
19035 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19036 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19041 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19042 signed attributes when verifying signatures (this would break them),
19043 the detached data encoding was wrong and public keys obtained using
19044 X509_get_pubkey() weren't freed.
19048 * Add text documentation for the BUFFER functions. Also added a work around
19049 to a Win95 console bug. This was triggered by the password read stuff: the
19050 last character typed gets carried over to the next fread(). If you were
19051 generating a new cert request using 'req' for example then the last
19052 character of the passphrase would be CR which would then enter the first
19057 * Added the new 'Includes OpenSSL Cryptography Software' button as
19058 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19059 button and can be used by applications based on OpenSSL to show the
19060 relationship to the OpenSSL project.
19062 *Ralf S. Engelschall*
19064 * Remove confusing variables in function signatures in files
19065 ssl/ssl_lib.c and ssl/ssl.h.
19067 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19069 * Don't install bss_file.c under PREFIX/include/
19071 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19073 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19074 functions that return function pointers and has support for NT specific
19075 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19076 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19077 unsigned to signed types: this was killing the Win32 compile.
19081 * Add new certificate file to stack functions,
19082 SSL_add_dir_cert_subjects_to_stack() and
19083 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19084 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19085 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19086 This means that Apache-SSL and similar packages don't have to mess around
19087 to add as many CAs as they want to the preferred list.
19091 * Experiment with doxygen documentation. Currently only partially applied to
19093 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19094 openssl.doxy as the configuration file.
19098 * Get rid of remaining C++-style comments which strict C compilers hate.
19100 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19102 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19103 compiled in by default: it has problems with large keys.
19107 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19108 DH private keys and/or callback functions which directly correspond to
19109 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19110 is needed for applications which have to configure certificates on a
19111 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19113 For the RSA certificate situation is makes no difference, but
19114 for the DSA certificate situation this fixes the "no shared cipher"
19115 problem where the OpenSSL cipher selection procedure failed because the
19116 temporary keys were not overtaken from the context and the API provided
19117 no way to reconfigure them.
19118 The new functions now let applications reconfigure the stuff and they
19119 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19120 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19121 non-public-API function ssl_cert_instantiate() is used as a helper
19122 function and also to reduce code redundancy inside ssl_rsa.c.
19124 *Ralf S. Engelschall*
19126 * Move s_server -dcert and -dkey options out of the undocumented feature
19127 area because they are useful for the DSA situation and should be
19128 recognized by the users.
19130 *Ralf S. Engelschall*
19132 * Fix the cipher decision scheme for export ciphers: the export bits are
19133 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19134 SSL_EXP_MASK. So, the original variable has to be used instead of the
19135 already masked variable.
19137 *Richard Levitte <levitte@stacken.kth.se>*
19139 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19141 *Richard Levitte <levitte@stacken.kth.se>*
19143 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19144 from `int` to `unsigned int` because it is a length and initialized by
19145 EVP_DigestFinal() which expects an `unsigned int *`.
19147 *Richard Levitte <levitte@stacken.kth.se>*
19149 * Don't hard-code path to Perl interpreter on shebang line of Configure
19150 script. Instead use the usual Shell->Perl transition trick.
19152 *Ralf S. Engelschall*
19154 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19155 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19156 -noout -modulus` as it's already the case for `openssl rsa -noout
19157 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19158 currently the public key is printed (a decision which was already done by
19159 `openssl dsa -modulus` in the past) which serves a similar purpose.
19160 Additionally the NO_RSA no longer completely removes the whole -modulus
19161 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19164 *Ralf S. Engelschall*
19166 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19167 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19169 *Arne Ansper <arne@ats.cyber.ee>*
19171 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19172 to be added. Now both 'req' and 'ca' can use new objects defined in the
19177 * Add cool BIO that does syslog (or event log on NT).
19179 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19181 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19182 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19183 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19184 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19188 * Add preliminary config info for new extension code.
19192 * Make RSA_NO_PADDING really use no padding.
19194 *Ulf Moeller <ulf@fitug.de>*
19196 * Generate errors when private/public key check is done.
19200 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19201 for some CRL extensions and new objects added.
19205 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19206 key usage extension and fuller support for authority key id.
19210 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19211 padding method for RSA, which is recommended for new applications in PKCS
19212 #1 v2.0 (RFC 2437, October 1998).
19213 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19214 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19215 against Bleichbacher's attack on RSA.
19216 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19219 * Updates to the new SSL compression code
19221 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19223 * Fix so that the version number in the master secret, when passed
19224 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19225 (because the server will not accept higher), that the version number
19226 is 0x03,0x01, not 0x03,0x00
19228 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19230 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19231 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19232 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19236 * Support for RAW extensions where an arbitrary extension can be
19237 created by including its DER encoding. See `apps/openssl.cnf` for
19242 * Make sure latest Perl versions don't interpret some generated C array
19243 code as Perl array code in the crypto/err/err_genc.pl script.
19245 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19247 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19248 not many people have the assembler. Various Win32 compilation fixes and
19249 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19250 build instructions.
19254 * Modify configure script 'Configure' to automatically create crypto/date.h
19255 file under Win32 and also build pem.h from pem.org. New script
19256 util/mkfiles.pl to create the MINFO file on environments that can't do a
19257 'make files': perl util/mkfiles.pl >MINFO should work.
19261 * Major rework of DES function declarations, in the pursuit of correctness
19262 and purity. As a result, many evil casts evaporated, and some weirdness,
19263 too. You may find this causes warnings in your code. Zapping your evil
19264 casts will probably fix them. Mostly.
19268 * Fix for a typo in asn1.h. Bug fix to object creation script
19269 obj_dat.pl. It considered a zero in an object definition to mean
19270 "end of object": none of the objects in objects.h have any zeros
19271 so it wasn't spotted.
19273 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19275 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19276 Masking (CBCM). In the absence of test vectors, the best I have been able
19277 to do is check that the decrypt undoes the encrypt, so far. Send me test
19278 vectors if you have them.
19282 * Correct calculation of key length for export ciphers (too much space was
19283 allocated for null ciphers). This has not been tested!
19287 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19288 message is now correct (it understands "crypto" and "ssl" on its
19289 command line). There is also now an "update" option. This will update
19290 the util/ssleay.num and util/libeay.num files with any new functions.
19292 perl util/mkdef.pl crypto ssl update
19293 it will update them.
19297 * Overhauled the Perl interface:
19298 - ported BN stuff to OpenSSL's different BN library
19299 - made the perl/ source tree CVS-aware
19300 - renamed the package from SSLeay to OpenSSL (the files still contain
19301 their history because I've copied them in the repository)
19302 - removed obsolete files (the test scripts will be replaced
19303 by better Test::Harness variants in the future)
19305 *Ralf S. Engelschall*
19307 * First cut for a very conservative source tree cleanup:
19308 1. merge various obsolete readme texts into doc/ssleay.txt
19309 where we collect the old documents and readme texts.
19310 2. remove the first part of files where I'm already sure that we no
19311 longer need them because of three reasons: either they are just temporary
19312 files which were left by Eric or they are preserved original files where
19313 I've verified that the diff is also available in the CVS via "cvs diff
19314 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19315 the crypto/md/ stuff).
19317 *Ralf S. Engelschall*
19319 * More extension code. Incomplete support for subject and issuer alt
19320 name, issuer and authority key id. Change the i2v function parameters
19321 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19322 what that's for :-) Fix to ASN1 macro which messed up
19323 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19327 * Preliminary support for ENUMERATED type. This is largely copied from the
19332 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19334 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19336 * Make sure `make rehash` target really finds the `openssl` program.
19338 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19340 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19341 like to hear about it if this slows down other processors.
19345 * Add CygWin32 platform information to Configure script.
19347 *Alan Batie <batie@aahz.jf.intel.com>*
19349 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19351 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19353 * New program nseq to manipulate netscape certificate sequences
19357 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19362 * Fixes to BN code. Previously the default was to define BN_RECURSION
19363 but the BN code had some problems that would cause failures when
19364 doing certificate verification and some other functions.
19366 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19368 * Add ASN1 and PEM code to support netscape certificate sequences.
19372 * Add ASN1 and PEM code to support netscape certificate sequences.
19376 * Add several PKIX and private extended key usage OIDs.
19380 * Modify the 'ca' program to handle the new extension code. Modify
19381 openssl.cnf for new extension format, add comments.
19385 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19386 and add a sample to openssl.cnf so req -x509 now adds appropriate
19391 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19392 error code, add initial support to X509_print() and x509 application.
19396 * Takes a deep breath and start adding X509 V3 extension support code. Add
19397 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19398 stuff is currently isolated and isn't even compiled yet.
19402 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19403 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19404 Removed the versions check from X509 routines when loading extensions:
19405 this allows certain broken certificates that don't set the version
19406 properly to be processed.
19410 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19411 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19412 can still be regenerated with "make depend".
19416 * Spelling mistake in C version of CAST-128.
19418 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19420 * Changes to the error generation code. The perl script err-code.pl
19421 now reads in the old error codes and retains the old numbers, only
19422 adding new ones if necessary. It also only changes the .err files if new
19423 codes are added. The makefiles have been modified to only insert errors
19424 when needed (to avoid needlessly modifying header files). This is done
19425 by only inserting errors if the .err file is newer than the auto generated
19426 C file. To rebuild all the error codes from scratch (the old behaviour)
19427 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19428 or delete all the .err files.
19432 * CAST-128 was incorrectly implemented for short keys. The C version has
19433 been fixed, but is untested. The assembler versions are also fixed, but
19434 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19435 to regenerate it if needed.
19436 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19437 Hagino <itojun@kame.net>*
19439 * File was opened incorrectly in randfile.c.
19441 *Ulf Möller <ulf@fitug.de>*
19443 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19444 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19445 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19446 al: it's just almost always a UTCTime. Note this patch adds new error
19447 codes so do a "make errors" if there are problems.
19451 * Correct Linux 1 recognition in config.
19453 *Ulf Möller <ulf@fitug.de>*
19455 * Remove pointless MD5 hash when using DSA keys in ca.
19457 *Anonymous <nobody@replay.com>*
19459 * Generate an error if given an empty string as a cert directory. Also
19460 generate an error if handed NULL (previously returned 0 to indicate an
19461 error, but didn't set one).
19463 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19465 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19469 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19470 parameters. This was causing a warning which killed off the Win32 compile.
19474 * Remove C++ style comments from crypto/bn/bn_local.h.
19476 *Neil Costigan <neil.costigan@celocom.com>*
19478 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19479 based on a text string, looking up short and long names and finally
19480 "dot" format. The "dot" format stuff didn't work. Added new function
19481 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19482 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19483 OID is not part of the table.
19487 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19488 X509_LOOKUP_by_alias().
19492 * Sort openssl functions by name.
19496 * Get the `gendsa` command working and add it to the `list` command. Remove
19497 encryption from sample DSA keys (in case anyone is interested the password
19502 * Make *all* `*_free` functions accept a NULL pointer.
19504 *Frans Heymans <fheymans@isaserver.be>*
19506 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19509 *Anonymous <nobody@replay.com>*
19511 * s_server should send the CAfile as acceptable CAs, not its own cert.
19513 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19515 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19517 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19519 * Temp key "for export" tests were wrong in s3_srvr.c.
19521 *Anonymous <nobody@replay.com>*
19523 * Add prototype for temp key callback functions
19524 SSL_CTX_set_tmp_{rsa,dh}_callback().
19528 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19529 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19533 * X509_name_add_entry() freed the wrong thing after an error.
19535 *Arne Ansper <arne@ats.cyber.ee>*
19537 * rsa_eay.c would attempt to free a NULL context.
19539 *Arne Ansper <arne@ats.cyber.ee>*
19541 * BIO_s_socket() had a broken should_retry() on Windoze.
19543 *Arne Ansper <arne@ats.cyber.ee>*
19545 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19547 *Arne Ansper <arne@ats.cyber.ee>*
19549 * Make sure the already existing X509_STORE->depth variable is initialized
19550 in X509_STORE_new(), but document the fact that this variable is still
19551 unused in the certificate verification process.
19553 *Ralf S. Engelschall*
19555 * Fix the various library and `apps/` files to free up pkeys obtained from
19556 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19560 * Fix reference counting in X509_PUBKEY_get(). This makes
19561 demos/maurice/example2.c work, amongst others, probably.
19563 *Steve Henson and Ben Laurie*
19565 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19566 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19567 are no longer created. This way we have a single and consistent command
19568 line interface `openssl <command>`, similar to `cvs <command>`.
19570 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19572 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19573 BIT STRING wrapper always have zero unused bits.
19577 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19581 * Make the top-level INSTALL documentation easier to understand.
19585 * Makefiles updated to exit if an error occurs in a sub-directory
19586 make (including if user presses ^C) [Paul Sutton]
19588 * Make Montgomery context stuff explicit in RSA data structure.
19592 * Fix build order of pem and err to allow for generated pem.h.
19596 * Fix renumbering bug in X509_NAME_delete_entry().
19600 * Enhanced the err-ins.pl script so it makes the error library number
19601 global and can add a library name. This is needed for external ASN1 and
19602 other error libraries.
19606 * Fixed sk_insert which never worked properly.
19610 * Fix ASN1 macros so they can handle indefinite length constructed
19611 EXPLICIT tags. Some non standard certificates use these: they can now
19616 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19617 into a single doc/ssleay.txt bundle. This way the information is still
19618 preserved but no longer messes up this directory. Now it's new room for
19619 the new set of documentation files.
19621 *Ralf S. Engelschall*
19623 * SETs were incorrectly DER encoded. This was a major pain, because they
19624 shared code with SEQUENCEs, which aren't coded the same. This means that
19625 almost everything to do with SETs or SEQUENCEs has either changed name or
19626 number of arguments.
19628 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19630 * Fix test data to work with the above.
19634 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19635 was already fixed by Eric for 0.9.1 it seems.
19637 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19639 * Autodetect FreeBSD3.
19643 * Fix various bugs in Configure. This affects the following platforms:
19647 unixware-2.0-pentium
19652 * Eliminate generated files from CVS. Reorder tests to regenerate files
19653 before they are needed.
19657 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19661 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19663 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19664 changed SSLeay to OpenSSL in version strings.
19666 *Ralf S. Engelschall*
19668 * Some fixups to the top-level documents.
19672 * Fixed the nasty bug where rsaref.h was not found under compile-time
19673 because the symlink to include/ was missing.
19675 *Ralf S. Engelschall*
19677 * Incorporated the popular no-RSA/DSA-only patches
19678 which allow to compile an RSA-free SSLeay.
19680 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19682 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19683 when "ssleay" is still not found.
19685 *Ralf S. Engelschall*
19687 * Added more platforms to Configure: Cray T3E, HPUX 11,
19689 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19691 * Updated the README file.
19693 *Ralf S. Engelschall*
19695 * Added various .cvsignore files in the CVS repository subdirs
19696 to make a "cvs update" really silent.
19698 *Ralf S. Engelschall*
19700 * Recompiled the error-definition header files and added
19701 missing symbols to the Win32 linker tables.
19703 *Ralf S. Engelschall*
19705 * Cleaned up the top-level documents;
19706 o new files: CHANGES and LICENSE
19707 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19708 o merged COPYRIGHT into LICENSE
19709 o removed obsolete TODO file
19710 o renamed MICROSOFT to INSTALL.W32
19712 *Ralf S. Engelschall*
19714 * Removed dummy files from the 0.9.1b source tree:
19715 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19716 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19717 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19718 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19719 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19721 *Ralf S. Engelschall*
19723 * Added various platform portability fixes.
19727 * The Genesis of the OpenSSL rpject:
19728 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19729 Young and Tim J. Hudson created while they were working for C2Net until
19732 *The OpenSSL Project*
19734 ### Changes between 0.9.0b and 0.9.1b [not released]
19736 * Updated a few CA certificates under certs/
19740 * Changed some BIGNUM api stuff.
19744 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19745 DGUX x86, Linux Alpha, etc.
19749 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19750 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19755 * Add -strparse option to asn1pars program which parses nested
19758 *Dr Stephen Henson <shenson@bigfoot.com>*
19760 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19764 * DSA fix for "ca" program.
19768 * Added "-genkey" option to "dsaparam" program.
19772 * Added RIPE MD160 (rmd160) message digest.
19776 * Added -a (all) option to "ssleay version" command.
19780 * Added PLATFORM define which is the id given to Configure.
19784 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19788 * Extended the ASN.1 parser routines.
19792 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19796 * Added a BN_CTX to the BN library.
19800 * Fixed the weak key values in DES library
19804 * Changed API in EVP library for cipher aliases.
19808 * Added support for RC2/64bit cipher.
19812 * Converted the lhash library to the crypto/mem.c functions.
19816 * Added more recognized ASN.1 object ids.
19820 * Added more RSA padding checks for SSL/TLS.
19824 * Added BIO proxy/filter functionality.
19828 * Added extra_certs to SSL_CTX which can be used
19829 send extra CA certificates to the client in the CA cert chain sending
19830 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19834 * Now Fortezza is denied in the authentication phase because
19835 this is key exchange mechanism is not supported by SSLeay at all.
19839 * Additional PKCS1 checks.
19843 * Support the string "TLSv1" for all TLS v1 ciphers.
19847 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19848 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19852 * Fixed a few memory leaks.
19856 * Fixed various code and comment typos.
19860 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19861 bytes sent in the client random.
19863 *Edward Bishop <ebishop@spyglass.com>*
19867 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
19868 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
19869 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19870 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19871 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19872 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19873 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19874 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19875 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19876 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19877 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19878 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19879 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19880 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19881 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19882 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19883 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19884 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19885 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19886 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19887 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19888 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19889 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19890 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19891 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19892 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19893 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19894 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19895 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19896 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19897 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19898 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19899 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19900 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19901 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19902 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19903 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19904 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19905 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19906 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19907 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19908 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19909 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19910 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19911 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19912 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19913 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19914 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19915 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19916 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19917 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19918 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19919 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19920 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19921 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19922 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19923 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19924 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19925 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19926 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19927 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19928 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19929 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19930 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19931 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19932 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19933 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19934 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19935 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19936 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19937 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19938 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19939 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19940 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19941 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19942 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19943 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19944 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19945 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19946 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19947 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19948 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19949 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19950 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19951 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19952 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19953 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19954 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19955 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19956 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19957 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19958 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19959 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19960 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19961 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19962 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19963 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19964 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19965 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19966 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19967 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19968 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19969 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19970 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19971 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19972 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19973 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19974 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19975 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19976 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19977 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19978 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19979 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19980 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19981 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19982 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19983 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19984 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19985 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19986 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19987 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19988 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19989 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19990 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19991 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19992 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19993 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19994 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19995 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19996 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19997 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19998 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19999 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20000 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20001 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20002 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20003 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20004 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20005 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20006 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20007 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20008 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20009 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20010 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20011 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20012 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20013 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20014 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20015 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20016 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20017 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20018 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20019 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20020 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20021 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20022 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20023 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20024 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20025 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20026 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20027 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20028 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20029 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20030 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20031 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20032 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20033 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20034 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20035 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20036 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20037 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20038 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20039 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20040 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20041 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20042 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20043 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20044 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20045 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20046 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20047 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20048 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20049 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20050 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655