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 3.1.4 and 3.1.5 [xx XXX xxxx]
27 * Fix excessive time spent in DH check / generation with large Q parameter
30 Applications that use the functions DH_generate_key() to generate an
31 X9.42 DH key may experience long delays. Likewise, applications that use
32 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
33 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
34 Where the key or parameters that are being checked have been obtained from
35 an untrusted source this may lead to a Denial of Service.
41 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
43 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
44 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
45 that alter the key or IV length ([CVE-2023-5363]).
49 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
51 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
53 The POLY1305 MAC (message authentication code) implementation in OpenSSL
54 does not save the contents of non-volatile XMM registers on Windows 64
55 platform when calculating the MAC of data larger than 64 bytes. Before
56 returning to the caller all the XMM registers are set to zero rather than
57 restoring their previous content. The vulnerable code is used only on newer
58 x86_64 processors supporting the AVX512-IFMA instructions.
60 The consequences of this kind of internal application state corruption can
61 be various - from no consequences, if the calling application does not
62 depend on the contents of non-volatile XMM registers at all, to the worst
63 consequences, where the attacker could get complete control of the
64 application process. However given the contents of the registers are just
65 zeroized so the attacker cannot put arbitrary values inside, the most likely
66 consequence, if any, would be an incorrect result of some application
67 dependent calculations or a crash leading to a denial of service.
73 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
75 * Fix excessive time spent checking DH q parameter value.
77 The function DH_check() performs various checks on DH parameters. After
78 fixing CVE-2023-3446 it was discovered that a large q parameter value can
79 also trigger an overly long computation during some of these checks.
80 A correct q value, if present, cannot be larger than the modulus p
81 parameter, thus it is unnecessary to perform these checks if q is larger
84 If DH_check() is called with such q parameter value,
85 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
86 intensive checks are skipped.
92 * Fix DH_check() excessive time with over sized modulus.
94 The function DH_check() performs various checks on DH parameters. One of
95 those checks confirms that the modulus ("p" parameter) is not too large.
96 Trying to use a very large modulus is slow and OpenSSL will not normally use
97 a modulus which is over 10,000 bits in length.
99 However the DH_check() function checks numerous aspects of the key or
100 parameters that have been supplied. Some of those checks use the supplied
101 modulus value even if it has already been found to be too large.
103 A new limit has been added to DH_check of 32,768 bits. Supplying a
104 key/parameters with a modulus over this size will simply cause DH_check() to
111 * Do not ignore empty associated data entries with AES-SIV.
113 The AES-SIV algorithm allows for authentication of multiple associated
114 data entries along with the encryption. To authenticate empty data the
115 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
116 with NULL pointer as the output buffer and 0 as the input buffer length.
117 The AES-SIV implementation in OpenSSL just returns success for such call
118 instead of performing the associated data authentication operation.
119 The empty data thus will not be authenticated. ([CVE-2023-2975])
121 Thanks to Juerg Wullschleger (Google) for discovering the issue.
123 The fix changes the authentication tag value and the ciphertext for
124 applications that use empty associated data entries with AES-SIV.
125 To decrypt data encrypted with previous versions of OpenSSL the application
126 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
131 * When building with the `enable-fips` option and using the resulting
132 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
133 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
134 not operate with truncated digests (FIPS 140-3 IG G.R).
138 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
140 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
141 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
143 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
144 numeric text form. For gigantic sub-identifiers, this would take a very
145 long time, the time complexity being O(n^2) where n is the size of that
146 sub-identifier. ([CVE-2023-2650])
148 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
149 IDENTIFIER to canonical numeric text form if the size of that OBJECT
150 IDENTIFIER is 586 bytes or less, and fail otherwise.
152 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
153 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
154 most 128 sub-identifiers, and that the maximum value that each sub-
155 identifier may have is 2^32-1 (4294967295 decimal).
157 For each byte of every sub-identifier, only the 7 lower bits are part of
158 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
159 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
164 * Multiple algorithm implementation fixes for ARM BE platforms.
168 * Added a -pedantic option to fipsinstall that adjusts the various
169 settings to ensure strict FIPS compliance rather than backwards
174 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
175 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
176 trigger a crash of an application using AES-XTS decryption if the memory
177 just after the buffer being decrypted is not mapped.
178 Thanks to Anton Romanov (Amazon) for discovering the issue.
183 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
184 The previous fix for this timing side channel turned out to cause
185 a severe 2-3x performance regression in the typical use case
186 compared to 3.0.7. The new fix uses existing constant time
187 code paths, and restores the previous performance level while
188 fully eliminating all existing timing side channels.
189 The fix was developed by Bernd Edlinger with testing support
194 * Add FIPS provider configuration option to disallow the use of
195 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
196 The option '-no_drbg_truncated_digests' can optionally be
197 supplied to 'openssl fipsinstall'.
201 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
202 that it does not enable policy checking. Thanks to David Benjamin for
203 discovering this issue.
208 * Fixed an issue where invalid certificate policies in leaf certificates are
209 silently ignored by OpenSSL and other certificate policy checks are skipped
210 for that certificate. A malicious CA could use this to deliberately assert
211 invalid certificate policies in order to circumvent policy checking on the
212 certificate altogether.
217 * Limited the number of nodes created in a policy tree to mitigate
218 against CVE-2023-0464. The default limit is set to 1000 nodes, which
219 should be sufficient for most installations. If required, the limit
220 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
221 time define to a desired maximum number of nodes or zero to allow
227 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
229 * Add FIPS provider configuration option to enforce the
230 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
231 The option '-ems-check' can optionally be supplied to
232 'openssl fipsinstall'.
236 * The FIPS provider includes a few non-approved algorithms for
237 backward compatibility purposes and the "fips=yes" property query
238 must be used for all algorithm fetches to ensure FIPS compliance.
240 The algorithms that are included but not approved are Triple DES ECB,
241 Triple DES CBC and EdDSA.
245 * Added support for KMAC in KBKDF.
249 * RNDR and RNDRRS support in provider functions to provide
250 random number generation for Arm CPUs (aarch64).
254 * s_client and s_server apps now explicitly say when the TLS version
255 does not include the renegotiation mechanism. This avoids confusion
256 between that scenario versus when the TLS version includes secure
257 renegotiation but the peer lacks support for it.
261 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
263 *Tomasz Kantecki, Andrey Matyukov*
265 * The various OBJ_* functions have been made thread safe.
269 * Parallel dual-prime 1536/2048-bit modular exponentiation for
270 AVX512_IFMA capable processors.
272 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
274 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
275 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
276 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
277 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
278 `OPENSSL_NO_DEPRECATED_3_1`.
280 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
281 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
282 definitions for these functions regardless of whether
283 `OPENSSL_NO_DEPRECATED_3_1` is defined.
285 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
286 functions regardless of whether they are using them. It is recommended that
287 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
291 * When generating safe-prime DH parameters set the recommended private key
292 length equivalent to minimum key lengths as in RFC 7919.
296 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
297 maximum size that is smaller or equal to the digest length to comply with
298 FIPS 186-4 section 5. This is implemented by a new option
299 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
300 `rsa_pss_saltlen` parameter, which is now the default. Signature
301 verification is not affected by this change and continues to work as before.
308 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
309 listed here are only a brief description.
310 The migration guide contains more detailed information related to new features,
311 breaking changes, and mappings for the large list of deprecated functions.
313 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
315 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
317 * Fixed NULL dereference during PKCS7 data verification.
319 A NULL pointer can be dereferenced when signatures are being
320 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
321 algorithm used for the signature is known to the OpenSSL library but
322 the implementation of the hash algorithm is not available the digest
323 initialization will fail. There is a missing check for the return
324 value from the initialization function which later leads to invalid
325 usage of the digest API most likely leading to a crash.
328 PKCS7 data is processed by the SMIME library calls and also by the
329 time stamp (TS) library calls. The TLS implementation in OpenSSL does
330 not call these functions however third party applications would be
331 affected if they call these functions to verify signatures on untrusted
336 * Fixed X.400 address type confusion in X.509 GeneralName.
338 There is a type confusion vulnerability relating to X.400 address processing
339 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
340 but the public structure definition for GENERAL_NAME incorrectly specified
341 the type of the x400Address field as ASN1_TYPE. This field is subsequently
342 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
345 When CRL checking is enabled (i.e. the application sets the
346 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
347 pass arbitrary pointers to a memcmp call, enabling them to read memory
348 contents or enact a denial of service.
353 * Fixed NULL dereference validating DSA public key.
355 An invalid pointer dereference on read can be triggered when an
356 application tries to check a malformed DSA public key by the
357 EVP_PKEY_public_check() function. This will most likely lead
358 to an application crash. This function can be called on public
359 keys supplied from untrusted sources which could allow an attacker
360 to cause a denial of service attack.
362 The TLS implementation in OpenSSL does not call this function
363 but applications might call the function if there are additional
364 security requirements imposed by standards such as FIPS 140-3.
367 *Shane Lontis, Tomáš Mráz*
369 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
371 An invalid pointer dereference on read can be triggered when an
372 application tries to load malformed PKCS7 data with the
373 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
375 The result of the dereference is an application crash which could
376 lead to a denial of service attack. The TLS implementation in OpenSSL
377 does not call this function however third party applications might
378 call these functions on untrusted data.
383 * Fixed Use-after-free following BIO_new_NDEF.
385 The public API function BIO_new_NDEF is a helper function used for
386 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
387 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
388 be called directly by end user applications.
390 The function receives a BIO from the caller, prepends a new BIO_f_asn1
391 filter BIO onto the front of it to form a BIO chain, and then returns
392 the new head of the BIO chain to the caller. Under certain conditions,
393 for example if a CMS recipient public key is invalid, the new filter BIO
394 is freed and the function returns a NULL result indicating a failure.
395 However, in this case, the BIO chain is not properly cleaned up and the
396 BIO passed by the caller still retains internal pointers to the previously
397 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
398 then a use-after-free will occur. This will most likely result in a crash.
401 *Viktor Dukhovni, Matt Caswell*
403 * Fixed Double free after calling PEM_read_bio_ex.
405 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
406 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
407 data. If the function succeeds then the "name_out", "header" and "data"
408 arguments are populated with pointers to buffers containing the relevant
409 decoded data. The caller is responsible for freeing those buffers. It is
410 possible to construct a PEM file that results in 0 bytes of payload data.
411 In this case PEM_read_bio_ex() will return a failure code but will populate
412 the header argument with a pointer to a buffer that has already been freed.
413 If the caller also frees this buffer then a double free will occur. This
414 will most likely lead to a crash.
416 The functions PEM_read_bio() and PEM_read() are simple wrappers around
417 PEM_read_bio_ex() and therefore these functions are also directly affected.
419 These functions are also called indirectly by a number of other OpenSSL
420 functions including PEM_X509_INFO_read_bio_ex() and
421 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
422 internal uses of these functions are not vulnerable because the caller does
423 not free the header argument if PEM_read_bio_ex() returns a failure code.
426 *Kurt Roeckx, Matt Caswell*
428 * Fixed Timing Oracle in RSA Decryption.
430 A timing based side channel exists in the OpenSSL RSA Decryption
431 implementation which could be sufficient to recover a plaintext across
432 a network in a Bleichenbacher style attack. To achieve a successful
433 decryption an attacker would have to be able to send a very large number
434 of trial messages for decryption. The vulnerability affects all RSA padding
435 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
438 *Dmitry Belyavsky, Hubert Kario*
440 * Fixed X.509 Name Constraints Read Buffer Overflow.
442 A read buffer overrun can be triggered in X.509 certificate verification,
443 specifically in name constraint checking. The read buffer overrun might
444 result in a crash which could lead to a denial of service attack.
445 In a TLS client, this can be triggered by connecting to a malicious
446 server. In a TLS server, this can be triggered if the server requests
447 client authentication and a malicious client connects.
452 * Fixed X.509 Policy Constraints Double Locking security issue.
454 If an X.509 certificate contains a malformed policy constraint and
455 policy processing is enabled, then a write lock will be taken twice
456 recursively. On some operating systems (most widely: Windows) this
457 results in a denial of service when the affected process hangs. Policy
458 processing being enabled on a publicly facing server is not considered
459 to be a common setup.
464 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
465 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
466 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
467 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
468 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
469 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
470 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
471 for legacy EC and SM2 keys is also changed similarly to honor the
472 equivalent conversion format flag as specified in the underlying
473 `EC_KEY` object being exported to a provider, when this function is
474 called through `EVP_PKEY_export()`.
478 * Fixed a type confusion vulnerability relating to X.400 address processing
479 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
480 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
481 vulnerability may allow an attacker who can provide a certificate chain and
482 CRL (neither of which need have a valid signature) to pass arbitrary pointers
483 to a `memcmp` call, creating a possible read primitive, subject to some
484 constraints. Refer to the advisory for more information. Thanks to David
485 Benjamin for discovering this issue. ([CVE-2023-0286])
487 This issue has been fixed by changing the public header file definition of
488 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
489 possible for any existing application to successfully use the existing
490 definition; however, if any application references the `x400Address` field
491 (e.g. in dead code), note that the type of this field has changed. There is
496 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
498 * Fixed two buffer overflows in punycode decoding functions.
500 A buffer overrun can be triggered in X.509 certificate verification,
501 specifically in name constraint checking. Note that this occurs after
502 certificate chain signature verification and requires either a CA to
503 have signed the malicious certificate or for the application to continue
504 certificate verification despite failure to construct a path to a trusted
507 In a TLS client, this can be triggered by connecting to a malicious
508 server. In a TLS server, this can be triggered if the server requests
509 client authentication and a malicious client connects.
511 An attacker can craft a malicious email address to overflow
512 an arbitrary number of bytes containing the `.` character (decimal 46)
513 on the stack. This buffer overflow could result in a crash (causing a
517 An attacker can craft a malicious email address to overflow four
518 attacker-controlled bytes on the stack. This buffer overflow could
519 result in a crash (causing a denial of service) or potentially remote code
520 execution depending on stack layout for any given platform/compiler.
525 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
526 parameters in OpenSSL code.
527 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
528 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
529 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
530 Using these invalid names may cause algorithms to use slower methods
531 that ignore the CRT parameters.
535 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
540 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
541 data to be signed before signing the certificate.
545 * Added RIPEMD160 to the default provider.
549 * Ensured that the key share group sent or accepted for the key exchange
550 is allowed for the protocol version.
554 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
556 * OpenSSL supports creating a custom cipher via the legacy
557 EVP_CIPHER_meth_new() function and associated function calls. This function
558 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
559 to use the new provider mechanism in order to implement custom ciphers.
561 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
562 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
563 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
564 and decryption initialisation functions). Instead of using the custom cipher
565 directly it incorrectly tries to fetch an equivalent cipher from the
566 available providers. An equivalent cipher is found based on the NID passed to
567 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
568 given cipher. However it is possible for an application to incorrectly pass
569 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
570 is used in this way the OpenSSL encryption/decryption initialisation function
571 will match the NULL cipher as being equivalent and will fetch this from the
572 available providers. This will succeed if the default provider has been
573 loaded (or if a third party provider has been loaded that offers this
574 cipher). Using the NULL cipher means that the plaintext is emitted as the
577 Applications are only affected by this issue if they call
578 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
579 encryption/decryption initialisation function. Applications that only use
580 SSL/TLS are not impacted by this issue.
585 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
590 * Fixed the linux-mips64 Configure target which was missing the
591 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
596 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
601 * Correctly handle a retransmitted ClientHello in DTLS
605 * Fixed detection of ktls support in cross-compile environment on Linux
609 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
614 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
615 report correct results in some cases
619 * Fix UWP builds by defining VirtualLock
623 * For known safe primes use the minimum key length according to RFC 7919.
624 Longer private key sizes unnecessarily raise the cycles needed to compute the
625 shared secret without any increase of the real security. This fixes a
626 regression from 1.1.1 where these shorter keys were generated for the known
631 * Added the loongarch64 target
635 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
636 only passed to the FIPS provider and not to the default or legacy provider.
640 * Fixed reported performance degradation on aarch64. Restored the
641 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
642 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
643 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
644 The new algorithm is still used for 32 bit targets.
648 * Added a missing header for memcmp that caused compilation failure on some
653 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
655 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
656 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
657 This issue makes the RSA implementation with 2048 bit private keys
658 incorrect on such machines and memory corruption will happen during
659 the computation. As a consequence of the memory corruption an attacker
660 may be able to trigger a remote code execution on the machine performing
663 SSL/TLS servers or other servers using 2048 bit RSA private keys running
664 on machines supporting AVX512IFMA instructions of the X86_64 architecture
665 are affected by this issue.
670 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
671 implementation would not encrypt the entirety of the data under some
672 circumstances. This could reveal sixteen bytes of data that was
673 preexisting in the memory that wasn't written. In the special case of
674 "in place" encryption, sixteen bytes of the plaintext would be revealed.
676 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
677 they are both unaffected.
680 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
682 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
684 * In addition to the c_rehash shell command injection identified in
685 CVE-2022-1292, further bugs where the c_rehash script does not
686 properly sanitise shell metacharacters to prevent command injection have been
689 When the CVE-2022-1292 was fixed it was not discovered that there
690 are other places in the script where the file names of certificates
691 being hashed were possibly passed to a command executed through the shell.
693 This script is distributed by some operating systems in a manner where
694 it is automatically executed. On such operating systems, an attacker
695 could execute arbitrary commands with the privileges of the script.
697 Use of the c_rehash script is considered obsolete and should be replaced
698 by the OpenSSL rehash command line tool.
701 *Daniel Fiala, Tomáš Mráz*
703 * Case insensitive string comparison no longer uses locales. It has instead
704 been directly implemented.
708 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
710 * Case insensitive string comparison is reimplemented via new locale-agnostic
711 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
712 comparison. The previous implementation had problems when the Turkish locale
717 * Fixed a bug in the c_rehash script which was not properly sanitising shell
718 metacharacters to prevent command injection. This script is distributed by
719 some operating systems in a manner where it is automatically executed. On
720 such operating systems, an attacker could execute arbitrary commands with the
721 privileges of the script.
723 Use of the c_rehash script is considered obsolete and should be replaced
724 by the OpenSSL rehash command line tool.
729 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
730 certificate on an OCSP response. The bug caused the function in the case
731 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
732 response (meaning a successful verification) even in the case where the
733 response signing certificate fails to verify.
735 It is anticipated that most users of `OCSP_basic_verify` will not use the
736 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
737 a negative value (indicating a fatal error) in the case of a certificate
738 verification failure. The normal expected return value in this case would be
741 This issue also impacts the command line OpenSSL "ocsp" application. When
742 verifying an ocsp response with the "-no_cert_checks" option the command line
743 application will report that the verification is successful even though it
744 has in fact failed. In this case the incorrect successful response will also
745 be accompanied by error messages showing the failure and contradicting the
746 apparently successful result.
751 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
752 AAD data as the MAC key. This made the MAC key trivially predictable.
754 An attacker could exploit this issue by performing a man-in-the-middle attack
755 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
756 that the modified data would still pass the MAC integrity check.
758 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
759 endpoint will always be rejected by the recipient and the connection will
760 fail at that point. Many application protocols require data to be sent from
761 the client to the server first. Therefore, in such a case, only an OpenSSL
762 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
764 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
765 sent in both directions. In this case both clients and servers could be
766 affected, regardless of the application protocol.
768 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
769 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
770 the handshake when using this ciphersuite.
772 The confidentiality of data is not impacted by this issue, i.e. an attacker
773 cannot decrypt data that has been encrypted using this ciphersuite - they can
776 In order for this attack to work both endpoints must legitimately negotiate
777 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
778 OpenSSL 3.0, and is not available within the default provider or the default
779 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
780 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
781 following must have occurred:
783 1) OpenSSL must have been compiled with the (non-default) compile time option
784 enable-weak-ssl-ciphers
786 2) OpenSSL must have had the legacy provider explicitly loaded (either
787 through application code or via configuration)
789 3) The ciphersuite must have been explicitly added to the ciphersuite list
791 4) The libssl security level must have been set to 0 (default is 1)
793 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
795 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
796 others that both endpoints have in common
801 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
802 occupied by the removed hash table entries.
804 This function is used when decoding certificates or keys. If a long lived
805 process periodically decodes certificates or keys its memory usage will
806 expand without bounds and the process might be terminated by the operating
807 system causing a denial of service. Also traversing the empty hash table
808 entries will take increasingly more time.
810 Typically such long lived processes might be TLS clients or TLS servers
811 configured to accept client certificate authentication.
814 *Hugo Landau, Aliaksei Levin*
816 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
817 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
818 statistics are no longer supported. For compatibility, these statistics are
819 still listed in the output but are now always reported as zero.
823 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
825 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
826 for non-prime moduli.
828 Internally this function is used when parsing certificates that contain
829 elliptic curve public keys in compressed form or explicit elliptic curve
830 parameters with a base point encoded in compressed form.
832 It is possible to trigger the infinite loop by crafting a certificate that
833 has invalid explicit curve parameters.
835 Since certificate parsing happens prior to verification of the certificate
836 signature, any process that parses an externally supplied certificate may thus
837 be subject to a denial of service attack. The infinite loop can also be
838 reached when parsing crafted private keys as they can contain explicit
839 elliptic curve parameters.
841 Thus vulnerable situations include:
843 - TLS clients consuming server certificates
844 - TLS servers consuming client certificates
845 - Hosting providers taking certificates or private keys from customers
846 - Certificate authorities parsing certification requests from subscribers
847 - Anything else which parses ASN.1 elliptic curve parameters
849 Also any other applications that use the BN_mod_sqrt() where the attacker
850 can control the parameter values are vulnerable to this DoS issue.
855 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
856 to the list of ciphersuites providing Perfect Forward Secrecy as
857 required by SECLEVEL >= 3.
859 *Dmitry Belyavskiy, Nicola Tuveri*
861 * Made the AES constant time code for no-asm configurations
862 optional due to the resulting 95% performance degradation.
863 The AES constant time code can be enabled, for no assembly
864 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
868 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
873 * The negative return value handling of the certificate verification callback
874 was reverted. The replacement is to set the verification retry state with
875 the SSL_set_retry_verify() function.
879 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
881 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
882 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
883 verify a certificate supplied by a server. That function may return a
884 negative return value to indicate an internal error (for example out of
885 memory). Such a negative return value is mishandled by OpenSSL and will cause
886 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
887 success and a subsequent call to SSL_get_error() to return the value
888 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
889 returned by OpenSSL if the application has previously called
890 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
891 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
892 totally unexpected and applications may not behave correctly as a result. The
893 exact behaviour will depend on the application but it could result in
894 crashes, infinite loops or other similar incorrect responses.
896 This issue is made more serious in combination with a separate bug in OpenSSL
897 3.0 that will cause X509_verify_cert() to indicate an internal error when
898 processing a certificate chain. This will occur where a certificate does not
899 include the Subject Alternative Name extension but where a Certificate
900 Authority has enforced name constraints. This issue can occur even with valid
906 * Corrected a few file name and file reference bugs in the build,
907 installation and setup scripts, which lead to installation verification
908 failures. Slightly enhanced the installation verification script.
912 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
917 * Fixed PVK encoder to properly query for the passphrase.
921 * Multiple fixes in the OSSL_HTTP API functions.
925 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
926 OSSL_PARAM_INTEGER data type and return error on negative numbers
927 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
928 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
932 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
936 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
940 * Multiple threading fixes.
944 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
948 * Allow fetching an operation from the provider that owns an unexportable key
949 as a fallback if that is still allowed by the property query.
953 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
955 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
960 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
961 S390X capability vector to zero. This simplifies testing of different code
962 paths on S390X architecture.
966 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
967 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
968 SP 800-38D". The communication will fail at this point.
972 * The EC_GROUP_clear_free() function is deprecated as there is nothing
973 confidential in EC_GROUP data.
977 * The byte order mark (BOM) character is ignored if encountered at the
978 beginning of a PEM-formatted file.
982 * Added CMS support for the Russian GOST algorithms.
986 * Due to move of the implementation of cryptographic operations
987 to the providers, validation of various operation parameters can
988 be postponed until the actual operation is executed where previously
989 it happened immediately when an operation parameter was set.
991 For example when setting an unsupported curve with
992 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
993 fail but later keygen operations with the EVP_PKEY_CTX will fail.
995 *OpenSSL team members and many third party contributors*
997 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
998 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
999 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1000 instead to retrieve these algorithms from a provider.
1004 * On build targets where the multilib postfix is set in the build
1005 configuration the libdir directory was changing based on whether
1006 the lib directory with the multilib postfix exists on the system
1007 or not. This unpredictable behavior was removed and eventual
1008 multilib postfix is now always added to the default libdir. Use
1009 `--libdir=lib` to override the libdir if adding the postfix is
1014 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1015 no longer interoperable with OpenSSL 1.1.1.
1019 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1020 function codes, this function can only cause problems for calling
1025 * Add a configurable flag to output date formats as ISO 8601. Does not
1026 change the default date format.
1030 * Version of MSVC earlier than 1300 could get link warnings, which could
1031 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1032 Support for this flag has been removed.
1036 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1037 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1038 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1039 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1040 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1044 * The signatures of the functions to get and set options on SSL and
1045 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1046 Some source code changes may be required.
1050 * The public definitions of conf_method_st and conf_st have been
1051 deprecated. They will be made opaque in a future release.
1053 *Rich Salz and Tomáš Mráz*
1055 * Client-initiated renegotiation is disabled by default. To allow it, use
1056 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1057 flag, or the "ClientRenegotiation" config parameter as appropriate.
1061 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1062 or modify relative pathname inclusion.
1066 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1067 validated. Please consult the README-FIPS and
1068 README-PROVIDERS files, as well as the migration guide.
1070 *OpenSSL team members and many third party contributors*
1072 * For the key types DH and DHX the allowed settable parameters are now different.
1076 * The openssl commands that read keys, certificates, and CRLs now
1077 automatically detect the PEM or DER format of the input files.
1079 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1081 * Added enhanced PKCS#12 APIs which accept a library context.
1085 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1089 * Added support for Kernel TLS (KTLS).
1091 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1093 * Support for RFC 5746 secure renegotiation is now required by default for
1094 SSL or TLS connections to succeed.
1098 * The signature of the `copy` functional parameter of the
1099 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1100 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1101 the signature of the `pub_decode` functional parameter of the
1102 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1103 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1107 * The error return values from some control calls (ctrl) have changed.
1111 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1115 * Many functions in the EVP_ namespace that are getters of values from
1116 implementations or contexts were renamed to include get or get0 in their
1117 names. Old names are provided as macro aliases for compatibility and
1122 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1123 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1124 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1129 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1132 * The output from the command line applications may have minor
1137 * The output from numerous "printing" may have minor changes.
1141 * Windows thread synchronization uses read/write primitives (SRWLock) when
1142 supported by the OS, otherwise CriticalSection continues to be used.
1146 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1147 work on read only BIO source/sinks that do not support these functions.
1148 This allows piping or redirection of a file BIO using stdin to be buffered
1149 into memory. This is used internally in OSSL_DECODER_from_bio().
1153 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1154 this function would return one of the values OSSL_STORE_INFO_NAME,
1155 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1156 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1157 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1158 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1159 using this function should be amended to handle the changed return value.
1163 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1164 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1165 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1166 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1167 contains more than one certificate identifier: This means that all
1168 certificates referenced there MUST be part of the validation chain.
1172 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1173 RC5, DESX and DES have been moved to the legacy provider.
1177 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1178 RIPEMD-160 have been moved to the legacy provider.
1182 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1187 * The deprecated functions EVP_PKEY_get0_RSA(),
1188 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1189 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1190 well as the similarly named "get1" functions behave differently in
1195 * A number of functions handling low-level keys or engines were deprecated
1196 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1197 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1198 EVP_PKEY_get0_siphash().
1202 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1203 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1204 will need to load the legacy crypto provider. This includes these PBE
1205 algorithms which use this KDF:
1206 - NID_pbeWithMD2AndDES_CBC
1207 - NID_pbeWithMD5AndDES_CBC
1208 - NID_pbeWithSHA1AndRC2_CBC
1209 - NID_pbeWithMD2AndRC2_CBC
1210 - NID_pbeWithMD5AndRC2_CBC
1211 - NID_pbeWithSHA1AndDES_CBC
1215 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1216 BIO_debug_callback() functions.
1220 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1221 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1225 * The RAND_METHOD APIs have been deprecated.
1229 * The SRP APIs have been deprecated.
1233 * Add a compile time option to prevent the caching of provider fetched
1234 algorithms. This is enabled by including the no-cached-fetch option
1235 at configuration time.
1239 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1240 count of PKCS12_DEFAULT_ITER.
1242 *Tomáš Mráz and Sahana Prasad*
1244 * The openssl speed command does not use low-level API calls anymore.
1248 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1251 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1253 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1257 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1258 providers may supply their own group implementations (using either the "key
1259 exchange" or the "key encapsulation" methods) which will automatically be
1260 detected and used by libssl.
1262 *Matt Caswell, Nicola Tuveri*
1264 * The undocumented function X509_certificate_type() has been deprecated;
1268 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1272 * Removed RSA padding mode for SSLv23 (which was only used for
1273 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1274 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1279 * Deprecated the obsolete X9.31 RSA key generation related functions.
1281 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1282 is not allowed to return a value > 1, this is no more taken as failure.
1284 *Viktor Dukhovni and David von Oheimb*
1286 * Deprecated the obsolete X9.31 RSA key generation related functions
1287 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1288 BN_X931_generate_prime_ex().
1292 * The default key generation method for the regular 2-prime RSA keys was
1293 changed to the FIPS 186-4 B.3.6 method.
1297 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1301 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1305 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1306 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1308 *Rich Salz, Richard Levitte, and David von Oheimb*
1310 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1314 * Deprecated `OCSP_parse_url()`.
1318 * Validation of SM2 keys has been separated from the validation of regular EC
1323 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1324 switches: a validation failure triggers an early exit, returning a failure
1325 exit status to the parent process.
1329 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1330 to ignore unknown ciphers.
1334 * The `-cipher-commands` and `-digest-commands` options
1335 of the command line utility `list` have been deprecated.
1336 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1340 * Added convenience functions for generating asymmetric key pairs:
1341 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1342 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1346 * All of the low-level EC_KEY functions have been deprecated.
1348 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1350 * Deprecated all the libcrypto and libssl error string loading
1355 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1356 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1361 * The `-crypt` option to the `passwd` command line tool has been removed.
1365 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1370 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1374 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1375 EVP_PKEY_get1_tls_encodedpoint().
1379 * The security callback, which can be customised by application code, supports
1380 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1381 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1385 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1386 interface. Their functionality remains unchanged.
1390 * Added new option for 'openssl list', '-providers', which will display the
1391 list of loaded providers, their names, version and status. It optionally
1392 displays their gettable parameters.
1396 * Removed EVP_PKEY_set_alias_type().
1400 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1401 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1405 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1406 parameters are still checked for type safety at compile time via helper
1411 * Remove the RAND_DRBG API
1413 *Paul Dale and Matthias St. Pierre*
1415 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1416 as well as actual hostnames.
1420 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1421 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1422 conversely, silently ignore DTLS protocol version bounds when configuring
1423 TLS-based contexts. The commands can be repeated to set bounds of both
1424 types. The same applies with the corresponding "min_protocol" and
1425 "max_protocol" command-line switches, in case some application uses both TLS
1428 SSL_CTX instances that are created for a fixed protocol version (e.g.
1429 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1430 attempts to apply bounds to these protocol versions would result in an
1431 error. Now only the "version-flexible" SSL_CTX instances are subject to
1432 limits in configuration files in command-line options.
1436 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1441 * Reworked the recorded ERR codes to make better space for system errors.
1442 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1443 given code is a system error (true) or an OpenSSL error (false).
1447 * Reworked the test perl framework to better allow parallel testing.
1449 *Nicola Tuveri and David von Oheimb*
1451 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1452 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1456 * 'Configure' has been changed to figure out the configuration target if
1457 none is given on the command line. Consequently, the 'config' script is
1458 now only a mere wrapper. All documentation is changed to only mention
1461 *Rich Salz and Richard Levitte*
1463 * Added a library context `OSSL_LIB_CTX` that applications as well as
1464 other libraries can use to form a separate context within which
1465 libcrypto operations are performed.
1469 * Added various `_ex` functions to the OpenSSL API that support using
1470 a non-default `OSSL_LIB_CTX`.
1474 * Handshake now fails if Extended Master Secret extension is dropped
1479 * Dropped interactive mode from the `openssl` program.
1483 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1485 *David von Oheimb and Shane Lontis*
1487 * Deprecated `EC_METHOD_get_field_type()`.
1491 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1492 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1493 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1497 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1501 * Add CAdES-BES signature verification support, mostly derived
1502 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1504 *Filipe Raimundo da Silva*
1506 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1510 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1511 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1515 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1519 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1520 EC_KEY_precompute_mult().
1524 * Deprecated EC_POINTs_mul().
1528 * Removed FIPS_mode() and FIPS_mode_set().
1532 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1536 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1537 EC_POINT_get_Jprojective_coordinates_GFp().
1541 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1542 arrays to be more easily constructed via a series of utility functions.
1543 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1544 the various push functions and finally convert to a passable OSSL_PARAM
1545 array using OSSL_PARAM_BLD_to_param().
1549 * The security strength of SHA1 and MD5 based signatures in TLS has been
1554 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1555 contain a provider side internal key.
1559 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1563 * Project text documents not yet having a proper file name extension
1564 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1565 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1569 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1570 have been converted to Markdown with the goal to produce documents
1571 which not only look pretty when viewed online in the browser, but
1572 remain well readable inside a plain text editor.
1574 To achieve this goal, a 'minimalistic' Markdown style has been applied
1575 which avoids formatting elements that interfere too much with the
1576 reading flow in the text file. For example, it
1578 * avoids [ATX headings][] and uses [setext headings][] instead
1579 (which works for `<h1>` and `<h2>` headings only).
1580 * avoids [inline links][] and uses [reference links][] instead.
1581 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1583 [ATX headings]: https://github.github.com/gfm/#atx-headings
1584 [setext headings]: https://github.github.com/gfm/#setext-headings
1585 [inline links]: https://github.github.com/gfm/#inline-link
1586 [reference links]: https://github.github.com/gfm/#reference-link
1587 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1588 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1590 *Matthias St. Pierre*
1592 * The test suite is changed to preserve results of each test recipe.
1593 A new directory test-runs/ with subdirectories named like the
1594 test recipes are created in the build tree for this purpose.
1598 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1599 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1600 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1602 *David von Oheimb, Martin Peylo*
1604 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1605 It supports arbitrary request and response content types, GET redirection,
1606 TLS, connections via HTTP(S) proxies, connections and exchange via
1607 user-defined BIOs (allowing implicit connections), persistent connections,
1608 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1609 The legacy OCSP-focused (and only partly documented) API
1610 is retained for backward compatibility, while most of it is deprecated.
1614 * Added `util/check-format.pl`, a tool for checking adherence to the
1615 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1616 The checks performed are incomplete and yield some false positives.
1617 Still the tool should be useful for detecting most typical glitches.
1621 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1622 If domain name resolution yields multiple IP addresses all of them are tried
1623 after `connect()` failures.
1627 * All of the low-level RSA functions have been deprecated.
1631 * X509 certificates signed using SHA1 are no longer allowed at security
1636 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1637 modified to use PKEY APIs. These commands are now in maintenance mode
1638 and no new features will be added to them.
1642 * The command line utility rsautl has been deprecated.
1646 * The command line utilities genrsa and rsa have been modified to use PKEY
1647 APIs. They now write PKCS#8 keys by default. These commands are now in
1648 maintenance mode and no new features will be added to them.
1652 * All of the low-level DH functions have been deprecated.
1654 *Paul Dale and Matt Caswell*
1656 * All of the low-level DSA functions have been deprecated.
1660 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1661 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1665 * Deprecated low-level ECDH and ECDSA functions.
1669 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1673 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1674 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1675 a new formulation to include all the things it can be used for,
1676 as well as words of caution.
1680 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1684 * All of the low-level HMAC functions have been deprecated.
1686 *Paul Dale and David von Oheimb*
1688 * Over two thousand fixes were made to the documentation, including:
1689 - Common options (such as -rand/-writerand, TLS version control, etc)
1690 were refactored and point to newly-enhanced descriptions in openssl.pod.
1691 - Added style conformance for all options (with help from Richard Levitte),
1692 documented all reported missing options, added a CI build to check
1693 that all options are documented and that no unimplemented options
1695 - Documented some internals, such as all use of environment variables.
1696 - Addressed all internal broken L<> references.
1700 * All of the low-level CMAC functions have been deprecated.
1704 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1705 functions have been deprecated.
1707 *Paul Dale and David von Oheimb*
1709 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1710 set of functions. The documentation mentioned negative values for some
1711 errors, but this was never the case, so the mention of negative values
1714 Code that followed the documentation and thereby check with something
1715 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1719 * All of the low-level cipher functions have been deprecated.
1721 *Matt Caswell and Paul Dale*
1723 * Removed include/openssl/opensslconf.h.in and replaced it with
1724 include/openssl/configuration.h.in, which differs in not including
1725 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1726 was added to include both.
1728 This allows internal hacks where one might need to modify the set
1729 of configured macros, for example this if deprecated symbols are
1730 still supposed to be available internally:
1732 #include <openssl/configuration.h>
1734 #undef OPENSSL_NO_DEPRECATED
1735 #define OPENSSL_SUPPRESS_DEPRECATED
1737 #include <openssl/macros.h>
1739 This should not be used by applications that use the exported
1740 symbols, as that will lead to linking errors.
1744 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1745 used in exponentiation with 512-bit moduli. No EC algorithms are
1746 affected. Analysis suggests that attacks against 2-prime RSA1024,
1747 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1748 difficult to perform and are not believed likely. Attacks against DH512
1749 are considered just feasible. However, for an attack the target would
1750 have to re-use the DH512 private key, which is not recommended anyway.
1751 Also applications directly using the low-level API BN_mod_exp may be
1752 affected if they use BN_FLG_CONSTTIME.
1757 * Most memory-debug features have been deprecated, and the functionality
1758 replaced with no-ops.
1762 * Added documentation for the STACK API.
1766 * Introduced a new method type and API, OSSL_ENCODER, to represent
1767 generic encoders. These do the same sort of job that PEM writers
1768 and d2i functions do, but with support for methods supplied by
1769 providers, and the possibility for providers to support other
1774 * Introduced a new method type and API, OSSL_DECODER, to represent
1775 generic decoders. These do the same sort of job that PEM readers
1776 and i2d functions do, but with support for methods supplied by
1777 providers, and the possibility for providers to support other
1782 * Added a .pragma directive to the syntax of configuration files, to
1783 allow varying behavior in a supported and predictable manner.
1784 Currently added pragma:
1788 This allows dollar signs to be a keyword character unless it's
1789 followed by a opening brace or parenthesis. This is useful for
1790 platforms where dollar signs are commonly used in names, such as
1791 volume names and system directory names on VMS.
1795 * Added functionality to create an EVP_PKEY from user data.
1799 * Change the interpretation of the '--api' configuration option to
1800 mean that this is a desired API compatibility level with no
1801 further meaning. The previous interpretation, that this would
1802 also mean to remove all deprecated symbols up to and including
1803 the given version, no requires that 'no-deprecated' is also used
1804 in the configuration.
1806 When building applications, the desired API compatibility level
1807 can be set with the OPENSSL_API_COMPAT macro like before. For
1808 API compatibility version below 3.0, the old style numerical
1809 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1810 For version 3.0 and on, the value is expected to be the decimal
1811 value calculated from the major and minor version like this:
1813 MAJOR * 10000 + MINOR * 100
1817 -DOPENSSL_API_COMPAT=30000 For 3.0
1818 -DOPENSSL_API_COMPAT=30200 For 3.2
1820 To hide declarations that are deprecated up to and including the
1821 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1822 given when building the application as well.
1826 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1827 access to certificate and CRL stores via URIs and OSSL_STORE
1830 This adds the following functions:
1832 - X509_LOOKUP_store()
1833 - X509_STORE_load_file()
1834 - X509_STORE_load_path()
1835 - X509_STORE_load_store()
1836 - SSL_add_store_cert_subjects_to_stack()
1837 - SSL_CTX_set_default_verify_store()
1838 - SSL_CTX_load_verify_file()
1839 - SSL_CTX_load_verify_dir()
1840 - SSL_CTX_load_verify_store()
1844 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1845 The presence of this system service is determined at run-time.
1849 * Added functionality to create an EVP_PKEY context based on data
1850 for methods from providers. This takes an algorithm name and a
1851 property query string and simply stores them, with the intent
1852 that any operation that uses this context will use those strings
1853 to fetch the needed methods implicitly, thereby making the port
1854 of application written for pre-3.0 OpenSSL easier.
1858 * The undocumented function NCONF_WIN32() has been deprecated; for
1859 conversion details see the HISTORY section of doc/man5/config.pod
1863 * Introduced the new functions EVP_DigestSignInit_ex() and
1864 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1865 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1866 pages for further details.
1870 * Over two thousand fixes were made to the documentation, including:
1871 adding missing command flags, better style conformance, documentation
1874 *Rich Salz, Richard Levitte*
1876 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1877 X25519, X448, Ed25519 and Ed448.
1881 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1886 * Deprecated the public definition of `ERR_STATE` as well as the function
1887 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1892 * Added ERR functionality to give callers access to the stored function
1893 names that have replaced the older function code based functions.
1895 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1896 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1897 ERR_peek_error_all() and ERR_peek_last_error_all().
1899 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1900 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1901 ERR_func_error_string().
1905 * Extended testing to be verbose for failing tests only. The make variables
1906 VERBOSE_FAILURE or VF can be used to enable this:
1908 $ make VF=1 test # Unix
1909 $ mms /macro=(VF=1) test ! OpenVMS
1910 $ nmake VF=1 test # Windows
1914 * Added the `-copy_extensions` option to the `x509` command for use with
1915 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1916 all extensions in the request are copied to the certificate or vice versa.
1918 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1920 * Added the `-copy_extensions` option to the `req` command for use with
1921 `-x509`. When given with the `copy` or `copyall` argument,
1922 all extensions in the certification request are copied to the certificate.
1926 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1927 they generate are by default RFC 5280 compliant in the following sense:
1928 There is a subjectKeyIdentifier extension with a hash value of the public key
1929 and for not self-signed certs there is an authorityKeyIdentifier extension
1930 with a keyIdentifier field or issuer information identifying the signing key.
1931 This is done unless some configuration overrides the new default behavior,
1932 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1936 * Added several checks to `X509_verify_cert()` according to requirements in
1937 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1938 (which may be done by using the CLI option `-x509_strict`):
1939 * The basicConstraints of CA certificates must be marked critical.
1940 * CA certificates must explicitly include the keyUsage extension.
1941 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1942 * The issuer name of any certificate must not be empty.
1943 * The subject name of CA certs, certs with keyUsage crlSign,
1944 and certs without subjectAlternativeName must not be empty.
1945 * If a subjectAlternativeName extension is given it must not be empty.
1946 * The signatureAlgorithm field and the cert signature must be consistent.
1947 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1948 must not be marked critical.
1949 * The authorityKeyIdentifier must be given for X.509v3 certs
1950 unless they are self-signed.
1951 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1955 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1956 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1960 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1961 used even when parsing explicit parameters, when loading a encoded key
1962 or calling `EC_GROUP_new_from_ecpkparameters()`/
1963 `EC_GROUP_new_from_ecparameters()`.
1964 This prevents bypass of security hardening and performance gains,
1965 especially for curves with specialized EC_METHODs.
1966 By default, if a key encoded with explicit parameters is loaded and later
1967 encoded, the output is still encoded with explicit parameters, even if
1968 internally a "named" EC_GROUP is used for computation.
1972 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1973 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1974 NULL. After this change, only the cofactor parameter can be NULL. It also
1975 does some minimal sanity checks on the passed order.
1980 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1981 An attack is simple, if the first CMS_recipientInfo is valid but the
1982 second CMS_recipientInfo is chosen ciphertext. If the second
1983 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1984 encryption key will be replaced by garbage, and the message cannot be
1985 decoded, but if the RSA decryption fails, the correct encryption key is
1986 used and the recipient will not notice the attack.
1987 As a work around for this potential attack the length of the decrypted
1988 key must be equal to the cipher default key length, in case the
1989 certificate is not given and all recipientInfo are tried out.
1990 The old behaviour can be re-enabled in the CMS code by setting the
1991 CMS_DEBUG_DECRYPT flag.
1995 * Early start up entropy quality from the DEVRANDOM seed source has been
1996 improved for older Linux systems. The RAND subsystem will wait for
1997 /dev/random to be producing output before seeding from /dev/urandom.
1998 The seeded state is stored for future library initialisations using
1999 a system global shared memory segment. The shared memory identifier
2000 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2001 the desired value. The default identifier is 114.
2005 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2006 when primes for RSA keys are computed.
2007 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2008 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2009 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2010 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2011 This avoids possible fingerprinting of newly generated RSA modules.
2015 * Correct the extended master secret constant on EBCDIC systems. Without this
2016 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2017 negotiate EMS will fail. Unfortunately this also means that TLS connections
2018 between EBCDIC systems with this fix, and EBCDIC systems without this
2019 fix will fail if they negotiate EMS.
2023 * Changed the library initialisation so that the config file is now loaded
2024 by default. This was already the case for libssl. It now occurs for both
2025 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2026 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2030 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2031 where the former acts as a replacement for `ERR_put_error()`, and the
2032 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2033 `ERR_raise_data()` adds more flexibility by taking a format string and
2034 an arbitrary number of arguments following it, to be processed with
2039 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2040 to check if a named provider is loaded and available. When called, it
2041 will also activate all fallback providers if such are still present.
2045 * Enforce a minimum DH modulus size of 512 bits.
2049 * Changed DH parameters to generate the order q subgroup instead of 2q.
2050 Previously generated DH parameters are still accepted by DH_check
2051 but DH_generate_key works around that by clearing bit 0 of the
2052 private key for those. This avoids leaking bit 0 of the private key.
2056 * Significantly reduce secure memory usage by the randomness pools.
2060 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2065 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2066 algorithms. An implementation of a key exchange algorithm can be obtained
2067 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2068 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2069 the older EVP_PKEY_derive_init() function. See the man pages for the new
2070 functions for further details.
2074 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2078 * Removed the function names from error messages and deprecated the
2083 * Removed NextStep support and the macro OPENSSL_UNISTD
2087 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2088 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2089 Also removed "export var as function" capability; we do not export
2090 variables, only functions.
2094 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2095 an error and 1 indicating success. In previous versions of OpenSSL this
2096 was a void type. If a key was set longer than the maximum possible this
2101 * Support SM2 signing and verification schemes with X509 certificate.
2105 * Use SHA256 as the default digest for TS query in the `ts` app.
2109 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2113 * Default cipher lists/suites are now available via a function, the
2114 #defines are deprecated.
2118 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2119 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2120 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2124 * Join the directories crypto/x509 and crypto/x509v3
2128 * Added command 'openssl kdf' that uses the EVP_KDF API.
2132 * Added command 'openssl mac' that uses the EVP_MAC API.
2136 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2137 as default directories. Also added the command 'openssl info'
2138 for scripting purposes.
2142 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2147 * Add prediction resistance to the DRBG reseeding process.
2151 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2152 mandated by IEEE Std 1619-2018.
2156 * Added newline escaping functionality to a filename when using openssl dgst.
2157 This output format is to replicate the output format found in the `*sum`
2158 checksum programs. This aims to preserve backward compatibility.
2160 *Matt Eaton, Richard Levitte, and Paul Dale*
2162 * Removed the heartbeat message in DTLS feature, as it has very
2163 little usage and doesn't seem to fulfill a valuable purpose.
2164 The configuration option is now deprecated.
2168 * Changed the output of 'openssl {digestname} < file' to display the
2169 digest name in its output.
2173 * Added a new generic trace API which provides support for enabling
2174 instrumentation through trace output.
2176 *Richard Levitte & Matthias St. Pierre*
2178 * Added build tests for C++. These are generated files that only do one
2179 thing, to include one public OpenSSL head file each. This tests that
2180 the public header files can be usefully included in a C++ application.
2182 This test isn't enabled by default. It can be enabled with the option
2183 'enable-buildtest-c++'.
2187 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2191 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2195 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2199 * Added KMAC to EVP_MAC.
2203 * Added property based algorithm implementation selection framework to
2208 * Added SCA hardening for modular field inversion in EC_GROUP through
2209 a new dedicated field_inv() pointer in EC_METHOD.
2210 This also addresses a leakage affecting conversions from projective
2211 to affine coordinates.
2213 *Billy Bob Brumley, Nicola Tuveri*
2215 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2216 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2217 those algorithms that were already supported through the EVP_PKEY API
2218 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2219 and scrypt are now wrappers that call EVP_KDF.
2223 * Build devcrypto engine as a dynamic engine.
2225 *Eneas U de Queiroz*
2227 * Add keyed BLAKE2 to EVP_MAC.
2231 * Fix a bug in the computation of the endpoint-pair shared secret used
2232 by DTLS over SCTP. This breaks interoperability with older versions
2233 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2234 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2235 interoperability with such broken implementations. However, enabling
2236 this switch breaks interoperability with correct implementations.
2238 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2239 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2243 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2247 * Changed the license to the Apache License v2.0.
2251 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2253 - Major releases (indicated by incrementing the MAJOR release number)
2254 may introduce incompatible API/ABI changes.
2255 - Minor releases (indicated by incrementing the MINOR release number)
2256 may introduce new features but retain API/ABI compatibility.
2257 - Patch releases (indicated by incrementing the PATCH number)
2258 are intended for bug fixes and other improvements of existing
2259 features only (like improving performance or adding documentation)
2260 and retain API/ABI compatibility.
2264 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2268 * Remove the 'dist' target and add a tarball building script. The
2269 'dist' target has fallen out of use, and it shouldn't be
2270 necessary to configure just to create a source distribution.
2274 * Recreate the OS390-Unix config target. It no longer relies on a
2275 special script like it did for OpenSSL pre-1.1.0.
2279 * Instead of having the source directories listed in Configure, add
2280 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2285 * Add GMAC to EVP_MAC.
2289 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2293 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2294 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2295 to facilitate the continued use of MACs through raw private keys in
2296 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2300 * Deprecate ECDH_KDF_X9_62().
2304 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2305 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2306 are retained for backwards compatibility.
2310 * AES-XTS mode now enforces that its two keys are different to mitigate
2311 the attacked described in "Efficient Instantiations of Tweakable
2312 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2313 Details of this attack can be obtained from:
2314 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2318 * Rename the object files, i.e. give them other names than in previous
2319 versions. Their names now include the name of the final product, as
2320 well as its type mnemonic (bin, lib, shlib).
2324 * Added new option for 'openssl list', '-objects', which will display the
2325 list of built in objects, i.e. OIDs with names.
2329 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2330 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2335 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2336 improves application performance by removing data copies and providing
2337 applications with zero-copy system calls such as sendfile and splice.
2341 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2345 * `PKCS12_parse` now maintains the order of the parsed certificates
2346 when outputting them via `*ca` (rather than reversing it).
2350 * Deprecated pthread fork support methods.
2354 * Added support for FFDHE key exchange in TLS 1.3.
2358 * Added a new concept for OpenSSL plugability: providers. This
2359 functionality is designed to replace the ENGINE API and ENGINE
2360 implementations, and to be much more dynamic, allowing provider
2361 authors to introduce new algorithms among other things, as long as
2362 there's an API that supports the algorithm type.
2364 With this concept comes a new core API for interaction between
2365 libcrypto and provider implementations. Public libcrypto functions
2366 that want to use providers do so through this core API.
2368 The main documentation for this core API is found in
2369 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2370 refer to other manuals describing the API specific for supported
2371 algorithm types (also called operations).
2378 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2380 * Avoid loading of a dynamic engine twice.
2384 * Prioritise DANE TLSA issuer certs over peer certs
2388 * Fixed random API for MacOS prior to 10.12
2390 These MacOS versions don't support the CommonCrypto APIs
2394 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2396 * Fixed an SM2 Decryption Buffer Overflow.
2398 In order to decrypt SM2 encrypted data an application is expected to
2399 call the API function EVP_PKEY_decrypt(). Typically an application will
2400 call this function twice. The first time, on entry, the "out" parameter
2401 can be NULL and, on exit, the "outlen" parameter is populated with the
2402 buffer size required to hold the decrypted plaintext. The application
2403 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2404 again, but this time passing a non-NULL value for the "out" parameter.
2406 A bug in the implementation of the SM2 decryption code means that the
2407 calculation of the buffer size required to hold the plaintext returned
2408 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2409 size required by the second call. This can lead to a buffer overflow
2410 when EVP_PKEY_decrypt() is called by the application a second time with
2411 a buffer that is too small.
2413 A malicious attacker who is able present SM2 content for decryption to
2414 an application could cause attacker chosen data to overflow the buffer
2415 by up to a maximum of 62 bytes altering the contents of other data held
2416 after the buffer, possibly changing application behaviour or causing
2417 the application to crash. The location of the buffer is application
2418 dependent but is typically heap allocated.
2423 * Fixed various read buffer overruns processing ASN.1 strings
2425 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2426 structure which contains a buffer holding the string data and a field
2427 holding the buffer length. This contrasts with normal C strings which
2428 are represented as a buffer for the string data which is terminated
2429 with a NUL (0) byte.
2431 Although not a strict requirement, ASN.1 strings that are parsed using
2432 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2433 well as any string whose value has been set with the ASN1_STRING_set()
2434 function will additionally NUL terminate the byte array in the
2435 ASN1_STRING structure.
2437 However, it is possible for applications to directly construct valid
2438 ASN1_STRING structures which do not NUL terminate the byte array by
2439 directly setting the "data" and "length" fields in the ASN1_STRING
2440 array. This can also happen by using the ASN1_STRING_set0() function.
2442 Numerous OpenSSL functions that print ASN.1 data have been found to
2443 assume that the ASN1_STRING byte array will be NUL terminated, even
2444 though this is not guaranteed for strings that have been directly
2445 constructed. Where an application requests an ASN.1 structure to be
2446 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2447 been directly constructed by the application without NUL terminating
2448 the "data" field, then a read buffer overrun can occur.
2450 The same thing can also occur during name constraints processing
2451 of certificates (for example if a certificate has been directly
2452 constructed by the application instead of loading it via the OpenSSL
2453 parsing functions, and the certificate contains non NUL terminated
2454 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2455 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2457 If a malicious actor can cause an application to directly construct an
2458 ASN1_STRING and then process it through one of the affected OpenSSL
2459 functions then this issue could be hit. This might result in a crash
2460 (causing a Denial of Service attack). It could also result in the
2461 disclosure of private memory contents (such as private keys, or
2462 sensitive plaintext).
2467 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2469 * Fixed a problem with verifying a certificate chain when using the
2470 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2471 the certificates present in a certificate chain. It is not set by default.
2473 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2474 the chain that have explicitly encoded elliptic curve parameters was added
2475 as an additional strict check.
2477 An error in the implementation of this check meant that the result of a
2478 previous check to confirm that certificates in the chain are valid CA
2479 certificates was overwritten. This effectively bypasses the check
2480 that non-CA certificates must not be able to issue other certificates.
2482 If a "purpose" has been configured then there is a subsequent opportunity
2483 for checks that the certificate is a valid CA. All of the named "purpose"
2484 values implemented in libcrypto perform this check. Therefore, where
2485 a purpose is set the certificate chain will still be rejected even when the
2486 strict flag has been used. A purpose is set by default in libssl client and
2487 server certificate verification routines, but it can be overridden or
2488 removed by an application.
2490 In order to be affected, an application must explicitly set the
2491 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2492 for the certificate verification or, in the case of TLS client or server
2493 applications, override the default purpose.
2498 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2499 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2500 renegotiation ClientHello omits the signature_algorithms extension (where it
2501 was present in the initial ClientHello), but includes a
2502 signature_algorithms_cert extension then a NULL pointer dereference will
2503 result, leading to a crash and a denial of service attack.
2505 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2506 (which is the default configuration). OpenSSL TLS clients are not impacted by
2510 *Peter Kästle and Samuel Sapalski*
2512 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2514 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2515 create a unique hash value based on the issuer and serial number data
2516 contained within an X509 certificate. However, it was failing to correctly
2517 handle any errors that may occur while parsing the issuer field (which might
2518 occur if the issuer field is maliciously constructed). This may subsequently
2519 result in a NULL pointer deref and a crash leading to a potential denial of
2525 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2526 padding mode to correctly check for rollback attacks. This is considered a
2527 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2532 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2533 functions. Previously they could overflow the output length argument in some
2534 cases where the input length is close to the maximum permissible length for
2535 an integer on the platform. In such cases the return value from the function
2536 call would be 1 (indicating success), but the output length value would be
2537 negative. This could cause applications to behave incorrectly or crash.
2542 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2543 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2544 could be exploited in a side channel attack to recover the password. Since
2545 the attack is local host only this is outside of the current OpenSSL
2546 threat model and therefore no CVE is assigned.
2548 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2553 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2555 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2556 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2557 If an attacker can control both items being compared then this could lead
2558 to a possible denial of service attack. OpenSSL itself uses the
2559 GENERAL_NAME_cmp function for two purposes:
2560 1) Comparing CRL distribution point names between an available CRL and a
2561 CRL distribution point embedded in an X509 certificate
2562 2) When verifying that a timestamp response token signer matches the
2563 timestamp authority name (exposed via the API functions
2564 TS_RESP_verify_response and TS_RESP_verify_token)
2569 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2571 * Certificates with explicit curve parameters are now disallowed in
2572 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2576 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2577 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2578 conversely, silently ignore DTLS protocol version bounds when configuring
2579 TLS-based contexts. The commands can be repeated to set bounds of both
2580 types. The same applies with the corresponding "min_protocol" and
2581 "max_protocol" command-line switches, in case some application uses both TLS
2584 SSL_CTX instances that are created for a fixed protocol version (e.g.
2585 TLSv1_server_method()) also silently ignore version bounds. Previously
2586 attempts to apply bounds to these protocol versions would result in an
2587 error. Now only the "version-flexible" SSL_CTX instances are subject to
2588 limits in configuration files in command-line options.
2592 * Handshake now fails if Extended Master Secret extension is dropped
2597 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2599 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2601 * Fixed segmentation fault in SSL_check_chain()
2602 Server or client applications that call the SSL_check_chain() function
2603 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2604 dereference as a result of incorrect handling of the
2605 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2606 or unrecognised signature algorithm is received from the peer. This could
2607 be exploited by a malicious peer in a Denial of Service attack.
2612 * Added AES consttime code for no-asm configurations
2613 an optional constant time support for AES was added
2614 when building openssl for no-asm.
2615 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2616 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2617 At this time this feature is by default disabled.
2618 It will be enabled by default in 3.0.
2622 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2624 * Revert the change of EOF detection while reading in libssl to avoid
2625 regressions in applications depending on the current way of reporting
2626 the EOF. As the existing method is not fully accurate the change to
2627 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2628 branch and will be present in the 3.0 release.
2632 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2633 when primes for RSA keys are computed.
2634 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2635 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2636 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2637 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2638 This avoids possible fingerprinting of newly generated RSA modules.
2642 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2644 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2645 while reading in libssl then we would report an error back to the
2646 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2647 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2648 therefore give a hint as to what went wrong.
2652 * Check that ed25519 and ed448 are allowed by the security level. Previously
2653 signature algorithms not using an MD were not being checked that they were
2654 allowed by the security level.
2658 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2659 was not quite right. The behaviour was not consistent between resumption
2660 and normal handshakes, and also not quite consistent with historical
2661 behaviour. The behaviour in various scenarios has been clarified and
2662 it has been updated to make it match historical behaviour as closely as
2667 * *[VMS only]* The header files that the VMS compilers include automatically,
2668 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2669 that the C++ compiler doesn't understand. This is a shortcoming in the
2670 compiler, but can be worked around with `__cplusplus` guards.
2672 C++ applications that use OpenSSL libraries must be compiled using the
2673 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2674 functions. Otherwise, only functions with symbols of less than 31
2675 characters can be used, as the linker will not be able to successfully
2676 resolve symbols with longer names.
2680 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2681 The presence of this system service is determined at run-time.
2685 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2690 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2692 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2693 number generator (RNG). This was intended to include protection in the
2694 event of a fork() system call in order to ensure that the parent and child
2695 processes did not share the same RNG state. However, this protection was not
2696 being used in the default case.
2698 A partial mitigation for this issue is that the output from a high
2699 precision timer is mixed into the RNG state so the likelihood of a parent
2700 and child process sharing state is significantly reduced.
2702 If an application already calls OPENSSL_init_crypto() explicitly using
2703 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2706 *Matthias St. Pierre*
2708 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2709 used even when parsing explicit parameters, when loading a encoded key
2710 or calling `EC_GROUP_new_from_ecpkparameters()`/
2711 `EC_GROUP_new_from_ecparameters()`.
2712 This prevents bypass of security hardening and performance gains,
2713 especially for curves with specialized EC_METHODs.
2714 By default, if a key encoded with explicit parameters is loaded and later
2715 encoded, the output is still encoded with explicit parameters, even if
2716 internally a "named" EC_GROUP is used for computation.
2720 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2721 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2722 NULL. After this change, only the cofactor parameter can be NULL. It also
2723 does some minimal sanity checks on the passed order.
2728 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2729 An attack is simple, if the first CMS_recipientInfo is valid but the
2730 second CMS_recipientInfo is chosen ciphertext. If the second
2731 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2732 encryption key will be replaced by garbage, and the message cannot be
2733 decoded, but if the RSA decryption fails, the correct encryption key is
2734 used and the recipient will not notice the attack.
2735 As a work around for this potential attack the length of the decrypted
2736 key must be equal to the cipher default key length, in case the
2737 certificate is not given and all recipientInfo are tried out.
2738 The old behaviour can be re-enabled in the CMS code by setting the
2739 CMS_DEBUG_DECRYPT flag.
2744 * Early start up entropy quality from the DEVRANDOM seed source has been
2745 improved for older Linux systems. The RAND subsystem will wait for
2746 /dev/random to be producing output before seeding from /dev/urandom.
2747 The seeded state is stored for future library initialisations using
2748 a system global shared memory segment. The shared memory identifier
2749 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2750 the desired value. The default identifier is 114.
2754 * Correct the extended master secret constant on EBCDIC systems. Without this
2755 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2756 negotiate EMS will fail. Unfortunately this also means that TLS connections
2757 between EBCDIC systems with this fix, and EBCDIC systems without this
2758 fix will fail if they negotiate EMS.
2762 * Use Windows installation paths in the mingw builds
2764 Mingw isn't a POSIX environment per se, which means that Windows
2765 paths should be used for installation.
2770 * Changed DH_check to accept parameters with order q and 2q subgroups.
2771 With order 2q subgroups the bit 0 of the private key is not secret
2772 but DH_generate_key works around that by clearing bit 0 of the
2773 private key for those. This avoids leaking bit 0 of the private key.
2777 * Significantly reduce secure memory usage by the randomness pools.
2781 * Revert the DEVRANDOM_WAIT feature for Linux systems
2783 The DEVRANDOM_WAIT feature added a select() call to wait for the
2784 /dev/random device to become readable before reading from the
2785 /dev/urandom device.
2787 It turned out that this change had negative side effects on
2788 performance which were not acceptable. After some discussion it
2789 was decided to revert this feature and leave it up to the OS
2790 resp. the platform maintainer to ensure a proper initialization
2791 during early boot time.
2793 *Matthias St. Pierre*
2795 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2797 * Add build tests for C++. These are generated files that only do one
2798 thing, to include one public OpenSSL head file each. This tests that
2799 the public header files can be usefully included in a C++ application.
2801 This test isn't enabled by default. It can be enabled with the option
2802 'enable-buildtest-c++'.
2806 * Enable SHA3 pre-hashing for ECDSA and DSA.
2810 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2811 This changes the size when using the `genpkey` command when no size is given.
2812 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2813 generation commands to use 2048 bits by default.
2817 * Reorganize the manual pages to consistently have RETURN VALUES,
2818 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2819 util/fix-doc-nits accordingly.
2821 *Paul Yang, Joshua Lock*
2823 * Add the missing accessor EVP_PKEY_get0_engine()
2827 * Have commands like `s_client` and `s_server` output the signature scheme
2828 along with other cipher suite parameters when debugging.
2832 * Make OPENSSL_config() error agnostic again.
2836 * Do the error handling in RSA decryption constant time.
2840 * Prevent over long nonces in ChaCha20-Poly1305.
2842 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2843 for every encryption operation. RFC 7539 specifies that the nonce value
2844 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2845 and front pads the nonce with 0 bytes if it is less than 12
2846 bytes. However it also incorrectly allows a nonce to be set of up to 16
2847 bytes. In this case only the last 12 bytes are significant and any
2848 additional leading bytes are ignored.
2850 It is a requirement of using this cipher that nonce values are
2851 unique. Messages encrypted using a reused nonce value are susceptible to
2852 serious confidentiality and integrity attacks. If an application changes
2853 the default nonce length to be longer than 12 bytes and then makes a
2854 change to the leading bytes of the nonce expecting the new value to be a
2855 new unique nonce then such an application could inadvertently encrypt
2856 messages with a reused nonce.
2858 Additionally the ignored bytes in a long nonce are not covered by the
2859 integrity guarantee of this cipher. Any application that relies on the
2860 integrity of these ignored leading bytes of a long nonce may be further
2861 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2862 is safe because no such use sets such a long nonce value. However user
2863 applications that use this cipher directly and set a non-default nonce
2864 length to be longer than 12 bytes may be vulnerable.
2866 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2872 * Add DEVRANDOM_WAIT feature for Linux systems
2874 On older Linux systems where the getrandom() system call is not available,
2875 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2876 Contrary to getrandom(), the /dev/urandom device will not block during
2877 early boot when the kernel CSPRNG has not been seeded yet.
2879 To mitigate this known weakness, use select() to wait for /dev/random to
2880 become readable before reading from /dev/urandom.
2882 * Ensure that SM2 only uses SM3 as digest algorithm
2886 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2888 * Change the info callback signals for the start and end of a post-handshake
2889 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2890 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2891 confused by this and assume that a TLSv1.2 renegotiation has started. This
2892 can break KeyUpdate handling. Instead we no longer signal the start and end
2893 of a post handshake message exchange (although the messages themselves are
2894 still signalled). This could break some applications that were expecting
2895 the old signals. However without this KeyUpdate is not usable for many
2900 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2902 * Timing vulnerability in DSA signature generation
2904 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2905 timing side channel attack. An attacker could use variations in the signing
2906 algorithm to recover the private key.
2908 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2913 * Timing vulnerability in ECDSA signature generation
2915 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2916 timing side channel attack. An attacker could use variations in the signing
2917 algorithm to recover the private key.
2919 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2924 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2925 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2926 of two gigabytes and the error handling improved.
2928 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2929 categorized as a normal bug, not a security issue, because the DRBG reseeds
2930 automatically and is fully functional even without additional randomness
2931 provided by the application.
2933 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2935 * Add a new ClientHello callback. Provides a callback interface that gives
2936 the application the ability to adjust the nascent SSL object at the
2937 earliest stage of ClientHello processing, immediately after extensions have
2938 been collected but before they have been processed. In particular, this
2939 callback can adjust the supported TLS versions in response to the contents
2944 * Add SM2 base algorithm support.
2948 * s390x assembly pack: add (improved) hardware-support for the following
2949 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2950 aes-cfb/cfb8, aes-ecb.
2954 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2955 parameter is no longer accepted, as it leads to a corrupt table. NULL
2956 pem_str is reserved for alias entries only.
2960 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2961 step for prime curves. The new implementation is based on formulae from
2962 differential addition-and-doubling in homogeneous projective coordinates
2963 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2964 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2965 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2966 to work in projective coordinates.
2968 *Billy Bob Brumley, Nicola Tuveri*
2970 * Change generating and checking of primes so that the error rate of not
2971 being prime depends on the intended use based on the size of the input.
2972 For larger primes this will result in more rounds of Miller-Rabin.
2973 The maximal error rate for primes with more than 1080 bits is lowered
2976 *Kurt Roeckx, Annie Yousar*
2978 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2982 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2983 moving between systems, and to avoid confusion when a Windows build is
2984 done with mingw vs with MSVC. For POSIX installs, there's still a
2985 symlink or copy named 'tsget' to avoid that confusion as well.
2989 * Revert blinding in ECDSA sign and instead make problematic addition
2990 length-invariant. Switch even to fixed-length Montgomery multiplication.
2994 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2995 step for binary curves. The new implementation is based on formulae from
2996 differential addition-and-doubling in mixed Lopez-Dahab projective
2997 coordinates, modified to independently blind the operands.
2999 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3001 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3002 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3003 EC_METHODs to implement their own specialized "ladder step", to take
3004 advantage of more favorable coordinate systems or more efficient
3005 differential addition-and-doubling algorithms.
3007 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3009 * Modified the random device based seed sources to keep the relevant
3010 file descriptors open rather than reopening them on each access.
3011 This allows such sources to operate in a chroot() jail without
3012 the associated device nodes being available. This behaviour can be
3013 controlled using RAND_keep_random_devices_open().
3017 * Numerous side-channel attack mitigations have been applied. This may have
3018 performance impacts for some algorithms for the benefit of improved
3019 security. Specific changes are noted in this change log by their respective
3024 * AIX shared library support overhaul. Switch to AIX "natural" way of
3025 handling shared libraries, which means collecting shared objects of
3026 different versions and bitnesses in one common archive. This allows to
3027 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3028 doesn't affect the way 3rd party applications are linked, only how
3029 multi-version installation is managed.
3033 * Make ec_group_do_inverse_ord() more robust and available to other
3034 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3035 mitigations are applied to the fallback BN_mod_inverse().
3036 When using this function rather than BN_mod_inverse() directly, new
3037 EC cryptosystem implementations are then safer-by-default.
3041 * Add coordinate blinding for EC_POINT and implement projective
3042 coordinate blinding for generic prime curves as a countermeasure to
3043 chosen point SCA attacks.
3045 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3047 * Add blinding to ECDSA and DSA signatures to protect against side channel
3048 attacks discovered by Keegan Ryan (NCC Group).
3052 * Enforce checking in the `pkeyutl` command to ensure that the input
3053 length does not exceed the maximum supported digest length when performing
3054 a sign, verify or verifyrecover operation.
3058 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3059 I/O in combination with something like select() or poll() will hang. This
3060 can be turned off again using SSL_CTX_clear_mode().
3061 Many applications do not properly handle non-application data records, and
3062 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3063 around the problems in those applications, but can also break some.
3064 It's recommended to read the manpages about SSL_read(), SSL_write(),
3065 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3066 SSL_CTX_set_read_ahead() again.
3070 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3071 now allow empty (zero character) pass phrases.
3075 * Apply blinding to binary field modular inversion and remove patent
3076 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3080 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3081 binary and prime elliptic curves.
3085 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3086 constant time fixed point multiplication.
3090 * Revise elliptic curve scalar multiplication with timing attack
3091 defenses: ec_wNAF_mul redirects to a constant time implementation
3092 when computing fixed point and variable point multiplication (which
3093 in OpenSSL are mostly used with secret scalars in keygen, sign,
3094 ECDH derive operations).
3095 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3098 * Updated CONTRIBUTING
3102 * Updated DRBG / RAND to request nonce and additional low entropy
3103 randomness from the system.
3105 *Matthias St. Pierre*
3107 * Updated 'openssl rehash' to use OpenSSL consistent default.
3111 * Moved the load of the ssl_conf module to libcrypto, which helps
3112 loading engines that libssl uses before libssl is initialised.
3116 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3120 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3122 *Ingo Schwarze, Rich Salz*
3124 * Added output of accepting IP address and port for 'openssl s_server'
3128 * Added a new API for TLSv1.3 ciphersuites:
3129 SSL_CTX_set_ciphersuites()
3130 SSL_set_ciphersuites()
3134 * Memory allocation failures consistently add an error to the error
3139 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3140 in libcrypto when run as setuid/setgid.
3144 * Load any config file by default when libssl is used.
3148 * Added new public header file <openssl/rand_drbg.h> and documentation
3149 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3151 *Matthias St. Pierre*
3153 * QNX support removed (cannot find contributors to get their approval
3154 for the license change).
3158 * TLSv1.3 replay protection for early data has been implemented. See the
3159 SSL_read_early_data() man page for further details.
3163 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3164 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3165 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3166 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3167 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3168 configuration has been separated out. See the ciphers man page or the
3169 SSL_CTX_set_ciphersuites() man page for more information.
3173 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3174 in responder mode now supports the new "-multi" option, which
3175 spawns the specified number of child processes to handle OCSP
3176 requests. The "-timeout" option now also limits the OCSP
3177 responder's patience to wait to receive the full client request
3178 on a newly accepted connection. Child processes are respawned
3179 as needed, and the CA index file is automatically reloaded
3180 when changed. This makes it possible to run the "ocsp" responder
3181 as a long-running service, making the OpenSSL CA somewhat more
3182 feature-complete. In this mode, most diagnostic messages logged
3183 after entering the event loop are logged via syslog(3) rather than
3188 * Added support for X448 and Ed448. Heavily based on original work by
3193 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3194 objects loaded. This adds the functions OSSL_STORE_expect() and
3195 OSSL_STORE_find() as well as needed tools to construct searches and
3196 get the search data out of them.
3200 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3201 version of OpenSSL should review their configuration settings to ensure
3202 that they are still appropriate for TLSv1.3. For further information see:
3203 <https://wiki.openssl.org/index.php/TLS1.3>
3207 * Grand redesign of the OpenSSL random generator
3209 The default RAND method now utilizes an AES-CTR DRBG according to
3210 NIST standard SP 800-90Ar1. The new random generator is essentially
3211 a port of the default random generator from the OpenSSL FIPS 2.0
3212 object module. It is a hybrid deterministic random bit generator
3213 using an AES-CTR bit stream and which seeds and reseeds itself
3214 automatically using trusted system entropy sources.
3216 Some of its new features are:
3217 - Support for multiple DRBG instances with seed chaining.
3218 - The default RAND method makes use of a DRBG.
3219 - There is a public and private DRBG instance.
3220 - The DRBG instances are fork-safe.
3221 - Keep all global DRBG instances on the secure heap if it is enabled.
3222 - The public and private DRBG instance are per thread for lock free
3225 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3227 * Changed Configure so it only says what it does and doesn't dump
3228 so much data. Instead, ./configdata.pm should be used as a script
3229 to display all sorts of configuration data.
3233 * Added processing of "make variables" to Configure.
3237 * Added SHA512/224 and SHA512/256 algorithm support.
3241 * The last traces of Netware support, first removed in 1.1.0, have
3246 * Get rid of Makefile.shared, and in the process, make the processing
3247 of certain files (rc.obj, or the .def/.map/.opt files produced from
3248 the ordinal files) more visible and hopefully easier to trace and
3249 debug (or make silent).
3253 * Make it possible to have environment variable assignments as
3254 arguments to config / Configure.
3258 * Add multi-prime RSA (RFC 8017) support.
3262 * Add SM3 implemented according to GB/T 32905-2016
3263 *Jack Lloyd <jack.lloyd@ribose.com>,*
3264 *Ronald Tse <ronald.tse@ribose.com>,*
3265 *Erick Borsboom <erick.borsboom@ribose.com>*
3267 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3268 as documented in RFC6066.
3269 Based on a patch from Tomasz Moń
3271 *Filipe Raimundo da Silva*
3273 * Add SM4 implemented according to GB/T 32907-2016.
3274 *Jack Lloyd <jack.lloyd@ribose.com>,*
3275 *Ronald Tse <ronald.tse@ribose.com>,*
3276 *Erick Borsboom <erick.borsboom@ribose.com>*
3278 * Reimplement -newreq-nodes and ERR_error_string_n; the
3279 original author does not agree with the license change.
3283 * Add ARIA AEAD TLS support.
3287 * Some macro definitions to support VS6 have been removed. Visual
3288 Studio 6 has not worked since 1.1.0
3292 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3293 without clearing the errors.
3297 * Add "atfork" functions. If building on a system that without
3298 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3299 requirements. The RAND facility now uses/requires this.
3307 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3308 not possible to disable entirely. However, it's still possible to
3309 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3312 To disable, configure with 'no-ui-console'. 'no-ui' is still
3313 possible to use as an alias. Check at compile time with the
3314 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3315 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3319 * Add a STORE module, which implements a uniform and URI based reader of
3320 stores that can contain keys, certificates, CRLs and numerous other
3321 objects. The main API is loosely based on a few stdio functions,
3322 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3323 OSSL_STORE_error and OSSL_STORE_close.
3324 The implementation uses backends called "loaders" to implement arbitrary
3325 URI schemes. There is one built in "loader" for the 'file' scheme.
3329 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3330 then adjusted to work on FreeBSD 8.4 as well.
3331 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3332 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3336 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3337 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3338 error code calls like this:
3340 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3342 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3343 that can be encoded in C. For the foreseeable future, this will only
3346 *Richard Levitte and Tim Hudson*
3348 * Removed BSD cryptodev engine.
3352 * Add a build target 'build_all_generated', to build all generated files
3353 and only that. This can be used to prepare everything that requires
3354 things like perl for a system that lacks perl and then move everything
3355 to that system and do the rest of the build there.
3359 * In the UI interface, make it possible to duplicate the user data. This
3360 can be used by engines that need to retain the data for a longer time
3361 than just the call where this user data is passed.
3365 * Ignore the '-named_curve auto' value for compatibility of applications
3368 *Tomáš Mráz <tmraz@fedoraproject.org>*
3370 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3371 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3372 alerts across multiple records (some of which could be empty). In practice
3373 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3374 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3375 support this at all. Supporting it adds significant complexity to the
3376 record layer, and its removal is unlikely to cause interoperability
3381 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3382 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3383 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3388 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3389 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3391 *Richard Levitte, Andy Polyakov*
3393 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3398 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3399 platform rather than 'mingw'.
3403 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3404 success if they are asked to add an object which already exists
3405 in the store. This change cascades to other functions which load
3406 certificates and CRLs.
3410 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3411 facilitate stack unwinding even from assembly subroutines.
3415 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3416 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3420 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3421 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3422 which is the minimum version we support.
3426 * Certificate time validation (X509_cmp_time) enforces stricter
3427 compliance with RFC 5280. Fractional seconds and timezone offsets
3428 are no longer allowed.
3432 * Add support for ARIA
3436 * s_client will now send the Server Name Indication (SNI) extension by
3437 default unless the new "-noservername" option is used. The server name is
3438 based on the host provided to the "-connect" option unless overridden by
3439 using "-servername".
3443 * Add support for SipHash
3447 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3448 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3449 prevent issues where no progress is being made and the peer continually
3450 sends unrecognised record types, using up resources processing them.
3454 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3455 using the algorithm defined in
3456 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3460 * Heartbeat support has been removed; the ABI is changed for now.
3462 *Richard Levitte, Rich Salz*
3464 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3468 * The RSA "null" method, which was partially supported to avoid patent
3469 issues, has been replaced to always returns NULL.
3476 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3478 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3479 used even when parsing explicit parameters, when loading a encoded key
3480 or calling `EC_GROUP_new_from_ecpkparameters()`/
3481 `EC_GROUP_new_from_ecparameters()`.
3482 This prevents bypass of security hardening and performance gains,
3483 especially for curves with specialized EC_METHODs.
3484 By default, if a key encoded with explicit parameters is loaded and later
3485 encoded, the output is still encoded with explicit parameters, even if
3486 internally a "named" EC_GROUP is used for computation.
3490 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3491 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3492 NULL. After this change, only the cofactor parameter can be NULL. It also
3493 does some minimal sanity checks on the passed order.
3498 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3499 An attack is simple, if the first CMS_recipientInfo is valid but the
3500 second CMS_recipientInfo is chosen ciphertext. If the second
3501 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3502 encryption key will be replaced by garbage, and the message cannot be
3503 decoded, but if the RSA decryption fails, the correct encryption key is
3504 used and the recipient will not notice the attack.
3505 As a work around for this potential attack the length of the decrypted
3506 key must be equal to the cipher default key length, in case the
3507 certificate is not given and all recipientInfo are tried out.
3508 The old behaviour can be re-enabled in the CMS code by setting the
3509 CMS_DEBUG_DECRYPT flag.
3514 * Use Windows installation paths in the mingw builds
3516 Mingw isn't a POSIX environment per se, which means that Windows
3517 paths should be used for installation.
3522 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3524 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3525 This changes the size when using the `genpkey` command when no size is given.
3526 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3527 generation commands to use 2048 bits by default.
3531 * Prevent over long nonces in ChaCha20-Poly1305.
3533 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3534 for every encryption operation. RFC 7539 specifies that the nonce value
3535 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3536 and front pads the nonce with 0 bytes if it is less than 12
3537 bytes. However it also incorrectly allows a nonce to be set of up to 16
3538 bytes. In this case only the last 12 bytes are significant and any
3539 additional leading bytes are ignored.
3541 It is a requirement of using this cipher that nonce values are
3542 unique. Messages encrypted using a reused nonce value are susceptible to
3543 serious confidentiality and integrity attacks. If an application changes
3544 the default nonce length to be longer than 12 bytes and then makes a
3545 change to the leading bytes of the nonce expecting the new value to be a
3546 new unique nonce then such an application could inadvertently encrypt
3547 messages with a reused nonce.
3549 Additionally the ignored bytes in a long nonce are not covered by the
3550 integrity guarantee of this cipher. Any application that relies on the
3551 integrity of these ignored leading bytes of a long nonce may be further
3552 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3553 is safe because no such use sets such a long nonce value. However user
3554 applications that use this cipher directly and set a non-default nonce
3555 length to be longer than 12 bytes may be vulnerable.
3557 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3563 * Added SCA hardening for modular field inversion in EC_GROUP through
3564 a new dedicated field_inv() pointer in EC_METHOD.
3565 This also addresses a leakage affecting conversions from projective
3566 to affine coordinates.
3568 *Billy Bob Brumley, Nicola Tuveri*
3570 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3571 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3575 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3579 * Remove the 'dist' target and add a tarball building script. The
3580 'dist' target has fallen out of use, and it shouldn't be
3581 necessary to configure just to create a source distribution.
3585 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3587 * Timing vulnerability in DSA signature generation
3589 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3590 timing side channel attack. An attacker could use variations in the signing
3591 algorithm to recover the private key.
3593 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3598 * Timing vulnerability in ECDSA signature generation
3600 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3601 timing side channel attack. An attacker could use variations in the signing
3602 algorithm to recover the private key.
3604 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3609 * Add coordinate blinding for EC_POINT and implement projective
3610 coordinate blinding for generic prime curves as a countermeasure to
3611 chosen point SCA attacks.
3613 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3615 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3617 * Client DoS due to large DH parameter
3619 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3620 malicious server can send a very large prime value to the client. This will
3621 cause the client to spend an unreasonably long period of time generating a
3622 key for this prime resulting in a hang until the client has finished. This
3623 could be exploited in a Denial Of Service attack.
3625 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3630 * Cache timing vulnerability in RSA Key Generation
3632 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3633 a cache timing side channel attack. An attacker with sufficient access to
3634 mount cache timing attacks during the RSA key generation process could
3635 recover the private key.
3637 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3638 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3643 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3644 parameter is no longer accepted, as it leads to a corrupt table. NULL
3645 pem_str is reserved for alias entries only.
3649 * Revert blinding in ECDSA sign and instead make problematic addition
3650 length-invariant. Switch even to fixed-length Montgomery multiplication.
3654 * Change generating and checking of primes so that the error rate of not
3655 being prime depends on the intended use based on the size of the input.
3656 For larger primes this will result in more rounds of Miller-Rabin.
3657 The maximal error rate for primes with more than 1080 bits is lowered
3660 *Kurt Roeckx, Annie Yousar*
3662 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3666 * Add blinding to ECDSA and DSA signatures to protect against side channel
3667 attacks discovered by Keegan Ryan (NCC Group).
3671 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3672 now allow empty (zero character) pass phrases.
3676 * Certificate time validation (X509_cmp_time) enforces stricter
3677 compliance with RFC 5280. Fractional seconds and timezone offsets
3678 are no longer allowed.
3682 * Fixed a text canonicalisation bug in CMS
3684 Where a CMS detached signature is used with text content the text goes
3685 through a canonicalisation process first prior to signing or verifying a
3686 signature. This process strips trailing space at the end of lines, converts
3687 line terminators to CRLF and removes additional trailing line terminators
3688 at the end of a file. A bug in the canonicalisation process meant that
3689 some characters, such as form-feed, were incorrectly treated as whitespace
3690 and removed. This is contrary to the specification (RFC5485). This fix
3691 could mean that detached text data signed with an earlier version of
3692 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3693 signed with a fixed OpenSSL may fail to verify with an earlier version of
3694 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3695 and use the "-binary" flag (for the "cms" command line application) or set
3696 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3700 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3702 * Constructed ASN.1 types with a recursive definition could exceed the stack
3704 Constructed ASN.1 types with a recursive definition (such as can be found
3705 in PKCS7) could eventually exceed the stack given malicious input with
3706 excessive recursion. This could result in a Denial Of Service attack. There
3707 are no such structures used within SSL/TLS that come from untrusted sources
3708 so this is considered safe.
3710 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3716 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3718 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3719 effectively reduced to only comparing the least significant bit of each
3720 byte. This allows an attacker to forge messages that would be considered as
3721 authenticated in an amount of tries lower than that guaranteed by the
3722 security claims of the scheme. The module can only be compiled by the
3723 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3725 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3731 * Add a build target 'build_all_generated', to build all generated files
3732 and only that. This can be used to prepare everything that requires
3733 things like perl for a system that lacks perl and then move everything
3734 to that system and do the rest of the build there.
3738 * Backport SSL_OP_NO_RENGOTIATION
3740 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3741 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3742 changes this is no longer possible in 1.1.0. Therefore, the new
3743 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3744 1.1.0 to provide equivalent functionality.
3746 Note that if an application built against 1.1.0h headers (or above) is run
3747 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3748 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3752 * Removed the OS390-Unix config target. It relied on a script that doesn't
3757 * rsaz_1024_mul_avx2 overflow bug on x86_64
3759 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3760 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3761 Analysis suggests that attacks against RSA and DSA as a result of this
3762 defect would be very difficult to perform and are not believed likely.
3763 Attacks against DH1024 are considered just feasible, because most of the
3764 work necessary to deduce information about a private key may be performed
3765 offline. The amount of resources required for such an attack would be
3766 significant. However, for an attack on TLS to be meaningful, the server
3767 would have to share the DH1024 private key among multiple clients, which is
3768 no longer an option since CVE-2016-0701.
3770 This only affects processors that support the AVX2 but not ADX extensions
3771 like Intel Haswell (4th generation).
3773 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3774 was originally found via the OSS-Fuzz project.
3779 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3781 * bn_sqrx8x_internal carry bug on x86_64
3783 There is a carry propagating bug in the x86_64 Montgomery squaring
3784 procedure. No EC algorithms are affected. Analysis suggests that attacks
3785 against RSA and DSA as a result of this defect would be very difficult to
3786 perform and are not believed likely. Attacks against DH are considered just
3787 feasible (although very difficult) because most of the work necessary to
3788 deduce information about a private key may be performed offline. The amount
3789 of resources required for such an attack would be very significant and
3790 likely only accessible to a limited number of attackers. An attacker would
3791 additionally need online access to an unpatched system using the target
3792 private key in a scenario with persistent DH parameters and a private
3793 key that is shared between multiple clients.
3795 This only affects processors that support the BMI1, BMI2 and ADX extensions
3796 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3798 This issue was reported to OpenSSL by the OSS-Fuzz project.
3803 * Malformed X.509 IPAddressFamily could cause OOB read
3805 If an X.509 certificate has a malformed IPAddressFamily extension,
3806 OpenSSL could do a one-byte buffer overread. The most likely result
3807 would be an erroneous display of the certificate in text format.
3809 This issue was reported to OpenSSL by the OSS-Fuzz project.
3814 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3816 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3817 platform rather than 'mingw'.
3821 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3822 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3823 which is the minimum version we support.
3827 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3829 * Encrypt-Then-Mac renegotiation crash
3831 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3832 negotiated where it was not in the original handshake (or vice-versa) then
3833 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3834 and servers are affected.
3836 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3841 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3843 * Truncated packet could crash via OOB read
3845 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3846 cipher is being used, then a truncated packet can cause that host to
3847 perform an out-of-bounds read, usually resulting in a crash.
3849 This issue was reported to OpenSSL by Robert Święcki of Google.
3854 * Bad (EC)DHE parameters cause a client crash
3856 If a malicious server supplies bad parameters for a DHE or ECDHE key
3857 exchange then this can result in the client attempting to dereference a
3858 NULL pointer leading to a client crash. This could be exploited in a Denial
3861 This issue was reported to OpenSSL by Guido Vranken.
3866 * BN_mod_exp may produce incorrect results on x86_64
3868 There is a carry propagating bug in the x86_64 Montgomery squaring
3869 procedure. No EC algorithms are affected. Analysis suggests that attacks
3870 against RSA and DSA as a result of this defect would be very difficult to
3871 perform and are not believed likely. Attacks against DH are considered just
3872 feasible (although very difficult) because most of the work necessary to
3873 deduce information about a private key may be performed offline. The amount
3874 of resources required for such an attack would be very significant and
3875 likely only accessible to a limited number of attackers. An attacker would
3876 additionally need online access to an unpatched system using the target
3877 private key in a scenario with persistent DH parameters and a private
3878 key that is shared between multiple clients. For example this can occur by
3879 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3880 similar to CVE-2015-3193 but must be treated as a separate problem.
3882 This issue was reported to OpenSSL by the OSS-Fuzz project.
3887 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3889 * ChaCha20/Poly1305 heap-buffer-overflow
3891 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3892 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3893 crash. This issue is not considered to be exploitable beyond a DoS.
3895 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3900 * CMS Null dereference
3902 Applications parsing invalid CMS structures can crash with a NULL pointer
3903 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3904 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3905 structure callback if an attempt is made to free certain invalid encodings.
3906 Only CHOICE structures using a callback which do not handle NULL value are
3909 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3914 * Montgomery multiplication may produce incorrect results
3916 There is a carry propagating bug in the Broadwell-specific Montgomery
3917 multiplication procedure that handles input lengths divisible by, but
3918 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3919 and DH private keys are impossible. This is because the subroutine in
3920 question is not used in operations with the private key itself and an input
3921 of the attacker's direct choice. Otherwise the bug can manifest itself as
3922 transient authentication and key negotiation failures or reproducible
3923 erroneous outcome of public-key operations with specially crafted input.
3924 Among EC algorithms only Brainpool P-512 curves are affected and one
3925 presumably can attack ECDH key negotiation. Impact was not analyzed in
3926 detail, because pre-requisites for attack are considered unlikely. Namely
3927 multiple clients have to choose the curve in question and the server has to
3928 share the private key among them, neither of which is default behaviour.
3929 Even then only clients that chose the curve will be affected.
3931 This issue was publicly reported as transient failures and was not
3932 initially recognized as a security issue. Thanks to Richard Morgan for
3933 providing reproducible case.
3938 * Removed automatic addition of RPATH in shared libraries and executables,
3939 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3943 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3945 * Fix Use After Free for large message sizes
3947 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3948 message larger than approx 16k is received then the underlying buffer to
3949 store the incoming message is reallocated and moved. Unfortunately a
3950 dangling pointer to the old location is left which results in an attempt to
3951 write to the previously freed location. This is likely to result in a
3952 crash, however it could potentially lead to execution of arbitrary code.
3954 This issue only affects OpenSSL 1.1.0a.
3956 This issue was reported to OpenSSL by Robert Święcki.
3961 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3963 * OCSP Status Request extension unbounded memory growth
3965 A malicious client can send an excessively large OCSP Status Request
3966 extension. If that client continually requests renegotiation, sending a
3967 large OCSP Status Request extension each time, then there will be unbounded
3968 memory growth on the server. This will eventually lead to a Denial Of
3969 Service attack through memory exhaustion. Servers with a default
3970 configuration are vulnerable even if they do not support OCSP. Builds using
3971 the "no-ocsp" build time option are not affected.
3973 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3978 * SSL_peek() hang on empty record
3980 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3981 sends an empty record. This could be exploited by a malicious peer in a
3982 Denial Of Service attack.
3984 This issue was reported to OpenSSL by Alex Gaynor.
3989 * Excessive allocation of memory in tls_get_message_header() and
3990 dtls1_preprocess_fragment()
3992 A (D)TLS message includes 3 bytes for its length in the header for the
3993 message. This would allow for messages up to 16Mb in length. Messages of
3994 this length are excessive and OpenSSL includes a check to ensure that a
3995 peer is sending reasonably sized messages in order to avoid too much memory
3996 being consumed to service a connection. A flaw in the logic of version
3997 1.1.0 means that memory for the message is allocated too early, prior to
3998 the excessive message length check. Due to way memory is allocated in
3999 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4000 to service a connection. This could lead to a Denial of Service through
4001 memory exhaustion. However, the excessive message length check still takes
4002 place, and this would cause the connection to immediately fail. Assuming
4003 that the application calls SSL_free() on the failed connection in a timely
4004 manner then the 21Mb of allocated memory will then be immediately freed
4005 again. Therefore, the excessive memory allocation will be transitory in
4006 nature. This then means that there is only a security impact if:
4008 1) The application does not call SSL_free() in a timely manner in the event
4009 that the connection fails
4011 2) The application is working in a constrained environment where there is
4012 very little free memory
4014 3) The attacker initiates multiple connection attempts such that there are
4015 multiple connections in a state where memory has been allocated for the
4016 connection; SSL_free() has not yet been called; and there is insufficient
4017 memory to service the multiple requests.
4019 Except in the instance of (1) above any Denial Of Service is likely to be
4020 transitory because as soon as the connection fails the memory is
4021 subsequently freed again in the SSL_free() call. However there is an
4022 increased risk during this period of application crashes due to the lack of
4023 memory - which would then mean a more serious Denial of Service.
4025 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4026 (CVE-2016-6307 and CVE-2016-6308)
4030 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4031 had to be removed. Primary reason is that vendor assembler can't
4032 assemble our modules with -KPIC flag. As result it, assembly
4033 support, was not even available as option. But its lack means
4034 lack of side-channel resistant code, which is incompatible with
4035 security by todays standards. Fortunately gcc is readily available
4036 prepackaged option, which we firmly point at...
4040 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4042 * Windows command-line tool supports UTF-8 opt-in option for arguments
4043 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4044 (to any value) allows Windows user to access PKCS#12 file generated
4045 with Windows CryptoAPI and protected with non-ASCII password, as well
4046 as files generated under UTF-8 locale on Linux also protected with
4051 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4052 have been disabled by default and removed from DEFAULT, just like RC4.
4053 See the RC4 item below to re-enable both.
4057 * The method for finding the storage location for the Windows RAND seed file
4058 has changed. First we check %RANDFILE%. If that is not set then we check
4059 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4060 all else fails we fall back to C:\.
4064 * The EVP_EncryptUpdate() function has had its return type changed from void
4065 to int. A return of 0 indicates and error while a return of 1 indicates
4070 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4071 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4072 off the constant time implementation for RSA, DSA and DH have been made
4073 no-ops and deprecated.
4077 * Windows RAND implementation was simplified to only get entropy by
4078 calling CryptGenRandom(). Various other RAND-related tickets
4081 *Joseph Wylie Yandle, Rich Salz*
4083 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4084 and `OPENSSL_LH_`, respectively. The old names are available
4085 with API compatibility. They new names are now completely documented.
4089 * Unify TYPE_up_ref(obj) methods signature.
4090 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4091 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4092 int (instead of void) like all others TYPE_up_ref() methods.
4093 So now these methods also check the return value of CRYPTO_atomic_add(),
4094 and the validity of object reference counter.
4096 *fdasilvayy@gmail.com*
4098 * With Windows Visual Studio builds, the .pdb files are installed
4099 alongside the installed libraries and executables. For a static
4100 library installation, ossl_static.pdb is the associate compiler
4101 generated .pdb file to be used when linking programs.
4105 * Remove openssl.spec. Packaging files belong with the packagers.
4109 * Automatic Darwin/OSX configuration has had a refresh, it will now
4110 recognise x86_64 architectures automatically. You can still decide
4111 to build for a different bitness with the environment variable
4112 KERNEL_BITS (can be 32 or 64), for example:
4114 KERNEL_BITS=32 ./config
4118 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4119 256 bit AES and HMAC with SHA256.
4123 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4127 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4131 * To enable users to have their own config files and build file templates,
4132 Configure looks in the directory indicated by the environment variable
4133 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4134 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4135 name and is used as is.
4139 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4140 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4141 X509_CERT_FILE_CTX was removed.
4145 * "shared" builds are now the default. To create only static libraries use
4146 the "no-shared" Configure option.
4150 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4151 All of these option have not worked for some while and are fundamental
4156 * Make various cleanup routines no-ops and mark them as deprecated. Most
4157 global cleanup functions are no longer required because they are handled
4158 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4159 Explicitly de-initing can cause problems (e.g. where a library that uses
4160 OpenSSL de-inits, but an application is still using it). The affected
4161 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4162 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4163 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4164 COMP_zlib_cleanup().
4168 * --strict-warnings no longer enables runtime debugging options
4169 such as REF_DEBUG. Instead, debug options are automatically
4170 enabled with '--debug' builds.
4172 *Andy Polyakov, Emilia Käsper*
4174 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4175 have been moved out of the public header files. New functions for managing
4176 these have been added.
4180 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4181 objects have been moved out of the public header files. New
4182 functions for managing these have been added.
4186 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4187 have been moved out of the public header files. New functions for managing
4188 these have been added.
4192 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4193 moved out of the public header files. New functions for managing these
4198 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4202 * Removed the mk1mf build scripts.
4206 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4207 it is always safe to #include a header now.
4211 * Removed the aged BC-32 config and all its supporting scripts
4215 * Removed support for Ultrix, Netware, and OS/2.
4219 * Add support for HKDF.
4221 *Alessandro Ghedini*
4223 * Add support for blake2b and blake2s
4227 * Added support for "pipelining". Ciphers that have the
4228 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4229 encryptions/decryptions simultaneously. There are currently no built-in
4230 ciphers with this property but the expectation is that engines will be able
4231 to offer it to significantly improve throughput. Support has been extended
4232 into libssl so that multiple records for a single connection can be
4233 processed in one go (for >=TLS 1.1).
4237 * Added the AFALG engine. This is an async capable engine which is able to
4238 offload work to the Linux kernel. In this initial version it only supports
4239 AES128-CBC. The kernel must be version 4.1.0 or greater.
4243 * OpenSSL now uses a new threading API. It is no longer necessary to
4244 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4245 are two supported threading models: pthreads and windows threads. It is
4246 also possible to configure OpenSSL at compile time for "no-threads". The
4247 old threading API should no longer be used. The functions have been
4248 replaced with "no-op" compatibility macros.
4250 *Alessandro Ghedini, Matt Caswell*
4252 * Modify behavior of ALPN to invoke callback after SNI/servername
4253 callback, such that updates to the SSL_CTX affect ALPN.
4257 * Add SSL_CIPHER queries for authentication and key-exchange.
4261 * Changes to the DEFAULT cipherlist:
4262 - Prefer (EC)DHE handshakes over plain RSA.
4263 - Prefer AEAD ciphers over legacy ciphers.
4264 - Prefer ECDSA over RSA when both certificates are available.
4265 - Prefer TLSv1.2 ciphers/PRF.
4266 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4271 * Change the ECC default curve list to be this, in order: x25519,
4272 secp256r1, secp521r1, secp384r1.
4276 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4277 disabled by default. They can be re-enabled using the
4278 enable-weak-ssl-ciphers option to Configure.
4282 * If the server has ALPN configured, but supports no protocols that the
4283 client advertises, send a fatal "no_application_protocol" alert.
4284 This behaviour is SHALL in RFC 7301, though it isn't universally
4285 implemented by other servers.
4289 * Add X25519 support.
4290 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4291 for public and private key encoding using the format documented in
4292 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4293 key generation and key derivation.
4295 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4300 * Deprecate SRP_VBASE_get_by_user.
4301 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4302 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4303 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4304 seed, even if the seed is configured.
4306 Users should use SRP_VBASE_get1_by_user instead. Note that in
4307 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4308 also that even though configuring the SRP seed attempts to hide
4309 invalid usernames by continuing the handshake with fake
4310 credentials, this behaviour is not constant time and no strong
4311 guarantees are made that the handshake is indistinguishable from
4312 that of a valid user.
4316 * Configuration change; it's now possible to build dynamic engines
4317 without having to build shared libraries and vice versa. This
4318 only applies to the engines in `engines/`, those in `crypto/engine/`
4319 will always be built into libcrypto (i.e. "static").
4321 Building dynamic engines is enabled by default; to disable, use
4322 the configuration option "disable-dynamic-engine".
4324 The only requirements for building dynamic engines are the
4325 presence of the DSO module and building with position independent
4326 code, so they will also automatically be disabled if configuring
4327 with "disable-dso" or "disable-pic".
4329 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4330 are also taken away from openssl/opensslconf.h, as they are
4335 * Configuration change; if there is a known flag to compile
4336 position independent code, it will always be applied on the
4337 libcrypto and libssl object files, and never on the application
4338 object files. This means other libraries that use routines from
4339 libcrypto / libssl can be made into shared libraries regardless
4340 of how OpenSSL was configured.
4342 If this isn't desirable, the configuration options "disable-pic"
4343 or "no-pic" can be used to disable the use of PIC. This will
4344 also disable building shared libraries and dynamic engines.
4348 * Removed JPAKE code. It was experimental and has no wide use.
4352 * The INSTALL_PREFIX Makefile variable has been renamed to
4353 DESTDIR. That makes for less confusion on what this variable
4354 is for. Also, the configuration option --install_prefix is
4359 * Heartbeat for TLS has been removed and is disabled by default
4360 for DTLS; configure with enable-heartbeats. Code that uses the
4361 old #define's might need to be updated.
4363 *Emilia Käsper, Rich Salz*
4365 * Rename REF_CHECK to REF_DEBUG.
4369 * New "unified" build system
4371 The "unified" build system is aimed to be a common system for all
4372 platforms we support. With it comes new support for VMS.
4374 This system builds supports building in a different directory tree
4375 than the source tree. It produces one Makefile (for unix family
4376 or lookalikes), or one descrip.mms (for VMS).
4378 The source of information to make the Makefile / descrip.mms is
4379 small files called 'build.info', holding the necessary
4380 information for each directory with source to compile, and a
4381 template in Configurations, like unix-Makefile.tmpl or
4384 With this change, the library names were also renamed on Windows
4385 and on VMS. They now have names that are closer to the standard
4386 on Unix, and include the major version number, and in certain
4387 cases, the architecture they are built for. See "Notes on shared
4388 libraries" in INSTALL.
4390 We rely heavily on the perl module Text::Template.
4394 * Added support for auto-initialisation and de-initialisation of the library.
4395 OpenSSL no longer requires explicit init or deinit routines to be called,
4396 except in certain circumstances. See the OPENSSL_init_crypto() and
4397 OPENSSL_init_ssl() man pages for further information.
4401 * The arguments to the DTLSv1_listen function have changed. Specifically the
4402 "peer" argument is now expected to be a BIO_ADDR object.
4404 * Rewrite of BIO networking library. The BIO library lacked consistent
4405 support of IPv6, and adding it required some more extensive
4406 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4407 which hold all types of addresses and chains of address information.
4408 It also introduces a new API, with functions like BIO_socket,
4409 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4410 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4411 have been adapted accordingly.
4415 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4420 * CRIME protection: disable compression by default, even if OpenSSL is
4421 compiled with zlib enabled. Applications can still enable compression
4422 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4423 using the SSL_CONF library to configure compression.
4427 * The signature of the session callback configured with
4428 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4429 was explicitly marked as `const unsigned char*` instead of
4434 * Always DPURIFY. Remove the use of uninitialized memory in the
4435 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4439 * Removed many obsolete configuration items, including
4440 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4441 MD2_CHAR, MD2_INT, MD2_LONG
4443 IDEA_SHORT, IDEA_LONG
4444 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4446 *Rich Salz, with advice from Andy Polyakov*
4448 * Many BN internals have been moved to an internal header file.
4450 *Rich Salz with help from Andy Polyakov*
4452 * Configuration and writing out the results from it has changed.
4453 Files such as Makefile include/openssl/opensslconf.h and are now
4454 produced through general templates, such as Makefile.in and
4455 crypto/opensslconf.h.in and some help from the perl module
4458 Also, the center of configuration information is no longer
4459 Makefile. Instead, Configure produces a perl module in
4460 configdata.pm which holds most of the config data (in the hash
4461 table %config), the target data that comes from the target
4462 configuration in one of the `Configurations/*.conf` files (in
4467 * To clarify their intended purposes, the Configure options
4468 --prefix and --openssldir change their semantics, and become more
4469 straightforward and less interdependent.
4471 --prefix shall be used exclusively to give the location INSTALLTOP
4472 where programs, scripts, libraries, include files and manuals are
4473 going to be installed. The default is now /usr/local.
4475 --openssldir shall be used exclusively to give the default
4476 location OPENSSLDIR where certificates, private keys, CRLs are
4477 managed. This is also where the default openssl.cnf gets
4479 If the directory given with this option is a relative path, the
4480 values of both the --prefix value and the --openssldir value will
4481 be combined to become OPENSSLDIR.
4482 The default for --openssldir is INSTALLTOP/ssl.
4484 Anyone who uses --openssldir to specify where OpenSSL is to be
4485 installed MUST change to use --prefix instead.
4489 * The GOST engine was out of date and therefore it has been removed. An up
4490 to date GOST engine is now being maintained in an external repository.
4491 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4492 support for GOST ciphersuites (these are only activated if a GOST engine
4497 * EGD is no longer supported by default; use enable-egd when
4500 *Ben Kaduk and Rich Salz*
4502 * The distribution now has Makefile.in files, which are used to
4503 create Makefile's when Configure is run. *Configure must be run
4504 before trying to build now.*
4508 * The return value for SSL_CIPHER_description() for error conditions
4513 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4515 Obtaining and performing DNSSEC validation of TLSA records is
4516 the application's responsibility. The application provides
4517 the TLSA records of its choice to OpenSSL, and these are then
4518 used to authenticate the peer.
4520 The TLSA records need not even come from DNS. They can, for
4521 example, be used to implement local end-entity certificate or
4522 trust-anchor "pinning", where the "pin" data takes the form
4523 of TLSA records, which can augment or replace verification
4524 based on the usual WebPKI public certification authorities.
4528 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4529 continues to support deprecated interfaces in default builds.
4530 However, applications are strongly advised to compile their
4531 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4532 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4533 or the 1.1.0 releases.
4535 In environments in which all applications have been ported to
4536 not use any deprecated interfaces OpenSSL's Configure script
4537 should be used with the --api=1.1.0 option to entirely remove
4538 support for the deprecated features from the library and
4539 unconditionally disable them in the installed headers.
4540 Essentially the same effect can be achieved with the "no-deprecated"
4541 argument to Configure, except that this will always restrict
4542 the build to just the latest API, rather than a fixed API
4545 As applications are ported to future revisions of the API,
4546 they should update their compile-time OPENSSL_API_COMPAT define
4547 accordingly, but in most cases should be able to continue to
4548 compile with later releases.
4550 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4551 0x10000000L and 0x00908000L, respectively. However those
4552 versions did not support the OPENSSL_API_COMPAT feature, and
4553 so applications are not typically tested for explicit support
4554 of just the undeprecated features of either release.
4558 * Add support for setting the minimum and maximum supported protocol.
4559 It can bet set via the SSL_set_min_proto_version() and
4560 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4561 MaxProtocol. It's recommended to use the new APIs to disable
4562 protocols instead of disabling individual protocols using
4563 SSL_set_options() or SSL_CONF's Protocol. This change also
4564 removes support for disabling TLS 1.2 in the OpenSSL TLS
4565 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4569 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4573 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4574 and integrates ECDSA and ECDH functionality into EC. Implementations can
4575 now redirect key generation and no longer need to convert to or from
4578 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4579 include the ec.h header file instead.
4583 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4584 ciphers who are no longer supported and drops support the ephemeral RSA key
4585 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4589 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4590 opaque. For HMAC_CTX, the following constructors and destructors
4593 HMAC_CTX *HMAC_CTX_new(void);
4594 void HMAC_CTX_free(HMAC_CTX *ctx);
4596 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4597 destroy such methods has been added. See EVP_MD_meth_new(3) and
4598 EVP_CIPHER_meth_new(3) for documentation.
4601 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4602 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4603 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4604 an already created structure.
4605 2) For consistency with the majority of our object creators and
4606 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4607 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4608 for deprecated builds.
4612 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4613 cryptographic operations to be performed asynchronously as long as an
4614 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4615 further details. Libssl has also had this capability integrated with the
4616 introduction of the new mode SSL_MODE_ASYNC and associated error
4617 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4618 pages. This work was developed in partnership with Intel Corp.
4622 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4623 always enabled now. If you want to disable the support you should
4624 exclude it using the list of supported ciphers. This also means that the
4625 "-no_ecdhe" option has been removed from s_server.
4629 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4630 SSL_{CTX_}set1_curves() which can set a list.
4634 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4635 curve you want to support using SSL_{CTX_}set1_curves().
4639 * State machine rewrite. The state machine code has been significantly
4640 refactored in order to remove much duplication of code and solve issues
4641 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4642 further details). This change does have some associated API changes.
4643 Notably the SSL_state() function has been removed and replaced by
4644 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4645 SSL_set_state() has been removed altogether. The previous handshake states
4646 defined in ssl.h and ssl3.h have also been removed.
4650 * All instances of the string "ssleay" in the public API were replaced
4651 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4652 Some error codes related to internal RSA_eay API's were renamed.
4656 * The demo files in crypto/threads were moved to demo/threads.
4660 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4663 *Matt Caswell, Rich Salz*
4665 * New ASN.1 embed macro.
4667 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4668 structure is not allocated: it is part of the parent. That is instead of
4676 This reduces memory fragmentation and make it impossible to accidentally
4677 set a mandatory field to NULL.
4679 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4680 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4681 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4686 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4690 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4691 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4692 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4693 DES and RC4 ciphersuites.
4697 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4698 This changes the decoding behaviour for some invalid messages,
4699 though the change is mostly in the more lenient direction, and
4700 legacy behaviour is preserved as much as possible.
4704 * Fix no-stdio build.
4705 *David Woodhouse <David.Woodhouse@intel.com> and also*
4706 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4708 * New testing framework
4709 The testing framework has been largely rewritten and is now using
4710 perl and the perl modules Test::Harness and an extended variant of
4711 Test::More called OpenSSL::Test to do its work. All test scripts in
4712 test/ have been rewritten into test recipes, and all direct calls to
4713 executables in test/Makefile have become individual recipes using the
4714 simplified testing OpenSSL::Test::Simple.
4716 For documentation on our testing modules, do:
4718 perldoc test/testlib/OpenSSL/Test/Simple.pm
4719 perldoc test/testlib/OpenSSL/Test.pm
4723 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4724 are used; the latter aborts on memory leaks (usually checked on exit).
4725 Some undocumented "set malloc, etc., hooks" functions were removed
4726 and others were changed. All are now documented.
4730 * In DSA_generate_parameters_ex, if the provided seed is too short,
4733 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4735 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4736 from RFC4279, RFC4785, RFC5487, RFC5489.
4738 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4739 original RSA_PSK patch.
4743 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4744 era flag was never set throughout the codebase (only read). Also removed
4745 SSL3_FLAGS_POP_BUFFER which was only used if
4746 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4750 * Changed the default name options in the "ca", "crl", "req" and "x509"
4751 to be "oneline" instead of "compat".
4755 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4756 not aware of clients that still exhibit this bug, and the workaround
4757 hasn't been working properly for a while.
4761 * The return type of BIO_number_read() and BIO_number_written() as well as
4762 the corresponding num_read and num_write members in the BIO structure has
4763 changed from unsigned long to uint64_t. On platforms where an unsigned
4764 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4769 * Given the pervasive nature of TLS extensions it is inadvisable to run
4770 OpenSSL without support for them. It also means that maintaining
4771 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4772 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4776 * Removed support for the two export grade static DH ciphersuites
4777 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4778 were newly added (along with a number of other static DH ciphersuites) to
4779 1.0.2. However the two export ones have *never* worked since they were
4780 introduced. It seems strange in any case to be adding new export
4781 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4785 * Version negotiation has been rewritten. In particular SSLv23_method(),
4786 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4787 and turned into macros which simply call the new preferred function names
4788 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4789 should use the new names instead. Also as part of this change the ssl23.h
4790 header file has been removed.
4794 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4795 code and the associated standard is no longer considered fit-for-purpose.
4799 * RT2547 was closed. When generating a private key, try to make the
4800 output file readable only by the owner. This behavior change might
4801 be noticeable when interacting with other software.
4803 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4808 * Added HTTP GET support to the ocsp command.
4812 * Changed default digest for the dgst and enc commands from MD5 to
4817 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4821 * Added support for TLS extended master secret from
4822 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4823 initial patch which was a great help during development.
4827 * All libssl internal structures have been removed from the public header
4828 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4829 now redundant). Users should not attempt to access internal structures
4830 directly. Instead they should use the provided API functions.
4834 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4835 Access to deprecated functions can be re-enabled by running config with
4836 "enable-deprecated". In addition applications wishing to use deprecated
4837 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4838 will, by default, disable some transitive includes that previously existed
4839 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4843 * Added support for OCB mode. OpenSSL has been granted a patent license
4844 compatible with the OpenSSL license for use of OCB. Details are available
4845 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4846 for OCB can be removed by calling config with no-ocb.
4850 * SSLv2 support has been removed. It still supports receiving an SSLv2
4851 compatible client hello.
4855 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4856 done while fixing the error code for the key-too-small case.
4858 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4860 * CA.sh has been removed; use CA.pl instead.
4864 * Removed old DES API.
4868 * Remove various unsupported platforms:
4874 Sinix/ReliantUNIX RM400
4879 16-bit platforms such as WIN16
4883 * Clean up OPENSSL_NO_xxx #define's
4884 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4885 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4886 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4887 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4888 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4889 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4890 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4891 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4892 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4893 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4897 * Cleaned up dead code
4898 Remove all but one '#ifdef undef' which is to be looked at.
4902 * Clean up calling of xxx_free routines.
4903 Just like free(), fix most of the xxx_free routines to accept
4904 NULL. Remove the non-null checks from callers. Save much code.
4908 * Add secure heap for storage of private keys (when possible).
4909 Add BIO_s_secmem(), CBIGNUM, etc.
4910 Contributed by Akamai Technologies under our Corporate CLA.
4914 * Experimental support for a new, fast, unbiased prime candidate generator,
4915 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4917 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4919 * New output format NSS in the sess_id command line tool. This allows
4920 exporting the session id and the master key in NSS keylog format.
4922 *Martin Kaiser <martin@kaiser.cx>*
4924 * Harmonize version and its documentation. -f flag is used to display
4927 *mancha <mancha1@zoho.com>*
4929 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4930 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4932 *mancha <mancha1@zoho.com>*
4934 * Fix some double frees. These are not thought to be exploitable.
4936 *mancha <mancha1@zoho.com>*
4938 * A missing bounds check in the handling of the TLS heartbeat extension
4939 can be used to reveal up to 64k of memory to a connected client or
4942 Thanks for Neel Mehta of Google Security for discovering this bug and to
4943 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4944 preparing the fix ([CVE-2014-0160])
4946 *Adam Langley, Bodo Moeller*
4948 * Fix for the attack described in the paper "Recovering OpenSSL
4949 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4950 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4951 <http://eprint.iacr.org/2014/140>
4953 Thanks to Yuval Yarom and Naomi Benger for discovering this
4954 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4956 *Yuval Yarom and Naomi Benger*
4958 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4959 this fixes a limitation in previous versions of OpenSSL.
4963 * Experimental encrypt-then-mac support.
4965 Experimental support for encrypt then mac from
4966 draft-gutmann-tls-encrypt-then-mac-02.txt
4968 To enable it set the appropriate extension number (0x42 for the test
4969 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4971 For non-compliant peers (i.e. just about everything) this should have no
4974 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4978 * Add EVP support for key wrapping algorithms, to avoid problems with
4979 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4980 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4981 algorithms and include tests cases.
4985 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4990 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4991 MGF1 digest and OAEP label.
4995 * Make openssl verify return errors.
4997 *Chris Palmer <palmer@google.com> and Ben Laurie*
4999 * New function ASN1_TIME_diff to calculate the difference between two
5000 ASN1_TIME structures or one structure and the current time.
5004 * Update fips_test_suite to support multiple command line options. New
5005 test to induce all self test errors in sequence and check expected
5010 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5011 sign or verify all in one operation.
5015 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5016 test programs and fips_test_suite. Includes functionality to parse
5017 the minimal script output of fipsalgest.pl directly.
5021 * Add authorisation parameter to FIPS_module_mode_set().
5025 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5029 * Use separate DRBG fields for internal and external flags. New function
5030 FIPS_drbg_health_check() to perform on demand health checking. Add
5031 generation tests to fips_test_suite with reduced health check interval to
5032 demonstrate periodic health checking. Add "nodh" option to
5033 fips_test_suite to skip very slow DH test.
5037 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5042 * More extensive health check for DRBG checking many more failure modes.
5043 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5044 combination: call this in fips_test_suite.
5048 * Add support for canonical generation of DSA parameter 'g'. See
5051 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5052 POST to handle HMAC cases.
5056 * Add functions FIPS_module_version() and FIPS_module_version_text()
5057 to return numerical and string versions of the FIPS module number.
5061 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5062 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5063 outside the validated module in the FIPS capable OpenSSL.
5067 * Minor change to DRBG entropy callback semantics. In some cases
5068 there is no multiple of the block length between min_len and
5069 max_len. Allow the callback to return more than max_len bytes
5070 of entropy but discard any extra: it is the callback's responsibility
5071 to ensure that the extra data discarded does not impact the
5072 requested amount of entropy.
5076 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5077 information in FIPS186-3, SP800-57 and SP800-131A.
5081 * CCM support via EVP. Interface is very similar to GCM case except we
5082 must supply all data in one chunk (i.e. no update, final) and the
5083 message length must be supplied if AAD is used. Add algorithm test
5088 * Initial version of POST overhaul. Add POST callback to allow the status
5089 of POST to be monitored and/or failures induced. Modify fips_test_suite
5090 to use callback. Always run all selftests even if one fails.
5094 * XTS support including algorithm test driver in the fips_gcmtest program.
5095 Note: this does increase the maximum key length from 32 to 64 bytes but
5096 there should be no binary compatibility issues as existing applications
5097 will never use XTS mode.
5101 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5102 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5103 performs algorithm blocking for unapproved PRNG types. Also do not
5104 set PRNG type in FIPS_mode_set(): leave this to the application.
5105 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5106 the standard OpenSSL PRNG: set additional data to a date time vector.
5110 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5111 This shouldn't present any incompatibility problems because applications
5112 shouldn't be using these directly and any that are will need to rethink
5113 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5117 * Extensive self tests and health checking required by SP800-90 DRBG.
5118 Remove strength parameter from FIPS_drbg_instantiate and always
5119 instantiate at maximum supported strength.
5123 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5127 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5131 * New function DH_compute_key_padded() to compute a DH key and pad with
5132 leading zeroes if needed: this complies with SP800-56A et al.
5136 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5137 anything, incomplete, subject to change and largely untested at present.
5141 * Modify fipscanisteronly build option to only build the necessary object
5142 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5146 * Add experimental option FIPSSYMS to give all symbols in
5147 fipscanister.o and FIPS or fips prefix. This will avoid
5148 conflicts with future versions of OpenSSL. Add perl script
5149 util/fipsas.pl to preprocess assembly language source files
5150 and rename any affected symbols.
5154 * Add selftest checks and algorithm block of non-fips algorithms in
5155 FIPS mode. Remove DES2 from selftests.
5159 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5160 return internal method without any ENGINE dependencies. Add new
5161 tiny fips sign and verify functions.
5165 * New build option no-ec2m to disable characteristic 2 code.
5169 * New build option "fipscanisteronly". This only builds fipscanister.o
5170 and (currently) associated fips utilities. Uses the file Makefile.fips
5171 instead of Makefile.org as the prototype.
5175 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5176 Update fips_gcmtest to use IV generator.
5180 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5181 setting output buffer to NULL. The `*Final` function must be
5182 called although it will not retrieve any additional data. The tag
5183 can be set or retrieved with a ctrl. The IV length is by default 12
5184 bytes (96 bits) but can be set to an alternative value. If the IV
5185 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5190 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5191 underlying do_cipher function handles all cipher semantics itself
5192 including padding and finalisation. This is useful if (for example)
5193 an ENGINE cipher handles block padding itself. The behaviour of
5194 do_cipher is subtly changed if this flag is set: the return value
5195 is the number of characters written to the output buffer (zero is
5196 no longer an error code) or a negative error code. Also if the
5197 input buffer is NULL and length 0 finalisation should be performed.
5201 * If a candidate issuer certificate is already part of the constructed
5202 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5206 * Improve forward-security support: add functions
5208 void SSL_CTX_set_not_resumable_session_callback(
5209 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5210 void SSL_set_not_resumable_session_callback(
5211 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5213 for use by SSL/TLS servers; the callback function will be called whenever a
5214 new session is created, and gets to decide whether the session may be
5215 cached to make it resumable (return 0) or not (return 1). (As by the
5216 SSL/TLS protocol specifications, the session_id sent by the server will be
5217 empty to indicate that the session is not resumable; also, the server will
5218 not generate RFC 4507 (RFC 5077) session tickets.)
5220 A simple reasonable callback implementation is to return is_forward_secure.
5221 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5222 by the SSL/TLS server library, indicating whether it can provide forward
5225 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5227 * New -verify_name option in command line utilities to set verification
5232 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5233 Add CMAC pkey methods.
5237 * Experimental renegotiation in s_server -www mode. If the client
5238 browses /reneg connection is renegotiated. If /renegcert it is
5239 renegotiated requesting a certificate.
5243 * Add an "external" session cache for debugging purposes to s_server. This
5244 should help trace issues which normally are only apparent in deployed
5245 multi-process servers.
5249 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5250 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5251 BIO_set_cipher() and some obscure PEM functions were changed so they
5252 can now return an error. The RAND changes required a change to the
5253 RAND_METHOD structure.
5257 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5258 a gcc attribute to warn if the result of a function is ignored. This
5259 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5260 whose return value is often ignored.
5264 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5265 These allow SCTs (signed certificate timestamps) to be requested and
5266 validated when establishing a connection.
5268 *Rob Percival <robpercival@google.com>*
5273 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5275 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5276 used even when parsing explicit parameters, when loading a encoded key
5277 or calling `EC_GROUP_new_from_ecpkparameters()`/
5278 `EC_GROUP_new_from_ecparameters()`.
5279 This prevents bypass of security hardening and performance gains,
5280 especially for curves with specialized EC_METHODs.
5281 By default, if a key encoded with explicit parameters is loaded and later
5282 encoded, the output is still encoded with explicit parameters, even if
5283 internally a "named" EC_GROUP is used for computation.
5287 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5288 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5289 NULL. After this change, only the cofactor parameter can be NULL. It also
5290 does some minimal sanity checks on the passed order.
5295 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5296 An attack is simple, if the first CMS_recipientInfo is valid but the
5297 second CMS_recipientInfo is chosen ciphertext. If the second
5298 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5299 encryption key will be replaced by garbage, and the message cannot be
5300 decoded, but if the RSA decryption fails, the correct encryption key is
5301 used and the recipient will not notice the attack.
5302 As a work around for this potential attack the length of the decrypted
5303 key must be equal to the cipher default key length, in case the
5304 certificate is not given and all recipientInfo are tried out.
5305 The old behaviour can be re-enabled in the CMS code by setting the
5306 CMS_DEBUG_DECRYPT flag.
5311 * Document issue with installation paths in diverse Windows builds
5313 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5314 binaries and run-time config file.
5319 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5321 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5322 This changes the size when using the `genpkey` command when no size is given.
5323 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5324 generation commands to use 2048 bits by default.
5328 * Add FIPS support for Android Arm 64-bit
5330 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5331 Module in Version 2.0.10. For some reason, the corresponding target
5332 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5333 built with FIPS support on Android Arm 64-bit. This omission has been
5336 *Matthias St. Pierre*
5338 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5340 * 0-byte record padding oracle
5342 If an application encounters a fatal protocol error and then calls
5343 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5344 then OpenSSL can respond differently to the calling application if a 0 byte
5345 record is received with invalid padding compared to if a 0 byte record is
5346 received with an invalid MAC. If the application then behaves differently
5347 based on that in a way that is detectable to the remote peer, then this
5348 amounts to a padding oracle that could be used to decrypt data.
5350 In order for this to be exploitable "non-stitched" ciphersuites must be in
5351 use. Stitched ciphersuites are optimised implementations of certain
5352 commonly used ciphersuites. Also the application must call SSL_shutdown()
5353 twice even if a protocol error has occurred (applications should not do
5354 this but some do anyway).
5356 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5357 Aviram, with additional investigation by Steven Collison and Andrew
5358 Hourselt. It was reported to OpenSSL on 10th December 2018.
5363 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5367 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5369 * Microarchitecture timing vulnerability in ECC scalar multiplication
5371 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5372 shown to be vulnerable to a microarchitecture timing side channel attack.
5373 An attacker with sufficient access to mount local timing attacks during
5374 ECDSA signature generation could recover the private key.
5376 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5377 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5383 * Timing vulnerability in DSA signature generation
5385 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5386 timing side channel attack. An attacker could use variations in the signing
5387 algorithm to recover the private key.
5389 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5394 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5395 Module, accidentally introduced while backporting security fixes from the
5396 development branch and hindering the use of ECC in FIPS mode.
5400 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5402 * Client DoS due to large DH parameter
5404 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5405 malicious server can send a very large prime value to the client. This will
5406 cause the client to spend an unreasonably long period of time generating a
5407 key for this prime resulting in a hang until the client has finished. This
5408 could be exploited in a Denial Of Service attack.
5410 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5415 * Cache timing vulnerability in RSA Key Generation
5417 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5418 a cache timing side channel attack. An attacker with sufficient access to
5419 mount cache timing attacks during the RSA key generation process could
5420 recover the private key.
5422 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5423 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5428 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5429 parameter is no longer accepted, as it leads to a corrupt table. NULL
5430 pem_str is reserved for alias entries only.
5434 * Revert blinding in ECDSA sign and instead make problematic addition
5435 length-invariant. Switch even to fixed-length Montgomery multiplication.
5439 * Change generating and checking of primes so that the error rate of not
5440 being prime depends on the intended use based on the size of the input.
5441 For larger primes this will result in more rounds of Miller-Rabin.
5442 The maximal error rate for primes with more than 1080 bits is lowered
5445 *Kurt Roeckx, Annie Yousar*
5447 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5451 * Add blinding to ECDSA and DSA signatures to protect against side channel
5452 attacks discovered by Keegan Ryan (NCC Group).
5456 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5457 now allow empty (zero character) pass phrases.
5461 * Certificate time validation (X509_cmp_time) enforces stricter
5462 compliance with RFC 5280. Fractional seconds and timezone offsets
5463 are no longer allowed.
5467 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5469 * Constructed ASN.1 types with a recursive definition could exceed the stack
5471 Constructed ASN.1 types with a recursive definition (such as can be found
5472 in PKCS7) could eventually exceed the stack given malicious input with
5473 excessive recursion. This could result in a Denial Of Service attack. There
5474 are no such structures used within SSL/TLS that come from untrusted sources
5475 so this is considered safe.
5477 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5483 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5485 * Read/write after SSL object in error state
5487 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5488 mechanism. The intent was that if a fatal error occurred during a handshake
5489 then OpenSSL would move into the error state and would immediately fail if
5490 you attempted to continue the handshake. This works as designed for the
5491 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5492 SSL_connect()), however due to a bug it does not work correctly if
5493 SSL_read() or SSL_write() is called directly. In that scenario, if the
5494 handshake fails then a fatal error will be returned in the initial function
5495 call. If SSL_read()/SSL_write() is subsequently called by the application
5496 for the same SSL object then it will succeed and the data is passed without
5497 being decrypted/encrypted directly from the SSL/TLS record layer.
5499 In order to exploit this issue an application bug would have to be present
5500 that resulted in a call to SSL_read()/SSL_write() being issued after having
5501 already received a fatal error.
5503 This issue was reported to OpenSSL by David Benjamin (Google).
5508 * rsaz_1024_mul_avx2 overflow bug on x86_64
5510 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5511 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5512 Analysis suggests that attacks against RSA and DSA as a result of this
5513 defect would be very difficult to perform and are not believed likely.
5514 Attacks against DH1024 are considered just feasible, because most of the
5515 work necessary to deduce information about a private key may be performed
5516 offline. The amount of resources required for such an attack would be
5517 significant. However, for an attack on TLS to be meaningful, the server
5518 would have to share the DH1024 private key among multiple clients, which is
5519 no longer an option since CVE-2016-0701.
5521 This only affects processors that support the AVX2 but not ADX extensions
5522 like Intel Haswell (4th generation).
5524 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5525 was originally found via the OSS-Fuzz project.
5530 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5532 * bn_sqrx8x_internal carry bug on x86_64
5534 There is a carry propagating bug in the x86_64 Montgomery squaring
5535 procedure. No EC algorithms are affected. Analysis suggests that attacks
5536 against RSA and DSA as a result of this defect would be very difficult to
5537 perform and are not believed likely. Attacks against DH are considered just
5538 feasible (although very difficult) because most of the work necessary to
5539 deduce information about a private key may be performed offline. The amount
5540 of resources required for such an attack would be very significant and
5541 likely only accessible to a limited number of attackers. An attacker would
5542 additionally need online access to an unpatched system using the target
5543 private key in a scenario with persistent DH parameters and a private
5544 key that is shared between multiple clients.
5546 This only affects processors that support the BMI1, BMI2 and ADX extensions
5547 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5549 This issue was reported to OpenSSL by the OSS-Fuzz project.
5554 * Malformed X.509 IPAddressFamily could cause OOB read
5556 If an X.509 certificate has a malformed IPAddressFamily extension,
5557 OpenSSL could do a one-byte buffer overread. The most likely result
5558 would be an erroneous display of the certificate in text format.
5560 This issue was reported to OpenSSL by the OSS-Fuzz project.
5564 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5566 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5567 platform rather than 'mingw'.
5571 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5573 * Truncated packet could crash via OOB read
5575 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5576 cipher is being used, then a truncated packet can cause that host to
5577 perform an out-of-bounds read, usually resulting in a crash.
5579 This issue was reported to OpenSSL by Robert Święcki of Google.
5584 * BN_mod_exp may produce incorrect results on x86_64
5586 There is a carry propagating bug in the x86_64 Montgomery squaring
5587 procedure. No EC algorithms are affected. Analysis suggests that attacks
5588 against RSA and DSA as a result of this defect would be very difficult to
5589 perform and are not believed likely. Attacks against DH are considered just
5590 feasible (although very difficult) because most of the work necessary to
5591 deduce information about a private key may be performed offline. The amount
5592 of resources required for such an attack would be very significant and
5593 likely only accessible to a limited number of attackers. An attacker would
5594 additionally need online access to an unpatched system using the target
5595 private key in a scenario with persistent DH parameters and a private
5596 key that is shared between multiple clients. For example this can occur by
5597 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5598 similar to CVE-2015-3193 but must be treated as a separate problem.
5600 This issue was reported to OpenSSL by the OSS-Fuzz project.
5605 * Montgomery multiplication may produce incorrect results
5607 There is a carry propagating bug in the Broadwell-specific Montgomery
5608 multiplication procedure that handles input lengths divisible by, but
5609 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5610 and DH private keys are impossible. This is because the subroutine in
5611 question is not used in operations with the private key itself and an input
5612 of the attacker's direct choice. Otherwise the bug can manifest itself as
5613 transient authentication and key negotiation failures or reproducible
5614 erroneous outcome of public-key operations with specially crafted input.
5615 Among EC algorithms only Brainpool P-512 curves are affected and one
5616 presumably can attack ECDH key negotiation. Impact was not analyzed in
5617 detail, because pre-requisites for attack are considered unlikely. Namely
5618 multiple clients have to choose the curve in question and the server has to
5619 share the private key among them, neither of which is default behaviour.
5620 Even then only clients that chose the curve will be affected.
5622 This issue was publicly reported as transient failures and was not
5623 initially recognized as a security issue. Thanks to Richard Morgan for
5624 providing reproducible case.
5629 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5630 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5631 prevent issues where no progress is being made and the peer continually
5632 sends unrecognised record types, using up resources processing them.
5636 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5638 * Missing CRL sanity check
5640 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5641 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5642 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5644 This issue only affects the OpenSSL 1.0.2i
5649 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5651 * OCSP Status Request extension unbounded memory growth
5653 A malicious client can send an excessively large OCSP Status Request
5654 extension. If that client continually requests renegotiation, sending a
5655 large OCSP Status Request extension each time, then there will be unbounded
5656 memory growth on the server. This will eventually lead to a Denial Of
5657 Service attack through memory exhaustion. Servers with a default
5658 configuration are vulnerable even if they do not support OCSP. Builds using
5659 the "no-ocsp" build time option are not affected.
5661 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5666 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5669 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5675 * OOB write in MDC2_Update()
5677 An overflow can occur in MDC2_Update() either if called directly or
5678 through the EVP_DigestUpdate() function using MDC2. If an attacker
5679 is able to supply very large amounts of input data after a previous
5680 call to EVP_EncryptUpdate() with a partial block then a length check
5681 can overflow resulting in a heap corruption.
5683 The amount of data needed is comparable to SIZE_MAX which is impractical
5686 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5691 * Malformed SHA512 ticket DoS
5693 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5694 DoS attack where a malformed ticket will result in an OOB read which will
5697 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5698 a custom server callback and ticket lookup mechanism.
5700 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5705 * OOB write in BN_bn2dec()
5707 The function BN_bn2dec() does not check the return value of BN_div_word().
5708 This can cause an OOB write if an application uses this function with an
5709 overly large BIGNUM. This could be a problem if an overly large certificate
5710 or CRL is printed out from an untrusted source. TLS is not affected because
5711 record limits will reject an oversized certificate before it is parsed.
5713 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5718 * OOB read in TS_OBJ_print_bio()
5720 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5721 the total length the OID text representation would use and not the amount
5722 of data written. This will result in OOB reads when large OIDs are
5725 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5730 * Pointer arithmetic undefined behaviour
5732 Avoid some undefined pointer arithmetic
5734 A common idiom in the codebase is to check limits in the following manner:
5737 Where "p" points to some malloc'd data of SIZE bytes and
5740 "len" here could be from some externally supplied data (e.g. from a TLS
5743 The rules of C pointer arithmetic are such that "p + len" is only well
5744 defined where len <= SIZE. Therefore the above idiom is actually
5745 undefined behaviour.
5747 For example this could cause problems if some malloc implementation
5748 provides an address for "p" such that "p + len" actually overflows for
5749 values of len that are too big and therefore p + len < limit.
5751 This issue was reported to OpenSSL by Guido Vranken
5756 * Constant time flag not preserved in DSA signing
5758 Operations in the DSA signing algorithm should run in constant time in
5759 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5760 implementation means that a non-constant time codepath is followed for
5761 certain operations. This has been demonstrated through a cache-timing
5762 attack to be sufficient for an attacker to recover the private DSA key.
5764 This issue was reported by César Pereida (Aalto University), Billy Brumley
5765 (Tampere University of Technology), and Yuval Yarom (The University of
5766 Adelaide and NICTA).
5771 * DTLS buffered message DoS
5773 In a DTLS connection where handshake messages are delivered out-of-order
5774 those messages that OpenSSL is not yet ready to process will be buffered
5775 for later use. Under certain circumstances, a flaw in the logic means that
5776 those messages do not get removed from the buffer even though the handshake
5777 has been completed. An attacker could force up to approx. 15 messages to
5778 remain in the buffer when they are no longer required. These messages will
5779 be cleared when the DTLS connection is closed. The default maximum size for
5780 a message is 100k. Therefore, the attacker could force an additional 1500k
5781 to be consumed per connection. By opening many simultaneous connections an
5782 attacker could cause a DoS attack through memory exhaustion.
5784 This issue was reported to OpenSSL by Quan Luo.
5789 * DTLS replay protection DoS
5791 A flaw in the DTLS replay attack protection mechanism means that records
5792 that arrive for future epochs update the replay protection "window" before
5793 the MAC for the record has been validated. This could be exploited by an
5794 attacker by sending a record for the next epoch (which does not have to
5795 decrypt or have a valid MAC), with a very large sequence number. This means
5796 that all subsequent legitimate packets are dropped causing a denial of
5797 service for a specific DTLS connection.
5799 This issue was reported to OpenSSL by the OCAP audit team.
5804 * Certificate message OOB reads
5806 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5807 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5808 theoretical DoS risk but this has not been observed in practice on common
5811 The messages affected are client certificate, client certificate request
5812 and server certificate. As a result the attack can only be performed
5813 against a client or a server which enables client authentication.
5815 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5820 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5822 * Prevent padding oracle in AES-NI CBC MAC check
5824 A MITM attacker can use a padding oracle attack to decrypt traffic
5825 when the connection uses an AES CBC cipher and the server support
5828 This issue was introduced as part of the fix for Lucky 13 padding
5829 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5830 constant time by making sure that always the same bytes are read and
5831 compared against either the MAC or padding bytes. But it no longer
5832 checked that there was enough data to have both the MAC and padding
5835 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5839 * Fix EVP_EncodeUpdate overflow
5841 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5842 Base64 encoding of binary data. If an attacker is able to supply very large
5843 amounts of input data then a length check can overflow resulting in a heap
5846 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5847 the `PEM_write_bio*` family of functions. These are mainly used within the
5848 OpenSSL command line applications, so any application which processes data
5849 from an untrusted source and outputs it as a PEM file should be considered
5850 vulnerable to this issue. User applications that call these APIs directly
5851 with large amounts of untrusted data may also be vulnerable.
5853 This issue was reported by Guido Vranken.
5858 * Fix EVP_EncryptUpdate overflow
5860 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5861 is able to supply very large amounts of input data after a previous call to
5862 EVP_EncryptUpdate() with a partial block then a length check can overflow
5863 resulting in a heap corruption. Following an analysis of all OpenSSL
5864 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5865 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5866 the first called function after an EVP_EncryptInit(), and therefore that
5867 specific call must be safe. The second form is where the length passed to
5868 EVP_EncryptUpdate() can be seen from the code to be some small value and
5869 therefore there is no possibility of an overflow. Since all instances are
5870 one of these two forms, it is believed that there can be no overflows in
5871 internal code due to this problem. It should be noted that
5872 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5873 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5874 of these calls have also been analysed too and it is believed there are no
5875 instances in internal usage where an overflow could occur.
5877 This issue was reported by Guido Vranken.
5882 * Prevent ASN.1 BIO excessive memory allocation
5884 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5885 a short invalid encoding can cause allocation of large amounts of memory
5886 potentially consuming excessive resources or exhausting memory.
5888 Any application parsing untrusted data through d2i BIO functions is
5889 affected. The memory based functions such as d2i_X509() are *not* affected.
5890 Since the memory based functions are used by the TLS library, TLS
5891 applications are not affected.
5893 This issue was reported by Brian Carpenter.
5900 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5901 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5902 in arbitrary stack data being returned in the buffer.
5904 This issue was reported by Guido Vranken.
5909 * Modify behavior of ALPN to invoke callback after SNI/servername
5910 callback, such that updates to the SSL_CTX affect ALPN.
5914 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5919 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5920 methods are enabled and ssl2 is disabled the methods return NULL.
5924 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5926 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5927 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5928 provide any "EXPORT" or "LOW" strength ciphers.
5932 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5933 is by default disabled at build-time. Builds that are not configured with
5934 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5935 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5936 will need to explicitly call either of:
5938 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5940 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5942 as appropriate. Even if either of those is used, or the application
5943 explicitly uses the version-specific SSLv2_method() or its client and
5944 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5945 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5946 ciphers, and SSLv2 56-bit DES are no longer available.
5951 * Fix a double-free in DSA code
5953 A double free bug was discovered when OpenSSL parses malformed DSA private
5954 keys and could lead to a DoS attack or memory corruption for applications
5955 that receive DSA private keys from untrusted sources. This scenario is
5958 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5964 * Disable SRP fake user seed to address a server memory leak.
5966 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5968 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5969 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5970 was changed to ignore the "fake user" SRP seed, even if the seed
5973 Users should use SRP_VBASE_get1_by_user instead. Note that in
5974 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5975 also that even though configuring the SRP seed attempts to hide
5976 invalid usernames by continuing the handshake with fake
5977 credentials, this behaviour is not constant time and no strong
5978 guarantees are made that the handshake is indistinguishable from
5979 that of a valid user.
5984 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5986 In the BN_hex2bn function the number of hex digits is calculated using an
5987 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5988 large values of `i` this can result in `bn_expand` not allocating any
5989 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5990 field as NULL leading to a subsequent NULL ptr deref. For very large values
5991 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5992 In this case memory is allocated to the internal BIGNUM data field, but it
5993 is insufficiently sized leading to heap corruption. A similar issue exists
5994 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5995 is ever called by user applications with very large untrusted hex/dec data.
5996 This is anticipated to be a rare occurrence.
5998 All OpenSSL internal usage of these functions use data that is not expected
5999 to be untrusted, e.g. config file data or application command line
6000 arguments. If user developed applications generate config file data based
6001 on untrusted data then it is possible that this could also lead to security
6002 consequences. This is also anticipated to be rare.
6004 This issue was reported to OpenSSL by Guido Vranken.
6009 * Fix memory issues in `BIO_*printf` functions
6011 The internal `fmtstr` function used in processing a "%s" format string in
6012 the `BIO_*printf` functions could overflow while calculating the length of a
6013 string and cause an OOB read when printing very long strings.
6015 Additionally the internal `doapr_outch` function can attempt to write to an
6016 OOB memory location (at an offset from the NULL pointer) in the event of a
6017 memory allocation failure. In 1.0.2 and below this could be caused where
6018 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6019 could be in processing a very long "%s" format string. Memory leaks can
6022 The first issue may mask the second issue dependent on compiler behaviour.
6023 These problems could enable attacks where large amounts of untrusted data
6024 is passed to the `BIO_*printf` functions. If applications use these functions
6025 in this way then they could be vulnerable. OpenSSL itself uses these
6026 functions when printing out human-readable dumps of ASN.1 data. Therefore
6027 applications that print this data could be vulnerable if the data is from
6028 untrusted sources. OpenSSL command line applications could also be
6029 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6030 as command line arguments.
6032 Libssl is not considered directly vulnerable. Additionally certificates etc
6033 received via remote connections via libssl are also unlikely to be able to
6034 trigger these issues because of message size limits enforced within libssl.
6036 This issue was reported to OpenSSL Guido Vranken.
6041 * Side channel attack on modular exponentiation
6043 A side-channel attack was found which makes use of cache-bank conflicts on
6044 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6045 of RSA keys. The ability to exploit this issue is limited as it relies on
6046 an attacker who has control of code in a thread running on the same
6047 hyper-threaded core as the victim thread which is performing decryptions.
6049 This issue was reported to OpenSSL by Yuval Yarom, The University of
6050 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6051 Nadia Heninger, University of Pennsylvania with more information at
6052 <http://cachebleed.info>.
6057 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6058 if no keysize is specified with default_bits. This fixes an
6059 omission in an earlier change that changed all RSA/DSA key generation
6060 commands to use 2048 bits by default.
6064 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6066 * DH small subgroups
6068 Historically OpenSSL only ever generated DH parameters based on "safe"
6069 primes. More recently (in version 1.0.2) support was provided for
6070 generating X9.42 style parameter files such as those required for RFC 5114
6071 support. The primes used in such files may not be "safe". Where an
6072 application is using DH configured with parameters based on primes that are
6073 not "safe" then an attacker could use this fact to find a peer's private
6074 DH exponent. This attack requires that the attacker complete multiple
6075 handshakes in which the peer uses the same private DH exponent. For example
6076 this could be used to discover a TLS server's private DH exponent if it's
6077 reusing the private DH exponent or it's using a static DH ciphersuite.
6079 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6080 TLS. It is not on by default. If the option is not set then the server
6081 reuses the same private DH exponent for the life of the server process and
6082 would be vulnerable to this attack. It is believed that many popular
6083 applications do set this option and would therefore not be at risk.
6085 The fix for this issue adds an additional check where a "q" parameter is
6086 available (as is the case in X9.42 based parameters). This detects the
6087 only known attack, and is the only possible defense for static DH
6088 ciphersuites. This could have some performance impact.
6090 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6091 default and cannot be disabled. This could have some performance impact.
6093 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6098 * SSLv2 doesn't block disabled ciphers
6100 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6101 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6102 been disabled, provided that the SSLv2 protocol was not also disabled via
6105 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6106 and Sebastian Schinzel.
6111 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6113 * BN_mod_exp may produce incorrect results on x86_64
6115 There is a carry propagating bug in the x86_64 Montgomery squaring
6116 procedure. No EC algorithms are affected. Analysis suggests that attacks
6117 against RSA and DSA as a result of this defect would be very difficult to
6118 perform and are not believed likely. Attacks against DH are considered just
6119 feasible (although very difficult) because most of the work necessary to
6120 deduce information about a private key may be performed offline. The amount
6121 of resources required for such an attack would be very significant and
6122 likely only accessible to a limited number of attackers. An attacker would
6123 additionally need online access to an unpatched system using the target
6124 private key in a scenario with persistent DH parameters and a private
6125 key that is shared between multiple clients. For example this can occur by
6126 default in OpenSSL DHE based SSL/TLS ciphersuites.
6128 This issue was reported to OpenSSL by Hanno Böck.
6133 * Certificate verify crash with missing PSS parameter
6135 The signature verification routines will crash with a NULL pointer
6136 dereference if presented with an ASN.1 signature using the RSA PSS
6137 algorithm and absent mask generation function parameter. Since these
6138 routines are used to verify certificate signature algorithms this can be
6139 used to crash any certificate verification operation and exploited in a
6140 DoS attack. Any application which performs certificate verification is
6141 vulnerable including OpenSSL clients and servers which enable client
6144 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6149 * X509_ATTRIBUTE memory leak
6151 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6152 memory. This structure is used by the PKCS#7 and CMS routines so any
6153 application which reads PKCS#7 or CMS data from untrusted sources is
6154 affected. SSL/TLS is not affected.
6156 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6162 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6163 This changes the decoding behaviour for some invalid messages,
6164 though the change is mostly in the more lenient direction, and
6165 legacy behaviour is preserved as much as possible.
6169 * In DSA_generate_parameters_ex, if the provided seed is too short,
6172 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6174 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6176 * Alternate chains certificate forgery
6178 During certificate verification, OpenSSL will attempt to find an
6179 alternative certificate chain if the first attempt to build such a chain
6180 fails. An error in the implementation of this logic can mean that an
6181 attacker could cause certain checks on untrusted certificates to be
6182 bypassed, such as the CA flag, enabling them to use a valid leaf
6183 certificate to act as a CA and "issue" an invalid certificate.
6185 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6190 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6192 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6193 incompatibility in the handling of HMAC. The previous ABI has now been
6198 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6200 * Malformed ECParameters causes infinite loop
6202 When processing an ECParameters structure OpenSSL enters an infinite loop
6203 if the curve specified is over a specially malformed binary polynomial
6206 This can be used to perform denial of service against any
6207 system which processes public keys, certificate requests or
6208 certificates. This includes TLS clients and TLS servers with
6209 client authentication enabled.
6211 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6216 * Exploitable out-of-bounds read in X509_cmp_time
6218 X509_cmp_time does not properly check the length of the ASN1_TIME
6219 string and can read a few bytes out of bounds. In addition,
6220 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6223 An attacker can use this to craft malformed certificates and CRLs of
6224 various sizes and potentially cause a segmentation fault, resulting in
6225 a DoS on applications that verify certificates or CRLs. TLS clients
6226 that verify CRLs are affected. TLS clients and servers with client
6227 authentication enabled may be affected if they use custom verification
6230 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6231 independently by Hanno Böck.
6236 * PKCS7 crash with missing EnvelopedContent
6238 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6239 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6240 with missing content and trigger a NULL pointer dereference on parsing.
6242 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6243 structures from untrusted sources are affected. OpenSSL clients and
6244 servers are not affected.
6246 This issue was reported to OpenSSL by Michal Zalewski (Google).
6251 * CMS verify infinite loop with unknown hash function
6253 When verifying a signedData message the CMS code can enter an infinite loop
6254 if presented with an unknown hash function OID. This can be used to perform
6255 denial of service against any system which verifies signedData messages using
6257 This issue was reported to OpenSSL by Johannes Bauer.
6262 * Race condition handling NewSessionTicket
6264 If a NewSessionTicket is received by a multi-threaded client when attempting to
6265 reuse a previous ticket then a race condition can occur potentially leading to
6266 a double free of the ticket data.
6271 * Only support 256-bit or stronger elliptic curves with the
6272 'ecdh_auto' setting (server) or by default (client). Of supported
6273 curves, prefer P-256 (both).
6277 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6279 * ClientHello sigalgs DoS fix
6281 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6282 invalid signature algorithms extension a NULL pointer dereference will
6283 occur. This can be exploited in a DoS attack against the server.
6285 This issue was was reported to OpenSSL by David Ramos of Stanford
6289 *Stephen Henson and Matt Caswell*
6291 * Multiblock corrupted pointer fix
6293 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6294 feature only applies on 64 bit x86 architecture platforms that support AES
6295 NI instructions. A defect in the implementation of "multiblock" can cause
6296 OpenSSL's internal write buffer to become incorrectly set to NULL when
6297 using non-blocking IO. Typically, when the user application is using a
6298 socket BIO for writing, this will only result in a failed connection.
6299 However if some other BIO is used then it is likely that a segmentation
6300 fault will be triggered, thus enabling a potential DoS attack.
6302 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6307 * Segmentation fault in DTLSv1_listen fix
6309 The DTLSv1_listen function is intended to be stateless and processes the
6310 initial ClientHello from many peers. It is common for user code to loop
6311 over the call to DTLSv1_listen until a valid ClientHello is received with
6312 an associated cookie. A defect in the implementation of DTLSv1_listen means
6313 that state is preserved in the SSL object from one invocation to the next
6314 that can lead to a segmentation fault. Errors processing the initial
6315 ClientHello can trigger this scenario. An example of such an error could be
6316 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6319 This issue was reported to OpenSSL by Per Allansson.
6324 * Segmentation fault in ASN1_TYPE_cmp fix
6326 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6327 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6328 certificate signature algorithm consistency this can be used to crash any
6329 certificate verification operation and exploited in a DoS attack. Any
6330 application which performs certificate verification is vulnerable including
6331 OpenSSL clients and servers which enable client authentication.
6336 * Segmentation fault for invalid PSS parameters fix
6338 The signature verification routines will crash with a NULL pointer
6339 dereference if presented with an ASN.1 signature using the RSA PSS
6340 algorithm and invalid parameters. Since these routines are used to verify
6341 certificate signature algorithms this can be used to crash any
6342 certificate verification operation and exploited in a DoS attack. Any
6343 application which performs certificate verification is vulnerable including
6344 OpenSSL clients and servers which enable client authentication.
6346 This issue was was reported to OpenSSL by Brian Carpenter.
6351 * ASN.1 structure reuse memory corruption fix
6353 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6354 memory corruption via an invalid write. Such reuse is and has been
6355 strongly discouraged and is believed to be rare.
6357 Applications that parse structures containing CHOICE or ANY DEFINED BY
6358 components may be affected. Certificate parsing (d2i_X509 and related
6359 functions) are however not affected. OpenSSL clients and servers are
6365 * PKCS7 NULL pointer dereferences fix
6367 The PKCS#7 parsing code does not handle missing outer ContentInfo
6368 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6369 missing content and trigger a NULL pointer dereference on parsing.
6371 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6372 otherwise parse PKCS#7 structures from untrusted sources are
6373 affected. OpenSSL clients and servers are not affected.
6375 This issue was reported to OpenSSL by Michal Zalewski (Google).
6380 * DoS via reachable assert in SSLv2 servers fix
6382 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6383 servers that both support SSLv2 and enable export cipher suites by sending
6384 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6386 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6387 (OpenSSL development team).
6392 * Empty CKE with client auth and DHE fix
6394 If client auth is used then a server can seg fault in the event of a DHE
6395 ciphersuite being selected and a zero length ClientKeyExchange message
6396 being sent by the client. This could be exploited in a DoS attack.
6401 * Handshake with unseeded PRNG fix
6403 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6404 with an unseeded PRNG. The conditions are:
6405 - The client is on a platform where the PRNG has not been seeded
6406 automatically, and the user has not seeded manually
6407 - A protocol specific client method version has been used (i.e. not
6408 SSL_client_methodv23)
6409 - A ciphersuite is used that does not require additional random data from
6410 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6412 If the handshake succeeds then the client random that has been used will
6413 have been generated from a PRNG with insufficient entropy and therefore the
6414 output may be predictable.
6416 For example using the following command with an unseeded openssl will
6417 succeed on an unpatched platform:
6419 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6424 * Use After Free following d2i_ECPrivatekey error fix
6426 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6427 could cause a use after free condition. This, in turn, could cause a double
6428 free in several private key parsing functions (such as d2i_PrivateKey
6429 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6430 for applications that receive EC private keys from untrusted
6431 sources. This scenario is considered rare.
6433 This issue was discovered by the BoringSSL project and fixed in their
6439 * X509_to_X509_REQ NULL pointer deref fix
6441 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6442 the certificate key is invalid. This function is rarely used in practice.
6444 This issue was discovered by Brian Carpenter.
6449 * Removed the export ciphers from the DEFAULT ciphers
6453 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6455 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6456 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6457 So far those who have to target multiple platforms would compromise
6458 and argue that binary targeting say ARMv5 would still execute on
6459 ARMv8. "Universal" build resolves this compromise by providing
6460 near-optimal performance even on newer platforms.
6464 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6465 (other platforms pending).
6467 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6469 * Add support for the SignedCertificateTimestampList certificate and
6470 OCSP response extensions from RFC6962.
6474 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6475 for corner cases. (Certain input points at infinity could lead to
6476 bogus results, with non-infinity inputs mapped to infinity too.)
6480 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6481 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6482 common cases are optimized and there still is room for further
6483 improvements. Vector Permutation AES for Altivec is also added.
6487 * Add support for little-endian ppc64 Linux target.
6489 *Marcelo Cerri (IBM)*
6491 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6492 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6493 are optimized and there still is room for further improvements.
6494 Both 32- and 64-bit modes are supported.
6496 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6498 * Improved ARMv7 NEON support.
6502 * Support for SPARC Architecture 2011 crypto extensions, first
6503 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6504 SHA256/512, MD5, GHASH and modular exponentiation.
6506 *Andy Polyakov, David Miller*
6508 * Accelerated modular exponentiation for Intel processors, a.k.a.
6511 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6513 * Support for new and upcoming Intel processors, including AVX2,
6514 BMI and SHA ISA extensions. This includes additional "stitched"
6515 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6518 This work was sponsored by Intel Corp.
6522 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6523 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6524 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6528 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6529 this fixes a limitation in previous versions of OpenSSL.
6533 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6534 MGF1 digest and OAEP label.
6538 * Add EVP support for key wrapping algorithms, to avoid problems with
6539 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6540 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6541 algorithms and include tests cases.
6545 * Add functions to allocate and set the fields of an ECDSA_METHOD
6548 *Douglas E. Engert, Steve Henson*
6550 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6551 difference in days and seconds between two tm or ASN1_TIME structures.
6555 * Add -rev test option to s_server to just reverse order of characters
6556 received by client and send back to server. Also prints an abbreviated
6557 summary of the connection parameters.
6561 * New option -brief for s_client and s_server to print out a brief summary
6562 of connection parameters.
6566 * Add callbacks for arbitrary TLS extensions.
6568 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6570 * New option -crl_download in several openssl utilities to download CRLs
6571 from CRLDP extension in certificates.
6575 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6579 * New function X509_CRL_diff to generate a delta CRL from the difference
6580 of two full CRLs. Add support to "crl" utility.
6584 * New functions to set lookup_crls function and to retrieve
6585 X509_STORE from X509_STORE_CTX.
6589 * Print out deprecated issuer and subject unique ID fields in
6594 * Extend OCSP I/O functions so they can be used for simple general purpose
6595 HTTP as well as OCSP. New wrapper function which can be used to download
6596 CRLs using the OCSP API.
6600 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6604 * `SSL_CONF*` functions. These provide a common framework for application
6605 configuration using configuration files or command lines.
6609 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6610 message callback and prints the results. Needs compile time option
6611 "enable-ssl-trace". New options to s_client and s_server to enable
6616 * New ctrl and macro to retrieve supported points extensions.
6617 Print out extension in s_server and s_client.
6621 * New functions to retrieve certificate signature and signature
6626 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6631 * New Suite B modes for TLS code. These use and enforce the requirements
6632 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6633 only use Suite B curves. The Suite B modes can be set by using the
6634 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6638 * New chain verification flags for Suite B levels of security. Check
6639 algorithms are acceptable when flags are set in X509_verify_cert.
6643 * Make tls1_check_chain return a set of flags indicating checks passed
6644 by a certificate chain. Add additional tests to handle client
6645 certificates: checks for matching certificate type and issuer name
6650 * If an attempt is made to use a signature algorithm not in the peer
6651 preference list abort the handshake. If client has no suitable
6652 signature algorithms in response to a certificate request do not
6653 use the certificate.
6657 * If server EC tmp key is not in client preference list abort handshake.
6661 * Add support for certificate stores in CERT structure. This makes it
6662 possible to have different stores per SSL structure or one store in
6663 the parent SSL_CTX. Include distinct stores for certificate chain
6664 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6665 to build and store a certificate chain in CERT structure: returning
6666 an error if the chain cannot be built: this will allow applications
6667 to test if a chain is correctly configured.
6669 Note: if the CERT based stores are not set then the parent SSL_CTX
6670 store is used to retain compatibility with existing behaviour.
6674 * New function ssl_set_client_disabled to set a ciphersuite disabled
6675 mask based on the current session, check mask when sending client
6676 hello and checking the requested ciphersuite.
6680 * New ctrls to retrieve and set certificate types in a certificate
6681 request message. Print out received values in s_client. If certificate
6682 types is not set with custom values set sensible values based on
6683 supported signature algorithms.
6687 * Support for distinct client and server supported signature algorithms.
6691 * Add certificate callback. If set this is called whenever a certificate
6692 is required by client or server. An application can decide which
6693 certificate chain to present based on arbitrary criteria: for example
6694 supported signature algorithms. Add very simple example to s_server.
6695 This fixes many of the problems and restrictions of the existing client
6696 certificate callback: for example you can now clear an existing
6697 certificate and specify the whole chain.
6701 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6702 the certificate can be used for (if anything). Set valid_flags field
6703 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6704 to have similar checks in it.
6706 Add new "cert_flags" field to CERT structure and include a "strict mode".
6707 This enforces some TLS certificate requirements (such as only permitting
6708 certificate signature algorithms contained in the supported algorithms
6709 extension) which some implementations ignore: this option should be used
6710 with caution as it could cause interoperability issues.
6714 * Update and tidy signature algorithm extension processing. Work out
6715 shared signature algorithms based on preferences and peer algorithms
6716 and print them out in s_client and s_server. Abort handshake if no
6717 shared signature algorithms.
6721 * Add new functions to allow customised supported signature algorithms
6722 for SSL and SSL_CTX structures. Add options to s_client and s_server
6727 * New function SSL_certs_clear() to delete all references to certificates
6728 from an SSL structure. Before this once a certificate had been added
6729 it couldn't be removed.
6733 * Integrate hostname, email address and IP address checking with certificate
6734 verification. New verify options supporting checking in openssl utility.
6738 * Fixes and wildcard matching support to hostname and email checking
6739 functions. Add manual page.
6741 *Florian Weimer (Red Hat Product Security Team)*
6743 * New functions to check a hostname email or IP address against a
6744 certificate. Add options x509 utility to print results of checks against
6749 * Fix OCSP checking.
6751 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6753 * Initial experimental support for explicitly trusted non-root CAs.
6754 OpenSSL still tries to build a complete chain to a root but if an
6755 intermediate CA has a trust setting included that is used. The first
6756 setting is used: whether to trust (e.g., -addtrust option to the x509
6761 * Add -trusted_first option which attempts to find certificates in the
6762 trusted store even if an untrusted chain is also supplied.
6766 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6767 platform support for Linux and Android.
6771 * Support for linux-x32, ILP32 environment in x86_64 framework.
6775 * Experimental multi-implementation support for FIPS capable OpenSSL.
6776 When in FIPS mode the approved implementations are used as normal,
6777 when not in FIPS mode the internal unapproved versions are used instead.
6778 This means that the FIPS capable OpenSSL isn't forced to use the
6779 (often lower performance) FIPS implementations outside FIPS mode.
6783 * Transparently support X9.42 DH parameters when calling
6784 PEM_read_bio_DHparameters. This means existing applications can handle
6785 the new parameter format automatically.
6789 * Initial experimental support for X9.42 DH parameter format: mainly
6790 to support use of 'q' parameter for RFC5114 parameters.
6794 * Add DH parameters from RFC5114 including test data to dhtest.
6798 * Support for automatic EC temporary key parameter selection. If enabled
6799 the most preferred EC parameters are automatically used instead of
6800 hardcoded fixed parameters. Now a server just has to call:
6801 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6802 support ECDH and use the most appropriate parameters.
6806 * Enhance and tidy EC curve and point format TLS extension code. Use
6807 static structures instead of allocation if default values are used.
6808 New ctrls to set curves we wish to support and to retrieve shared curves.
6809 Print out shared curves in s_server. New options to s_server and s_client
6810 to set list of supported curves.
6814 * New ctrls to retrieve supported signature algorithms and
6815 supported curve values as an array of NIDs. Extend openssl utility
6816 to print out received values.
6820 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6821 between NIDs and the more common NIST names such as "P-256". Enhance
6822 ecparam utility and ECC method to recognise the NIST names for curves.
6826 * Enhance SSL/TLS certificate chain handling to support different
6827 chains for each certificate instead of one chain in the parent SSL_CTX.
6831 * Support for fixed DH ciphersuite client authentication: where both
6832 server and client use DH certificates with common parameters.
6836 * Support for fixed DH ciphersuites: those requiring DH server
6841 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6843 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6844 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6845 X509_CINF_get_signature were reverted post internal team review.
6850 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6852 * OCSP Status Request extension unbounded memory growth
6854 A malicious client can send an excessively large OCSP Status Request
6855 extension. If that client continually requests renegotiation, sending a
6856 large OCSP Status Request extension each time, then there will be unbounded
6857 memory growth on the server. This will eventually lead to a Denial Of
6858 Service attack through memory exhaustion. Servers with a default
6859 configuration are vulnerable even if they do not support OCSP. Builds using
6860 the "no-ocsp" build time option are not affected.
6862 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6867 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6870 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6876 * OOB write in MDC2_Update()
6878 An overflow can occur in MDC2_Update() either if called directly or
6879 through the EVP_DigestUpdate() function using MDC2. If an attacker
6880 is able to supply very large amounts of input data after a previous
6881 call to EVP_EncryptUpdate() with a partial block then a length check
6882 can overflow resulting in a heap corruption.
6884 The amount of data needed is comparable to SIZE_MAX which is impractical
6887 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6892 * Malformed SHA512 ticket DoS
6894 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6895 DoS attack where a malformed ticket will result in an OOB read which will
6898 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6899 a custom server callback and ticket lookup mechanism.
6901 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6906 * OOB write in BN_bn2dec()
6908 The function BN_bn2dec() does not check the return value of BN_div_word().
6909 This can cause an OOB write if an application uses this function with an
6910 overly large BIGNUM. This could be a problem if an overly large certificate
6911 or CRL is printed out from an untrusted source. TLS is not affected because
6912 record limits will reject an oversized certificate before it is parsed.
6914 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6919 * OOB read in TS_OBJ_print_bio()
6921 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6922 the total length the OID text representation would use and not the amount
6923 of data written. This will result in OOB reads when large OIDs are
6926 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6931 * Pointer arithmetic undefined behaviour
6933 Avoid some undefined pointer arithmetic
6935 A common idiom in the codebase is to check limits in the following manner:
6938 Where "p" points to some malloc'd data of SIZE bytes and
6941 "len" here could be from some externally supplied data (e.g. from a TLS
6944 The rules of C pointer arithmetic are such that "p + len" is only well
6945 defined where len <= SIZE. Therefore, the above idiom is actually
6946 undefined behaviour.
6948 For example this could cause problems if some malloc implementation
6949 provides an address for "p" such that "p + len" actually overflows for
6950 values of len that are too big and therefore p + len < limit.
6952 This issue was reported to OpenSSL by Guido Vranken
6957 * Constant time flag not preserved in DSA signing
6959 Operations in the DSA signing algorithm should run in constant time in
6960 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6961 implementation means that a non-constant time codepath is followed for
6962 certain operations. This has been demonstrated through a cache-timing
6963 attack to be sufficient for an attacker to recover the private DSA key.
6965 This issue was reported by César Pereida (Aalto University), Billy Brumley
6966 (Tampere University of Technology), and Yuval Yarom (The University of
6967 Adelaide and NICTA).
6972 * DTLS buffered message DoS
6974 In a DTLS connection where handshake messages are delivered out-of-order
6975 those messages that OpenSSL is not yet ready to process will be buffered
6976 for later use. Under certain circumstances, a flaw in the logic means that
6977 those messages do not get removed from the buffer even though the handshake
6978 has been completed. An attacker could force up to approx. 15 messages to
6979 remain in the buffer when they are no longer required. These messages will
6980 be cleared when the DTLS connection is closed. The default maximum size for
6981 a message is 100k. Therefore, the attacker could force an additional 1500k
6982 to be consumed per connection. By opening many simultaneous connections an
6983 attacker could cause a DoS attack through memory exhaustion.
6985 This issue was reported to OpenSSL by Quan Luo.
6990 * DTLS replay protection DoS
6992 A flaw in the DTLS replay attack protection mechanism means that records
6993 that arrive for future epochs update the replay protection "window" before
6994 the MAC for the record has been validated. This could be exploited by an
6995 attacker by sending a record for the next epoch (which does not have to
6996 decrypt or have a valid MAC), with a very large sequence number. This means
6997 that all subsequent legitimate packets are dropped causing a denial of
6998 service for a specific DTLS connection.
7000 This issue was reported to OpenSSL by the OCAP audit team.
7005 * Certificate message OOB reads
7007 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7008 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7009 theoretical DoS risk but this has not been observed in practice on common
7012 The messages affected are client certificate, client certificate request
7013 and server certificate. As a result the attack can only be performed
7014 against a client or a server which enables client authentication.
7016 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7021 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7023 * Prevent padding oracle in AES-NI CBC MAC check
7025 A MITM attacker can use a padding oracle attack to decrypt traffic
7026 when the connection uses an AES CBC cipher and the server support
7029 This issue was introduced as part of the fix for Lucky 13 padding
7030 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7031 constant time by making sure that always the same bytes are read and
7032 compared against either the MAC or padding bytes. But it no longer
7033 checked that there was enough data to have both the MAC and padding
7036 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7041 * Fix EVP_EncodeUpdate overflow
7043 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7044 Base64 encoding of binary data. If an attacker is able to supply very large
7045 amounts of input data then a length check can overflow resulting in a heap
7048 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7049 the `PEM_write_bio*` family of functions. These are mainly used within the
7050 OpenSSL command line applications, so any application which processes data
7051 from an untrusted source and outputs it as a PEM file should be considered
7052 vulnerable to this issue. User applications that call these APIs directly
7053 with large amounts of untrusted data may also be vulnerable.
7055 This issue was reported by Guido Vranken.
7060 * Fix EVP_EncryptUpdate overflow
7062 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7063 is able to supply very large amounts of input data after a previous call to
7064 EVP_EncryptUpdate() with a partial block then a length check can overflow
7065 resulting in a heap corruption. Following an analysis of all OpenSSL
7066 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7067 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7068 the first called function after an EVP_EncryptInit(), and therefore that
7069 specific call must be safe. The second form is where the length passed to
7070 EVP_EncryptUpdate() can be seen from the code to be some small value and
7071 therefore there is no possibility of an overflow. Since all instances are
7072 one of these two forms, it is believed that there can be no overflows in
7073 internal code due to this problem. It should be noted that
7074 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7075 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7076 of these calls have also been analysed too and it is believed there are no
7077 instances in internal usage where an overflow could occur.
7079 This issue was reported by Guido Vranken.
7084 * Prevent ASN.1 BIO excessive memory allocation
7086 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7087 a short invalid encoding can casuse allocation of large amounts of memory
7088 potentially consuming excessive resources or exhausting memory.
7090 Any application parsing untrusted data through d2i BIO functions is
7091 affected. The memory based functions such as d2i_X509() are *not* affected.
7092 Since the memory based functions are used by the TLS library, TLS
7093 applications are not affected.
7095 This issue was reported by Brian Carpenter.
7102 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7103 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7104 in arbitrary stack data being returned in the buffer.
7106 This issue was reported by Guido Vranken.
7111 * Modify behavior of ALPN to invoke callback after SNI/servername
7112 callback, such that updates to the SSL_CTX affect ALPN.
7116 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7121 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7122 methods are enabled and ssl2 is disabled the methods return NULL.
7126 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7128 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7129 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7130 provide any "EXPORT" or "LOW" strength ciphers.
7134 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7135 is by default disabled at build-time. Builds that are not configured with
7136 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7137 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7138 will need to explicitly call either of:
7140 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7142 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7144 as appropriate. Even if either of those is used, or the application
7145 explicitly uses the version-specific SSLv2_method() or its client and
7146 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7147 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7148 ciphers, and SSLv2 56-bit DES are no longer available.
7153 * Fix a double-free in DSA code
7155 A double free bug was discovered when OpenSSL parses malformed DSA private
7156 keys and could lead to a DoS attack or memory corruption for applications
7157 that receive DSA private keys from untrusted sources. This scenario is
7160 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7166 * Disable SRP fake user seed to address a server memory leak.
7168 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7170 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7171 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7172 was changed to ignore the "fake user" SRP seed, even if the seed
7175 Users should use SRP_VBASE_get1_by_user instead. Note that in
7176 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7177 also that even though configuring the SRP seed attempts to hide
7178 invalid usernames by continuing the handshake with fake
7179 credentials, this behaviour is not constant time and no strong
7180 guarantees are made that the handshake is indistinguishable from
7181 that of a valid user.
7186 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7188 In the BN_hex2bn function the number of hex digits is calculated using an
7189 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7190 large values of `i` this can result in `bn_expand` not allocating any
7191 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7192 field as NULL leading to a subsequent NULL ptr deref. For very large values
7193 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7194 In this case memory is allocated to the internal BIGNUM data field, but it
7195 is insufficiently sized leading to heap corruption. A similar issue exists
7196 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7197 is ever called by user applications with very large untrusted hex/dec data.
7198 This is anticipated to be a rare occurrence.
7200 All OpenSSL internal usage of these functions use data that is not expected
7201 to be untrusted, e.g. config file data or application command line
7202 arguments. If user developed applications generate config file data based
7203 on untrusted data then it is possible that this could also lead to security
7204 consequences. This is also anticipated to be rare.
7206 This issue was reported to OpenSSL by Guido Vranken.
7211 * Fix memory issues in `BIO_*printf` functions
7213 The internal `fmtstr` function used in processing a "%s" format string in
7214 the `BIO_*printf` functions could overflow while calculating the length of a
7215 string and cause an OOB read when printing very long strings.
7217 Additionally the internal `doapr_outch` function can attempt to write to an
7218 OOB memory location (at an offset from the NULL pointer) in the event of a
7219 memory allocation failure. In 1.0.2 and below this could be caused where
7220 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7221 could be in processing a very long "%s" format string. Memory leaks can
7224 The first issue may mask the second issue dependent on compiler behaviour.
7225 These problems could enable attacks where large amounts of untrusted data
7226 is passed to the `BIO_*printf` functions. If applications use these functions
7227 in this way then they could be vulnerable. OpenSSL itself uses these
7228 functions when printing out human-readable dumps of ASN.1 data. Therefore
7229 applications that print this data could be vulnerable if the data is from
7230 untrusted sources. OpenSSL command line applications could also be
7231 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7232 as command line arguments.
7234 Libssl is not considered directly vulnerable. Additionally certificates etc
7235 received via remote connections via libssl are also unlikely to be able to
7236 trigger these issues because of message size limits enforced within libssl.
7238 This issue was reported to OpenSSL Guido Vranken.
7243 * Side channel attack on modular exponentiation
7245 A side-channel attack was found which makes use of cache-bank conflicts on
7246 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7247 of RSA keys. The ability to exploit this issue is limited as it relies on
7248 an attacker who has control of code in a thread running on the same
7249 hyper-threaded core as the victim thread which is performing decryptions.
7251 This issue was reported to OpenSSL by Yuval Yarom, The University of
7252 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7253 Nadia Heninger, University of Pennsylvania with more information at
7254 <http://cachebleed.info>.
7259 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7260 if no keysize is specified with default_bits. This fixes an
7261 omission in an earlier change that changed all RSA/DSA key generation
7262 commands to use 2048 bits by default.
7266 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7268 * Protection for DH small subgroup attacks
7270 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7271 switched on by default and cannot be disabled. This could have some
7276 * SSLv2 doesn't block disabled ciphers
7278 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7279 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7280 been disabled, provided that the SSLv2 protocol was not also disabled via
7283 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7284 and Sebastian Schinzel.
7289 * Reject DH handshakes with parameters shorter than 1024 bits.
7293 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7295 * Certificate verify crash with missing PSS parameter
7297 The signature verification routines will crash with a NULL pointer
7298 dereference if presented with an ASN.1 signature using the RSA PSS
7299 algorithm and absent mask generation function parameter. Since these
7300 routines are used to verify certificate signature algorithms this can be
7301 used to crash any certificate verification operation and exploited in a
7302 DoS attack. Any application which performs certificate verification is
7303 vulnerable including OpenSSL clients and servers which enable client
7306 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7311 * X509_ATTRIBUTE memory leak
7313 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7314 memory. This structure is used by the PKCS#7 and CMS routines so any
7315 application which reads PKCS#7 or CMS data from untrusted sources is
7316 affected. SSL/TLS is not affected.
7318 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7324 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7325 This changes the decoding behaviour for some invalid messages,
7326 though the change is mostly in the more lenient direction, and
7327 legacy behaviour is preserved as much as possible.
7331 * In DSA_generate_parameters_ex, if the provided seed is too short,
7332 use a random seed, as already documented.
7334 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7336 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7338 * Alternate chains certificate forgery
7340 During certificate verfification, OpenSSL will attempt to find an
7341 alternative certificate chain if the first attempt to build such a chain
7342 fails. An error in the implementation of this logic can mean that an
7343 attacker could cause certain checks on untrusted certificates to be
7344 bypassed, such as the CA flag, enabling them to use a valid leaf
7345 certificate to act as a CA and "issue" an invalid certificate.
7347 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7353 * Race condition handling PSK identify hint
7355 If PSK identity hints are received by a multi-threaded client then
7356 the values are wrongly updated in the parent SSL_CTX structure. This can
7357 result in a race condition potentially leading to a double free of the
7363 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7365 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7366 incompatibility in the handling of HMAC. The previous ABI has now been
7369 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7371 * Malformed ECParameters causes infinite loop
7373 When processing an ECParameters structure OpenSSL enters an infinite loop
7374 if the curve specified is over a specially malformed binary polynomial
7377 This can be used to perform denial of service against any
7378 system which processes public keys, certificate requests or
7379 certificates. This includes TLS clients and TLS servers with
7380 client authentication enabled.
7382 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7387 * Exploitable out-of-bounds read in X509_cmp_time
7389 X509_cmp_time does not properly check the length of the ASN1_TIME
7390 string and can read a few bytes out of bounds. In addition,
7391 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7394 An attacker can use this to craft malformed certificates and CRLs of
7395 various sizes and potentially cause a segmentation fault, resulting in
7396 a DoS on applications that verify certificates or CRLs. TLS clients
7397 that verify CRLs are affected. TLS clients and servers with client
7398 authentication enabled may be affected if they use custom verification
7401 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7402 independently by Hanno Böck.
7407 * PKCS7 crash with missing EnvelopedContent
7409 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7410 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7411 with missing content and trigger a NULL pointer dereference on parsing.
7413 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7414 structures from untrusted sources are affected. OpenSSL clients and
7415 servers are not affected.
7417 This issue was reported to OpenSSL by Michal Zalewski (Google).
7422 * CMS verify infinite loop with unknown hash function
7424 When verifying a signedData message the CMS code can enter an infinite loop
7425 if presented with an unknown hash function OID. This can be used to perform
7426 denial of service against any system which verifies signedData messages using
7428 This issue was reported to OpenSSL by Johannes Bauer.
7433 * Race condition handling NewSessionTicket
7435 If a NewSessionTicket is received by a multi-threaded client when attempting to
7436 reuse a previous ticket then a race condition can occur potentially leading to
7437 a double free of the ticket data.
7442 * Reject DH handshakes with parameters shorter than 768 bits.
7444 *Kurt Roeckx and Emilia Kasper*
7446 * dhparam: generate 2048-bit parameters by default.
7448 *Kurt Roeckx and Emilia Kasper*
7450 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7452 * Segmentation fault in ASN1_TYPE_cmp fix
7454 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7455 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7456 certificate signature algorithm consistency this can be used to crash any
7457 certificate verification operation and exploited in a DoS attack. Any
7458 application which performs certificate verification is vulnerable including
7459 OpenSSL clients and servers which enable client authentication.
7464 * ASN.1 structure reuse memory corruption fix
7466 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7467 memory corruption via an invalid write. Such reuse is and has been
7468 strongly discouraged and is believed to be rare.
7470 Applications that parse structures containing CHOICE or ANY DEFINED BY
7471 components may be affected. Certificate parsing (d2i_X509 and related
7472 functions) are however not affected. OpenSSL clients and servers are
7478 * PKCS7 NULL pointer dereferences fix
7480 The PKCS#7 parsing code does not handle missing outer ContentInfo
7481 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7482 missing content and trigger a NULL pointer dereference on parsing.
7484 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7485 otherwise parse PKCS#7 structures from untrusted sources are
7486 affected. OpenSSL clients and servers are not affected.
7488 This issue was reported to OpenSSL by Michal Zalewski (Google).
7493 * DoS via reachable assert in SSLv2 servers fix
7495 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7496 servers that both support SSLv2 and enable export cipher suites by sending
7497 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7499 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7500 (OpenSSL development team).
7505 * Use After Free following d2i_ECPrivatekey error fix
7507 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7508 could cause a use after free condition. This, in turn, could cause a double
7509 free in several private key parsing functions (such as d2i_PrivateKey
7510 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7511 for applications that receive EC private keys from untrusted
7512 sources. This scenario is considered rare.
7514 This issue was discovered by the BoringSSL project and fixed in their
7520 * X509_to_X509_REQ NULL pointer deref fix
7522 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7523 the certificate key is invalid. This function is rarely used in practice.
7525 This issue was discovered by Brian Carpenter.
7530 * Removed the export ciphers from the DEFAULT ciphers
7534 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7536 * Build fixes for the Windows and OpenVMS platforms
7538 *Matt Caswell and Richard Levitte*
7540 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7542 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7543 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7544 dereference. This could lead to a Denial Of Service attack. Thanks to
7545 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7550 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7551 dtls1_buffer_record function under certain conditions. In particular this
7552 could occur if an attacker sent repeated DTLS records with the same
7553 sequence number but for the next epoch. The memory leak could be exploited
7554 by an attacker in a Denial of Service attack through memory exhaustion.
7555 Thanks to Chris Mueller for reporting this issue.
7560 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7561 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7562 method would be set to NULL which could later result in a NULL pointer
7563 dereference. Thanks to Frank Schmirler for reporting this issue.
7568 * Abort handshake if server key exchange message is omitted for ephemeral
7571 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7572 reporting this issue.
7577 * Remove non-export ephemeral RSA code on client and server. This code
7578 violated the TLS standard by allowing the use of temporary RSA keys in
7579 non-export ciphersuites and could be used by a server to effectively
7580 downgrade the RSA key length used to a value smaller than the server
7581 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7582 INRIA or reporting this issue.
7587 * Fixed issue where DH client certificates are accepted without verification.
7588 An OpenSSL server will accept a DH certificate for client authentication
7589 without the certificate verify message. This effectively allows a client to
7590 authenticate without the use of a private key. This only affects servers
7591 which trust a client certificate authority which issues certificates
7592 containing DH keys: these are extremely rare and hardly ever encountered.
7593 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7599 * Ensure that the session ID context of an SSL is updated when its
7600 SSL_CTX is updated via SSL_set_SSL_CTX.
7602 The session ID context is typically set from the parent SSL_CTX,
7603 and can vary with the CTX.
7607 * Fix various certificate fingerprint issues.
7609 By using non-DER or invalid encodings outside the signed portion of a
7610 certificate the fingerprint can be changed without breaking the signature.
7611 Although no details of the signed portion of the certificate can be changed
7612 this can cause problems with some applications: e.g. those using the
7613 certificate fingerprint for blacklists.
7615 1. Reject signatures with non zero unused bits.
7617 If the BIT STRING containing the signature has non zero unused bits reject
7618 the signature. All current signature algorithms require zero unused bits.
7620 2. Check certificate algorithm consistency.
7622 Check the AlgorithmIdentifier inside TBS matches the one in the
7623 certificate signature. NB: this will result in signature failure
7624 errors for some broken certificates.
7626 Thanks to Konrad Kraszewski from Google for reporting this issue.
7628 3. Check DSA/ECDSA signatures use DER.
7630 Re-encode DSA/ECDSA signatures and compare with the original received
7631 signature. Return an error if there is a mismatch.
7633 This will reject various cases including garbage after signature
7634 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7635 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7636 (negative or with leading zeroes).
7638 Further analysis was conducted and fixes were developed by Stephen Henson
7639 of the OpenSSL core team.
7645 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7646 results on some platforms, including x86_64. This bug occurs at random
7647 with a very low probability, and is not known to be exploitable in any
7648 way, though its exact impact is difficult to determine. Thanks to Pieter
7649 Wuille (Blockstream) who reported this issue and also suggested an initial
7650 fix. Further analysis was conducted by the OpenSSL development team and
7651 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7652 the OpenSSL core team.
7657 * Do not resume sessions on the server if the negotiated protocol
7658 version does not match the session's version. Resuming with a different
7659 version, while not strictly forbidden by the RFC, is of questionable
7660 sanity and breaks all known clients.
7662 *David Benjamin, Emilia Käsper*
7664 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7665 early CCS messages during renegotiation. (Note that because
7666 renegotiation is encrypted, this early CCS was not exploitable.)
7670 * Tighten client-side session ticket handling during renegotiation:
7671 ensure that the client only accepts a session ticket if the server sends
7672 the extension anew in the ServerHello. Previously, a TLS client would
7673 reuse the old extension state and thus accept a session ticket if one was
7674 announced in the initial ServerHello.
7676 Similarly, ensure that the client requires a session ticket if one
7677 was advertised in the ServerHello. Previously, a TLS client would
7678 ignore a missing NewSessionTicket message.
7682 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7686 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7687 sends a carefully crafted handshake message, to cause OpenSSL to fail
7688 to free up to 64k of memory causing a memory leak. This could be
7689 exploited in a Denial Of Service attack. This issue affects OpenSSL
7690 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7691 whether SRTP is used or configured. Implementations of OpenSSL that
7692 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7694 The fix was developed by the OpenSSL team.
7699 * Session Ticket Memory Leak.
7701 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7702 integrity of that ticket is first verified. In the event of a session
7703 ticket integrity check failing, OpenSSL will fail to free memory
7704 causing a memory leak. By sending a large number of invalid session
7705 tickets an attacker could exploit this issue in a Denial Of Service
7711 * Build option no-ssl3 is incomplete.
7713 When OpenSSL is configured with "no-ssl3" as a build option, servers
7714 could accept and complete an SSL 3.0 handshake, and clients could be
7715 configured to send them.
7718 *Akamai and the OpenSSL team*
7720 * Add support for TLS_FALLBACK_SCSV.
7721 Client applications doing fallback retries should call
7722 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7725 *Adam Langley, Bodo Moeller*
7727 * Add additional DigestInfo checks.
7729 Re-encode DigestInto in DER and check against the original when
7730 verifying RSA signature: this will reject any improperly encoded
7731 DigestInfo structures.
7733 Note: this is a precautionary measure and no attacks are currently known.
7737 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7739 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7740 SRP code can be overrun an internal buffer. Add sanity check that
7741 g, A, B < N to SRP code.
7743 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7744 Group for discovering this issue.
7749 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7750 TLS 1.0 instead of higher protocol versions when the ClientHello message
7751 is badly fragmented. This allows a man-in-the-middle attacker to force a
7752 downgrade to TLS 1.0 even if both the server and the client support a
7753 higher protocol version, by modifying the client's TLS records.
7755 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7756 researching this issue.
7761 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7762 to a denial of service attack. A malicious server can crash the client
7763 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7764 ciphersuite and sending carefully crafted handshake messages.
7766 Thanks to Felix Gröbert (Google) for discovering and researching this
7772 * By sending carefully crafted DTLS packets an attacker could cause openssl
7773 to leak memory. This can be exploited through a Denial of Service attack.
7774 Thanks to Adam Langley for discovering and researching this issue.
7779 * An attacker can force openssl to consume large amounts of memory whilst
7780 processing DTLS handshake messages. This can be exploited through a
7781 Denial of Service attack.
7782 Thanks to Adam Langley for discovering and researching this issue.
7787 * An attacker can force an error condition which causes openssl to crash
7788 whilst processing DTLS packets due to memory being freed twice. This
7789 can be exploited through a Denial of Service attack.
7790 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7796 * If a multithreaded client connects to a malicious server using a resumed
7797 session and the server sends an ec point format extension it could write
7798 up to 255 bytes to freed memory.
7800 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7806 * A malicious server can crash an OpenSSL client with a null pointer
7807 dereference (read) by specifying an SRP ciphersuite even though it was not
7808 properly negotiated with the client. This can be exploited through a
7809 Denial of Service attack.
7811 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7812 discovering and researching this issue.
7817 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7818 X509_name_oneline, X509_name_print_ex et al. to leak some information
7819 from the stack. Applications may be affected if they echo pretty printing
7820 output to the attacker.
7822 Thanks to Ivan Fratric (Google) for discovering this issue.
7825 *Emilia Käsper, and Steve Henson*
7827 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7828 for corner cases. (Certain input points at infinity could lead to
7829 bogus results, with non-infinity inputs mapped to infinity too.)
7833 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7835 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7836 handshake can force the use of weak keying material in OpenSSL
7837 SSL/TLS clients and servers.
7839 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7840 researching this issue. ([CVE-2014-0224])
7842 *KIKUCHI Masashi, Steve Henson*
7844 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7845 OpenSSL DTLS client the code can be made to recurse eventually crashing
7848 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7851 *Imre Rad, Steve Henson*
7853 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7854 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7855 client or server. This is potentially exploitable to run arbitrary
7856 code on a vulnerable client or server.
7858 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7860 *Jüri Aedla, Steve Henson*
7862 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7863 are subject to a denial of service attack.
7865 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7866 this issue. ([CVE-2014-3470])
7868 *Felix Gröbert, Ivan Fratric, Steve Henson*
7870 * Harmonize version and its documentation. -f flag is used to display
7873 *mancha <mancha1@zoho.com>*
7875 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7876 in i2d_ECPrivateKey.
7878 *mancha <mancha1@zoho.com>*
7880 * Fix some double frees. These are not thought to be exploitable.
7882 *mancha <mancha1@zoho.com>*
7884 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7886 * A missing bounds check in the handling of the TLS heartbeat extension
7887 can be used to reveal up to 64k of memory to a connected client or
7890 Thanks for Neel Mehta of Google Security for discovering this bug and to
7891 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7892 preparing the fix ([CVE-2014-0160])
7894 *Adam Langley, Bodo Moeller*
7896 * Fix for the attack described in the paper "Recovering OpenSSL
7897 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7898 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7899 <http://eprint.iacr.org/2014/140>
7901 Thanks to Yuval Yarom and Naomi Benger for discovering this
7902 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7904 *Yuval Yarom and Naomi Benger*
7906 * TLS pad extension: draft-agl-tls-padding-03
7908 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7909 TLS client Hello record length value would otherwise be > 255 and
7910 less that 512 pad with a dummy extension containing zeroes so it
7911 is at least 512 bytes long.
7913 *Adam Langley, Steve Henson*
7915 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7917 * Fix for TLS record tampering bug. A carefully crafted invalid
7918 handshake could crash OpenSSL with a NULL pointer exception.
7919 Thanks to Anton Johansson for reporting this issues.
7922 * Keep original DTLS digest and encryption contexts in retransmission
7923 structures so we can use the previous session parameters if they need
7924 to be resent. ([CVE-2013-6450])
7928 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7929 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7930 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7931 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7932 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7933 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7935 *Rob Stradling, Adam Langley*
7937 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7939 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7940 supporting platforms or when small records were transferred.
7942 *Andy Polyakov, Steve Henson*
7944 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7946 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7948 This addresses the flaw in CBC record processing discovered by
7949 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7950 at: <http://www.isg.rhul.ac.uk/tls/>
7952 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7953 Security Group at Royal Holloway, University of London
7954 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7955 Emilia Käsper for the initial patch.
7958 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7960 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7961 ciphersuites which can be exploited in a denial of service attack.
7962 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7963 and detecting this bug and to Wolfgang Ettlinger
7964 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7969 * Return an error when checking OCSP signatures when key is NULL.
7970 This fixes a DoS attack. ([CVE-2013-0166])
7974 * Make openssl verify return errors.
7976 *Chris Palmer <palmer@google.com> and Ben Laurie*
7978 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7979 the right response is stapled. Also change SSL_get_certificate()
7980 so it returns the certificate actually sent.
7981 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7983 *Rob Stradling <rob.stradling@comodo.com>*
7985 * Fix possible deadlock when decoding public keys.
7989 * Don't use TLS 1.0 record version number in initial client hello
7994 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7996 * Sanity check record length before skipping explicit IV in TLS
7997 1.2, 1.1 and DTLS to fix DoS attack.
7999 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8000 fuzzing as a service testing platform.
8005 * Initialise tkeylen properly when encrypting CMS messages.
8006 Thanks to Solar Designer of Openwall for reporting this issue.
8010 * In FIPS mode don't try to use composite ciphers as they are not
8015 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8017 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8018 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8019 mean any application compiled against OpenSSL 1.0.0 headers setting
8020 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8021 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8022 0x10000000L Any application which was previously compiled against
8023 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8024 will need to be recompiled as a result. Letting be results in
8025 inability to disable specifically TLS 1.1 and in client context,
8026 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8030 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8031 disable just protocol X, but all protocols above X *if* there are
8032 protocols *below* X still enabled. In more practical terms it means
8033 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8034 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8035 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8040 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8042 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8043 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8044 in CRYPTO_realloc_clean.
8046 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8047 issue and to Adam Langley <agl@chromium.org> for fixing it.
8050 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8052 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8056 * Workarounds for some broken servers that "hang" if a client hello
8057 record length exceeds 255 bytes.
8059 1. Do not use record version number > TLS 1.0 in initial client
8060 hello: some (but not all) hanging servers will now work.
8061 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8062 the number of ciphers sent in the client hello. This should be
8063 set to an even number, such as 50, for example by passing:
8064 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8065 Most broken servers should now work.
8066 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8067 TLS 1.2 client support entirely.
8071 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8075 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8077 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8078 STRING form instead of a DigestInfo.
8082 * The format used for MDC2 RSA signatures is inconsistent between EVP
8083 and the RSA_sign/RSA_verify functions. This was made more apparent when
8084 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8085 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8086 the correct format in RSA_verify so both forms transparently work.
8090 * Some servers which support TLS 1.0 can choke if we initially indicate
8091 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8092 encrypted premaster secret. As a workaround use the maximum permitted
8093 client version in client hello, this should keep such servers happy
8094 and still work with previous versions of OpenSSL.
8098 * Add support for TLS/DTLS heartbeats.
8100 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8102 * Add support for SCTP.
8104 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8106 * Improved PRNG seeding for VOS.
8108 *Paul Green <Paul.Green@stratus.com>*
8110 * Extensive assembler packs updates, most notably:
8112 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8113 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8114 - x86_64: bit-sliced AES implementation;
8115 - ARM: NEON support, contemporary platforms optimizations;
8116 - s390x: z196 support;
8117 - `*`: GHASH and GF(2^m) multiplication implementations;
8121 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8122 (removal of unnecessary code)
8124 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8126 * Add TLS key material exporter from RFC 5705.
8130 * Add DTLS-SRTP negotiation from RFC 5764.
8134 * Add Next Protocol Negotiation,
8135 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8136 disabled with a no-npn flag to config or Configure. Code donated
8139 *Adam Langley <agl@google.com> and Ben Laurie*
8141 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8142 NIST-P256, NIST-P521, with constant-time single point multiplication on
8143 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8144 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8145 Code made available under Apache License version 2.0.
8147 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8148 line to include this in your build of OpenSSL, and run "make depend" (or
8149 "make update"). This enables the following EC_METHODs:
8151 EC_GFp_nistp224_method()
8152 EC_GFp_nistp256_method()
8153 EC_GFp_nistp521_method()
8155 EC_GROUP_new_by_curve_name() will automatically use these (while
8156 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8159 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8161 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8162 all platforms. Move ssize_t definition from e_os.h to the public
8163 header file e_os2.h as it now appears in public header file cms.h
8167 * New -sigopt option to the ca, req and x509 utilities. Additional
8168 signature parameters can be passed using this option and in
8173 * Add RSA PSS signing function. This will generate and set the
8174 appropriate AlgorithmIdentifiers for PSS based on those in the
8175 corresponding EVP_MD_CTX structure. No application support yet.
8179 * Support for companion algorithm specific ASN1 signing routines.
8180 New function ASN1_item_sign_ctx() signs a pre-initialised
8181 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8182 the appropriate parameters.
8186 * Add new algorithm specific ASN1 verification initialisation function
8187 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8188 handling will be the same no matter what EVP_PKEY_METHOD is used.
8189 Add a PSS handler to support verification of PSS signatures: checked
8190 against a number of sample certificates.
8194 * Add signature printing for PSS. Add PSS OIDs.
8196 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8198 * Add algorithm specific signature printing. An individual ASN1 method
8199 can now print out signatures instead of the standard hex dump.
8201 More complex signatures (e.g. PSS) can print out more meaningful
8202 information. Include DSA version that prints out the signature
8207 * Password based recipient info support for CMS library: implementing
8212 * Split password based encryption into PBES2 and PBKDF2 functions. This
8213 neatly separates the code into cipher and PBE sections and is required
8214 for some algorithms that split PBES2 into separate pieces (such as
8215 password based CMS).
8219 * Session-handling fixes:
8220 - Fix handling of connections that are resuming with a session ID,
8221 but also support Session Tickets.
8222 - Fix a bug that suppressed issuing of a new ticket if the client
8223 presented a ticket with an expired session.
8224 - Try to set the ticket lifetime hint to something reasonable.
8225 - Make tickets shorter by excluding irrelevant information.
8226 - On the client side, don't ignore renewed tickets.
8228 *Adam Langley, Bodo Moeller (Google)*
8230 * Fix PSK session representation.
8234 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8236 This work was sponsored by Intel.
8240 * Add GCM support to TLS library. Some custom code is needed to split
8241 the IV between the fixed (from PRF) and explicit (from TLS record)
8242 portions. This adds all GCM ciphersuites supported by RFC5288 and
8243 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8244 add a special AESGCM string for GCM only.
8248 * Expand range of ctrls for AES GCM. Permit setting invocation
8249 field on decrypt and retrieval of invocation field only on encrypt.
8253 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8254 As required by RFC5289 these ciphersuites cannot be used if for
8255 versions of TLS earlier than 1.2.
8259 * For FIPS capable OpenSSL interpret a NULL default public key method
8260 as unset and return the appropriate default but do *not* set the default.
8261 This means we can return the appropriate method in applications that
8262 switch between FIPS and non-FIPS modes.
8266 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8267 ENGINE is used then we cannot handle that in the FIPS module so we
8268 keep original code iff non-FIPS operations are allowed.
8272 * Add -attime option to openssl utilities.
8274 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8276 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8280 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8281 FIPS EC methods unconditionally for now.
8285 * New build option no-ec2m to disable characteristic 2 code.
8289 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8290 all cases can be covered as some introduce binary incompatibilities.
8294 * Redirect RSA operations to FIPS module including keygen,
8295 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8299 * Add similar low-level API blocking to ciphers.
8303 * low-level digest APIs are not approved in FIPS mode: any attempt
8304 to use these will cause a fatal error. Applications that *really* want
8305 to use them can use the `private_*` version instead.
8309 * Redirect cipher operations to FIPS module for FIPS builds.
8313 * Redirect digest operations to FIPS module for FIPS builds.
8317 * Update build system to add "fips" flag which will link in fipscanister.o
8318 for static and shared library builds embedding a signature if needed.
8322 * Output TLS supported curves in preference order instead of numerical
8323 order. This is currently hardcoded for the highest order curves first.
8324 This should be configurable so applications can judge speed vs strength.
8328 * Add TLS v1.2 server support for client authentication.
8332 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8337 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8338 FIPS modules versions.
8342 * Add TLS v1.2 client side support for client authentication. Keep cache
8343 of handshake records longer as we don't know the hash algorithm to use
8344 until after the certificate request message is received.
8348 * Initial TLS v1.2 client support. Add a default signature algorithms
8349 extension including all the algorithms we support. Parse new signature
8350 format in client key exchange. Relax some ECC signing restrictions for
8351 TLS v1.2 as indicated in RFC5246.
8355 * Add server support for TLS v1.2 signature algorithms extension. Switch
8356 to new signature format when needed using client digest preference.
8357 All server ciphersuites should now work correctly in TLS v1.2. No client
8358 support yet and no support for client certificates.
8362 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8363 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8364 ciphersuites. At present only RSA key exchange ciphersuites work with
8365 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8366 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8367 and version checking.
8371 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8372 with this defined it will not be affected by any changes to ssl internal
8373 structures. Add several utility functions to allow openssl application
8374 to work with OPENSSL_NO_SSL_INTERN defined.
8378 * A long standing patch to add support for SRP from EdelWeb (Peter
8379 Sylvester and Christophe Renou) was integrated.
8380 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8381 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8384 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8388 * Permit abbreviated handshakes when renegotiating using the function
8389 SSL_renegotiate_abbreviated().
8391 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8393 * Add call to ENGINE_register_all_complete() to
8394 ENGINE_load_builtin_engines(), so some implementations get used
8395 automatically instead of needing explicit application support.
8399 * Add support for TLS key exporter as described in RFC5705.
8401 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8403 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8404 a few changes are required:
8406 Add SSL_OP_NO_TLSv1_1 flag.
8407 Add TLSv1_1 methods.
8408 Update version checking logic to handle version 1.1.
8409 Add explicit IV handling (ported from DTLS code).
8410 Add command line options to s_client/s_server.
8417 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8419 * X509_ATTRIBUTE memory leak
8421 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8422 memory. This structure is used by the PKCS#7 and CMS routines so any
8423 application which reads PKCS#7 or CMS data from untrusted sources is
8424 affected. SSL/TLS is not affected.
8426 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8432 * Race condition handling PSK identify hint
8434 If PSK identity hints are received by a multi-threaded client then
8435 the values are wrongly updated in the parent SSL_CTX structure. This can
8436 result in a race condition potentially leading to a double free of the
8442 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8444 * Malformed ECParameters causes infinite loop
8446 When processing an ECParameters structure OpenSSL enters an infinite loop
8447 if the curve specified is over a specially malformed binary polynomial
8450 This can be used to perform denial of service against any
8451 system which processes public keys, certificate requests or
8452 certificates. This includes TLS clients and TLS servers with
8453 client authentication enabled.
8455 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8460 * Exploitable out-of-bounds read in X509_cmp_time
8462 X509_cmp_time does not properly check the length of the ASN1_TIME
8463 string and can read a few bytes out of bounds. In addition,
8464 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8467 An attacker can use this to craft malformed certificates and CRLs of
8468 various sizes and potentially cause a segmentation fault, resulting in
8469 a DoS on applications that verify certificates or CRLs. TLS clients
8470 that verify CRLs are affected. TLS clients and servers with client
8471 authentication enabled may be affected if they use custom verification
8474 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8475 independently by Hanno Böck.
8480 * PKCS7 crash with missing EnvelopedContent
8482 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8483 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8484 with missing content and trigger a NULL pointer dereference on parsing.
8486 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8487 structures from untrusted sources are affected. OpenSSL clients and
8488 servers are not affected.
8490 This issue was reported to OpenSSL by Michal Zalewski (Google).
8495 * CMS verify infinite loop with unknown hash function
8497 When verifying a signedData message the CMS code can enter an infinite loop
8498 if presented with an unknown hash function OID. This can be used to perform
8499 denial of service against any system which verifies signedData messages using
8501 This issue was reported to OpenSSL by Johannes Bauer.
8506 * Race condition handling NewSessionTicket
8508 If a NewSessionTicket is received by a multi-threaded client when attempting to
8509 reuse a previous ticket then a race condition can occur potentially leading to
8510 a double free of the ticket data.
8515 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8517 * Segmentation fault in ASN1_TYPE_cmp fix
8519 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8520 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8521 certificate signature algorithm consistency this can be used to crash any
8522 certificate verification operation and exploited in a DoS attack. Any
8523 application which performs certificate verification is vulnerable including
8524 OpenSSL clients and servers which enable client authentication.
8529 * ASN.1 structure reuse memory corruption fix
8531 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8532 memory corruption via an invalid write. Such reuse is and has been
8533 strongly discouraged and is believed to be rare.
8535 Applications that parse structures containing CHOICE or ANY DEFINED BY
8536 components may be affected. Certificate parsing (d2i_X509 and related
8537 functions) are however not affected. OpenSSL clients and servers are
8543 * PKCS7 NULL pointer dereferences fix
8545 The PKCS#7 parsing code does not handle missing outer ContentInfo
8546 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8547 missing content and trigger a NULL pointer dereference on parsing.
8549 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8550 otherwise parse PKCS#7 structures from untrusted sources are
8551 affected. OpenSSL clients and servers are not affected.
8553 This issue was reported to OpenSSL by Michal Zalewski (Google).
8558 * DoS via reachable assert in SSLv2 servers fix
8560 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8561 servers that both support SSLv2 and enable export cipher suites by sending
8562 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8564 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8565 (OpenSSL development team).
8570 * Use After Free following d2i_ECPrivatekey error fix
8572 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8573 could cause a use after free condition. This, in turn, could cause a double
8574 free in several private key parsing functions (such as d2i_PrivateKey
8575 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8576 for applications that receive EC private keys from untrusted
8577 sources. This scenario is considered rare.
8579 This issue was discovered by the BoringSSL project and fixed in their
8585 * X509_to_X509_REQ NULL pointer deref fix
8587 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8588 the certificate key is invalid. This function is rarely used in practice.
8590 This issue was discovered by Brian Carpenter.
8595 * Removed the export ciphers from the DEFAULT ciphers
8599 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8601 * Build fixes for the Windows and OpenVMS platforms
8603 *Matt Caswell and Richard Levitte*
8605 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8607 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8608 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8609 dereference. This could lead to a Denial Of Service attack. Thanks to
8610 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8615 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8616 dtls1_buffer_record function under certain conditions. In particular this
8617 could occur if an attacker sent repeated DTLS records with the same
8618 sequence number but for the next epoch. The memory leak could be exploited
8619 by an attacker in a Denial of Service attack through memory exhaustion.
8620 Thanks to Chris Mueller for reporting this issue.
8625 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8626 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8627 method would be set to NULL which could later result in a NULL pointer
8628 dereference. Thanks to Frank Schmirler for reporting this issue.
8633 * Abort handshake if server key exchange message is omitted for ephemeral
8636 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8637 reporting this issue.
8642 * Remove non-export ephemeral RSA code on client and server. This code
8643 violated the TLS standard by allowing the use of temporary RSA keys in
8644 non-export ciphersuites and could be used by a server to effectively
8645 downgrade the RSA key length used to a value smaller than the server
8646 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8647 INRIA or reporting this issue.
8652 * Fixed issue where DH client certificates are accepted without verification.
8653 An OpenSSL server will accept a DH certificate for client authentication
8654 without the certificate verify message. This effectively allows a client to
8655 authenticate without the use of a private key. This only affects servers
8656 which trust a client certificate authority which issues certificates
8657 containing DH keys: these are extremely rare and hardly ever encountered.
8658 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8664 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8665 results on some platforms, including x86_64. This bug occurs at random
8666 with a very low probability, and is not known to be exploitable in any
8667 way, though its exact impact is difficult to determine. Thanks to Pieter
8668 Wuille (Blockstream) who reported this issue and also suggested an initial
8669 fix. Further analysis was conducted by the OpenSSL development team and
8670 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8671 the OpenSSL core team.
8676 * Fix various certificate fingerprint issues.
8678 By using non-DER or invalid encodings outside the signed portion of a
8679 certificate the fingerprint can be changed without breaking the signature.
8680 Although no details of the signed portion of the certificate can be changed
8681 this can cause problems with some applications: e.g. those using the
8682 certificate fingerprint for blacklists.
8684 1. Reject signatures with non zero unused bits.
8686 If the BIT STRING containing the signature has non zero unused bits reject
8687 the signature. All current signature algorithms require zero unused bits.
8689 2. Check certificate algorithm consistency.
8691 Check the AlgorithmIdentifier inside TBS matches the one in the
8692 certificate signature. NB: this will result in signature failure
8693 errors for some broken certificates.
8695 Thanks to Konrad Kraszewski from Google for reporting this issue.
8697 3. Check DSA/ECDSA signatures use DER.
8699 Reencode DSA/ECDSA signatures and compare with the original received
8700 signature. Return an error if there is a mismatch.
8702 This will reject various cases including garbage after signature
8703 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8704 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8705 (negative or with leading zeroes).
8707 Further analysis was conducted and fixes were developed by Stephen Henson
8708 of the OpenSSL core team.
8714 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8716 * Session Ticket Memory Leak.
8718 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8719 integrity of that ticket is first verified. In the event of a session
8720 ticket integrity check failing, OpenSSL will fail to free memory
8721 causing a memory leak. By sending a large number of invalid session
8722 tickets an attacker could exploit this issue in a Denial Of Service
8728 * Build option no-ssl3 is incomplete.
8730 When OpenSSL is configured with "no-ssl3" as a build option, servers
8731 could accept and complete an SSL 3.0 handshake, and clients could be
8732 configured to send them.
8735 *Akamai and the OpenSSL team*
8737 * Add support for TLS_FALLBACK_SCSV.
8738 Client applications doing fallback retries should call
8739 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8742 *Adam Langley, Bodo Moeller*
8744 * Add additional DigestInfo checks.
8746 Reencode DigestInto in DER and check against the original when
8747 verifying RSA signature: this will reject any improperly encoded
8748 DigestInfo structures.
8750 Note: this is a precautionary measure and no attacks are currently known.
8754 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8756 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8757 to a denial of service attack. A malicious server can crash the client
8758 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8759 ciphersuite and sending carefully crafted handshake messages.
8761 Thanks to Felix Gröbert (Google) for discovering and researching this
8767 * By sending carefully crafted DTLS packets an attacker could cause openssl
8768 to leak memory. This can be exploited through a Denial of Service attack.
8769 Thanks to Adam Langley for discovering and researching this issue.
8774 * An attacker can force openssl to consume large amounts of memory whilst
8775 processing DTLS handshake messages. This can be exploited through a
8776 Denial of Service attack.
8777 Thanks to Adam Langley for discovering and researching this issue.
8782 * An attacker can force an error condition which causes openssl to crash
8783 whilst processing DTLS packets due to memory being freed twice. This
8784 can be exploited through a Denial of Service attack.
8785 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8791 * If a multithreaded client connects to a malicious server using a resumed
8792 session and the server sends an ec point format extension it could write
8793 up to 255 bytes to freed memory.
8795 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8801 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8802 X509_name_oneline, X509_name_print_ex et al. to leak some information
8803 from the stack. Applications may be affected if they echo pretty printing
8804 output to the attacker.
8806 Thanks to Ivan Fratric (Google) for discovering this issue.
8809 *Emilia Käsper, and Steve Henson*
8811 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8812 for corner cases. (Certain input points at infinity could lead to
8813 bogus results, with non-infinity inputs mapped to infinity too.)
8817 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8819 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8820 handshake can force the use of weak keying material in OpenSSL
8821 SSL/TLS clients and servers.
8823 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8824 researching this issue. ([CVE-2014-0224])
8826 *KIKUCHI Masashi, Steve Henson*
8828 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8829 OpenSSL DTLS client the code can be made to recurse eventually crashing
8832 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8835 *Imre Rad, Steve Henson*
8837 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8838 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8839 client or server. This is potentially exploitable to run arbitrary
8840 code on a vulnerable client or server.
8842 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8844 *Jüri Aedla, Steve Henson*
8846 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8847 are subject to a denial of service attack.
8849 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8850 this issue. ([CVE-2014-3470])
8852 *Felix Gröbert, Ivan Fratric, Steve Henson*
8854 * Harmonize version and its documentation. -f flag is used to display
8857 *mancha <mancha1@zoho.com>*
8859 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8860 in i2d_ECPrivateKey.
8862 *mancha <mancha1@zoho.com>*
8864 * Fix some double frees. These are not thought to be exploitable.
8866 *mancha <mancha1@zoho.com>*
8868 * Fix for the attack described in the paper "Recovering OpenSSL
8869 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8870 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8871 <http://eprint.iacr.org/2014/140>
8873 Thanks to Yuval Yarom and Naomi Benger for discovering this
8874 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8876 *Yuval Yarom and Naomi Benger*
8878 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8880 * Keep original DTLS digest and encryption contexts in retransmission
8881 structures so we can use the previous session parameters if they need
8882 to be resent. ([CVE-2013-6450])
8886 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8887 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8888 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8889 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8890 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8891 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8893 *Rob Stradling, Adam Langley*
8895 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8897 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8899 This addresses the flaw in CBC record processing discovered by
8900 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8901 at: <http://www.isg.rhul.ac.uk/tls/>
8903 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8904 Security Group at Royal Holloway, University of London
8905 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8906 Emilia Käsper for the initial patch.
8909 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8911 * Return an error when checking OCSP signatures when key is NULL.
8912 This fixes a DoS attack. ([CVE-2013-0166])
8916 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8917 the right response is stapled. Also change SSL_get_certificate()
8918 so it returns the certificate actually sent.
8919 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8920 (This is a backport)
8922 *Rob Stradling <rob.stradling@comodo.com>*
8924 * Fix possible deadlock when decoding public keys.
8928 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8930 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8933 * Sanity check record length before skipping explicit IV in DTLS
8936 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8937 fuzzing as a service testing platform.
8942 * Initialise tkeylen properly when encrypting CMS messages.
8943 Thanks to Solar Designer of Openwall for reporting this issue.
8947 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8949 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8950 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8951 in CRYPTO_realloc_clean.
8953 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8954 issue and to Adam Langley <agl@chromium.org> for fixing it.
8957 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8959 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8961 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8962 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8963 content decryption and always return the same error. Note: this attack
8964 needs on average 2^20 messages so it only affects automated senders. The
8965 old behaviour can be re-enabled in the CMS code by setting the
8966 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8967 an MMA defence is not necessary.
8968 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8969 this issue. ([CVE-2012-0884])
8973 * Fix CVE-2011-4619: make sure we really are receiving a
8974 client hello before rejecting multiple SGC restarts. Thanks to
8975 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8979 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8981 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8982 Thanks to Antonio Martin, Enterprise Secure Access Research and
8983 Development, Cisco Systems, Inc. for discovering this bug and
8984 preparing a fix. ([CVE-2012-0050])
8988 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8990 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8991 of the Vaudenay padding oracle attack on CBC mode encryption
8992 which enables an efficient plaintext recovery attack against
8993 the OpenSSL implementation of DTLS. Their attack exploits timing
8994 differences arising during decryption processing. A research
8995 paper describing this attack can be found at:
8996 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8997 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8998 Security Group at Royal Holloway, University of London
8999 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9000 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9001 for preparing the fix. ([CVE-2011-4108])
9003 *Robin Seggelmann, Michael Tuexen*
9005 * Clear bytes used for block padding of SSL 3.0 records.
9008 *Adam Langley (Google)*
9010 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9011 Kadianakis <desnacked@gmail.com> for discovering this issue and
9012 Adam Langley for preparing the fix. ([CVE-2011-4619])
9014 *Adam Langley (Google)*
9016 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9018 *Andrey Kulikov <amdeich@gmail.com>*
9020 * Prevent malformed RFC3779 data triggering an assertion failure.
9021 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9022 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9024 *Rob Austein <sra@hactrn.net>*
9026 * Improved PRNG seeding for VOS.
9028 *Paul Green <Paul.Green@stratus.com>*
9030 * Fix ssl_ciph.c set-up race.
9032 *Adam Langley (Google)*
9034 * Fix spurious failures in ecdsatest.c.
9036 *Emilia Käsper (Google)*
9038 * Fix the BIO_f_buffer() implementation (which was mixing different
9039 interpretations of the `..._len` fields).
9041 *Adam Langley (Google)*
9043 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9044 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9045 threads won't reuse the same blinding coefficients.
9047 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9048 lock to call BN_BLINDING_invert_ex, and avoids one use of
9049 BN_BLINDING_update for each BN_BLINDING structure (previously,
9050 the last update always remained unused).
9052 *Emilia Käsper (Google)*
9054 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9056 *Bob Buckholz (Google)*
9058 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9060 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9061 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9063 *Kaspar Brand <ossl@velox.ch>*
9065 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9066 for multi-threaded use of ECDH. ([CVE-2011-3210])
9068 *Adam Langley (Google)*
9070 * Fix x509_name_ex_d2i memory leak on bad inputs.
9074 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9075 signature public key algorithm by using OID xref utilities instead.
9076 Before this you could only use some ECC ciphersuites with SHA1 only.
9080 * Add protection against ECDSA timing attacks as mentioned in the paper
9081 by Billy Bob Brumley and Nicola Tuveri, see:
9082 <http://eprint.iacr.org/2011/232.pdf>
9084 *Billy Bob Brumley and Nicola Tuveri*
9086 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9088 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9090 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9092 * Fix bug in string printing code: if *any* escaping is enabled we must
9093 escape the escape character (backslash) or the resulting string is
9098 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9100 * Disable code workaround for ancient and obsolete Netscape browsers
9101 and servers: an attacker can use it in a ciphersuite downgrade attack.
9102 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9106 * Fixed J-PAKE implementation error, originally discovered by
9107 Sebastien Martini, further info and confirmation from Stefan
9108 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9112 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9114 * Fix extension code to avoid race conditions which can result in a buffer
9115 overrun vulnerability: resumed sessions must not be modified as they can
9116 be shared by multiple threads. CVE-2010-3864
9120 * Fix WIN32 build system to correctly link an ENGINE directory into
9125 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9127 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9130 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9132 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9134 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9135 context. The operation can be customised via the ctrl mechanism in
9136 case ENGINEs want to include additional functionality.
9140 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9144 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9145 output hashes compatible with older versions of OpenSSL.
9147 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9149 * Fix compression algorithm handling: if resuming a session use the
9150 compression algorithm of the resumed session instead of determining
9151 it from client hello again. Don't allow server to change algorithm.
9155 * Add load_crls() function to commands tidying load_certs() too. Add option
9156 to verify utility to allow additional CRLs to be included.
9160 * Update OCSP request code to permit adding custom headers to the request:
9161 some responders need this.
9165 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9168 *Julia Lawall <julia@diku.dk>*
9170 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9171 needlessly dereferenced structures, used obsolete functions and
9172 didn't handle all updated verify codes correctly.
9176 * Disable MD2 in the default configuration.
9180 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9181 indicate the initial BIO being pushed or popped. This makes it possible
9182 to determine whether the BIO is the one explicitly called or as a result
9183 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9184 it handles reference counts correctly and doesn't zero out the I/O bio
9185 when it is not being explicitly popped. WARNING: applications which
9186 included workarounds for the old buggy behaviour will need to be modified
9187 or they could free up already freed BIOs.
9191 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9192 renaming to all platforms (within the 0.9.8 branch, this was
9193 done conditionally on Netware platforms to avoid a name clash).
9195 *Guenter <lists@gknw.net>*
9197 * Add ECDHE and PSK support to DTLS.
9199 *Michael Tuexen <tuexen@fh-muenster.de>*
9201 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9206 * Add "missing" function EVP_MD_flags() (without this the only way to
9207 retrieve a digest flags is by accessing the structure directly. Update
9208 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9209 or cipher is registered as in the "from" argument. Print out all
9210 registered digests in the dgst usage message instead of manually
9211 attempting to work them out.
9215 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9216 this allows the use of compression and extensions. Change default cipher
9217 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9218 by default unless an application cipher string requests it.
9222 * Alter match criteria in PKCS12_parse(). It used to try to use local
9223 key ids to find matching certificates and keys but some PKCS#12 files
9224 don't follow the (somewhat unwritten) rules and this strategy fails.
9225 Now just gather all certificates together and the first private key
9226 then look for the first certificate that matches the key.
9230 * Support use of registered digest and cipher names for dgst and cipher
9231 commands instead of having to add each one as a special case. So now
9238 openssl dgst -sha256 foo
9240 and this works for ENGINE based algorithms too.
9244 * Update Gost ENGINE to support parameter files.
9246 *Victor B. Wagner <vitus@cryptocom.ru>*
9248 * Support GeneralizedTime in ca utility.
9250 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9252 * Enhance the hash format used for certificate directory links. The new
9253 form uses the canonical encoding (meaning equivalent names will work
9254 even if they aren't identical) and uses SHA1 instead of MD5. This form
9255 is incompatible with the older format and as a result c_rehash should
9256 be used to rebuild symbolic links.
9260 * Make PKCS#8 the default write format for private keys, replacing the
9261 traditional format. This form is standardised, more secure and doesn't
9262 include an implicit MD5 dependency.
9266 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9267 committed to OpenSSL should pass this lot as a minimum.
9271 * Add session ticket override functionality for use by EAP-FAST.
9273 *Jouni Malinen <j@w1.fi>*
9275 * Modify HMAC functions to return a value. Since these can be implemented
9276 in an ENGINE errors can occur.
9280 * Type-checked OBJ_bsearch_ex.
9284 * Type-checked OBJ_bsearch. Also some constification necessitated
9285 by type-checking. Still to come: TXT_DB, bsearch(?),
9286 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9291 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9292 seconds to a tm structure directly, instead of going through OS
9293 specific date routines. This avoids any issues with OS routines such
9294 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9295 and X509_time_adj_ex() to cover the extended range. The existing
9296 X509_time_adj() is still usable and will no longer have any date issues.
9300 * Delta CRL support. New use deltas option which will attempt to locate
9301 and search any appropriate delta CRLs available.
9303 This work was sponsored by Google.
9307 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9308 code and add additional score elements. Validate alternate CRL paths
9309 as part of the CRL checking and indicate a new error "CRL path validation
9310 error" in this case. Applications wanting additional details can use
9311 the verify callback and check the new "parent" field. If this is not
9312 NULL CRL path validation is taking place. Existing applications won't
9313 see this because it requires extended CRL support which is off by
9316 This work was sponsored by Google.
9320 * Support for freshest CRL extension.
9322 This work was sponsored by Google.
9326 * Initial indirect CRL support. Currently only supported in the CRLs
9327 passed directly and not via lookup. Process certificate issuer
9328 CRL entry extension and lookup CRL entries by bother issuer name
9329 and serial number. Check and process CRL issuer entry in IDP extension.
9331 This work was sponsored by Google.
9335 * Add support for distinct certificate and CRL paths. The CRL issuer
9336 certificate is validated separately in this case. Only enabled if
9337 an extended CRL support flag is set: this flag will enable additional
9338 CRL functionality in future.
9340 This work was sponsored by Google.
9344 * Add support for policy mappings extension.
9346 This work was sponsored by Google.
9350 * Fixes to pathlength constraint, self issued certificate handling,
9351 policy processing to align with RFC3280 and PKITS tests.
9353 This work was sponsored by Google.
9357 * Support for name constraints certificate extension. DN, email, DNS
9358 and URI types are currently supported.
9360 This work was sponsored by Google.
9364 * To cater for systems that provide a pointer-based thread ID rather
9365 than numeric, deprecate the current numeric thread ID mechanism and
9366 replace it with a structure and associated callback type. This
9367 mechanism allows a numeric "hash" to be extracted from a thread ID in
9368 either case, and on platforms where pointers are larger than 'long',
9369 mixing is done to help ensure the numeric 'hash' is usable even if it
9370 can't be guaranteed unique. The default mechanism is to use "&errno"
9371 as a pointer-based thread ID to distinguish between threads.
9373 Applications that want to provide their own thread IDs should now use
9374 CRYPTO_THREADID_set_callback() to register a callback that will call
9375 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9377 Note that ERR_remove_state() is now deprecated, because it is tied
9378 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9379 to free the current thread's error state should be replaced by
9380 ERR_remove_thread_state(NULL).
9382 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9383 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9384 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9385 application was previously providing a numeric thread callback that
9386 was inappropriate for distinguishing threads, then uniqueness might
9387 have been obtained with &errno that happened immediately in the
9388 intermediate development versions of OpenSSL; this is no longer the
9389 case, the numeric thread callback will now override the automatic use
9392 *Geoff Thorpe, with help from Bodo Moeller*
9394 * Initial support for different CRL issuing certificates. This covers a
9395 simple case where the self issued certificates in the chain exist and
9396 the real CRL issuer is higher in the existing chain.
9398 This work was sponsored by Google.
9402 * Removed effectively defunct crypto/store from the build.
9406 * Revamp of STACK to provide stronger type-checking. Still to come:
9407 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9408 ASN1_STRING, CONF_VALUE.
9412 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9413 RAM on SSL connections. This option can save about 34k per idle SSL.
9417 * Revamp of LHASH to provide stronger type-checking. Still to come:
9418 STACK, TXT_DB, bsearch, qsort.
9422 * Initial support for Cryptographic Message Syntax (aka CMS) based
9423 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9424 support for data, signedData, compressedData, digestedData and
9425 encryptedData, envelopedData types included. Scripts to check against
9426 RFC4134 examples draft and interop and consistency checks of many
9427 content types and variants.
9431 * Add options to enc utility to support use of zlib compression BIO.
9435 * Extend mk1mf to support importing of options and assembly language
9436 files from Configure script, currently only included in VC-WIN32.
9437 The assembly language rules can now optionally generate the source
9438 files from the associated perl scripts.
9442 * Implement remaining functionality needed to support GOST ciphersuites.
9443 Interop testing has been performed using CryptoPro implementations.
9445 *Victor B. Wagner <vitus@cryptocom.ru>*
9447 * s390x assembler pack.
9451 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9456 * Implement Opaque PRF Input TLS extension as specified in
9457 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9458 official specification yet and no extension type assignment by
9459 IANA exists, this extension (for now) will have to be explicitly
9460 enabled when building OpenSSL by providing the extension number
9461 to use. For example, specify an option
9463 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9465 to the "config" or "Configure" script to enable the extension,
9466 assuming extension number 0x9527 (which is a completely arbitrary
9467 and unofficial assignment based on the MD5 hash of the Internet
9468 Draft). Note that by doing so, you potentially lose
9469 interoperability with other TLS implementations since these might
9470 be using the same extension number for other purposes.
9472 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9473 opaque PRF input value to use in the handshake. This will create
9474 an internal copy of the length-'len' string at 'src', and will
9475 return non-zero for success.
9477 To get more control and flexibility, provide a callback function
9480 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9481 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9485 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9488 Callback function 'cb' will be called in handshakes, and is
9489 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9490 Argument 'arg' is for application purposes (the value as given to
9491 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9492 be provided to the callback function). The callback function
9493 has to return non-zero to report success: usually 1 to use opaque
9494 PRF input just if possible, or 2 to enforce use of the opaque PRF
9495 input. In the latter case, the library will abort the handshake
9496 if opaque PRF input is not successfully negotiated.
9498 Arguments 'peerinput' and 'len' given to the callback function
9499 will always be NULL and 0 in the case of a client. A server will
9500 see the client's opaque PRF input through these variables if
9501 available (NULL and 0 otherwise). Note that if the server
9502 provides an opaque PRF input, the length must be the same as the
9503 length of the client's opaque PRF input.
9505 Note that the callback function will only be called when creating
9506 a new session (session resumption can resume whatever was
9507 previously negotiated), and will not be called in SSL 2.0
9508 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9509 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9510 for applications that need to enforce opaque PRF input.
9514 * Update ssl code to support digests other than SHA1+MD5 for handshake
9517 *Victor B. Wagner <vitus@cryptocom.ru>*
9519 * Add RFC4507 support to OpenSSL. This includes the corrections in
9520 RFC4507bis. The encrypted ticket format is an encrypted encoded
9521 SSL_SESSION structure, that way new session features are automatically
9524 If a client application caches session in an SSL_SESSION structure
9525 support is transparent because tickets are now stored in the encoded
9528 The SSL_CTX structure automatically generates keys for ticket
9529 protection in servers so again support should be possible
9530 with no application modification.
9532 If a client or server wishes to disable RFC4507 support then the option
9533 SSL_OP_NO_TICKET can be set.
9535 Add a TLS extension debugging callback to allow the contents of any client
9536 or server extensions to be examined.
9538 This work was sponsored by Google.
9542 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9543 OpenSSL should now compile cleanly on gcc 4.2
9545 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9547 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9548 support including streaming MAC support: this is required for GOST
9549 ciphersuite support.
9551 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9553 * Add option -stream to use PKCS#7 streaming in smime utility. New
9554 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9555 to output in BER and PEM format.
9559 * Experimental support for use of HMAC via EVP_PKEY interface. This
9560 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9561 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9562 ENGINE support for HMAC keys which are unextractable. New -mac and
9563 -macopt options to dgst utility.
9567 * New option -sigopt to dgst utility. Update dgst to use
9568 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9569 alternative signing parameters such as X9.31 or PSS in the dgst
9574 * Change ssl_cipher_apply_rule(), the internal function that does
9575 the work each time a ciphersuite string requests enabling
9576 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9577 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9578 the order of disabled ciphersuites such that those ciphersuites
9579 that most recently went from enabled to disabled not only stay
9580 in order with respect to each other, but also have higher priority
9581 than other disabled ciphersuites the next time ciphersuites are
9584 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9585 the same ciphersuites as with "HIGH" alone, but in a specific
9586 order where the PSK ciphersuites come first (since they are the
9587 most recently disabled ciphersuites when "HIGH" is parsed).
9589 Also, change ssl_create_cipher_list() (using this new
9590 functionality) such that between otherwise identical
9591 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9596 * Change ssl_create_cipher_list() so that it automatically
9597 arranges the ciphersuites in reasonable order before starting
9598 to process the rule string. Thus, the definition for "DEFAULT"
9599 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9600 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9601 This makes it much easier to arrive at a reasonable default order
9602 in applications for which anonymous ciphers are OK (meaning
9603 that you can't actually use DEFAULT).
9605 *Bodo Moeller; suggested by Victor Duchovni*
9607 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9608 processing) into multiple integers instead of setting
9609 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9610 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9611 (These masks as well as the individual bit definitions are hidden
9612 away into the non-exported interface ssl/ssl_locl.h, so this
9613 change to the definition of the SSL_CIPHER structure shouldn't
9614 affect applications.) This give us more bits for each of these
9615 categories, so there is no longer a need to coagulate AES128 and
9616 AES256 into a single algorithm bit, and to coagulate Camellia128
9617 and Camellia256 into a single algorithm bit, which has led to all
9620 Thus, among other things, the kludge introduced in 0.9.7m and
9621 0.9.8e for masking out AES256 independently of AES128 or masking
9622 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9624 With the change, we also introduce new ciphersuite aliases that
9625 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9630 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9631 Use the leftmost N bytes of the signature input if the input is
9632 larger than the prime q (with N being the size in bytes of q).
9636 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9637 it yet and it is largely untested.
9641 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9645 * Initial incomplete changes to avoid need for function casts in OpenSSL
9646 some compilers (gcc 4.2 and later) reject their use. Safestack is
9647 reimplemented. Update ASN1 to avoid use of legacy functions.
9651 * Win32/64 targets are linked with Winsock2.
9655 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9656 to external functions. This can be used to increase CRL handling
9657 efficiency especially when CRLs are very large by (for example) storing
9658 the CRL revoked certificates in a database.
9662 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9663 new CRLs added to a directory can be used. New command line option
9664 -verify_return_error to s_client and s_server. This causes real errors
9665 to be returned by the verify callback instead of carrying on no matter
9666 what. This reflects the way a "real world" verify callback would behave.
9670 * GOST engine, supporting several GOST algorithms and public key formats.
9671 Kindly donated by Cryptocom.
9675 * Partial support for Issuing Distribution Point CRL extension. CRLs
9676 partitioned by DP are handled but no indirect CRL or reason partitioning
9677 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9678 selected via a scoring technique which handles IDP and AKID in CRLs.
9682 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9683 will ultimately be used for all verify operations: this will remove the
9684 X509_STORE dependency on certificate verification and allow alternative
9685 lookup methods. X509_STORE based implementations of these two callbacks.
9689 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9690 Modify get_crl() to find a valid (unexpired) CRL if possible.
9694 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9695 this would be called X509_CRL_cmp() but that name is already used by
9696 a function that just compares CRL issuer names. Cache several CRL
9697 extensions in X509_CRL structure and cache CRLDP in X509.
9701 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9702 this maps equivalent X509_NAME structures into a consistent structure.
9703 Name comparison can then be performed rapidly using memcmp().
9707 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9712 * Allow digests to supply their own micalg string for S/MIME type using
9713 the ctrl EVP_MD_CTRL_MICALG.
9717 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9718 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9719 ctrl. It can then customise the structure before and/or after signing
9724 * New function OBJ_add_sigid() to allow application defined signature OIDs
9725 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9726 to free up any added signature OIDs.
9730 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9731 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9732 digest and cipher tables. New options added to openssl utility:
9733 list-message-digest-algorithms and list-cipher-algorithms.
9737 * Change the array representation of binary polynomials: the list
9738 of degrees of non-zero coefficients is now terminated with -1.
9739 Previously it was terminated with 0, which was also part of the
9740 value; thus, the array representation was not applicable to
9741 polynomials where t^0 has coefficient zero. This change makes
9742 the array representation useful in a more general context.
9746 * Various modifications and fixes to SSL/TLS cipher string
9747 handling. For ECC, the code now distinguishes between fixed ECDH
9748 with RSA certificates on the one hand and with ECDSA certificates
9749 on the other hand, since these are separate ciphersuites. The
9750 unused code for Fortezza ciphersuites has been removed.
9752 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9753 (not "ECDHE"). For consistency with the code for DH
9754 certificates, use of ECDH certificates is now considered ECDH
9755 authentication, not RSA or ECDSA authentication (the latter is
9756 merely the CA's signing algorithm and not actively used in the
9759 The temporary ciphersuite alias "ECCdraft" is no longer
9760 available, and ECC ciphersuites are no longer excluded from "ALL"
9761 and "DEFAULT". The following aliases now exist for RFC 4492
9762 ciphersuites, most of these by analogy with the DH case:
9764 kECDHr - ECDH cert, signed with RSA
9765 kECDHe - ECDH cert, signed with ECDSA
9766 kECDH - ECDH cert (signed with either RSA or ECDSA)
9767 kEECDH - ephemeral ECDH
9768 ECDH - ECDH cert or ephemeral ECDH
9774 AECDH - anonymous ECDH
9775 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9779 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9780 Use correct micalg parameters depending on digest(s) in signed message.
9784 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9785 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9789 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9790 an engine to register a method. Add ENGINE lookups for methods and
9791 functional reference processing.
9795 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9796 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9801 * New -resign option to smime utility. This adds one or more signers
9802 to an existing PKCS#7 signedData structure. Also -md option to use an
9803 alternative message digest algorithm for signing.
9807 * Tidy up PKCS#7 routines and add new functions to make it easier to
9808 create PKCS7 structures containing multiple signers. Update smime
9809 application to support multiple signers.
9813 * New -macalg option to pkcs12 utility to allow setting of an alternative
9818 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9819 Reorganize PBE internals to lookup from a static table using NIDs,
9820 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9821 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9822 PRF which will be automatically used with PBES2.
9826 * Replace the algorithm specific calls to generate keys in "req" with the
9831 * Update PKCS#7 enveloped data routines to use new API. This is now
9832 supported by any public key method supporting the encrypt operation. A
9833 ctrl is added to allow the public key algorithm to examine or modify
9834 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9839 * Add a ctrl to asn1 method to allow a public key algorithm to express
9840 a default digest type to use. In most cases this will be SHA1 but some
9841 algorithms (such as GOST) need to specify an alternative digest. The
9842 return value indicates how strong the preference is 1 means optional and
9843 2 is mandatory (that is it is the only supported type). Modify
9844 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9845 use the default md. Update openssl utilities to use the default digest
9846 type for signing if it is not explicitly indicated.
9850 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9851 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9852 signing method from the key type. This effectively removes the link
9853 between digests and public key types.
9857 * Add an OID cross reference table and utility functions. Its purpose is to
9858 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9859 rsaEncryption. This will allow some of the algorithm specific hackery
9860 needed to use the correct OID to be removed.
9864 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9865 structures for PKCS7_sign(). They are now set up by the relevant public
9870 * Add provisional EC pkey method with support for ECDSA and ECDH.
9874 * Add support for key derivation (agreement) in the API, DH method and
9879 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9880 public and private key formats. As a side effect these add additional
9881 command line functionality not previously available: DSA signatures can be
9882 generated and verified using pkeyutl and DH key support and generation in
9889 *Oliver Tappe <zooey@hirschkaefer.de>*
9891 * New make target "install_html_docs" installs HTML renditions of the
9894 *Oliver Tappe <zooey@hirschkaefer.de>*
9896 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9897 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9898 support key and parameter generation and add initial key generation
9899 functionality for RSA.
9903 * Add functions for main EVP_PKEY_method operations. The undocumented
9904 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9905 `EVP_PKEY_{encrypt,decrypt}_old`.
9909 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9910 key API, doesn't do much yet.
9914 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9915 public key algorithms. New option to openssl utility:
9916 "list-public-key-algorithms" to print out info.
9920 * Implement the Supported Elliptic Curves Extension for
9921 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9925 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9926 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9930 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9931 utilities such as rsa, dsa, dsaparam etc except they process any key
9936 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9937 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9938 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9943 * Initial support for pluggable public key ASN1.
9944 De-spaghettify the public key ASN1 handling. Move public and private
9945 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9946 algorithm specific handling to a single module within the relevant
9947 algorithm directory. Add functions to allow (near) opaque processing
9948 of public and private key structures.
9952 * Implement the Supported Point Formats Extension for
9953 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9957 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9958 for the psk identity [hint] and the psk callback functions to the
9959 SSL_SESSION, SSL and SSL_CTX structure.
9962 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9966 SSL_CTX_use_psk_identity_hint
9967 SSL_get_psk_identity_hint
9968 SSL_get_psk_identity
9969 SSL_use_psk_identity_hint
9971 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9973 * Add RFC 3161 compliant time stamp request creation, response generation
9974 and response verification functionality.
9976 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9978 * Add initial support for TLS extensions, specifically for the server_name
9979 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9980 have new members for a hostname. The SSL data structure has an
9981 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9982 stored in that context to allow for session resumption, even after the
9983 SSL has been switched to a new SSL_CTX in reaction to a client's
9984 server_name extension.
9986 New functions (subject to change):
9988 SSL_get_servername()
9989 SSL_get_servername_type()
9992 New CTRL codes and macros (subject to change):
9994 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9995 - SSL_CTX_set_tlsext_servername_callback()
9996 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9997 - SSL_CTX_set_tlsext_servername_arg()
9998 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10000 openssl s_client has a new '-servername ...' option.
10002 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10003 '-key2 ...', '-servername_fatal' (subject to change). This allows
10004 testing the HostName extension for a specific single hostname ('-cert'
10005 and '-key' remain fallbacks for handshakes without HostName
10006 negotiation). If the unrecognized_name alert has to be sent, this by
10007 default is a warning; it becomes fatal with the '-servername_fatal'
10010 *Peter Sylvester, Remy Allais, Christophe Renou*
10012 * Whirlpool hash implementation is added.
10016 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10017 bn(64,32). Because of instruction set limitations it doesn't have
10018 any negative impact on performance. This was done mostly in order
10019 to make it possible to share assembler modules, such as bn_mul_mont
10020 implementations, between 32- and 64-bit builds without hassle.
10024 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10025 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10030 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10031 dedicated Montgomery multiplication procedure, is introduced.
10032 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10033 "64-bit" performance on certain 32-bit targets.
10037 * New option SSL_OP_NO_COMP to disable use of compression selectively
10038 in SSL structures. New SSL ctrl to set maximum send fragment size.
10039 Save memory by setting the I/O buffer sizes dynamically instead of
10040 using the maximum available value.
10044 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10045 in addition to the text details.
10049 * Very, very preliminary EXPERIMENTAL support for printing of general
10050 ASN1 structures. This currently produces rather ugly output and doesn't
10051 handle several customised structures at all.
10055 * Integrated support for PVK file format and some related formats such
10056 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10057 these in the 'rsa' and 'dsa' utilities.
10061 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10065 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10066 place for the (very old) "NETSCAPE" format certificates which are now
10067 handled using new ASN1 code equivalents.
10071 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10072 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10073 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10077 * Modify CRL distribution points extension code to print out previously
10078 unsupported fields. Enhance extension setting code to allow setting of
10083 * Add print and set support for Issuing Distribution Point CRL extension.
10087 * Change 'Configure' script to enable Camellia by default.
10094 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10096 * When rejecting SSL/TLS records due to an incorrect version number, never
10097 update s->server with a new major version number. As of
10098 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10099 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10100 the previous behavior could result in a read attempt at NULL when
10101 receiving specific incorrect SSL/TLS records once record payload
10102 protection is active. ([CVE-2010-0740])
10104 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10106 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10107 could be crashed if the relevant tables were not present (e.g. chrooted).
10109 *Tomas Hoger <thoger@redhat.com>*
10111 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10113 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10115 *Martin Olsson, Neel Mehta*
10117 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10118 accommodate for stack sorting, always a write lock!).
10122 * On some versions of WIN32 Heap32Next is very slow. This can cause
10123 excessive delays in the RAND_poll(): over a minute. As a workaround
10124 include a time check in the inner Heap32Next loop too.
10128 * The code that handled flushing of data in SSL/TLS originally used the
10129 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10130 the problem outlined in PR#1949. The fix suggested there however can
10131 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10132 of Apache). So instead simplify the code to flush unconditionally.
10133 This should be fine since flushing with no data to flush is a no op.
10137 * Handle TLS versions 2.0 and later properly and correctly use the
10138 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10139 off ancient servers have a habit of sticking around for a while...
10143 * Modify compression code so it frees up structures without using the
10144 ex_data callbacks. This works around a problem where some applications
10145 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10146 restarting) then use compression (e.g. SSL with compression) later.
10147 This results in significant per-connection memory leaks and
10148 has caused some security issues including CVE-2008-1678 and
10153 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10154 change when encrypting or decrypting.
10158 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10159 connect and renegotiate with servers which do not support RI.
10160 Until RI is more widely deployed this option is enabled by default.
10164 * Add "missing" ssl ctrls to clear options and mode.
10168 * If client attempts to renegotiate and doesn't support RI respond with
10169 a no_renegotiation alert as required by RFC5746. Some renegotiating
10170 TLS clients will continue a connection gracefully when they receive
10171 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10172 waiting for a server hello which it will never receive. Now we treat a
10173 received no_renegotiation alert as a fatal error. This is because
10174 applications requesting a renegotiation might well expect it to succeed
10175 and would have no code in place to handle the server denying it so the
10176 only safe thing to do is to terminate the connection.
10180 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10181 peer supports secure renegotiation and 0 otherwise. Print out peer
10182 renegotiation support in s_client/s_server.
10186 * Replace the highly broken and deprecated SPKAC certification method with
10187 the updated NID creation version. This should correctly handle UTF8.
10191 * Implement RFC5746. Re-enable renegotiation but require the extension
10192 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10193 turns out to be a bad idea. It has been replaced by
10194 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10195 SSL_CTX_set_options(). This is really not recommended unless you
10196 know what you are doing.
10198 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10200 * Fixes to stateless session resumption handling. Use initial_ctx when
10201 issuing and attempting to decrypt tickets in case it has changed during
10202 servername handling. Use a non-zero length session ID when attempting
10203 stateless session resumption: this makes it possible to determine if
10204 a resumption has occurred immediately after receiving server hello
10205 (several places in OpenSSL subtly assume this) instead of later in
10210 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10211 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10212 fixes for a few places where the return code is not checked
10215 *Julia Lawall <julia@diku.dk>*
10217 * Add --strict-warnings option to Configure script to include devteam
10218 warnings in other configurations.
10222 * Add support for --libdir option and LIBDIR variable in makefiles. This
10223 makes it possible to install openssl libraries in locations which
10224 have names other than "lib", for example "/usr/lib64" which some
10227 *Steve Henson, based on patch from Jeremy Utley*
10229 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10230 X690 8.9.12 and can produce some misleading textual output of OIDs.
10232 *Steve Henson, reported by Dan Kaminsky*
10234 * Delete MD2 from algorithm tables. This follows the recommendation in
10235 several standards that it is not used in new applications due to
10236 several cryptographic weaknesses. For binary compatibility reasons
10237 the MD2 API is still compiled in by default.
10241 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10246 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10247 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10250 *Guenter <lists@gknw.net>*
10252 * Fix the server certificate chain building code to use X509_verify_cert(),
10253 it used to have an ad-hoc builder which was unable to cope with anything
10254 other than a simple chain.
10256 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10258 * Don't check self signed certificate signatures in X509_verify_cert()
10259 by default (a flag can override this): it just wastes time without
10260 adding any security. As a useful side effect self signed root CAs
10261 with non-FIPS digests are now usable in FIPS mode.
10265 * In dtls1_process_out_of_seq_message() the check if the current message
10266 is already buffered was missing. For every new message was memory
10267 allocated, allowing an attacker to perform an denial of service attack
10268 with sending out of seq handshake messages until there is no memory
10269 left. Additionally every future message was buffered, even if the
10270 sequence number made no sense and would be part of another handshake.
10271 So only messages with sequence numbers less than 10 in advance will be
10272 buffered. ([CVE-2009-1378])
10274 *Robin Seggelmann, discovered by Daniel Mentz*
10276 * Records are buffered if they arrive with a future epoch to be
10277 processed after finishing the corresponding handshake. There is
10278 currently no limitation to this buffer allowing an attacker to perform
10279 a DOS attack with sending records with future epochs until there is no
10280 memory left. This patch adds the pqueue_size() function to determine
10281 the size of a buffer and limits the record buffer to 100 entries.
10284 *Robin Seggelmann, discovered by Daniel Mentz*
10286 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10287 parent structure is freed. ([CVE-2009-1379])
10291 * Handle non-blocking I/O properly in SSL_shutdown() call.
10293 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10295 * Add `2.5.4.*` OIDs
10297 *Ilya O. <vrghost@gmail.com>*
10299 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10301 * Disable renegotiation completely - this fixes a severe security
10302 problem ([CVE-2009-3555]) at the cost of breaking all
10303 renegotiation. Renegotiation can be re-enabled by setting
10304 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10305 run-time. This is really not recommended unless you know what
10310 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10312 * Don't set val to NULL when freeing up structures, it is freed up by
10313 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10314 zeroing past the valid field. ([CVE-2009-0789])
10316 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10318 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10319 checked correctly. This would allow some invalid signed attributes to
10320 appear to verify correctly. ([CVE-2009-0591])
10322 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10324 * Reject UniversalString and BMPString types with invalid lengths. This
10325 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10326 a legal length. ([CVE-2009-0590])
10330 * Set S/MIME signing as the default purpose rather than setting it
10331 unconditionally. This allows applications to override it at the store
10336 * Permit restricted recursion of ASN1 strings. This is needed in practice
10337 to handle some structures.
10341 * Improve efficiency of mem_gets: don't search whole buffer each time
10344 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10346 * New -hex option for openssl rand.
10350 * Print out UTF8String and NumericString when parsing ASN1.
10354 * Support NumericString type for name components.
10358 * Allow CC in the environment to override the automatically chosen
10359 compiler. Note that nothing is done to ensure flags work with the
10364 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10366 * Properly check EVP_VerifyFinal() and similar return values
10369 *Ben Laurie, Bodo Moeller, Google Security Team*
10371 * Enable TLS extensions by default.
10375 * Allow the CHIL engine to be loaded, whether the application is
10376 multithreaded or not. (This does not release the developer from the
10377 obligation to set up the dynamic locking callbacks.)
10379 *Sander Temme <sander@temme.net>*
10381 * Use correct exit code if there is an error in dgst command.
10383 *Steve Henson; problem pointed out by Roland Dirlewanger*
10385 * Tweak Configure so that you need to say "experimental-jpake" to enable
10386 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10390 * Add experimental JPAKE support, including demo authentication in
10391 s_client and s_server.
10395 * Set the comparison function in v3_addr_canonize().
10397 *Rob Austein <sra@hactrn.net>*
10399 * Add support for XMPP STARTTLS in s_client.
10401 *Philip Paeps <philip@freebsd.org>*
10403 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10404 to ensure that even with this option, only ciphersuites in the
10405 server's preference list will be accepted. (Note that the option
10406 applies only when resuming a session, so the earlier behavior was
10407 just about the algorithm choice for symmetric cryptography.)
10411 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10413 * Fix NULL pointer dereference if a DTLS server received
10414 ChangeCipherSpec as first record ([CVE-2009-1386]).
10418 * Fix a state transition in s3_srvr.c and d1_srvr.c
10419 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10421 *Nagendra Modadugu*
10423 * The fix in 0.9.8c that supposedly got rid of unsafe
10424 double-checked locking was incomplete for RSA blinding,
10425 addressing just one layer of what turns out to have been
10426 doubly unsafe triple-checked locking.
10428 So now fix this for real by retiring the MONT_HELPER macro
10429 in crypto/rsa/rsa_eay.c.
10431 *Bodo Moeller; problem pointed out by Marius Schilder*
10433 * Various precautionary measures:
10435 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10437 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10438 (NB: This would require knowledge of the secret session ticket key
10439 to exploit, in which case you'd be SOL either way.)
10441 - Change bn_nist.c so that it will properly handle input BIGNUMs
10442 outside the expected range.
10444 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10447 *Neel Mehta, Bodo Moeller*
10449 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10450 the load fails. Useful for distros.
10452 *Ben Laurie and the FreeBSD team*
10454 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10458 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10462 * Expand ENGINE to support engine supplied SSL client certificate functions.
10464 This work was sponsored by Logica.
10468 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10469 keystores. Support for SSL/TLS client authentication too.
10470 Not compiled unless enable-capieng specified to Configure.
10472 This work was sponsored by Logica.
10476 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10477 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10478 attribute creation routines such as certificate requests and PKCS#12
10483 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10485 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10486 handshake which could lead to a client crash as found using the
10487 Codenomicon TLS test suite ([CVE-2008-1672])
10489 *Steve Henson, Mark Cox*
10491 * Fix double free in TLS server name extensions which could lead to
10492 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10496 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10498 Clear the error queue to ensure that error entries left from
10499 older function calls do not interfere with the correct operation.
10501 *Lutz Jaenicke, Erik de Castro Lopo*
10503 * Remove root CA certificates of commercial CAs:
10505 The OpenSSL project does not recommend any specific CA and does not
10506 have any policy with respect to including or excluding any CA.
10507 Therefore, it does not make any sense to ship an arbitrary selection
10508 of root CA certificates with the OpenSSL software.
10512 * RSA OAEP patches to fix two separate invalid memory reads.
10513 The first one involves inputs when 'lzero' is greater than
10514 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10515 before the beginning of from). The second one involves inputs where
10516 the 'db' section contains nothing but zeroes (there is a one-byte
10517 invalid read after the end of 'db').
10519 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10521 * Partial backport from 0.9.9-dev:
10523 Introduce bn_mul_mont (dedicated Montgomery multiplication
10524 procedure) as a candidate for BIGNUM assembler implementation.
10525 While 0.9.9-dev uses assembler for various architectures, only
10526 x86_64 is available by default here in the 0.9.8 branch, and
10527 32-bit x86 is available through a compile-time setting.
10529 To try the 32-bit x86 assembler implementation, use Configure
10530 option "enable-montasm" (which exists only for this backport).
10532 As "enable-montasm" for 32-bit x86 disclaims code stability
10533 anyway, in this constellation we activate additional code
10534 backported from 0.9.9-dev for further performance improvements,
10535 namely BN_from_montgomery_word. (To enable this otherwise,
10536 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10538 *Andy Polyakov (backport partially by Bodo Moeller)*
10540 * Add TLS session ticket callback. This allows an application to set
10541 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10542 values. This is useful for key rollover for example where several key
10543 sets may exist with different names.
10547 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10548 This was broken until now in 0.9.8 releases, such that the only way
10549 a registered ENGINE could be used (assuming it initialises
10550 successfully on the host) was to explicitly set it as the default
10551 for the relevant algorithms. This is in contradiction with 0.9.7
10552 behaviour and the documentation. With this fix, when an ENGINE is
10553 registered into a given algorithm's table of implementations, the
10554 'uptodate' flag is reset so that auto-discovery will be used next
10555 time a new context for that algorithm attempts to select an
10558 *Ian Lister (tweaked by Geoff Thorpe)*
10560 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10561 implementation in the following ways:
10563 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10566 Lack of BER streaming support means one pass streaming processing is
10567 only supported if data is detached: setting the streaming flag is
10568 ignored for embedded content.
10570 CMS support is disabled by default and must be explicitly enabled
10571 with the enable-cms configuration option.
10575 * Update the GMP engine glue to do direct copies between BIGNUM and
10576 mpz_t when openssl and GMP use the same limb size. Otherwise the
10577 existing "conversion via a text string export" trick is still used.
10579 *Paul Sheer <paulsheer@gmail.com>*
10581 * Zlib compression BIO. This is a filter BIO which compressed and
10582 uncompresses any data passed through it.
10586 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10587 RFC3394 compatible AES key wrapping.
10591 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10592 sets string data without copying. X509_ALGOR_set0() and
10593 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10594 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10595 from an X509_ATTRIBUTE structure optionally checking it occurs only
10596 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10601 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10602 to get the expected BN_FLG_CONSTTIME behavior.
10604 *Bodo Moeller (Google)*
10608 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10609 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10610 - added some more tests to do_tests.pl
10611 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10612 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10613 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10614 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10615 - various changes to netware.pl to enable gcc-cross builds on Win32
10617 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10618 - various changes to fix missing prototype warnings
10619 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10620 - added AES, WHIRLPOOL and CPUID assembler code to build files
10621 - added missing AES assembler make rules to mk1mf.pl
10622 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10624 *Guenter Knauf <eflash@gmx.net>*
10626 * Implement certificate status request TLS extension defined in RFC3546.
10627 A client can set the appropriate parameters and receive the encoded
10628 OCSP response via a callback. A server can query the supplied parameters
10629 and set the encoded OCSP response in the callback. Add simplified examples
10630 to s_client and s_server.
10634 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10636 * Fix various bugs:
10637 + Binary incompatibility of ssl_ctx_st structure
10638 + DTLS interoperation with non-compliant servers
10639 + Don't call get_session_cb() without proposed session
10640 + Fix ia64 assembler code
10642 *Andy Polyakov, Steve Henson*
10644 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10646 * DTLS Handshake overhaul. There were longstanding issues with
10647 OpenSSL DTLS implementation, which were making it impossible for
10648 RFC 4347 compliant client to communicate with OpenSSL server.
10649 Unfortunately just fixing these incompatibilities would "cut off"
10650 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10651 server keeps tolerating non RFC compliant syntax. The opposite is
10652 not true, 0.9.8f client can not communicate with earlier server.
10653 This update even addresses CVE-2007-4995.
10657 * Changes to avoid need for function casts in OpenSSL: some compilers
10658 (gcc 4.2 and later) reject their use.
10659 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10662 * Add RFC4507 support to OpenSSL. This includes the corrections in
10663 RFC4507bis. The encrypted ticket format is an encrypted encoded
10664 SSL_SESSION structure, that way new session features are automatically
10667 If a client application caches session in an SSL_SESSION structure
10668 support is transparent because tickets are now stored in the encoded
10671 The SSL_CTX structure automatically generates keys for ticket
10672 protection in servers so again support should be possible
10673 with no application modification.
10675 If a client or server wishes to disable RFC4507 support then the option
10676 SSL_OP_NO_TICKET can be set.
10678 Add a TLS extension debugging callback to allow the contents of any client
10679 or server extensions to be examined.
10681 This work was sponsored by Google.
10685 * Add initial support for TLS extensions, specifically for the server_name
10686 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10687 have new members for a hostname. The SSL data structure has an
10688 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10689 stored in that context to allow for session resumption, even after the
10690 SSL has been switched to a new SSL_CTX in reaction to a client's
10691 server_name extension.
10693 New functions (subject to change):
10695 SSL_get_servername()
10696 SSL_get_servername_type()
10699 New CTRL codes and macros (subject to change):
10701 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10702 - SSL_CTX_set_tlsext_servername_callback()
10703 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10704 - SSL_CTX_set_tlsext_servername_arg()
10705 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10707 openssl s_client has a new '-servername ...' option.
10709 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10710 '-key2 ...', '-servername_fatal' (subject to change). This allows
10711 testing the HostName extension for a specific single hostname ('-cert'
10712 and '-key' remain fallbacks for handshakes without HostName
10713 negotiation). If the unrecognized_name alert has to be sent, this by
10714 default is a warning; it becomes fatal with the '-servername_fatal'
10717 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10719 * Add AES and SSE2 assembly language support to VC++ build.
10723 * Mitigate attack on final subtraction in Montgomery reduction.
10727 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10728 (which previously caused an internal error).
10732 * Squeeze another 10% out of IGE mode when in != out.
10736 * AES IGE mode speedup.
10738 *Dean Gaudet (Google)*
10740 * Add the Korean symmetric 128-bit cipher SEED (see
10741 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10742 add SEED ciphersuites from RFC 4162:
10744 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10745 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10746 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10747 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10749 To minimize changes between patchlevels in the OpenSSL 0.9.8
10750 series, SEED remains excluded from compilation unless OpenSSL
10751 is configured with 'enable-seed'.
10753 *KISA, Bodo Moeller*
10755 * Mitigate branch prediction attacks, which can be practical if a
10756 single processor is shared, allowing a spy process to extract
10757 information. For detailed background information, see
10758 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10759 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10760 and Necessary Software Countermeasures"). The core of the change
10761 are new versions BN_div_no_branch() and
10762 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10763 respectively, which are slower, but avoid the security-relevant
10764 conditional branches. These are automatically called by BN_div()
10765 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10766 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10767 remove a conditional branch.
10769 BN_FLG_CONSTTIME is the new name for the previous
10770 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10771 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10772 in the exponent causes BN_mod_exp_mont() to use the alternative
10773 implementation in BN_mod_exp_mont_consttime().) The old name
10774 remains as a deprecated alias.
10776 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10777 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10778 constant-time implementations for more than just exponentiation.
10779 Here too the old name is kept as a deprecated alias.
10781 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10782 the BN_BLINDING structure gets an independent copy of the
10783 modulus. This means that the previous `BIGNUM *m` argument to
10784 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10785 essentially becomes `const BIGNUM *m`, although we can't actually
10786 change this in the header file before 0.9.9. It allows
10787 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10788 enable BN_FLG_CONSTTIME.
10790 *Matthew D Wood (Intel Corp)*
10792 * In the SSL/TLS server implementation, be strict about session ID
10793 context matching (which matters if an application uses a single
10794 external cache for different purposes). Previously,
10795 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10796 set. This did ensure strict client verification, but meant that,
10797 with applications using a single external cache for quite
10798 different requirements, clients could circumvent ciphersuite
10799 restrictions for a given session ID context by starting a session
10800 in a different context.
10804 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10805 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10806 authentication-only ciphersuites.
10810 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10811 not complete and could lead to a possible single byte overflow
10812 ([CVE-2007-5135]) [Ben Laurie]
10814 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10816 * Since AES128 and AES256 (and similarly Camellia128 and
10817 Camellia256) share a single mask bit in the logic of
10818 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10819 kludge to work properly if AES128 is available and AES256 isn't
10820 (or if Camellia128 is available and Camellia256 isn't).
10824 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10825 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10826 When a point or a seed is encoded in a BIT STRING, we need to
10827 prevent the removal of trailing zero bits to get the proper DER
10828 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10829 of a NamedBitList, for which trailing 0 bits need to be removed.)
10833 * Have SSL/TLS server implementation tolerate "mismatched" record
10834 protocol version while receiving ClientHello even if the
10835 ClientHello is fragmented. (The server can't insist on the
10836 particular protocol version it has chosen before the ServerHello
10837 message has informed the client about his choice.)
10841 * Add RFC 3779 support.
10843 *Rob Austein for ARIN, Ben Laurie*
10845 * Load error codes if they are not already present instead of using a
10846 static variable. This allows them to be cleanly unloaded and reloaded.
10847 Improve header file function name parsing.
10851 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10852 or CAPABILITY handshake as required by RFCs.
10854 *Goetz Babin-Ebell*
10856 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10858 * Introduce limits to prevent malicious keys being able to
10859 cause a denial of service. ([CVE-2006-2940])
10861 *Steve Henson, Bodo Moeller*
10863 * Fix ASN.1 parsing of certain invalid structures that can result
10864 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10866 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10867 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10869 * Fix SSL client code which could crash if connecting to a
10870 malicious SSLv2 server. ([CVE-2006-4343])
10872 *Tavis Ormandy and Will Drewry, Google Security Team*
10874 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10875 match only those. Before that, "AES256-SHA" would be interpreted
10876 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10877 the same strength classification in 0.9.7h) as we currently only
10878 have a single AES bit in the ciphersuite description bitmap.
10879 That change, however, also applied to ciphersuite strings such as
10880 "RC4-MD5" that intentionally matched multiple ciphersuites --
10881 namely, SSL 2.0 ciphersuites in addition to the more common ones
10882 from SSL 3.0/TLS 1.0.
10884 So we change the selection algorithm again: Naming an explicit
10885 ciphersuite selects this one ciphersuite, and any other similar
10886 ciphersuite (same bitmap) from *other* protocol versions.
10887 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10888 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10890 Since SSL 2.0 does not have any ciphersuites for which the
10891 128/256 bit distinction would be relevant, this works for now.
10892 The proper fix will be to use different bits for AES128 and
10893 AES256, which would have avoided the problems from the beginning;
10894 however, bits are scarce, so we can only do this in a new release
10895 (not just a patchlevel) when we can change the SSL_CIPHER
10896 definition to split the single 'unsigned long mask' bitmap into
10897 multiple values to extend the available space.
10901 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10903 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10904 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10906 * Add AES IGE and biIGE modes.
10910 * Change the Unix randomness entropy gathering to use poll() when
10911 possible instead of select(), since the latter has some
10912 undesirable limitations.
10914 *Darryl Miles via Richard Levitte and Bodo Moeller*
10916 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10917 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10918 cannot be implicitly activated as part of, e.g., the "AES" alias.
10919 However, please upgrade to OpenSSL 0.9.9[-dev] for
10920 non-experimental use of the ECC ciphersuites to get TLS extension
10921 support, which is required for curve and point format negotiation
10922 to avoid potential handshake problems.
10926 * Disable rogue ciphersuites:
10928 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10929 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10930 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10932 The latter two were purportedly from
10933 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10936 Also deactivate the remaining ciphersuites from
10937 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10938 unofficial, and the ID has long expired.
10942 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10943 dual-core machines) and other potential thread-safety issues.
10947 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10948 versions), which is now available for royalty-free use
10949 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10950 Also, add Camellia TLS ciphersuites from RFC 4132.
10952 To minimize changes between patchlevels in the OpenSSL 0.9.8
10953 series, Camellia remains excluded from compilation unless OpenSSL
10954 is configured with 'enable-camellia'.
10958 * Disable the padding bug check when compression is in use. The padding
10959 bug check assumes the first packet is of even length, this is not
10960 necessarily true if compression is enabled and can result in false
10961 positives causing handshake failure. The actual bug test is ancient
10962 code so it is hoped that implementations will either have fixed it by
10963 now or any which still have the bug do not support compression.
10967 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10969 * When applying a cipher rule check to see if string match is an explicit
10970 cipher suite and only match that one cipher suite if it is.
10974 * Link in manifests for VC++ if needed.
10976 *Austin Ziegler <halostatue@gmail.com>*
10978 * Update support for ECC-based TLS ciphersuites according to
10979 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10980 TLS extensions, which are supported starting with the 0.9.9
10981 branch, not in the OpenSSL 0.9.8 branch).
10985 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10986 opaque EVP_CIPHER_CTX handling.
10990 * Fixes and enhancements to zlib compression code. We now only use
10991 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10992 to conform with the standards mentioned here:
10993 <http://www.zlib.net/DLL_FAQ.txt>
10994 Static zlib linking now works on Windows and the new --with-zlib-include
10995 --with-zlib-lib options to Configure can be used to supply the location
10996 of the headers and library. Gracefully handle case where zlib library
11001 * Several fixes and enhancements to the OID generation code. The old code
11002 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11003 handle numbers larger than ULONG_MAX, truncated printing and had a
11004 non standard OBJ_obj2txt() behaviour.
11008 * Add support for building of engines under engine/ as shared libraries
11009 under VC++ build system.
11013 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11014 Hopefully, we will not see any false combination of paths any more.
11018 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11020 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11021 (part of SSL_OP_ALL). This option used to disable the
11022 countermeasure against man-in-the-middle protocol-version
11023 rollback in the SSL 2.0 server implementation, which is a bad
11024 idea. ([CVE-2005-2969])
11026 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11027 for Information Security, National Institute of Advanced Industrial
11028 Science and Technology [AIST], Japan)*
11030 * Add two function to clear and return the verify parameter flags.
11034 * Keep cipherlists sorted in the source instead of sorting them at
11035 runtime, thus removing the need for a lock.
11039 * Avoid some small subgroup attacks in Diffie-Hellman.
11041 *Nick Mathewson and Ben Laurie*
11043 * Add functions for well-known primes.
11047 * Extended Windows CE support.
11049 *Satoshi Nakamura and Andy Polyakov*
11051 * Initialize SSL_METHOD structures at compile time instead of during
11052 runtime, thus removing the need for a lock.
11056 * Make PKCS7_decrypt() work even if no certificate is supplied by
11057 attempting to decrypt each encrypted key in turn. Add support to
11062 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11064 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11067 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11071 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11072 key into the same file any more.
11076 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11080 * Add -utf8 command line and config file option to 'ca'.
11082 *Stefan <stf@udoma.org*
11084 * Removed the macro des_crypt(), as it seems to conflict with some
11085 libraries. Use DES_crypt().
11089 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11090 involves renaming the source and generated shared-libs for
11091 both. The engines will accept the corrected or legacy ids
11092 ('ncipher' and '4758_cca' respectively) when binding. NB,
11093 this only applies when building 'shared'.
11095 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11097 * Add attribute functions to EVP_PKEY structure. Modify
11098 PKCS12_create() to recognize a CSP name attribute and
11099 use it. Make -CSP option work again in pkcs12 utility.
11103 * Add new functionality to the bn blinding code:
11104 - automatic re-creation of the BN_BLINDING parameters after
11105 a fixed number of uses (currently 32)
11106 - add new function for parameter creation
11107 - introduce flags to control the update behaviour of the
11108 BN_BLINDING parameters
11109 - hide BN_BLINDING structure
11110 Add a second BN_BLINDING slot to the RSA structure to improve
11111 performance when a single RSA object is shared among several
11116 * Add support for DTLS.
11118 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11120 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11121 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11125 * Remove buggy and incomplete DH cert support from
11126 ssl/ssl_rsa.c and ssl/s3_both.c
11130 * Use SHA-1 instead of MD5 as the default digest algorithm for
11131 the `apps/openssl` commands.
11135 * Compile clean with "-Wall -Wmissing-prototypes
11136 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11137 DEBUG_SAFESTACK must also be set.
11141 * Change ./Configure so that certain algorithms can be disabled by default.
11142 The new counterpiece to "no-xxx" is "enable-xxx".
11144 The patented RC5 and MDC2 algorithms will now be disabled unless
11145 "enable-rc5" and "enable-mdc2", respectively, are specified.
11147 (IDEA remains enabled despite being patented. This is because IDEA
11148 is frequently required for interoperability, and there is no license
11149 fee for non-commercial use. As before, "no-idea" can be used to
11150 avoid this algorithm.)
11154 * Add processing of proxy certificates (see RFC 3820). This work was
11155 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11156 EGEE (Enabling Grids for E-science in Europe).
11160 * RC4 performance overhaul on modern architectures/implementations, such
11161 as Intel P4, IA-64 and AMD64.
11165 * New utility extract-section.pl. This can be used specify an alternative
11166 section number in a pod file instead of having to treat each file as
11167 a separate case in Makefile. This can be done by adding two lines to the
11170 =for comment openssl_section:XXX
11172 The blank line is mandatory.
11176 * New arguments -certform, -keyform and -pass for s_client and s_server
11177 to allow alternative format key and certificate files and passphrase
11182 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11183 update associated structures and add various utility functions.
11185 Add new policy related verify parameters, include policy checking in
11186 standard verify code. Enhance 'smime' application with extra parameters
11187 to support policy checking and print out.
11191 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11192 Nehemiah processors. These extensions support AES encryption in hardware
11193 as well as RNG (though RNG support is currently disabled).
11195 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11197 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11201 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11203 *Andy Polyakov and a number of other people*
11205 * Improved PowerPC platform support. Most notably BIGNUM assembler
11206 implementation contributed by IBM.
11208 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11210 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11211 exponent rather than 'unsigned long'. There is a corresponding change to
11212 the new 'rsa_keygen' element of the RSA_METHOD structure.
11214 *Jelte Jansen, Geoff Thorpe*
11216 * Functionality for creating the initial serial number file is now
11217 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11219 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11220 number file to 1, which is bound to cause problems. To avoid
11221 the problems while respecting compatibility between different 0.9.7
11222 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11223 CA.pl for serial number initialization. With the new release 0.9.8,
11224 we can fix the problem directly in the 'ca' utility.)
11228 * Reduced header interdependencies by declaring more opaque objects in
11229 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11230 give fewer recursive includes, which could break lazy source code - so
11231 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11232 developers should define this symbol when building and using openssl to
11233 ensure they track the recommended behaviour, interfaces, [etc], but
11234 backwards-compatible behaviour prevails when this isn't defined.
11238 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11242 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11243 This will generate a random key of the appropriate length based on the
11244 cipher context. The EVP_CIPHER can provide its own random key generation
11245 routine to support keys of a specific form. This is used in the des and
11246 3des routines to generate a key of the correct parity. Update S/MIME
11247 code to use new functions and hence generate correct parity DES keys.
11248 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11249 valid (weak or incorrect parity).
11253 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11254 as looking them up. This is useful when the verified structure may contain
11255 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11256 present unless the new PKCS7_NO_CRL flag is asserted.
11260 * Extend ASN1 oid configuration module. It now additionally accepts the
11263 shortName = some long name, 1.2.3.4
11267 * Reimplemented the BN_CTX implementation. There is now no more static
11268 limitation on the number of variables it can handle nor the depth of the
11269 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11270 information can now expand as required, and rather than having a single
11271 static array of bignums, BN_CTX now uses a linked-list of such arrays
11272 allowing it to expand on demand whilst maintaining the usefulness of
11273 BN_CTX's "bundling".
11277 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11278 to allow all RSA operations to function using a single BN_CTX.
11282 * Preliminary support for certificate policy evaluation and checking. This
11283 is initially intended to pass the tests outlined in "Conformance Testing
11284 of Relying Party Client Certificate Path Processing Logic" v1.07.
11288 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11289 remained unused and not that useful. A variety of other little bignum
11290 tweaks and fixes have also been made continuing on from the audit (see
11295 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11296 associated ASN1, EVP and SSL functions and old ASN1 macros.
11300 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11301 and this should never fail. So the return value from the use of
11302 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11303 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11307 * BN_CTX_get() should return zero-valued bignums, providing the same
11308 initialised value as BN_new().
11310 *Geoff Thorpe, suggested by Ulf Möller*
11312 * Support for inhibitAnyPolicy certificate extension.
11316 * An audit of the BIGNUM code is underway, for which debugging code is
11317 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11318 is considered valid when processing BIGNUMs, and causes execution to
11319 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11320 further steps are taken to deliberately pollute unused data in BIGNUM
11321 structures to try and expose faulty code further on. For now, openssl will
11322 (in its default mode of operation) continue to tolerate the inconsistent
11323 forms that it has tolerated in the past, but authors and packagers should
11324 consider trying openssl and their own applications when compiled with
11325 these debugging symbols defined. It will help highlight potential bugs in
11326 their own code, and will improve the test coverage for OpenSSL itself. At
11327 some point, these tighter rules will become openssl's default to improve
11328 maintainability, though the assert()s and other overheads will remain only
11329 in debugging configurations. See bn.h for more details.
11331 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11333 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11334 that can only be obtained through BN_CTX_new() (which implicitly
11335 initialises it). The presence of this function only made it possible
11336 to overwrite an existing structure (and cause memory leaks).
11340 * Because of the callback-based approach for implementing LHASH as a
11341 template type, lh_insert() adds opaque objects to hash-tables and
11342 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11343 to clean up those corresponding objects before destroying the hash table
11344 (and losing the object pointers). So some over-zealous constifications in
11345 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11346 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11347 prototyped to have "const" restrictions on the object pointers they are
11348 given (and so aren't required to cast them away any more).
11352 * The tmdiff.h API was so ugly and minimal that our own timing utility
11353 (speed) prefers to use its own implementation. The two implementations
11354 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11355 its object type properly exposed (MS_TM) instead of casting to/from
11356 `char *`. This may still change yet if someone realises MS_TM and
11358 aren't necessarily the greatest nomenclatures - but this is what was used
11359 internally to the implementation so I've used that for now.
11363 * Ensure that deprecated functions do not get compiled when
11364 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11365 the self-tests were still using deprecated key-generation functions so
11366 these have been updated also.
11370 * Reorganise PKCS#7 code to separate the digest location functionality
11371 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11372 New function PKCS7_set_digest() to set the digest type for PKCS#7
11373 digestedData type. Add additional code to correctly generate the
11374 digestedData type and add support for this type in PKCS7 initialization
11379 * New function PKCS7_set0_type_other() this initializes a PKCS7
11380 structure of type "other".
11384 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11385 sure the loop does correctly stop and breaking ("division by zero")
11386 modulus operations are not performed. The (pre-generated) prime
11387 table crypto/bn/bn_prime.h was already correct, but it could not be
11388 re-generated on some platforms because of the "division by zero"
11389 situation in the script.
11391 *Ralf S. Engelschall*
11393 * Update support for ECC-based TLS ciphersuites according to
11394 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11395 SHA-1 now is only used for "small" curves (where the
11396 representation of a field element takes up to 24 bytes); for
11397 larger curves, the field element resulting from ECDH is directly
11398 used as premaster secret.
11400 *Douglas Stebila (Sun Microsystems Laboratories)*
11402 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11403 curve secp160r1 to the tests.
11405 *Douglas Stebila (Sun Microsystems Laboratories)*
11407 * Add the possibility to load symbols globally with DSO.
11409 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11411 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11412 control of the error stack.
11416 * Add support for STORE in ENGINE.
11420 * Add the STORE type. The intention is to provide a common interface
11421 to certificate and key stores, be they simple file-based stores, or
11422 HSM-type store, or LDAP stores, or...
11423 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11427 * Add a generic structure called OPENSSL_ITEM. This can be used to
11428 pass a list of arguments to any function as well as provide a way
11429 for a function to pass data back to the caller.
11433 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11434 works like BUF_strdup() but can be used to duplicate a portion of
11435 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11440 * Add the function sk_find_ex() which works like sk_find(), but will
11441 return an index to an element even if an exact match couldn't be
11442 found. The index is guaranteed to point at the element where the
11443 searched-for key would be inserted to preserve sorting order.
11447 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11448 takes an extra flags argument for optional functionality. Currently,
11449 the following flags are defined:
11451 OBJ_BSEARCH_VALUE_ON_NOMATCH
11452 This one gets OBJ_bsearch_ex() to return a pointer to the first
11453 element where the comparing function returns a negative or zero
11456 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11457 This one gets OBJ_bsearch_ex() to return a pointer to the first
11458 element where the comparing function returns zero. This is useful
11459 if there are more than one element where the comparing function
11464 * Make it possible to create self-signed certificates with 'openssl ca'
11465 in such a way that the self-signed certificate becomes part of the
11466 CA database and uses the same mechanisms for serial number generation
11467 as all other certificate signing. The new flag '-selfsign' enables
11468 this functionality. Adapt CA.sh and CA.pl.in.
11472 * Add functionality to check the public key of a certificate request
11473 against a given private. This is useful to check that a certificate
11474 request can be signed by that key (self-signing).
11478 * Make it possible to have multiple active certificates with the same
11479 subject in the CA index file. This is done only if the keyword
11480 'unique_subject' is set to 'no' in the main CA section (default
11481 if 'CA_default') of the configuration file. The value is saved
11482 with the database itself in a separate index attribute file,
11483 named like the index file with '.attr' appended to the name.
11487 * Generate multi-valued AVAs using '+' notation in config files for
11492 * Support for nameConstraints certificate extension.
11496 * Support for policyConstraints certificate extension.
11500 * Support for policyMappings certificate extension.
11504 * Make sure the default DSA_METHOD implementation only uses its
11505 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11506 and change its own handlers to be NULL so as to remove unnecessary
11507 indirection. This lets alternative implementations fallback to the
11508 default implementation more easily.
11512 * Support for directoryName in GeneralName related extensions
11517 * Make it possible to link applications using Makefile.shared.
11518 Make that possible even when linking against static libraries!
11522 * Support for single pass processing for S/MIME signing. This now
11523 means that S/MIME signing can be done from a pipe, in addition
11524 cleartext signing (multipart/signed type) is effectively streaming
11525 and the signed data does not need to be all held in memory.
11527 This is done with a new flag PKCS7_STREAM. When this flag is set
11528 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11529 is done after the data is output (and digests calculated) in
11530 SMIME_write_PKCS7().
11534 * Add full support for -rpath/-R, both in shared libraries and
11535 applications, at least on the platforms where it's known how
11540 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11541 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11542 will now compute a table of multiples of the generator that
11543 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11544 faster (notably in the case of a single point multiplication,
11545 scalar * generator).
11547 *Nils Larsch, Bodo Moeller*
11549 * IPv6 support for certificate extensions. The various extensions
11550 which use the IP:a.b.c.d can now take IPv6 addresses using the
11551 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11556 * Added an ENGINE that implements RSA by performing private key
11557 exponentiations with the GMP library. The conversions to and from
11558 GMP's mpz_t format aren't optimised nor are any montgomery forms
11559 cached, and on x86 it appears OpenSSL's own performance has caught up.
11560 However there are likely to be other architectures where GMP could
11561 provide a boost. This ENGINE is not built in by default, but it can be
11562 specified at Configure time and should be accompanied by the necessary
11563 linker additions, eg;
11564 ./config -DOPENSSL_USE_GMP -lgmp
11568 * "openssl engine" will not display ENGINE/DSO load failure errors when
11569 testing availability of engines with "-t" - the old behaviour is
11570 produced by increasing the feature's verbosity with "-tt".
11574 * ECDSA routines: under certain error conditions uninitialized BN objects
11575 could be freed. Solution: make sure initialization is performed early
11576 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11581 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11582 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11583 software implementations. For DSA and DH, parameter generation can
11584 also be overridden by providing the appropriate method callbacks.
11588 * Change the "progress" mechanism used in key-generation and
11589 primality testing to functions that take a new BN_GENCB pointer in
11590 place of callback/argument pairs. The new API functions have `_ex`
11591 postfixes and the older functions are reimplemented as wrappers for
11592 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11593 declarations of the old functions to help (graceful) attempts to
11594 migrate to the new functions. Also, the new key-generation API
11595 functions operate on a caller-supplied key-structure and return
11596 success/failure rather than returning a key or NULL - this is to
11597 help make "keygen" another member function of RSA_METHOD etc.
11599 Example for using the new callback interface:
11601 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11602 void *my_arg = ...;
11605 BN_GENCB_set(&my_cb, my_callback, my_arg);
11607 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11608 /* For the meaning of a, b in calls to my_callback(), see the
11609 * documentation of the function that calls the callback.
11610 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11611 * my_callback should return 1 if it wants BN_is_prime_ex()
11612 * to continue, or 0 to stop.
11617 * Change the ZLIB compression method to be stateful, and make it
11618 available to TLS with the number defined in
11619 draft-ietf-tls-compression-04.txt.
11623 * Add the ASN.1 structures and functions for CertificatePair, which
11624 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11626 CertificatePair ::= SEQUENCE {
11627 forward [0] Certificate OPTIONAL,
11628 reverse [1] Certificate OPTIONAL,
11629 -- at least one of the pair shall be present -- }
11631 Also implement the PEM functions to read and write certificate
11632 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11634 This needed to be defined, mostly for the sake of the LDAP
11635 attribute crossCertificatePair, but may prove useful elsewhere as
11640 * Make it possible to inhibit symlinking of shared libraries in
11641 Makefile.shared, for Cygwin's sake.
11645 * Extend the BIGNUM API by creating a function
11646 void BN_set_negative(BIGNUM *a, int neg);
11647 and a macro that behave like
11648 int BN_is_negative(const BIGNUM *a);
11650 to avoid the need to access 'a->neg' directly in applications.
11654 * Implement fast modular reduction for pseudo-Mersenne primes
11655 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11656 EC_GROUP_new_curve_GFp() will now automatically use this
11659 *Nils Larsch <nla@trustcenter.de>*
11661 * Add new lock type (CRYPTO_LOCK_BN).
11665 * Change the ENGINE framework to automatically load engines
11666 dynamically from specific directories unless they could be
11667 found to already be built in or loaded. Move all the
11668 current engines except for the cryptodev one to a new
11669 directory engines/.
11670 The engines in engines/ are built as shared libraries if
11671 the "shared" options was given to ./Configure or ./config.
11672 Otherwise, they are inserted in libcrypto.a.
11673 /usr/local/ssl/engines is the default directory for dynamic
11674 engines, but that can be overridden at configure time through
11675 the usual use of --prefix and/or --openssldir, and at run
11676 time with the environment variable OPENSSL_ENGINES.
11678 *Geoff Thorpe and Richard Levitte*
11680 * Add Makefile.shared, a helper makefile to build shared
11681 libraries. Adapt Makefile.org.
11685 * Add version info to Win32 DLLs.
11687 *Peter 'Luna' Runestig" <peter@runestig.com>*
11689 * Add new 'medium level' PKCS#12 API. Certificates and keys
11690 can be added using this API to created arbitrary PKCS#12
11691 files while avoiding the low-level API.
11693 New options to PKCS12_create(), key or cert can be NULL and
11694 will then be omitted from the output file. The encryption
11695 algorithm NIDs can be set to -1 for no encryption, the mac
11696 iteration count can be set to 0 to omit the mac.
11698 Enhance pkcs12 utility by making the -nokeys and -nocerts
11699 options work when creating a PKCS#12 file. New option -nomac
11700 to omit the mac, NONE can be set for an encryption algorithm.
11701 New code is modified to use the enhanced PKCS12_create()
11702 instead of the low-level API.
11706 * Extend ASN1 encoder to support indefinite length constructed
11707 encoding. This can output sequences tags and octet strings in
11708 this form. Modify pk7_asn1.c to support indefinite length
11709 encoding. This is experimental and needs additional code to
11710 be useful, such as an ASN1 bio and some enhanced streaming
11713 Extend template encode functionality so that tagging is passed
11714 down to the template encoder.
11718 * Let 'openssl req' fail if an argument to '-newkey' is not
11719 recognized instead of using RSA as a default.
11723 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11724 As these are not official, they are not included in "ALL";
11725 the "ECCdraft" ciphersuite group alias can be used to select them.
11727 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11729 * Add ECDH engine support.
11731 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11733 * Add ECDH in new directory crypto/ecdh/.
11735 *Douglas Stebila (Sun Microsystems Laboratories)*
11737 * Let BN_rand_range() abort with an error after 100 iterations
11738 without success (which indicates a broken PRNG).
11742 * Change BN_mod_sqrt() so that it verifies that the input value
11743 is really the square of the return value. (Previously,
11744 BN_mod_sqrt would show GIGO behaviour.)
11748 * Add named elliptic curves over binary fields from X9.62, SECG,
11749 and WAP/WTLS; add OIDs that were still missing.
11751 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11753 * Extend the EC library for elliptic curves over binary fields
11754 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11757 EC_GF2m_simple_method
11761 EC_GROUP_new_curve_GF2m
11762 EC_GROUP_set_curve_GF2m
11763 EC_GROUP_get_curve_GF2m
11764 EC_POINT_set_affine_coordinates_GF2m
11765 EC_POINT_get_affine_coordinates_GF2m
11766 EC_POINT_set_compressed_coordinates_GF2m
11768 Point compression for binary fields is disabled by default for
11769 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11772 As binary polynomials are represented as BIGNUMs, various members
11773 of the EC_GROUP and EC_POINT data structures can be shared
11774 between the implementations for prime fields and binary fields;
11775 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11776 are essentially identical to their `..._GFp` counterparts.
11777 (For simplicity, the `..._GFp` prefix has been dropped from
11778 various internal method names.)
11780 An internal 'field_div' method (similar to 'field_mul' and
11781 'field_sqr') has been added; this is used only for binary fields.
11783 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11785 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11786 through methods ('mul', 'precompute_mult').
11788 The generic implementations (now internally called 'ec_wNAF_mul'
11789 and 'ec_wNAF_precomputed_mult') remain the default if these
11790 methods are undefined.
11792 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11794 * New function EC_GROUP_get_degree, which is defined through
11795 EC_METHOD. For curves over prime fields, this returns the bit
11796 length of the modulus.
11798 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11800 * New functions EC_GROUP_dup, EC_POINT_dup.
11801 (These simply call ..._new and ..._copy).
11803 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11805 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11806 Polynomials are represented as BIGNUMs (where the sign bit is not
11807 used) in the following functions [macros]:
11810 BN_GF2m_sub [= BN_GF2m_add]
11811 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11812 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11813 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11815 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11816 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11817 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11818 BN_GF2m_cmp [= BN_ucmp]
11820 (Note that only the 'mod' functions are actually for fields GF(2^m).
11821 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11823 For some functions, an the irreducible polynomial defining a
11824 field can be given as an 'unsigned int[]' with strictly
11825 decreasing elements giving the indices of those bits that are set;
11826 i.e., p[] represents the polynomial
11827 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11829 p[0] > p[1] > ... > p[k] = 0.
11830 This applies to the following functions:
11833 BN_GF2m_mod_mul_arr
11834 BN_GF2m_mod_sqr_arr
11835 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11836 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11837 BN_GF2m_mod_exp_arr
11838 BN_GF2m_mod_sqrt_arr
11839 BN_GF2m_mod_solve_quad_arr
11843 Conversion can be performed by the following functions:
11848 bntest.c has additional tests for binary polynomial arithmetic.
11850 Two implementations for BN_GF2m_mod_div() are available.
11851 The default algorithm simply uses BN_GF2m_mod_inv() and
11852 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11853 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11854 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11856 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11858 * Add new error code 'ERR_R_DISABLED' that can be used when some
11859 functionality is disabled at compile-time.
11861 *Douglas Stebila <douglas.stebila@sun.com>*
11863 * Change default behaviour of 'openssl asn1parse' so that more
11864 information is visible when viewing, e.g., a certificate:
11866 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11867 mode the content of non-printable OCTET STRINGs is output in a
11868 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11869 avoid the appearance of a printable string.
11871 *Nils Larsch <nla@trustcenter.de>*
11873 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11875 EC_GROUP_set_asn1_flag()
11876 EC_GROUP_get_asn1_flag()
11877 EC_GROUP_set_point_conversion_form()
11878 EC_GROUP_get_point_conversion_form()
11879 These control ASN1 encoding details:
11880 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11881 has been set to OPENSSL_EC_NAMED_CURVE.
11882 - Points are encoded in uncompressed form by default; options for
11883 asn1_for are as for point2oct, namely
11884 POINT_CONVERSION_COMPRESSED
11885 POINT_CONVERSION_UNCOMPRESSED
11886 POINT_CONVERSION_HYBRID
11888 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11890 EC_GROUP_set_seed()
11891 EC_GROUP_get0_seed()
11892 EC_GROUP_get_seed_len()
11893 This is used only for ASN1 purposes (so far).
11895 *Nils Larsch <nla@trustcenter.de>*
11897 * Add 'field_type' member to EC_METHOD, which holds the NID
11898 of the appropriate field type OID. The new function
11899 EC_METHOD_get_field_type() returns this value.
11901 *Nils Larsch <nla@trustcenter.de>*
11904 EC_POINT_point2bn()
11905 EC_POINT_bn2point()
11906 EC_POINT_point2hex()
11907 EC_POINT_hex2point()
11908 providing useful interfaces to EC_POINT_point2oct() and
11909 EC_POINT_oct2point().
11911 *Nils Larsch <nla@trustcenter.de>*
11913 * Change internals of the EC library so that the functions
11914 EC_GROUP_set_generator()
11915 EC_GROUP_get_generator()
11916 EC_GROUP_get_order()
11917 EC_GROUP_get_cofactor()
11918 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11919 to methods, which would lead to unnecessary code duplication when
11920 adding different types of curves.
11922 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11924 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11925 arithmetic, and such that modified wNAFs are generated
11926 (which avoid length expansion in many cases).
11930 * Add a function EC_GROUP_check_discriminant() (defined via
11931 EC_METHOD) that verifies that the curve discriminant is non-zero.
11933 Add a function EC_GROUP_check() that makes some sanity tests
11934 on a EC_GROUP, its generator and order. This includes
11935 EC_GROUP_check_discriminant().
11937 *Nils Larsch <nla@trustcenter.de>*
11939 * Add ECDSA in new directory crypto/ecdsa/.
11941 Add applications 'openssl ecparam' and 'openssl ecdsa'
11942 (these are based on 'openssl dsaparam' and 'openssl dsa').
11944 ECDSA support is also included in various other files across the
11945 library. Most notably,
11946 - 'openssl req' now has a '-newkey ecdsa:file' option;
11947 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11948 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11949 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11950 them suitable for ECDSA where domain parameters must be
11951 extracted before the specific public key;
11952 - ECDSA engine support has been added.
11954 *Nils Larsch <nla@trustcenter.de>*
11956 * Include some named elliptic curves, and add OIDs from X9.62,
11957 SECG, and WAP/WTLS. Each curve can be obtained from the new
11959 EC_GROUP_new_by_curve_name(),
11960 and the list of available named curves can be obtained with
11961 EC_get_builtin_curves().
11962 Also add a 'curve_name' member to EC_GROUP objects, which can be
11964 EC_GROUP_set_curve_name()
11965 EC_GROUP_get_curve_name()
11967 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11969 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11970 was actually never needed) and in BN_mul(). The removal in BN_mul()
11971 required a small change in bn_mul_part_recursive() and the addition
11972 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11973 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11974 bn_sub_words() and bn_add_words() except they take arrays with
11979 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11981 * Cleanse PEM buffers before freeing them since they may contain
11984 *Benjamin Bennett <ben@psc.edu>*
11986 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11987 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11988 authentication-only ciphersuites.
11992 * Since AES128 and AES256 share a single mask bit in the logic of
11993 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11994 kludge to work properly if AES128 is available and AES256 isn't.
11998 * Expand security boundary to match 1.1.1 module.
12002 * Remove redundant features: hash file source, editing of test vectors
12003 modify fipsld to use external fips_premain.c signature.
12007 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12008 run algorithm test programs.
12012 * Make algorithm test programs more tolerant of whitespace.
12016 * Have SSL/TLS server implementation tolerate "mismatched" record
12017 protocol version while receiving ClientHello even if the
12018 ClientHello is fragmented. (The server can't insist on the
12019 particular protocol version it has chosen before the ServerHello
12020 message has informed the client about his choice.)
12024 * Load error codes if they are not already present instead of using a
12025 static variable. This allows them to be cleanly unloaded and reloaded.
12029 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12031 * Introduce limits to prevent malicious keys being able to
12032 cause a denial of service. ([CVE-2006-2940])
12034 *Steve Henson, Bodo Moeller*
12036 * Fix ASN.1 parsing of certain invalid structures that can result
12037 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12039 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12040 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12042 * Fix SSL client code which could crash if connecting to a
12043 malicious SSLv2 server. ([CVE-2006-4343])
12045 *Tavis Ormandy and Will Drewry, Google Security Team*
12047 * Change ciphersuite string processing so that an explicit
12048 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12049 will no longer include "AES128-SHA"), and any other similar
12050 ciphersuite (same bitmap) from *other* protocol versions (so that
12051 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12052 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12053 changes from 0.9.8b and 0.9.8d.
12057 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12059 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12060 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12062 * Change the Unix randomness entropy gathering to use poll() when
12063 possible instead of select(), since the latter has some
12064 undesirable limitations.
12066 *Darryl Miles via Richard Levitte and Bodo Moeller*
12068 * Disable rogue ciphersuites:
12070 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12071 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12072 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12074 The latter two were purportedly from
12075 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12078 Also deactivate the remaining ciphersuites from
12079 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12080 unofficial, and the ID has long expired.
12084 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12085 dual-core machines) and other potential thread-safety issues.
12089 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12091 * Adapt fipsld and the build system to link against the validated FIPS
12092 module in FIPS mode.
12096 * Fixes for VC++ 2005 build under Windows.
12100 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12101 from a Windows bash shell such as MSYS. It is autodetected from the
12102 "config" script when run from a VC++ environment. Modify standard VC++
12103 build to use fipscanister.o from the GNU make build.
12107 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12109 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12110 The value now differs depending on if you build for FIPS or not.
12111 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12112 safely run with a non-FIPSed libcrypto, as it may crash because of
12113 the difference induced by this change.
12117 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12119 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12120 (part of SSL_OP_ALL). This option used to disable the
12121 countermeasure against man-in-the-middle protocol-version
12122 rollback in the SSL 2.0 server implementation, which is a bad
12123 idea. ([CVE-2005-2969])
12125 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12126 for Information Security, National Institute of Advanced Industrial
12127 Science and Technology [AIST, Japan)]*
12129 * Minimal support for X9.31 signatures and PSS padding modes. This is
12130 mainly for FIPS compliance and not fully integrated at this stage.
12134 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12135 the exponentiation using a fixed-length exponent. (Otherwise,
12136 the information leaked through timing could expose the secret key
12137 after many signatures; cf. Bleichenbacher's attack on DSA with
12142 * Make a new fixed-window mod_exp implementation the default for
12143 RSA, DSA, and DH private-key operations so that the sequence of
12144 squares and multiplies and the memory access pattern are
12145 independent of the particular secret key. This will mitigate
12146 cache-timing and potential related attacks.
12148 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12149 and this is automatically used by BN_mod_exp_mont() if the new flag
12150 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12151 will use this BN flag for private exponents unless the flag
12152 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12153 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12155 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12157 * Change the client implementation for SSLv23_method() and
12158 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12159 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12160 (Previously, the SSL 2.0 backwards compatible Client Hello
12161 message format would be used even with SSL_OP_NO_SSLv2.)
12165 * Add support for smime-type MIME parameter in S/MIME messages which some
12170 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12171 a threadsafe manner. Modify rsa code to use new function and add calls
12172 to dsa and dh code (which had race conditions before).
12176 * Include the fixed error library code in the C error file definitions
12177 instead of fixing them up at runtime. This keeps the error code
12178 structures constant.
12182 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12184 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12187 * Fixes for newer kerberos headers. NB: the casts are needed because
12188 the 'length' field is signed on one version and unsigned on another
12189 with no (?) obvious way to tell the difference, without these VC++
12190 complains. Also the "definition" of FAR (blank) is no longer included
12191 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12192 some needed definitions.
12196 * Undo Cygwin change.
12200 * Added support for proxy certificates according to RFC 3820.
12201 Because they may be a security thread to unaware applications,
12202 they must be explicitly allowed in run-time. See
12203 docs/HOWTO/proxy_certificates.txt for further information.
12207 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12209 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12210 server and client random values. Previously
12211 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12212 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12214 This change has negligible security impact because:
12216 1. Server and client random values still have 24 bytes of pseudo random
12219 2. Server and client random values are sent in the clear in the initial
12222 3. The master secret is derived using the premaster secret (48 bytes in
12223 size for static RSA ciphersuites) as well as client server and random
12226 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12229 *Stephen Henson, reported by UK NISCC*
12231 * Use Windows randomness collection on Cygwin.
12235 * Fix hang in EGD/PRNGD query when communication socket is closed
12236 prematurely by EGD/PRNGD.
12238 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12240 * Prompt for pass phrases when appropriate for PKCS12 input format.
12244 * Back-port of selected performance improvements from development
12245 branch, as well as improved support for PowerPC platforms.
12249 * Add lots of checks for memory allocation failure, error codes to indicate
12250 failure and freeing up memory if a failure occurs.
12252 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12254 * Add new -passin argument to dgst.
12258 * Perform some character comparisons of different types in X509_NAME_cmp:
12259 this is needed for some certificates that re-encode DNs into UTF8Strings
12260 (in violation of RFC3280) and can't or won't issue name rollover
12265 * Make an explicit check during certificate validation to see that
12266 the CA setting in each certificate on the chain is correct. As a
12267 side effect always do the following basic checks on extensions,
12268 not just when there's an associated purpose to the check:
12270 - if there is an unhandled critical extension (unless the user
12271 has chosen to ignore this fault)
12272 - if the path length has been exceeded (if one is set at all)
12273 - that certain extensions fit the associated purpose (if one has
12278 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12280 * Avoid a race condition when CRLs are checked in a multi threaded
12281 environment. This would happen due to the reordering of the revoked
12282 entries during signature checking and serial number lookup. Now the
12283 encoding is cached and the serial number sort performed under a lock.
12284 Add new STACK function sk_is_sorted().
12288 * Add Delta CRL to the extension code.
12292 * Various fixes to s3_pkt.c so alerts are sent properly.
12294 *David Holmes <d.holmes@f5.com>*
12296 * Reduce the chances of duplicate issuer name and serial numbers (in
12297 violation of RFC3280) using the OpenSSL certificate creation utilities.
12298 This is done by creating a random 64 bit value for the initial serial
12299 number when a serial number file is created or when a self signed
12300 certificate is created using 'openssl req -x509'. The initial serial
12301 number file is created using 'openssl x509 -next_serial' in CA.pl
12302 rather than being initialized to 1.
12306 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12308 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12309 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12311 *Joe Orton, Steve Henson*
12313 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12316 *Joe Orton, Steve Henson*
12318 * Make it possible to have multiple active certificates with the same
12319 subject in the CA index file. This is done only if the keyword
12320 'unique_subject' is set to 'no' in the main CA section (default
12321 if 'CA_default') of the configuration file. The value is saved
12322 with the database itself in a separate index attribute file,
12323 named like the index file with '.attr' appended to the name.
12327 * X509 verify fixes. Disable broken certificate workarounds when
12328 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12329 keyUsage extension present. Don't accept CRLs with unhandled critical
12330 extensions: since verify currently doesn't process CRL extensions this
12331 rejects a CRL with *any* critical extensions. Add new verify error codes
12336 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12337 A clarification of RFC2560 will require the use of OCTET STRINGs and
12338 some implementations cannot handle the current raw format. Since OpenSSL
12339 copies and compares OCSP nonces as opaque blobs without any attempt at
12340 parsing them this should not create any compatibility issues.
12344 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12345 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12346 this HMAC (and other) operations are several times slower than OpenSSL
12351 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12353 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12355 * Use the correct content when signing type "other".
12359 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12361 * Fix various bugs revealed by running the NISCC test suite:
12363 Stop out of bounds reads in the ASN1 code when presented with
12364 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12366 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12368 If verify callback ignores invalid public key errors don't try to check
12369 certificate signature with the NULL public key.
12373 * New -ignore_err option in ocsp application to stop the server
12374 exiting on the first error in a request.
12378 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12379 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12384 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12385 extra data after the compression methods not only for TLS 1.0
12386 but also for SSL 3.0 (as required by the specification).
12388 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12390 * Change X509_certificate_type() to mark the key as exported/exportable
12391 when it's 512 *bits* long, not 512 bytes.
12395 * Change AES_cbc_encrypt() so it outputs exact multiple of
12396 blocks during encryption.
12400 * Various fixes to base64 BIO and non blocking I/O. On write
12401 flushes were not handled properly if the BIO retried. On read
12402 data was not being buffered properly and had various logic bugs.
12403 This also affects blocking I/O when the data being decoded is a
12408 * Various S/MIME bugfixes and compatibility changes:
12409 output correct application/pkcs7 MIME type if
12410 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12411 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12412 of files as .eml work). Correctly handle very long lines in MIME
12417 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12419 * Countermeasure against the Klima-Pokorny-Rosa extension of
12420 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12421 a protocol version number mismatch like a decryption error
12422 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12426 * Turn on RSA blinding by default in the default implementation
12427 to avoid a timing attack. Applications that don't want it can call
12428 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12429 They would be ill-advised to do so in most cases.
12431 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12433 * Change RSA blinding code so that it works when the PRNG is not
12434 seeded (in this case, the secret RSA exponent is abused as
12435 an unpredictable seed -- if it is not unpredictable, there
12436 is no point in blinding anyway). Make RSA blinding thread-safe
12437 by remembering the creator's thread ID in rsa->blinding and
12438 having all other threads use local one-time blinding factors
12439 (this requires more computation than sharing rsa->blinding, but
12440 avoids excessive locking; and if an RSA object is not shared
12441 between threads, blinding will still be very fast).
12445 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12446 ENGINE as defaults for all supported algorithms irrespective of
12447 the 'flags' parameter. 'flags' is now honoured, so applications
12448 should make sure they are passing it correctly.
12452 * Target "mingw" now allows native Windows code to be generated in
12453 the Cygwin environment as well as with the MinGW compiler.
12457 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12459 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12460 via timing by performing a MAC computation even if incorrect
12461 block cipher padding has been found. This is a countermeasure
12462 against active attacks where the attacker has to distinguish
12463 between bad padding and a MAC verification error. ([CVE-2003-0078])
12465 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12466 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12467 Martin Vuagnoux (EPFL, Ilion)*
12469 * Make the no-err option work as intended. The intention with no-err
12470 is not to have the whole error stack handling routines removed from
12471 libcrypto, it's only intended to remove all the function name and
12472 reason texts, thereby removing some of the footprint that may not
12473 be interesting if those errors aren't displayed anyway.
12475 NOTE: it's still possible for any application or module to have its
12476 own set of error texts inserted. The routines are there, just not
12477 used by default when no-err is given.
12481 * Add support for FreeBSD on IA64.
12483 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12485 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12486 Kerberos function mit_des_cbc_cksum(). Before this change,
12487 the value returned by DES_cbc_cksum() was like the one from
12488 mit_des_cbc_cksum(), except the bytes were swapped.
12490 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12492 * Allow an application to disable the automatic SSL chain building.
12493 Before this a rather primitive chain build was always performed in
12494 ssl3_output_cert_chain(): an application had no way to send the
12495 correct chain if the automatic operation produced an incorrect result.
12497 Now the chain builder is disabled if either:
12499 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12501 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12503 The reasoning behind this is that an application would not want the
12504 auto chain building to take place if extra chain certificates are
12505 present and it might also want a means of sending no additional
12506 certificates (for example the chain has two certificates and the
12511 * Add the possibility to build without the ENGINE framework.
12513 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12515 * Under Win32 gmtime() can return NULL: check return value in
12516 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12520 * DSA routines: under certain error conditions uninitialized BN objects
12521 could be freed. Solution: make sure initialization is performed early
12522 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12523 Nils Larsch <nla@trustcenter.de> via PR#459)
12527 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12528 checked on reconnect on the client side, therefore session resumption
12529 could still fail with a "ssl session id is different" error. This
12530 behaviour is masked when SSL_OP_ALL is used due to
12531 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12532 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12533 followup to PR #377.
12537 * IA-32 assembler support enhancements: unified ELF targets, support
12538 for SCO/Caldera platforms, fix for Cygwin shared build.
12542 * Add support for FreeBSD on sparc64. As a consequence, support for
12543 FreeBSD on non-x86 processors is separate from x86 processors on
12544 the config script, much like the NetBSD support.
12546 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12548 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12550 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12553 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12554 code (06) was taken as the first octet of the session ID and the last
12555 octet was ignored consequently. As a result SSLv2 client side session
12556 caching could not have worked due to the session ID mismatch between
12558 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12563 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12564 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12569 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12570 seems that in spite of existing for more than a year, many application
12571 author have done nothing to provide the necessary callbacks, which
12572 means that this particular engine will not work properly anywhere.
12573 This is a very unfortunate situation which forces us, in the name
12574 of usability, to give the hw_ncipher.c a static lock, which is part
12576 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12577 appear in 0.9.8 or later. We EXPECT application authors to have
12578 dealt properly with this when 0.9.8 is released (unless we actually
12579 make such changes in the libcrypto locking code that changes will
12580 have to be made anyway).
12584 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12585 octets have been read, EOF or an error occurs. Without this change
12586 some truncated ASN1 structures will not produce an error.
12590 * Disable Heimdal support, since it hasn't been fully implemented.
12591 Still give the possibility to force the use of Heimdal, but with
12592 warnings and a request that patches get sent to openssl-dev.
12596 * Add the VC-CE target, introduce the WINCE sysname, and add
12597 INSTALL.WCE and appropriate conditionals to make it build.
12599 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12601 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12602 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12603 edit numbers of the version.
12605 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12607 * Introduce safe string copy and catenation functions
12608 (BUF_strlcpy() and BUF_strlcat()).
12610 *Ben Laurie (CHATS) and Richard Levitte*
12612 * Avoid using fixed-size buffers for one-line DNs.
12614 *Ben Laurie (CHATS)*
12616 * Add BUF_MEM_grow_clean() to avoid information leakage when
12617 resizing buffers containing secrets, and use where appropriate.
12619 *Ben Laurie (CHATS)*
12621 * Avoid using fixed size buffers for configuration file location.
12623 *Ben Laurie (CHATS)*
12625 * Avoid filename truncation for various CA files.
12627 *Ben Laurie (CHATS)*
12629 * Use sizeof in preference to magic numbers.
12631 *Ben Laurie (CHATS)*
12633 * Avoid filename truncation in cert requests.
12635 *Ben Laurie (CHATS)*
12637 * Add assertions to check for (supposedly impossible) buffer
12640 *Ben Laurie (CHATS)*
12642 * Don't cache truncated DNS entries in the local cache (this could
12643 potentially lead to a spoofing attack).
12645 *Ben Laurie (CHATS)*
12647 * Fix various buffers to be large enough for hex/decimal
12648 representations in a platform independent manner.
12650 *Ben Laurie (CHATS)*
12652 * Add CRYPTO_realloc_clean() to avoid information leakage when
12653 resizing buffers containing secrets, and use where appropriate.
12655 *Ben Laurie (CHATS)*
12657 * Add BIO_indent() to avoid much slightly worrying code to do
12660 *Ben Laurie (CHATS)*
12662 * Convert sprintf()/BIO_puts() to BIO_printf().
12664 *Ben Laurie (CHATS)*
12666 * buffer_gets() could terminate with the buffer only half
12669 *Ben Laurie (CHATS)*
12671 * Add assertions to prevent user-supplied crypto functions from
12672 overflowing internal buffers by having large block sizes, etc.
12674 *Ben Laurie (CHATS)*
12676 * New OPENSSL_assert() macro (similar to assert(), but enabled
12679 *Ben Laurie (CHATS)*
12681 * Eliminate unused copy of key in RC4.
12683 *Ben Laurie (CHATS)*
12685 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12687 *Ben Laurie (CHATS)*
12689 * Fix off-by-one error in EGD path.
12691 *Ben Laurie (CHATS)*
12693 * If RANDFILE path is too long, ignore instead of truncating.
12695 *Ben Laurie (CHATS)*
12697 * Eliminate unused and incorrectly sized X.509 structure
12700 *Ben Laurie (CHATS)*
12702 * Eliminate unused and dangerous function knumber().
12704 *Ben Laurie (CHATS)*
12706 * Eliminate unused and dangerous structure, KSSL_ERR.
12708 *Ben Laurie (CHATS)*
12710 * Protect against overlong session ID context length in an encoded
12711 session object. Since these are local, this does not appear to be
12714 *Ben Laurie (CHATS)*
12716 * Change from security patch (see 0.9.6e below) that did not affect
12717 the 0.9.6 release series:
12719 Remote buffer overflow in SSL3 protocol - an attacker could
12720 supply an oversized master key in Kerberos-enabled versions.
12723 *Ben Laurie (CHATS)*
12725 * Change the SSL kerb5 codes to match RFC 2712.
12729 * Make -nameopt work fully for req and add -reqopt switch.
12731 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12733 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12735 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12737 * Make sure tests can be performed even if the corresponding algorithms
12738 have been removed entirely. This was also the last step to make
12739 OpenSSL compilable with DJGPP under all reasonable conditions.
12741 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12743 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12744 to allow version independent disabling of normally unselected ciphers,
12745 which may be activated as a side-effect of selecting a single cipher.
12747 (E.g., cipher list string "RSA" enables ciphersuites that are left
12748 out of "ALL" because they do not provide symmetric encryption.
12749 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12751 *Lutz Jaenicke, Bodo Moeller*
12753 * Add appropriate support for separate platform-dependent build
12754 directories. The recommended way to make a platform-dependent
12755 build directory is the following (tested on Linux), maybe with
12758 # Place yourself outside of the OpenSSL source tree. In
12759 # this example, the environment variable OPENSSL_SOURCE
12760 # is assumed to contain the absolute OpenSSL source directory.
12761 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12762 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12763 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12764 mkdir -p `dirname $F`
12765 ln -s $OPENSSL_SOURCE/$F $F
12768 To be absolutely sure not to disturb the source tree, a "make clean"
12769 is a good thing. If it isn't successful, don't worry about it,
12770 it probably means the source directory is very clean.
12774 * Make sure any ENGINE control commands make local copies of string
12775 pointers passed to them whenever necessary. Otherwise it is possible
12776 the caller may have overwritten (or deallocated) the original string
12777 data when a later ENGINE operation tries to use the stored values.
12779 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12781 * Improve diagnostics in file reading and command-line digests.
12783 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12785 * Add AES modes CFB and OFB to the object database. Correct an
12786 error in AES-CFB decryption.
12790 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12791 allows existing EVP_CIPHER_CTX structures to be reused after
12792 calling `EVP_*Final()`. This behaviour is used by encryption
12793 BIOs and some applications. This has the side effect that
12794 applications must explicitly clean up cipher contexts with
12795 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12799 * Check the values of dna and dnb in bn_mul_recursive before calling
12800 bn_mul_comba (a non zero value means the a or b arrays do not contain
12801 n2 elements) and fallback to bn_mul_normal if either is not zero.
12805 * Fix escaping of non-ASCII characters when using the -subj option
12806 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12810 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12811 form for "surname", serialNumber has no short form.
12812 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12813 therefore remove "mail" short name for "internet 7".
12814 The OID for unique identifiers in X509 certificates is
12815 x500UniqueIdentifier, not uniqueIdentifier.
12816 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12820 * Add an "init" command to the ENGINE config module and auto initialize
12821 ENGINEs. Without any "init" command the ENGINE will be initialized
12822 after all ctrl commands have been executed on it. If init=1 the
12823 ENGINE is initialized at that point (ctrls before that point are run
12824 on the uninitialized ENGINE and after on the initialized one). If
12825 init=0 then the ENGINE will not be initialized at all.
12829 * Fix the 'app_verify_callback' interface so that the user-defined
12830 argument is actually passed to the callback: In the
12831 SSL_CTX_set_cert_verify_callback() prototype, the callback
12832 declaration has been changed from
12835 int (*cb)(X509_STORE_CTX *,void *);
12836 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12837 i=s->ctx->app_verify_callback(&ctx)
12838 has been changed into
12839 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12841 To update applications using SSL_CTX_set_cert_verify_callback(),
12842 a dummy argument can be added to their callback functions.
12844 *D. K. Smetters <smetters@parc.xerox.com>*
12846 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12848 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12850 * Add and OPENSSL_LOAD_CONF define which will cause
12851 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12852 This allows older applications to transparently support certain
12853 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12854 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12855 load the config file and OPENSSL_add_all_algorithms_conf() which will
12856 always load it have also been added.
12860 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12861 Adjust NIDs and EVP layer.
12863 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12865 * Config modules support in openssl utility.
12867 Most commands now load modules from the config file,
12868 though in a few (such as version) this isn't done
12869 because it couldn't be used for anything.
12871 In the case of ca and req the config file used is
12872 the same as the utility itself: that is the -config
12873 command line option can be used to specify an
12878 * Move default behaviour from OPENSSL_config(). If appname is NULL
12879 use "openssl_conf" if filename is NULL use default openssl config file.
12883 * Add an argument to OPENSSL_config() to allow the use of an alternative
12884 config section name. Add a new flag to tolerate a missing config file
12885 and move code to CONF_modules_load_file().
12889 * Support for crypto accelerator cards from Accelerated Encryption
12890 Processing, www.aep.ie. (Use engine 'aep')
12891 The support was copied from 0.9.6c [engine] and adapted/corrected
12892 to work with the new engine framework.
12894 *AEP Inc. and Richard Levitte*
12896 * Support for SureWare crypto accelerator cards from Baltimore
12897 Technologies. (Use engine 'sureware')
12898 The support was copied from 0.9.6c [engine] and adapted
12899 to work with the new engine framework.
12903 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12904 make the newer ENGINE framework commands for the CHIL engine work.
12906 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12908 * Make it possible to produce shared libraries on ReliantUNIX.
12910 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12912 * Add the configuration target debug-linux-ppro.
12913 Make 'openssl rsa' use the general key loading routines
12914 implemented in `apps.c`, and make those routines able to
12915 handle the key format FORMAT_NETSCAPE and the variant
12918 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12920 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12922 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12924 * Add -keyform to rsautl, and document -engine.
12926 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12928 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12929 BIO_R_NO_SUCH_FILE error code rather than the generic
12930 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12934 * Add new functions
12935 ERR_peek_last_error
12936 ERR_peek_last_error_line
12937 ERR_peek_last_error_line_data.
12938 These are similar to
12940 ERR_peek_error_line
12941 ERR_peek_error_line_data,
12942 but report on the latest error recorded rather than the first one
12943 still in the error queue.
12945 *Ben Laurie, Bodo Moeller*
12947 * default_algorithms option in ENGINE config module. This allows things
12949 default_algorithms = ALL
12950 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12954 * Preliminary ENGINE config module.
12958 * New experimental application configuration code.
12962 * Change the AES code to follow the same name structure as all other
12963 symmetric ciphers, and behave the same way. Move everything to
12964 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12966 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12968 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12970 *Ben Laurie and Theo de Raadt*
12972 * Add option to output public keys in req command.
12974 *Massimiliano Pala madwolf@openca.org*
12976 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12977 (up to about 10% better than before for P-192 and P-224).
12981 * New functions/macros
12983 SSL_CTX_set_msg_callback(ctx, cb)
12984 SSL_CTX_set_msg_callback_arg(ctx, arg)
12985 SSL_set_msg_callback(ssl, cb)
12986 SSL_set_msg_callback_arg(ssl, arg)
12988 to request calling a callback function
12990 void cb(int write_p, int version, int content_type,
12991 const void *buf, size_t len, SSL *ssl, void *arg)
12993 whenever a protocol message has been completely received
12994 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12995 protocol version according to which the SSL library interprets
12996 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12997 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12998 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12999 specification (change_cipher_spec(20), alert(21), handshake(22)).
13000 'buf' and 'len' point to the actual message, 'ssl' to the
13001 SSL object, and 'arg' is the application-defined value set by
13002 SSL[_CTX]_set_msg_callback_arg().
13004 'openssl s_client' and 'openssl s_server' have new '-msg' options
13005 to enable a callback that displays all protocol messages.
13009 * Change the shared library support so shared libraries are built as
13010 soon as the corresponding static library is finished, and thereby get
13011 openssl and the test programs linked against the shared library.
13012 This still only happens when the keyword "shard" has been given to
13013 the configuration scripts.
13015 NOTE: shared library support is still an experimental thing, and
13016 backward binary compatibility is still not guaranteed.
13018 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13020 * Add support for Subject Information Access extension.
13022 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13024 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13025 additional bytes when new memory had to be allocated, not just
13026 when reusing an existing buffer.
13030 * New command line and configuration option 'utf8' for the req command.
13031 This allows field values to be specified as UTF8 strings.
13035 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13036 runs for the former and machine-readable output for the latter.
13040 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13041 of the e-mail address in the DN (i.e., it will go into a certificate
13042 extension only). The new configuration file option 'email_in_dn = no'
13043 has the same effect.
13045 *Massimiliano Pala madwolf@openca.org*
13047 * Change all functions with names starting with `des_` to be starting
13048 with `DES_` instead. Add wrappers that are compatible with libdes,
13049 but are named `_ossl_old_des_*`. Finally, add macros that map the
13050 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13051 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13052 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13055 Since we provide two compatibility mappings, the user needs to
13056 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13057 compatibility is desired. The default (i.e., when that macro
13058 isn't defined) is OpenSSL 0.9.6c compatibility.
13060 There are also macros that enable and disable the support of old
13061 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13062 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13063 are defined, the default will apply: to support the old des routines.
13065 In either case, one must include openssl/des.h to get the correct
13066 definitions. Do not try to just include openssl/des_old.h, that
13069 NOTE: This is a major break of an old API into a new one. Software
13070 authors are encouraged to switch to the `DES_` style functions. Some
13071 time in the future, des_old.h and the libdes compatibility functions
13072 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13073 default), and then completely removed.
13077 * Test for certificates which contain unsupported critical extensions.
13078 If such a certificate is found during a verify operation it is
13079 rejected by default: this behaviour can be overridden by either
13080 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13081 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13082 X509_supported_extension() has also been added which returns 1 if a
13083 particular extension is supported.
13087 * Modify the behaviour of EVP cipher functions in similar way to digests
13088 to retain compatibility with existing code.
13092 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13093 compatibility with existing code. In particular the 'ctx' parameter does
13094 not have to be to be initialized before the call to EVP_DigestInit() and
13095 it is tidied up after a call to EVP_DigestFinal(). New function
13096 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13097 EVP_MD_CTX_copy() changed to not require the destination to be
13098 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13099 requires the destination to be valid.
13101 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13102 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13106 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13107 so that complete 'Handshake' protocol structures are kept in memory
13108 instead of overwriting 'msg_type' and 'length' with 'body' data.
13112 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13114 *Massimo Santin via Richard Levitte*
13116 * Major restructuring to the underlying ENGINE code. This includes
13117 reduction of linker bloat, separation of pure "ENGINE" manipulation
13118 (initialisation, etc) from functionality dealing with implementations
13119 of specific crypto interfaces. This change also introduces integrated
13120 support for symmetric ciphers and digest implementations - so ENGINEs
13121 can now accelerate these by providing EVP_CIPHER and EVP_MD
13122 implementations of their own. This is detailed in
13123 [crypto/engine/README.md](crypto/engine/README.md)
13124 as it couldn't be adequately described here. However, there are a few
13125 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13126 were changed in the original introduction of ENGINE code have now
13127 reverted back - the hooking from this code to ENGINE is now a good
13128 deal more passive and at run-time, operations deal directly with
13129 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13130 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13131 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13132 they were not being used by the framework as there is no concept of a
13133 BIGNUM_METHOD and they could not be generalised to the new
13134 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13135 ENGINE_cpy() has been removed as it cannot be consistently defined in
13140 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13144 * Change mkdef.pl to sort symbols that get the same entry number,
13145 and make sure the automatically generated functions `ERR_load_*`
13146 become part of libeay.num as well.
13150 * New function SSL_renegotiate_pending(). This returns true once
13151 renegotiation has been requested (either SSL_renegotiate() call
13152 or HelloRequest/ClientHello received from the peer) and becomes
13153 false once a handshake has been completed.
13154 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13155 sends a HelloRequest, but does not ensure that a handshake takes
13156 place. SSL_renegotiate_pending() is useful for checking if the
13157 client has followed the request.)
13161 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13162 By default, clients may request session resumption even during
13163 renegotiation (if session ID contexts permit); with this option,
13164 session resumption is possible only in the first handshake.
13166 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13167 more bits available for options that should not be part of
13168 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13172 * Add some demos for certificate and certificate request creation.
13176 * Make maximum certificate chain size accepted from the peer application
13177 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13178 "Douglas E. Engert" <deengert@anl.gov>.
13182 * Add support for shared libraries for Unixware-7
13183 (Boyd Lynn Gerber <gerberb@zenez.com>).
13187 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13188 be done prior to destruction. Use this to unload error strings from
13189 ENGINEs that load their own error strings. NB: This adds two new API
13190 functions to "get" and "set" this destroy handler in an ENGINE.
13194 * Alter all existing ENGINE implementations (except "openssl" and
13195 "openbsd") to dynamically instantiate their own error strings. This
13196 makes them more flexible to be built both as statically-linked ENGINEs
13197 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13198 Also, add stub code to each that makes building them as self-contained
13199 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13203 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13204 implementations into applications that are completely implemented in
13205 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13206 commands that can be used to configure what shared-library to load and
13207 to control aspects of the way it is handled. Also, made an update to
13208 the [README-Engine.md](README-Engine.md) file
13209 that brings its information up-to-date and
13210 provides some information and instructions on the "dynamic" ENGINE
13211 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13215 * Make it possible to unload ranges of ERR strings with a new
13216 "ERR_unload_strings" function.
13220 * Add a copy() function to EVP_MD.
13224 * Make EVP_MD routines take a context pointer instead of just the
13225 md_data void pointer.
13229 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13230 that the digest can only process a single chunk of data
13231 (typically because it is provided by a piece of
13232 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13233 is only going to provide a single chunk of data, and hence the
13234 framework needn't accumulate the data for oneshot drivers.
13238 * As with "ERR", make it possible to replace the underlying "ex_data"
13239 functions. This change also alters the storage and management of global
13240 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13241 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13242 index counters. The API functions that use this state have been changed
13243 to take a "class_index" rather than pointers to the class's local STACK
13244 and counter, and there is now an API function to dynamically create new
13245 classes. This centralisation allows us to (a) plug a lot of the
13246 thread-safety problems that existed, and (b) makes it possible to clean
13247 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13248 such data would previously have always leaked in application code and
13249 workarounds were in place to make the memory debugging turn a blind eye
13250 to it. Application code that doesn't use this new function will still
13251 leak as before, but their memory debugging output will announce it now
13252 rather than letting it slide.
13254 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13255 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13256 has a return value to indicate success or failure.
13260 * Make it possible to replace the underlying "ERR" functions such that the
13261 global state (2 LHASH tables and 2 locks) is only used by the "default"
13262 implementation. This change also adds two functions to "get" and "set"
13263 the implementation prior to it being automatically set the first time
13264 any other ERR function takes place. Ie. an application can call "get",
13265 pass the return value to a module it has just loaded, and that module
13266 can call its own "set" function using that value. This means the
13267 module's "ERR" operations will use (and modify) the error state in the
13268 application and not in its own statically linked copy of OpenSSL code.
13272 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13273 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13274 the operation, and provides a more encapsulated way for external code
13275 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13276 to use these functions rather than manually incrementing the counts.
13278 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13282 * Add EVP test program.
13286 * Add symmetric cipher support to ENGINE. Expect the API to change!
13290 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13291 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13292 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13293 These allow a CRL to be built without having to access X509_CRL fields
13294 directly. Modify 'ca' application to use new functions.
13298 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13299 bug workarounds. Rollback attack detection is a security feature.
13300 The problem will only arise on OpenSSL servers when TLSv1 is not
13301 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13302 Software authors not wanting to support TLSv1 will have special reasons
13303 for their choice and can explicitly enable this option.
13305 *Bodo Moeller, Lutz Jaenicke*
13307 * Rationalise EVP so it can be extended: don't include a union of
13308 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13309 (similar to those existing for EVP_CIPHER_CTX).
13314 EVP_MD_CTX_init(&md); /* new function call */
13315 EVP_DigestInit(&md, EVP_sha1());
13316 EVP_DigestUpdate(&md, in, len);
13317 EVP_DigestFinal(&md, out, NULL);
13318 EVP_MD_CTX_cleanup(&md); /* new function call */
13322 * Make DES key schedule conform to the usual scheme, as well as
13323 correcting its structure. This means that calls to DES functions
13324 now have to pass a pointer to a des_key_schedule instead of a
13325 plain des_key_schedule (which was actually always a pointer
13328 des_key_schedule ks;
13330 des_set_key_checked(..., &ks);
13331 des_ncbc_encrypt(..., &ks, ...);
13333 (Note that a later change renames 'des_...' into 'DES_...'.)
13337 * Initial reduction of linker bloat: the use of some functions, such as
13338 PEM causes large amounts of unused functions to be linked in due to
13339 poor organisation. For example pem_all.c contains every PEM function
13340 which has a knock on effect of linking in large amounts of (unused)
13341 ASN1 code. Grouping together similar functions and splitting unrelated
13342 functions prevents this.
13346 * Cleanup of EVP macros.
13350 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13351 correct `_ecb suffix`.
13355 * Add initial OCSP responder support to ocsp application. The
13356 revocation information is handled using the text based index
13357 use by the ca application. The responder can either handle
13358 requests generated internally, supplied in files (for example
13359 via a CGI script) or using an internal minimal server.
13363 * Add configuration choices to get zlib compression for TLS.
13367 * Changes to Kerberos SSL for RFC 2712 compliance:
13368 1. Implemented real KerberosWrapper, instead of just using
13369 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13370 2. Implemented optional authenticator field of KerberosWrapper.
13372 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13373 and authenticator structs; see crypto/krb5/.
13375 Generalized Kerberos calls to support multiple Kerberos libraries.
13376 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13377 via Richard Levitte*
13379 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13380 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13381 values for each of the key sizes rather than having just
13382 parameters (and 'speed' generating keys each time).
13386 * Speed up EVP routines.
13389 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13390 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13391 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13392 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13394 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13395 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13396 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13399 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13401 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13405 * Added the OS2-EMX target.
13407 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13409 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13410 New functions to support `NCONF` routines in extension code.
13411 New function `CONF_set_nconf()`
13412 to allow functions which take an `NCONF` to also handle the old `LHASH`
13413 structure: this means that the old `CONF` compatible routines can be
13414 retained (in particular w.rt. extensions) without having to duplicate the
13415 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13419 * Enhance the general user interface with mechanisms for inner control
13420 and with possibilities to have yes/no kind of prompts.
13424 * Change all calls to low-level digest routines in the library and
13425 applications to use EVP. Add missing calls to HMAC_cleanup() and
13426 don't assume HMAC_CTX can be copied using memcpy().
13428 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13430 * Add the possibility to control engines through control names but with
13431 arbitrary arguments instead of just a string.
13432 Change the key loaders to take a UI_METHOD instead of a callback
13433 function pointer. NOTE: this breaks binary compatibility with earlier
13434 versions of OpenSSL [engine].
13435 Adapt the nCipher code for these new conditions and add a card insertion
13440 * Enhance the general user interface with mechanisms to better support
13441 dialog box interfaces, application-defined prompts, the possibility
13442 to use defaults (for example default passwords from somewhere else)
13443 and interrupts/cancellations.
13447 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13448 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13452 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13453 tidy up some unnecessarily weird code in 'sk_new()').
13455 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13457 * Change the key loading routines for ENGINEs to use the same kind
13458 callback (pem_password_cb) as all other routines that need this
13463 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13464 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13465 than this minimum value is recommended.
13469 * New random seeder for OpenVMS, using the system process statistics
13470 that are easily reachable.
13474 * Windows apparently can't transparently handle global
13475 variables defined in DLLs. Initialisations such as:
13477 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13479 won't compile. This is used by the any applications that need to
13480 declare their own ASN1 modules. This was fixed by adding the option
13481 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13482 needed for static libraries under Win32.
13486 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13487 setting of purpose and trust fields. New X509_STORE trust and
13488 purpose functions and tidy up setting in other SSL functions.
13492 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13493 structure. These are inherited by X509_STORE_CTX when it is
13494 initialised. This allows various defaults to be set in the
13495 X509_STORE structure (such as flags for CRL checking and custom
13496 purpose or trust settings) for functions which only use X509_STORE_CTX
13497 internally such as S/MIME.
13499 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13500 trust settings if they are not set in X509_STORE. This allows X509_STORE
13501 purposes and trust (in S/MIME for example) to override any set by default.
13503 Add command line options for CRL checking to smime, s_client and s_server
13508 * Initial CRL based revocation checking. If the CRL checking flag(s)
13509 are set then the CRL is looked up in the X509_STORE structure and
13510 its validity and signature checked, then if the certificate is found
13511 in the CRL the verify fails with a revoked error.
13513 Various new CRL related callbacks added to X509_STORE_CTX structure.
13515 Command line options added to 'verify' application to support this.
13517 This needs some additional work, such as being able to handle multiple
13518 CRLs with different times, extension based lookup (rather than just
13519 by subject name) and ultimately more complete V2 CRL extension
13524 * Add a general user interface API (crypto/ui/). This is designed
13525 to replace things like des_read_password and friends (backward
13526 compatibility functions using this new API are provided).
13527 The purpose is to remove prompting functions from the DES code
13528 section as well as provide for prompting through dialog boxes in
13529 a window system and the like.
13533 * Add "ex_data" support to ENGINE so implementations can add state at a
13534 per-structure level rather than having to store it globally.
13538 * Make it possible for ENGINE structures to be copied when retrieved by
13539 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13540 This causes the "original" ENGINE structure to act like a template,
13541 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13542 operational state can be localised to each ENGINE structure, despite the
13543 fact they all share the same "methods". New ENGINE structures returned in
13544 this case have no functional references and the return value is the single
13545 structural reference. This matches the single structural reference returned
13546 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13551 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13552 needs to match any other type at all we need to manually clear the
13557 * Changes to the "openssl engine" utility to include;
13558 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13559 about an ENGINE's available control commands.
13560 - executing control commands from command line arguments using the
13561 '-pre' and '-post' switches. '-post' is only used if '-t' is
13562 specified and the ENGINE is successfully initialised. The syntax for
13563 the individual commands are colon-separated, for example;
13564 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13568 * New dynamic control command support for ENGINEs. ENGINEs can now
13569 declare their own commands (numbers), names (strings), descriptions,
13570 and input types for run-time discovery by calling applications. A
13571 subset of these commands are implicitly classed as "executable"
13572 depending on their input type, and only these can be invoked through
13573 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13574 can be based on user input, config files, etc). The distinction is
13575 that "executable" commands cannot return anything other than a boolean
13576 result and can only support numeric or string input, whereas some
13577 discoverable commands may only be for direct use through
13578 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13579 pointers, or other custom uses. The "executable" commands are to
13580 support parameterisations of ENGINE behaviour that can be
13581 unambiguously defined by ENGINEs and used consistently across any
13582 OpenSSL-based application. Commands have been added to all the
13583 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13584 control over shared-library paths without source code alterations.
13588 * Changed all ENGINE implementations to dynamically allocate their
13589 ENGINEs rather than declaring them statically. Apart from this being
13590 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13591 this also allows the implementations to compile without using the
13592 internal engine_int.h header.
13596 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13597 'const' value. Any code that should be able to modify a RAND_METHOD
13598 should already have non-const pointers to it (ie. they should only
13599 modify their own ones).
13603 * Made a variety of little tweaks to the ENGINE code.
13604 - "atalla" and "ubsec" string definitions were moved from header files
13605 to C code. "nuron" string definitions were placed in variables
13606 rather than hard-coded - allowing parameterisation of these values
13607 later on via ctrl() commands.
13608 - Removed unused "#if 0"'d code.
13609 - Fixed engine list iteration code so it uses ENGINE_free() to release
13610 structural references.
13611 - Constified the RAND_METHOD element of ENGINE structures.
13612 - Constified various get/set functions as appropriate and added
13613 missing functions (including a catch-all ENGINE_cpy that duplicates
13614 all ENGINE values onto a new ENGINE except reference counts/state).
13615 - Removed NULL parameter checks in get/set functions. Setting a method
13616 or function to NULL is a way of cancelling out a previously set
13617 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13618 and doesn't justify the extra error symbols and code.
13619 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13620 flags from engine_int.h to engine.h.
13621 - Changed prototypes for ENGINE handler functions (init(), finish(),
13622 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13626 * Implement binary inversion algorithm for BN_mod_inverse in addition
13627 to the algorithm using long division. The binary algorithm can be
13628 used only if the modulus is odd. On 32-bit systems, it is faster
13629 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13630 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13631 up to 450 bits. In 64-bit environments, the binary algorithm
13632 appears to be advantageous for much longer moduli; here we use it
13633 for moduli up to 2048 bits.
13637 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13638 could not support the combine flag in choice fields.
13642 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13643 extensions from a certificate request to the certificate.
13647 * Allow multiple 'certopt' and 'nameopt' options to be separated
13648 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13649 file: this allows the display of the certificate about to be
13650 signed to be customised, to allow certain fields to be included
13651 or excluded and extension details. The old system didn't display
13652 multicharacter strings properly, omitted fields not in the policy
13653 and couldn't display additional details such as extensions.
13657 * Function EC_POINTs_mul for multiple scalar multiplication
13658 of an arbitrary number of elliptic curve points
13659 \sum scalars[i]*points[i],
13660 optionally including the generator defined for the EC_GROUP:
13661 scalar*generator + \sum scalars[i]*points[i].
13663 EC_POINT_mul is a simple wrapper function for the typical case
13664 that the point list has just one item (besides the optional
13669 * First EC_METHODs for curves over GF(p):
13671 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13672 operations and provides various method functions that can also
13673 operate with faster implementations of modular arithmetic.
13675 EC_GFp_mont_method() reuses most functions that are part of
13676 EC_GFp_simple_method, but uses Montgomery arithmetic.
13678 *Bodo Moeller; point addition and point doubling
13679 implementation directly derived from source code provided by
13680 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13682 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13683 crypto/ec/ec_lib.c):
13685 Curves are EC_GROUP objects (with an optional group generator)
13686 based on EC_METHODs that are built into the library.
13688 Points are EC_POINT objects based on EC_GROUP objects.
13690 Most of the framework would be able to handle curves over arbitrary
13691 finite fields, but as there are no obvious types for fields other
13692 than GF(p), some functions are limited to that for now.
13696 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13697 that the file contains a complete HTTP response.
13701 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13702 change the def and num file printf format specifier from "%-40sXXX"
13703 to "%-39s XXX". The latter will always guarantee a space after the
13704 field while the former will cause them to run together if the field
13705 is 40 of more characters long.
13709 * Constify the cipher and digest 'method' functions and structures
13710 and modify related functions to take constant EVP_MD and EVP_CIPHER
13715 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13716 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13720 * Modify `EVP_Digest*()` routines so they now return values. Although the
13721 internal software routines can never fail additional hardware versions
13726 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13728 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13729 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13732 ERR_R_NESTED_ASN1_ERROR
13734 ERR_R_MISSING_ASN1_EOS
13735 were 4 .. 9, conflicting with
13736 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13738 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13739 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13741 Add new error code 'ERR_R_INTERNAL_ERROR'.
13745 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13750 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13751 sets the subject name for a new request or supersedes the
13752 subject name in a given request. Formats that can be parsed are
13753 'CN=Some Name, OU=myOU, C=IT'
13755 'CN=Some Name/OU=myOU/C=IT'.
13757 Add options '-batch' and '-verbose' to 'openssl req'.
13759 *Massimiliano Pala <madwolf@hackmasters.net>*
13761 * Introduce the possibility to access global variables through
13762 functions on platform were that's the best way to handle exporting
13763 global variables in shared libraries. To enable this functionality,
13764 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13765 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13766 is normally done by Configure or something similar).
13768 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13769 in the source file (foo.c) like this:
13771 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13772 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13774 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13775 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13777 OPENSSL_DECLARE_GLOBAL(int,foo);
13778 #define foo OPENSSL_GLOBAL_REF(foo)
13779 OPENSSL_DECLARE_GLOBAL(double,bar);
13780 #define bar OPENSSL_GLOBAL_REF(bar)
13782 The #defines are very important, and therefore so is including the
13783 header file everywhere where the defined globals are used.
13785 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13786 of ASN.1 items, but that structure is a bit different.
13788 The largest change is in util/mkdef.pl which has been enhanced with
13789 better and easier to understand logic to choose which symbols should
13790 go into the Windows .def files as well as a number of fixes and code
13791 cleanup (among others, algorithm keywords are now sorted
13792 lexicographically to avoid constant rewrites).
13796 * In BN_div() keep a copy of the sign of 'num' before writing the
13797 result to 'rm' because if rm==num the value will be overwritten
13798 and produce the wrong result if 'num' is negative: this caused
13799 problems with BN_mod() and BN_nnmod().
13803 * Function OCSP_request_verify(). This checks the signature on an
13804 OCSP request and verifies the signer certificate. The signer
13805 certificate is just checked for a generic purpose and OCSP request
13810 * Add OCSP_check_validity() function to check the validity of OCSP
13811 responses. OCSP responses are prepared in real time and may only
13812 be a few seconds old. Simply checking that the current time lies
13813 between thisUpdate and nextUpdate max reject otherwise valid responses
13814 caused by either OCSP responder or client clock inaccuracy. Instead
13815 we allow thisUpdate and nextUpdate to fall within a certain period of
13816 the current time. The age of the response can also optionally be
13817 checked. Two new options -validity_period and -status_age added to
13822 * If signature or public key algorithm is unrecognized print out its
13823 OID rather that just UNKNOWN.
13827 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13828 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13829 ID to be generated from the issuer certificate alone which can then be
13830 passed to OCSP_id_issuer_cmp().
13834 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13835 ASN1 modules to export functions returning ASN1_ITEM pointers
13836 instead of the ASN1_ITEM structures themselves. This adds several
13837 new macros which allow the underlying ASN1 function/structure to
13838 be accessed transparently. As a result code should not use ASN1_ITEM
13839 references directly (such as &X509_it) but instead use the relevant
13840 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13841 use of the new ASN1 code on platforms where exporting structures
13842 is problematical (for example in shared libraries) but exporting
13843 functions returning pointers to structures is not.
13847 * Add support for overriding the generation of SSL/TLS session IDs.
13848 These callbacks can be registered either in an SSL_CTX or per SSL.
13849 The purpose of this is to allow applications to control, if they wish,
13850 the arbitrary values chosen for use as session IDs, particularly as it
13851 can be useful for session caching in multiple-server environments. A
13852 command-line switch for testing this (and any client code that wishes
13853 to use such a feature) has been added to "s_server".
13855 *Geoff Thorpe, Lutz Jaenicke*
13857 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13858 of the form `#if defined(...) || defined(...) || ...` and
13859 `#if !defined(...) && !defined(...) && ...`. This also avoids
13860 the growing number of special cases it was previously handling.
13864 * Make all configuration macros available for application by making
13865 sure they are available in opensslconf.h, by giving them names starting
13866 with `OPENSSL_` to avoid conflicts with other packages and by making
13867 sure e_os2.h will cover all platform-specific cases together with
13869 Additionally, it is now possible to define configuration/platform-
13870 specific names (called "system identities"). In the C code, these
13871 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13872 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13873 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13878 * New option -set_serial to 'req' and 'x509' this allows the serial
13879 number to use to be specified on the command line. Previously self
13880 signed certificates were hard coded with serial number 0 and the
13881 CA options of 'x509' had to use a serial number in a file which was
13886 * New options to 'ca' utility to support V2 CRL entry extensions.
13887 Currently CRL reason, invalidity date and hold instruction are
13888 supported. Add new CRL extensions to V3 code and some new objects.
13892 * New function EVP_CIPHER_CTX_set_padding() this is used to
13893 disable standard block padding (aka PKCS#5 padding) in the EVP
13894 API, which was previously mandatory. This means that the data is
13895 not padded in any way and so the total length much be a multiple
13896 of the block size, otherwise an error occurs.
13900 * Initial (incomplete) OCSP SSL support.
13904 * New function OCSP_parse_url(). This splits up a URL into its host,
13905 port and path components: primarily to parse OCSP URLs. New -url
13906 option to ocsp utility.
13910 * New nonce behavior. The return value of OCSP_check_nonce() now
13911 reflects the various checks performed. Applications can decide
13912 whether to tolerate certain situations such as an absent nonce
13913 in a response when one was present in a request: the ocsp application
13914 just prints out a warning. New function OCSP_add1_basic_nonce()
13915 this is to allow responders to include a nonce in a response even if
13916 the request is nonce-less.
13920 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13921 skipped when using openssl x509 multiple times on a single input file,
13922 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13926 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13927 set string type: to handle setting ASN1_TIME structures. Fix ca
13928 utility to correctly initialize revocation date of CRLs.
13932 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13933 the clients preferred ciphersuites and rather use its own preferences.
13934 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13935 Internet Explorer by ensuring unchanged hash method during stepup.
13936 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13940 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13941 to aes and add a new 'exist' option to print out symbols that don't
13946 * Additional options to ocsp utility to allow flags to be set and
13947 additional certificates supplied.
13951 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13952 OCSP client a number of certificate to only verify the response
13957 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13958 handle the new API. Currently only ECB, CBC modes supported. Add new
13961 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13962 Encryption Standard (AES) Ciphersuites for Transport Layer
13963 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13964 not enabled by default and were not part of the "ALL" ciphersuite
13965 alias because they were not yet official; they could be
13966 explicitly requested by specifying the "AESdraft" ciphersuite
13967 group alias. In the final release of OpenSSL 0.9.7, the group
13968 alias is called "AES" and is part of "ALL".)
13970 *Ben Laurie, Steve Henson, Bodo Moeller*
13972 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13973 request to response.
13977 * Functions for OCSP responders. OCSP_request_onereq_count(),
13978 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13979 extract information from a certificate request. OCSP_response_create()
13980 creates a response and optionally adds a basic response structure.
13981 OCSP_basic_add1_status() adds a complete single response to a basic
13982 response and returns the OCSP_SINGLERESP structure just added (to allow
13983 extensions to be included for example). OCSP_basic_add1_cert() adds a
13984 certificate to a basic response and OCSP_basic_sign() signs a basic
13985 response with various flags. New helper functions ASN1_TIME_check()
13986 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13987 (converts ASN1_TIME to GeneralizedTime).
13991 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13992 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13993 structure from a certificate. X509_pubkey_digest() digests the public_key
13994 contents: this is used in various key identifiers.
13998 * Make sk_sort() tolerate a NULL argument.
14000 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14002 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14003 passed by the function are trusted implicitly. If any of them signed the
14004 response then it is assumed to be valid and is not verified.
14008 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14009 to data. This was previously part of the PKCS7 ASN1 code. This
14010 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14011 *Steve Henson, reported by Kenneth R. Robinette
14012 <support@securenetterm.com>*
14014 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14015 routines: without these tracing memory leaks is very painful.
14016 Fix leaks in PKCS12 and PKCS7 routines.
14020 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14021 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14022 effectively meant GeneralizedTime would never be used. Now it
14023 is initialised to -1 but X509_time_adj() now has to check the value
14024 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14025 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14026 *Steve Henson, reported by Kenneth R. Robinette
14027 <support@securenetterm.com>*
14029 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14030 result in a zero length in the ASN1_INTEGER structure which was
14031 not consistent with the structure when d2i_ASN1_INTEGER() was used
14032 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14033 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14034 where it did not print out a minus for negative ASN1_INTEGER.
14038 * Add summary printout to ocsp utility. The various functions which
14039 convert status values to strings have been renamed to:
14040 OCSP_response_status_str(), OCSP_cert_status_str() and
14041 OCSP_crl_reason_str() and are no longer static. New options
14042 to verify nonce values and to disable verification. OCSP response
14043 printout format cleaned up.
14047 * Add additional OCSP certificate checks. These are those specified
14048 in RFC2560. This consists of two separate checks: the CA of the
14049 certificate being checked must either be the OCSP signer certificate
14050 or the issuer of the OCSP signer certificate. In the latter case the
14051 OCSP signer certificate must contain the OCSP signing extended key
14052 usage. This check is performed by attempting to match the OCSP
14053 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14054 in the OCSP_CERTID structures of the response.
14058 * Initial OCSP certificate verification added to OCSP_basic_verify()
14059 and related routines. This uses the standard OpenSSL certificate
14060 verify routines to perform initial checks (just CA validity) and
14061 to obtain the certificate chain. Then additional checks will be
14062 performed on the chain. Currently the root CA is checked to see
14063 if it is explicitly trusted for OCSP signing. This is used to set
14064 a root CA as a global signing root: that is any certificate that
14065 chains to that CA is an acceptable OCSP signing certificate.
14069 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14070 extensions from a separate configuration file.
14071 As when reading extensions from the main configuration file,
14072 the '-extensions ...' option may be used for specifying the
14075 *Massimiliano Pala <madwolf@comune.modena.it>*
14077 * New OCSP utility. Allows OCSP requests to be generated or
14078 read. The request can be sent to a responder and the output
14079 parsed, outputted or printed in text form. Not complete yet:
14080 still needs to check the OCSP response validity.
14084 * New subcommands for 'openssl ca':
14085 `openssl ca -status <serial>` prints the status of the cert with
14086 the given serial number (according to the index file).
14087 `openssl ca -updatedb` updates the expiry status of certificates
14090 *Massimiliano Pala <madwolf@comune.modena.it>*
14092 * New '-newreq-nodes' command option to CA.pl. This is like
14093 '-newreq', but calls 'openssl req' with the '-nodes' option
14094 so that the resulting key is not encrypted.
14096 *Damien Miller <djm@mindrot.org>*
14098 * New configuration for the GNU Hurd.
14100 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14102 * Initial code to implement OCSP basic response verify. This
14103 is currently incomplete. Currently just finds the signer's
14104 certificate and verifies the signature on the response.
14108 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14109 value of OPENSSLDIR. This is available via the new '-d' option
14110 to 'openssl version', and is also included in 'openssl version -a'.
14114 * Allowing defining memory allocation callbacks that will be given
14115 file name and line number information in additional arguments
14116 (a `const char*` and an int). The basic functionality remains, as
14117 well as the original possibility to just replace malloc(),
14118 realloc() and free() by functions that do not know about these
14119 additional arguments. To register and find out the current
14120 settings for extended allocation functions, the following
14121 functions are provided:
14123 CRYPTO_set_mem_ex_functions
14124 CRYPTO_set_locked_mem_ex_functions
14125 CRYPTO_get_mem_ex_functions
14126 CRYPTO_get_locked_mem_ex_functions
14128 These work the same way as CRYPTO_set_mem_functions and friends.
14129 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14130 extended allocation function is enabled.
14131 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14132 a conventional allocation function is enabled.
14134 *Richard Levitte, Bodo Moeller*
14136 * Finish off removing the remaining LHASH function pointer casts.
14137 There should no longer be any prototype-casting required when using
14138 the LHASH abstraction, and any casts that remain are "bugs". See
14139 the callback types and macros at the head of lhash.h for details
14140 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14144 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14145 If /dev/[u]random devices are not available or do not return enough
14146 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14148 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14149 /etc/entropy will be queried once each in this sequence, querying stops
14150 when enough entropy was collected without querying more sockets.
14154 * Change the Unix RAND_poll() variant to be able to poll several
14155 random devices, as specified by DEVRANDOM, until a sufficient amount
14156 of data has been collected. We spend at most 10 ms on each file
14157 (select timeout) and read in non-blocking mode. DEVRANDOM now
14158 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14159 (previously it was just the string "/dev/urandom"), so on typical
14160 platforms the 10 ms delay will never occur.
14161 Also separate out the Unix variant to its own file, rand_unix.c.
14162 For VMS, there's a currently-empty rand_vms.c.
14166 * Move OCSP client related routines to ocsp_cl.c. These
14167 provide utility functions which an application needing
14168 to issue a request to an OCSP responder and analyse the
14169 response will typically need: as opposed to those which an
14170 OCSP responder itself would need which will be added later.
14172 OCSP_request_sign() signs an OCSP request with an API similar
14173 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14174 response. OCSP_response_get1_basic() extracts basic response
14175 from response. OCSP_resp_find_status(): finds and extracts status
14176 information from an OCSP_CERTID structure (which will be created
14177 when the request structure is built). These are built from lower
14178 level functions which work on OCSP_SINGLERESP structures but
14179 won't normally be used unless the application wishes to examine
14180 extensions in the OCSP response for example.
14182 Replace nonce routines with a pair of functions.
14183 OCSP_request_add1_nonce() adds a nonce value and optionally
14184 generates a random value. OCSP_check_nonce() checks the
14185 validity of the nonce in an OCSP response.
14189 * Change function OCSP_request_add() to OCSP_request_add0_id().
14190 This doesn't copy the supplied OCSP_CERTID and avoids the
14191 need to free up the newly created id. Change return type
14192 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14193 This can then be used to add extensions to the request.
14194 Deleted OCSP_request_new(), since most of its functionality
14195 is now in OCSP_REQUEST_new() (and the case insensitive name
14196 clash) apart from the ability to set the request name which
14197 will be added elsewhere.
14201 * Update OCSP API. Remove obsolete extensions argument from
14202 various functions. Extensions are now handled using the new
14203 OCSP extension code. New simple OCSP HTTP function which
14204 can be used to send requests and parse the response.
14208 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14209 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14210 uses the special reorder version of SET OF to sort the attributes
14211 and reorder them to match the encoded order. This resolves a long
14212 standing problem: a verify on a PKCS7 structure just after signing
14213 it used to fail because the attribute order did not match the
14214 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14215 it uses the received order. This is necessary to tolerate some broken
14216 software that does not order SET OF. This is handled by encoding
14217 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14218 to produce the required SET OF.
14222 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14223 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14224 files to get correct declarations of the ASN.1 item variables.
14228 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14229 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14230 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14231 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14232 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14233 ASN1_ITEM and no wrapper functions.
14237 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14238 replace the old function pointer based I/O routines. Change most of
14239 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14243 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14244 lines, recognize more "algorithms" that can be deselected, and make
14245 it complain about algorithm deselection that isn't recognised.
14249 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14250 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14251 to use new functions. Add NO_ASN1_OLD which can be set to remove
14252 some old style ASN1 functions: this can be used to determine if old
14253 code will still work when these eventually go away.
14257 * New extension functions for OCSP structures, these follow the
14258 same conventions as certificates and CRLs.
14262 * New function X509V3_add1_i2d(). This automatically encodes and
14263 adds an extension. Its behaviour can be customised with various
14264 flags to append, replace or delete. Various wrappers added for
14265 certificates and CRLs.
14269 * Fix to avoid calling the underlying ASN1 print routine when
14270 an extension cannot be parsed. Correct a typo in the
14271 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14275 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14276 entries for variables.
14280 * Add functionality to `apps/openssl.c` for detecting locking
14281 problems: As the program is single-threaded, all we have
14282 to do is register a locking callback using an array for
14283 storing which locks are currently held by the program.
14287 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14288 SSL_get_ex_data_X509_STORE_idx(), which is used in
14289 ssl_verify_cert_chain() and thus can be called at any time
14290 during TLS/SSL handshakes so that thread-safety is essential.
14291 Unfortunately, the ex_data design is not at all suited
14292 for multi-threaded use, so it probably should be abolished.
14296 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14298 *Broadcom, tweaked and integrated by Geoff Thorpe*
14300 * Move common extension printing code to new function
14301 X509V3_print_extensions(). Reorganise OCSP print routines and
14302 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14306 * New function X509_signature_print() to remove duplication in some
14311 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14312 set (this was treated exactly the same as SET OF previously). This
14313 is used to reorder the STACK representing the structure to match the
14314 encoding. This will be used to get round a problem where a PKCS7
14315 structure which was signed could not be verified because the STACK
14316 order did not reflect the encoded order.
14320 * Reimplement the OCSP ASN1 module using the new code.
14324 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14325 for its ASN1 operations. The old style function pointers still exist
14326 for now but they will eventually go away.
14330 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14331 completely replaces the old ASN1 functionality with a table driven
14332 encoder and decoder which interprets an ASN1_ITEM structure describing
14333 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14334 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14335 has also been converted to the new form.
14339 * Change BN_mod_exp_recp so that negative moduli are tolerated
14340 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14341 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14342 for negative moduli.
14346 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14347 of not touching the result's sign bit.
14351 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14356 * Changed the LHASH code to use prototypes for callbacks, and created
14357 macros to declare and implement thin (optionally static) functions
14358 that provide type-safety and avoid function pointer casting for the
14359 type-specific callbacks.
14363 * Added Kerberos Cipher Suites to be used with TLS, as written in
14365 *Veers Staats <staatsvr@asc.hpc.mil>,
14366 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14368 * Reformat the FAQ so the different questions and answers can be divided
14369 in sections depending on the subject.
14373 * Have the zlib compression code load ZLIB.DLL dynamically under
14378 * New function BN_mod_sqrt for computing square roots modulo a prime
14379 (using the probabilistic Tonelli-Shanks algorithm unless
14380 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14381 be handled deterministically).
14383 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14385 * Make BN_mod_inverse faster by explicitly handling small quotients
14386 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14387 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14391 * New function BN_kronecker.
14395 * Fix BN_gcd so that it works on negative inputs; the result is
14396 positive unless both parameters are zero.
14397 Previously something reasonably close to an infinite loop was
14398 possible because numbers could be growing instead of shrinking
14399 in the implementation of Euclid's algorithm.
14403 * Fix BN_is_word() and BN_is_one() macros to take into account the
14404 sign of the number in question.
14406 Fix BN_is_word(a,w) to work correctly for w == 0.
14408 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14409 because its test if the absolute value of 'a' equals 'w'.
14410 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14411 it exists mostly for use in the implementations of BN_is_zero(),
14412 BN_is_one(), and BN_is_word().
14416 * New function BN_swap.
14420 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14421 the exponentiation functions are more likely to produce reasonable
14422 results on negative inputs.
14426 * Change BN_mod_mul so that the result is always non-negative.
14427 Previously, it could be negative if one of the factors was negative;
14428 I don't think anyone really wanted that behaviour.
14432 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14433 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14434 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14435 and add new functions:
14444 BN_mod_lshift1_quick
14446 BN_mod_lshift_quick
14448 These functions always generate non-negative results.
14450 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14451 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14453 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14454 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14455 be reduced modulo `m`.
14457 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14460 The following entry accidentally appeared in the CHANGES file
14461 distributed with OpenSSL 0.9.7. The modifications described in
14462 it do *not* apply to OpenSSL 0.9.7.
14464 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14465 was actually never needed) and in BN_mul(). The removal in BN_mul()
14466 required a small change in bn_mul_part_recursive() and the addition
14467 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14468 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14469 bn_sub_words() and bn_add_words() except they take arrays with
14475 * In 'openssl passwd', verify passwords read from the terminal
14476 unless the '-salt' option is used (which usually means that
14477 verification would just waste user's time since the resulting
14478 hash is going to be compared with some given password hash)
14479 or the new '-noverify' option is used.
14481 This is an incompatible change, but it does not affect
14482 non-interactive use of 'openssl passwd' (passwords on the command
14483 line, '-stdin' option, '-in ...' option) and thus should not
14484 cause any problems.
14488 * Remove all references to RSAref, since there's no more need for it.
14492 * Make DSO load along a path given through an environment variable
14493 (SHLIB_PATH) with shl_load().
14497 * Constify the ENGINE code as a result of BIGNUM constification.
14498 Also constify the RSA code and most things related to it. In a
14499 few places, most notable in the depth of the ASN.1 code, ugly
14500 casts back to non-const were required (to be solved at a later
14505 * Make it so the openssl application has all engines loaded by default.
14509 * Constify the BIGNUM routines a little more.
14513 * Add the following functions:
14515 ENGINE_load_cswift()
14517 ENGINE_load_atalla()
14518 ENGINE_load_nuron()
14519 ENGINE_load_builtin_engines()
14521 That way, an application can itself choose if external engines that
14522 are built-in in OpenSSL shall ever be used or not. The benefit is
14523 that applications won't have to be linked with libdl or other dso
14524 libraries unless it's really needed.
14526 Changed 'openssl engine' to load all engines on demand.
14527 Changed the engine header files to avoid the duplication of some
14528 declarations (they differed!).
14532 * 'openssl engine' can now list capabilities.
14536 * Better error reporting in 'openssl engine'.
14540 * Never call load_dh_param(NULL) in s_server.
14544 * Add engine application. It can currently list engines by name and
14545 identity, and test if they are actually available.
14549 * Improve RPM specification file by forcing symbolic linking and making
14550 sure the installed documentation is also owned by root.root.
14552 *Damien Miller <djm@mindrot.org>*
14554 * Give the OpenSSL applications more possibilities to make use of
14555 keys (public as well as private) handled by engines.
14559 * Add OCSP code that comes from CertCo.
14563 * Add VMS support for the Rijndael code.
14567 * Added untested support for Nuron crypto accelerator.
14571 * Add support for external cryptographic devices. This code was
14572 previously distributed separately as the "engine" branch.
14574 *Geoff Thorpe, Richard Levitte*
14576 * Rework the filename-translation in the DSO code. It is now possible to
14577 have far greater control over how a "name" is turned into a filename
14578 depending on the operating environment and any oddities about the
14579 different shared library filenames on each system.
14583 * Support threads on FreeBSD-elf in Configure.
14587 * Fix for SHA1 assembly problem with MASM: it produces
14588 warnings about corrupt line number information when assembling
14589 with debugging information. This is caused by the overlapping
14592 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14595 NCONF_get_number() has no error checking at all. As a replacement,
14596 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14597 promoted strongly. The old NCONF_get_number is kept around for
14598 binary backward compatibility.
14599 Make it possible for methods to load from something other than a BIO,
14600 by providing a function pointer that is given a name instead of a BIO.
14601 For example, this could be used to load configuration data from an
14606 * Fix for non blocking accept BIOs. Added new I/O special reason
14607 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14608 with non blocking I/O was not possible because no retry code was
14609 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14614 * Added the beginnings of Rijndael support.
14618 * Fix for bug in DirectoryString mask setting. Add support for
14619 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14620 to allow certificate printing to more controllable, additional
14621 'certopt' option to 'x509' to allow new printing options to be
14626 * Clean old EAY MD5 hack from e_os.h.
14630 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14632 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14633 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14635 *Joe Orton, Steve Henson*
14637 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14639 * Fix additional bug revealed by the NISCC test suite:
14641 Stop bug triggering large recursion when presented with
14642 certain ASN.1 tags ([CVE-2003-0851])
14646 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14648 * Fix various bugs revealed by running the NISCC test suite:
14650 Stop out of bounds reads in the ASN1 code when presented with
14651 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14653 If verify callback ignores invalid public key errors don't try to check
14654 certificate signature with the NULL public key.
14658 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14659 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14664 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14665 extra data after the compression methods not only for TLS 1.0
14666 but also for SSL 3.0 (as required by the specification).
14668 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14670 * Change X509_certificate_type() to mark the key as exported/exportable
14671 when it's 512 *bits* long, not 512 bytes.
14675 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14677 * Countermeasure against the Klima-Pokorny-Rosa extension of
14678 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14679 a protocol version number mismatch like a decryption error
14680 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14684 * Turn on RSA blinding by default in the default implementation
14685 to avoid a timing attack. Applications that don't want it can call
14686 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14687 They would be ill-advised to do so in most cases.
14689 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14691 * Change RSA blinding code so that it works when the PRNG is not
14692 seeded (in this case, the secret RSA exponent is abused as
14693 an unpredictable seed -- if it is not unpredictable, there
14694 is no point in blinding anyway). Make RSA blinding thread-safe
14695 by remembering the creator's thread ID in rsa->blinding and
14696 having all other threads use local one-time blinding factors
14697 (this requires more computation than sharing rsa->blinding, but
14698 avoids excessive locking; and if an RSA object is not shared
14699 between threads, blinding will still be very fast).
14703 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14705 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14706 via timing by performing a MAC computation even if incorrect
14707 block cipher padding has been found. This is a countermeasure
14708 against active attacks where the attacker has to distinguish
14709 between bad padding and a MAC verification error. ([CVE-2003-0078])
14711 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14712 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14713 Martin Vuagnoux (EPFL, Ilion)*
14715 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14717 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14718 memory from its contents. This is done with a counter that will
14719 place alternating values in each byte. This can be used to solve
14720 two issues: 1) the removal of calls to memset() by highly optimizing
14721 compilers, and 2) cleansing with other values than 0, since those can
14722 be read through on certain media, for example a swap space on disk.
14726 * Bugfix: client side session caching did not work with external caching,
14727 because the session->cipher setting was not restored when reloading
14728 from the external cache. This problem was masked, when
14729 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14730 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14734 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14735 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14737 *Zeev Lieber <zeev-l@yahoo.com>*
14739 * Undo an undocumented change introduced in 0.9.6e which caused
14740 repeated calls to OpenSSL_add_all_ciphers() and
14741 OpenSSL_add_all_digests() to be ignored, even after calling
14746 * Change the default configuration reader to deal with last line not
14747 being properly terminated.
14751 * Change X509_NAME_cmp() so it applies the special rules on handling
14752 DN values that are of type PrintableString, as well as RDNs of type
14753 emailAddress where the value has the type ia5String.
14755 *stefank@valicert.com via Richard Levitte*
14757 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14758 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14759 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14760 the bitwise-OR of the two for use by the majority of applications
14761 wanting this behaviour, and update the docs. The documented
14762 behaviour and actual behaviour were inconsistent and had been
14763 changing anyway, so this is more a bug-fix than a behavioural
14766 *Geoff Thorpe, diagnosed by Nadav Har'El*
14768 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14769 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14773 * Fix initialization code race conditions in
14774 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14775 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14776 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14777 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14778 ssl2_get_cipher_by_char(),
14779 ssl3_get_cipher_by_char().
14781 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14783 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14784 the cached sessions are flushed, as the remove_cb() might use ex_data
14785 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14786 (see [openssl.org #212]).
14788 *Geoff Thorpe, Lutz Jaenicke*
14790 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14791 length, instead of the encoding length to d2i_ASN1_OBJECT.
14795 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14797 * [In 0.9.6g-engine release:]
14798 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14800 *Lynn Gazis <lgazis@rainbow.com>*
14802 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14804 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14805 and get fix the header length calculation.
14806 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14807 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14809 * Use proper error handling instead of 'assertions' in buffer
14810 overflow checks added in 0.9.6e. This prevents DoS (the
14811 assertions could call abort()).
14813 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14815 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14817 * Add various sanity checks to asn1_get_length() to reject
14818 the ASN1 length bytes if they exceed sizeof(long), will appear
14819 negative or the content length exceeds the length of the
14822 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14824 * Fix cipher selection routines: ciphers without encryption had no flags
14825 for the cipher strength set and where therefore not handled correctly
14826 by the selection routines (PR #130).
14830 * Fix EVP_dsa_sha macro.
14835 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14836 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14837 that was added in OpenSSL 0.9.6d.
14839 As the countermeasure turned out to be incompatible with some
14840 broken SSL implementations, the new option is part of SSL_OP_ALL.
14841 SSL_OP_ALL is usually employed when compatibility with weird SSL
14842 implementations is desired (e.g. '-bugs' option to 's_client' and
14843 's_server'), so the new option is automatically set in many
14848 * Changes in security patch:
14850 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14851 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14852 Air Force Materiel Command, USAF, under agreement number
14855 * Add various sanity checks to asn1_get_length() to reject
14856 the ASN1 length bytes if they exceed sizeof(long), will appear
14857 negative or the content length exceeds the length of the
14858 supplied buffer. ([CVE-2002-0659])
14860 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14862 * Assertions for various potential buffer overflows, not known to
14863 happen in practice.
14865 *Ben Laurie (CHATS)*
14867 * Various temporary buffers to hold ASCII versions of integers were
14868 too small for 64 bit platforms. ([CVE-2002-0655])
14869 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14871 * Remote buffer overflow in SSL3 protocol - an attacker could
14872 supply an oversized session ID to a client. ([CVE-2002-0656])
14874 *Ben Laurie (CHATS)*
14876 * Remote buffer overflow in SSL2 protocol - an attacker could
14877 supply an oversized client master key. ([CVE-2002-0656])
14879 *Ben Laurie (CHATS)*
14881 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14883 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14884 encoded as NULL) with id-dsa-with-sha1.
14886 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14888 * Check various `X509_...()` return values in `apps/req.c`.
14890 *Nils Larsch <nla@trustcenter.de>*
14892 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14893 an end-of-file condition would erroneously be flagged, when the CRLF
14894 was just at the end of a processed block. The bug was discovered when
14895 processing data through a buffering memory BIO handing the data to a
14896 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14897 <ptsekov@syntrex.com> and Nedelcho Stanev.
14901 * Implement a countermeasure against a vulnerability recently found
14902 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14903 before application data chunks to avoid the use of known IVs
14904 with data potentially chosen by the attacker.
14908 * Fix length checks in ssl3_get_client_hello().
14912 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14913 to prevent ssl3_read_internal() from incorrectly assuming that
14914 ssl3_read_bytes() found application data while handshake
14915 processing was enabled when in fact s->s3->in_read_app_data was
14916 merely automatically cleared during the initial handshake.
14918 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14920 * Fix object definitions for Private and Enterprise: they were not
14921 recognized in their shortname (=lowercase) representation. Extend
14922 obj_dat.pl to issue an error when using undefined keywords instead
14923 of silently ignoring the problem (Svenning Sorensen
14924 <sss@sss.dnsalias.net>).
14928 * Fix DH_generate_parameters() so that it works for 'non-standard'
14929 generators, i.e. generators other than 2 and 5. (Previously, the
14930 code did not properly initialise the 'add' and 'rem' values to
14931 BN_generate_prime().)
14933 In the new general case, we do not insist that 'generator' is
14934 actually a primitive root: This requirement is rather pointless;
14935 a generator of the order-q subgroup is just as good, if not
14940 * Map new X509 verification errors to alerts. Discovered and submitted by
14941 Tom Wu <tom@arcot.com>.
14945 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14946 returning non-zero before the data has been completely received
14947 when using non-blocking I/O.
14949 *Bodo Moeller; problem pointed out by John Hughes*
14951 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14953 *Ben Laurie, Lutz Jaenicke*
14955 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14956 Yoram Zahavi <YoramZ@gilian.com>).
14960 * Add information about CygWin 1.3 and on, and preserve proper
14961 configuration for the versions before that.
14963 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14965 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14966 check whether we deal with a copy of a session and do not delete from
14967 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14968 <izhar@checkpoint.com>.
14972 * Do not store session data into the internal session cache, if it
14973 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14974 flag is set). Proposed by Aslam <aslam@funk.com>.
14978 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14983 * [In 0.9.6d-engine release:]
14984 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14986 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14988 * Add the configuration target linux-s390x.
14990 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14992 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14993 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14994 variable as an indication that a ClientHello message has been
14995 received. As the flag value will be lost between multiple
14996 invocations of ssl3_accept when using non-blocking I/O, the
14997 function may not be aware that a handshake has actually taken
14998 place, thus preventing a new session from being added to the
15001 To avoid this problem, we now set s->new_session to 2 instead of
15002 using a local variable.
15004 *Lutz Jaenicke, Bodo Moeller*
15006 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15007 if the SSL_R_LENGTH_MISMATCH error is detected.
15009 *Geoff Thorpe, Bodo Moeller*
15011 * New 'shared_ldflag' column in Configure platform table.
15015 * Fix EVP_CIPHER_mode macro.
15017 *"Dan S. Camper" <dan@bti.net>*
15019 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15020 type, we must throw them away by setting rr->length to 0.
15022 *D P Chang <dpc@qualys.com>*
15024 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15026 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15027 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15028 worked incorrectly for those cases where range = `10..._2` and
15029 `3*range` is two bits longer than range.)
15033 * Only add signing time to PKCS7 structures if it is not already
15038 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15039 OBJ_ld_ce should be OBJ_id_ce.
15040 Also some ip-pda OIDs in crypto/objects/objects.txt were
15041 incorrect (cf. RFC 3039).
15043 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15045 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15046 returns early because it has nothing to do.
15048 *Andy Schneider <andy.schneider@bjss.co.uk>*
15050 * [In 0.9.6c-engine release:]
15051 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15053 *Andy Schneider <andy.schneider@bjss.co.uk>*
15055 * [In 0.9.6c-engine release:]
15056 Add support for Cryptographic Appliance's keyserver technology.
15057 (Use engine 'keyclient')
15059 *Cryptographic Appliances and Geoff Thorpe*
15061 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15062 is called via tools/c89.sh because arguments have to be
15063 rearranged (all '-L' options must appear before the first object
15066 *Richard Shapiro <rshapiro@abinitio.com>*
15068 * [In 0.9.6c-engine release:]
15069 Add support for Broadcom crypto accelerator cards, backported
15072 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15074 * [In 0.9.6c-engine release:]
15075 Add support for SureWare crypto accelerator cards from
15076 Baltimore Technologies. (Use engine 'sureware')
15078 *Baltimore Technologies and Mark Cox*
15080 * [In 0.9.6c-engine release:]
15081 Add support for crypto accelerator cards from Accelerated
15082 Encryption Processing, www.aep.ie. (Use engine 'aep')
15084 *AEP Inc. and Mark Cox*
15086 * Add a configuration entry for gcc on UnixWare.
15088 *Gary Benson <gbenson@redhat.com>*
15090 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15091 messages are stored in a single piece (fixed-length part and
15092 variable-length part combined) and fix various bugs found on the way.
15096 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15097 instead. BIO_gethostbyname() does not know what timeouts are
15098 appropriate, so entries would stay in cache even when they have
15100 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15102 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15103 faced with a pathologically small ClientHello fragment that does
15104 not contain client_version: Instead of aborting with an error,
15105 simply choose the highest available protocol version (i.e.,
15106 TLS 1.0 unless it is disabled). In practice, ClientHello
15107 messages are never sent like this, but this change gives us
15108 strictly correct behaviour at least for TLS.
15112 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15113 never resets s->method to s->ctx->method when called from within
15114 one of the SSL handshake functions.
15116 *Bodo Moeller; problem pointed out by Niko Baric*
15118 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15119 (sent using the client's version number) if client_version is
15120 smaller than the protocol version in use. Also change
15121 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15122 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15123 the client will at least see that alert.
15127 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15132 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15133 client receives HelloRequest while in a handshake.
15135 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15137 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15138 should end in 'break', not 'goto end' which circumvents various
15139 cleanups done in state SSL_ST_OK. But session related stuff
15140 must be disabled for SSL_ST_OK in the case that we just sent a
15143 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15144 before just sending a HelloRequest.
15146 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15148 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15149 reveal whether illegal block cipher padding was found or a MAC
15150 verification error occurred. (Neither SSLerr() codes nor alerts
15151 are directly visible to potential attackers, but the information
15152 may leak via logfiles.)
15154 Similar changes are not required for the SSL 2.0 implementation
15155 because the number of padding bytes is sent in clear for SSL 2.0,
15156 and the extra bytes are just ignored. However ssl/s2_pkt.c
15157 failed to verify that the purported number of padding bytes is in
15162 * Add OpenUNIX-8 support including shared libraries
15163 (Boyd Lynn Gerber <gerberb@zenez.com>).
15167 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15168 'wristwatch attack' using huge encoding parameters (cf.
15169 James H. Manger's CRYPTO 2001 paper). Note that the
15170 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15171 encoding parameters and hence was not vulnerable.
15175 * BN_sqr() bug fix.
15177 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15179 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15180 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15181 followed by modular reduction.
15183 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15185 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15186 equivalent based on BN_pseudo_rand() instead of BN_rand().
15190 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15191 This function was broken, as the check for a new client hello message
15192 to handle SGC did not allow these large messages.
15193 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15197 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15201 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15202 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15206 * Rework the configuration and shared library support for Tru64 Unix.
15207 The configuration part makes use of modern compiler features and
15208 still retains old compiler behavior for those that run older versions
15209 of the OS. The shared library support part includes a variant that
15210 uses the RPATH feature, and is available through the special
15211 configuration target "alpha-cc-rpath", which will never be selected
15214 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15216 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15217 with the same message size as in ssl3_get_certificate_request().
15218 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15219 messages might inadvertently be reject as too long.
15221 *Petr Lampa <lampa@fee.vutbr.cz>*
15223 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15227 * Modified SSL library such that the verify_callback that has been set
15228 specifically for an SSL object with SSL_set_verify() is actually being
15229 used. Before the change, a verify_callback set with this function was
15230 ignored and the verify_callback() set in the SSL_CTX at the time of
15231 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15232 to allow the necessary settings.
15236 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15237 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15238 done automatically (in contradiction to the requirements of the C
15239 standard). This made problems when used from OpenSSH.
15243 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15244 dh->length and always used
15246 BN_rand_range(priv_key, dh->p).
15248 BN_rand_range() is not necessary for Diffie-Hellman, and this
15249 specific range makes Diffie-Hellman unnecessarily inefficient if
15250 dh->length (recommended exponent length) is much smaller than the
15251 length of dh->p. We could use BN_rand_range() if the order of
15252 the subgroup was stored in the DH structure, but we only have
15257 BN_rand(priv_key, l, ...)
15259 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15266 RSA_eay_public_encrypt
15267 RSA_eay_private_decrypt
15268 RSA_eay_private_encrypt (signing)
15269 RSA_eay_public_decrypt (signature verification)
15271 (default implementations for RSA_public_encrypt,
15272 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15273 always reject numbers >= n.
15277 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15278 to synchronize access to 'locking_thread'. This is necessary on
15279 systems where access to 'locking_thread' (an 'unsigned long'
15280 variable) is not atomic.
15284 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15285 *before* setting the 'crypto_lock_rand' flag. The previous code had
15286 a race condition if 0 is a valid thread ID.
15288 *Travis Vitek <vitek@roguewave.com>*
15290 * Add support for shared libraries under Irix.
15292 *Albert Chin-A-Young <china@thewrittenword.com>*
15294 * Add configuration option to build on Linux on both big-endian and
15295 little-endian MIPS.
15297 *Ralf Baechle <ralf@uni-koblenz.de>*
15299 * Add the possibility to create shared libraries on HP-UX.
15303 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15305 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15306 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15307 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15308 PRNG state recovery was possible based on the output of
15309 one PRNG request appropriately sized to gain knowledge on
15310 'md' followed by enough consecutive 1-byte PRNG requests
15311 to traverse all of 'state'.
15313 1. When updating 'md_local' (the current thread's copy of 'md')
15314 during PRNG output generation, hash all of the previous
15315 'md_local' value, not just the half used for PRNG output.
15317 2. Make the number of bytes from 'state' included into the hash
15318 independent from the number of PRNG bytes requested.
15320 The first measure alone would be sufficient to avoid
15321 Markku-Juhani's attack. (Actually it had never occurred
15322 to me that the half of 'md_local' used for chaining was the
15323 half from which PRNG output bytes were taken -- I had always
15324 assumed that the secret half would be used.) The second
15325 measure makes sure that additional data from 'state' is never
15326 mixed into 'md_local' in small portions; this heuristically
15327 further strengthens the PRNG.
15331 * Fix crypto/bn/asm/mips3.s.
15335 * When only the key is given to "enc", the IV is undefined. Print out
15336 an error message in this case.
15340 * Handle special case when X509_NAME is empty in X509 printing routines.
15344 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15345 positive and less than q.
15349 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15350 used: it isn't thread safe and the add_lock_callback should handle
15353 *Paul Rose <Paul.Rose@bridge.com>*
15355 * Verify that incoming data obeys the block size in
15356 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15362 *Ulf Möller, Bodo Möller*
15364 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15365 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15366 when fixing the server behaviour for backwards-compatible 'client
15367 hello' messages. (Note that the attack is impractical against
15368 SSL 3.0 and TLS 1.0 anyway because length and version checking
15369 means that the probability of guessing a valid ciphertext is
15370 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15373 Before 0.9.5, the countermeasure (hide the error by generating a
15374 random 'decryption result') did not work properly because
15375 ERR_clear_error() was missing, meaning that SSL_get_error() would
15376 detect the supposedly ignored error.
15378 Both problems are now fixed.
15382 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15383 (previously it was 1024).
15387 * Fix for compatibility mode trust settings: ignore trust settings
15388 unless some valid trust or reject settings are present.
15392 * Fix for blowfish EVP: its a variable length cipher.
15396 * Fix various bugs related to DSA S/MIME verification. Handle missing
15397 parameters in DSA public key structures and return an error in the
15398 DSA routines if parameters are absent.
15402 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15403 in the current directory if neither $RANDFILE nor $HOME was set.
15404 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15405 caused some confusion to Windows users who haven't defined $HOME.
15406 Thus RAND_file_name() is changed again: e_os.h can define a
15407 DEFAULT_HOME, which will be used if $HOME is not set.
15408 For Windows, we use "C:"; on other platforms, we still require
15409 environment variables.
15411 * Move 'if (!initialized) RAND_poll()' into regions protected by
15412 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15413 having multiple threads call RAND_poll() concurrently.
15417 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15418 combination of a flag and a thread ID variable.
15419 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15420 flag), *other* threads can enter ssleay_add_bytes without obeying
15421 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15422 that they do not hold after the first thread unsets add_do_not_lock).
15426 * Change bctest again: '-x' expressions are not available in all
15427 versions of 'test'.
15431 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15433 * Fix a couple of memory leaks in PKCS7_dataDecode()
15435 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15437 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15438 the default extension for executables, if any. Also, make the perl
15439 scripts that use symlink() to test if it really exists and use "cp"
15440 if it doesn't. All this made OpenSSL compilable and installable in
15445 * Fix for asn1_GetSequence() for indefinite length constructed data.
15446 If SEQUENCE is length is indefinite just set c->slen to the total
15447 amount of data available.
15449 *Steve Henson, reported by shige@FreeBSD.org*
15451 *This change does not apply to 0.9.7.*
15453 * Change bctest to avoid here-documents inside command substitution
15454 (workaround for FreeBSD /bin/sh bug).
15455 For compatibility with Ultrix, avoid shell functions (introduced
15456 in the bctest version that searches along $PATH).
15460 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15461 with des_encrypt() defined on some operating systems, like Solaris
15466 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15467 On the Importance of Eliminating Errors in Cryptographic
15468 Computations, J. Cryptology 14 (2001) 2, 101-119,
15469 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15473 * MIPS assembler BIGNUM division bug fix.
15477 * Disabled incorrect Alpha assembler code.
15481 * Fix PKCS#7 decode routines so they correctly update the length
15482 after reading an EOC for the EXPLICIT tag.
15486 *This change does not apply to 0.9.7.*
15488 * Fix bug in PKCS#12 key generation routines. This was triggered
15489 if a 3DES key was generated with a 0 initial byte. Include
15490 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15491 (but broken) behaviour.
15495 * Enhance bctest to search for a working bc along $PATH and print
15498 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15500 * Fix memory leaks in err.c: free err_data string if necessary;
15501 don't write to the wrong index in ERR_set_error_data.
15505 * Implement ssl23_peek (analogous to ssl23_read), which previously
15510 * Replace rdtsc with `_emit` statements for VC++ version 5.
15512 *Jeremy Cooper <jeremy@baymoo.org>*
15514 * Make it possible to reuse SSLv2 sessions.
15518 * In copy_email() check for >= 0 as a return value for
15519 X509_NAME_get_index_by_NID() since 0 is a valid index.
15521 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15523 * Avoid coredump with unsupported or invalid public keys by checking if
15524 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15525 PKCS7_verify() fails with non detached data.
15529 * Don't use getenv in library functions when run as setuid/setgid.
15530 New function OPENSSL_issetugid().
15534 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15535 due to incorrect handling of multi-threading:
15537 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15539 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15541 3. Count how many times MemCheck_off() has been called so that
15542 nested use can be treated correctly. This also avoids
15543 inband-signalling in the previous code (which relied on the
15544 assumption that thread ID 0 is impossible).
15548 * Add "-rand" option also to s_client and s_server.
15552 * Fix CPU detection on Irix 6.x.
15553 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15554 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15556 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15561 *This change does not apply to 0.9.7.*
15563 * Use the cached encoding of an X509_NAME structure rather than
15564 copying it. This is apparently the reason for the libsafe "errors"
15565 but the code is actually correct.
15569 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15570 Bleichenbacher's DSA attack.
15571 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15572 to be set and top=0 forces the highest bit to be set; top=-1 is new
15573 and leaves the highest bit random.
15575 *Ulf Moeller, Bodo Moeller*
15577 * In the `NCONF_...`-based implementations for `CONF_...` queries
15578 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15579 a temporary CONF structure with the data component set to NULL
15580 (which gives segmentation faults in lh_retrieve).
15581 Instead, use NULL for the CONF pointer in CONF_get_string and
15582 CONF_get_number (which may use environment variables) and directly
15583 return NULL from CONF_get_section.
15587 * Fix potential buffer overrun for EBCDIC.
15591 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15592 keyUsage if basicConstraints absent for a CA.
15596 * Make SMIME_write_PKCS7() write mail header values with a format that
15597 is more generally accepted (no spaces before the semicolon), since
15598 some programs can't parse those values properly otherwise. Also make
15599 sure BIO's that break lines after each write do not create invalid
15604 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15605 macros previously used would not encode an empty SEQUENCE OF
15606 and break the signature.
15610 *This change does not apply to 0.9.7.*
15612 * Zero the premaster secret after deriving the master secret in
15617 * Add some EVP_add_digest_alias registrations (as found in
15618 OpenSSL_add_all_digests()) to SSL_library_init()
15619 aka OpenSSL_add_ssl_algorithms(). This provides improved
15620 compatibility with peers using X.509 certificates
15621 with unconventional AlgorithmIdentifier OIDs.
15625 * Fix for Irix with NO_ASM.
15627 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15629 * ./config script fixes.
15631 *Ulf Moeller, Richard Levitte*
15633 * Fix 'openssl passwd -1'.
15637 * Change PKCS12_key_gen_asc() so it can cope with non null
15638 terminated strings whose length is passed in the passlen
15639 parameter, for example from PEM callbacks. This was done
15640 by adding an extra length parameter to asc2uni().
15642 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15644 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15645 call failed, free the DSA structure.
15649 * Fix to uni2asc() to cope with zero length Unicode strings.
15650 These are present in some PKCS#12 files.
15654 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15655 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15656 when writing a 32767 byte record.
15658 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15660 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15661 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15663 (RSA objects have a reference count access to which is protected
15664 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15665 so they are meant to be shared between threads.)
15666 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15667 "Reddie, Steven" <Steven.Reddie@ca.com>*
15669 * Fix a deadlock in CRYPTO_mem_leaks().
15673 * Use better test patterns in bntest.
15677 * rand_win.c fix for Borland C.
15681 * BN_rshift bugfix for n == 0.
15685 * Add a 'bctest' script that checks for some known 'bc' bugs
15686 so that 'make test' does not abort just because 'bc' is broken.
15690 * Store verify_result within SSL_SESSION also for client side to
15691 avoid potential security hole. (Re-used sessions on the client side
15692 always resulted in verify_result==X509_V_OK, not using the original
15693 result of the server certificate verification.)
15697 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15698 SSL3_RT_APPLICATION_DATA, return 0.
15699 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15704 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15705 releases, have been re-implemented by renaming the previous
15706 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15707 and ssl3_read_internal, respectively, and adding 'peek' parameters
15708 to them. The new ssl[23]_{read,peek} functions are calls to
15709 ssl[23]_read_internal with the 'peek' flag set appropriately.
15710 A 'peek' parameter has also been added to ssl3_read_bytes, which
15711 does the actual work for ssl3_read_internal.
15715 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15716 the method-specific "init()" handler. Also clean up ex_data after
15717 calling the method-specific "finish()" handler. Previously, this was
15718 happening the other way round.
15722 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15723 The previous value, 12, was not always sufficient for BN_mod_exp().
15727 * Make sure that shared libraries get the internal name engine with
15728 the full version number and not just 0. This should mark the
15729 shared libraries as not backward compatible. Of course, this should
15730 be changed again when we can guarantee backward binary compatibility.
15734 * Fix typo in get_cert_by_subject() in by_dir.c
15736 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15738 * Rework the system to generate shared libraries:
15740 - Make note of the expected extension for the shared libraries and
15741 if there is a need for symbolic links from for example libcrypto.so.0
15742 to libcrypto.so.0.9.7. There is extended info in Configure for
15745 - Make as few rebuilds of the shared libraries as possible.
15747 - Still avoid linking the OpenSSL programs with the shared libraries.
15749 - When installing, install the shared libraries separately from the
15754 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15756 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15757 and not in SSL_clear because the latter is also used by the
15758 accept/connect functions; previously, the settings made by
15759 SSL_set_read_ahead would be lost during the handshake.
15761 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15763 * Correct util/mkdef.pl to be selective about disabled algorithms.
15764 Previously, it would create entries for disabled algorithms no
15769 * Added several new manual pages for SSL_* function.
15773 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15775 * In ssl23_get_client_hello, generate an error message when faced
15776 with an initial SSL 3.0/TLS record that is too small to contain the
15777 first two bytes of the ClientHello message, i.e. client_version.
15778 (Note that this is a pathologic case that probably has never happened
15779 in real life.) The previous approach was to use the version number
15780 from the record header as a substitute; but our protocol choice
15781 should not depend on that one because it is not authenticated
15782 by the Finished messages.
15786 * More robust randomness gathering functions for Windows.
15788 *Jeffrey Altman <jaltman@columbia.edu>*
15790 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15791 not set then we don't setup the error code for issuer check errors
15792 to avoid possibly overwriting other errors which the callback does
15793 handle. If an application does set the flag then we assume it knows
15794 what it is doing and can handle the new informational codes
15799 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15800 a general "ANY" type, as such it should be able to decode anything
15801 including tagged types. However it didn't check the class so it would
15802 wrongly interpret tagged types in the same way as their universal
15803 counterpart and unknown types were just rejected. Changed so that the
15804 tagged and unknown types are handled in the same way as a SEQUENCE:
15805 that is the encoding is stored intact. There is also a new type
15806 "V_ASN1_OTHER" which is used when the class is not universal, in this
15807 case we have no idea what the actual type is so we just lump them all
15812 * On VMS, stdout may very well lead to a file that is written to
15813 in a record-oriented fashion. That means that every write() will
15814 write a separate record, which will be read separately by the
15815 programs trying to read from it. This can be very confusing.
15817 The solution is to put a BIO filter in the way that will buffer
15818 text until a linefeed is reached, and then write everything a
15819 line at a time, so every record written will be an actual line,
15820 not chunks of lines and not (usually doesn't happen, but I've
15821 seen it once) several lines in one record. BIO_f_linebuffer() is
15824 Currently, it's a VMS-only method, because that's where it has
15825 been tested well enough.
15829 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15830 it can return incorrect results.
15831 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15832 but it was in 0.9.6-beta[12].)
15836 * Disable the check for content being present when verifying detached
15837 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15838 include zero length content when signing messages.
15842 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15843 BIO_ctrl (for BIO pairs).
15847 * Add DSO method for VMS.
15851 * Bug fix: Montgomery multiplication could produce results with the
15856 * Add RPM specification openssl.spec and modify it to build three
15857 packages. The default package contains applications, application
15858 documentation and run-time libraries. The devel package contains
15859 include files, static libraries and function documentation. The
15860 doc package contains the contents of the doc directory. The original
15861 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15865 * Add a large number of documentation files for many SSL routines.
15867 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15869 * Add a configuration entry for Sony News 4.
15871 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15873 * Don't set the two most significant bits to one when generating a
15874 random number < q in the DSA library.
15878 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15879 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15880 the underlying transport is blocking) if a handshake took place.
15881 (The default behaviour is needed by applications such as s_client
15882 and s_server that use select() to determine when to use SSL_read;
15883 but for applications that know in advance when to expect data, it
15884 just makes things more complicated.)
15888 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15893 * Add a few more EBCDIC conditionals that make `req` and `x509`
15894 work better on such systems.
15896 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15898 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15899 Update PKCS12_parse() so it copies the friendlyName and the
15900 keyid to the certificates aux info.
15904 * Fix bug in PKCS7_verify() which caused an infinite loop
15905 if there was more than one signature.
15907 *Sven Uszpelkat <su@celocom.de>*
15909 * Major change in util/mkdef.pl to include extra information
15910 about each symbol, as well as presenting variables as well
15911 as functions. This change means that there's n more need
15912 to rebuild the .num files when some algorithms are excluded.
15916 * Allow the verify time to be set by an application,
15917 rather than always using the current time.
15921 * Phase 2 verify code reorganisation. The certificate
15922 verify code now looks up an issuer certificate by a
15923 number of criteria: subject name, authority key id
15924 and key usage. It also verifies self signed certificates
15925 by the same criteria. The main comparison function is
15926 X509_check_issued() which performs these checks.
15928 Lot of changes were necessary in order to support this
15929 without completely rewriting the lookup code.
15931 Authority and subject key identifier are now cached.
15933 The LHASH 'certs' is X509_STORE has now been replaced
15934 by a STACK_OF(X509_OBJECT). This is mainly because an
15935 LHASH can't store or retrieve multiple objects with
15936 the same hash value.
15938 As a result various functions (which were all internal
15939 use only) have changed to handle the new X509_STORE
15940 structure. This will break anything that messed round
15941 with X509_STORE internally.
15943 The functions X509_STORE_add_cert() now checks for an
15944 exact match, rather than just subject name.
15946 The X509_STORE API doesn't directly support the retrieval
15947 of multiple certificates matching a given criteria, however
15948 this can be worked round by performing a lookup first
15949 (which will fill the cache with candidate certificates)
15950 and then examining the cache for matches. This is probably
15951 the best we can do without throwing out X509_LOOKUP
15952 entirely (maybe later...).
15954 The X509_VERIFY_CTX structure has been enhanced considerably.
15956 All certificate lookup operations now go via a get_issuer()
15957 callback. Although this currently uses an X509_STORE it
15958 can be replaced by custom lookups. This is a simple way
15959 to bypass the X509_STORE hackery necessary to make this
15960 work and makes it possible to use more efficient techniques
15961 in future. A very simple version which uses a simple
15962 STACK for its trusted certificate store is also provided
15963 using X509_STORE_CTX_trusted_stack().
15965 The verify_cb() and verify() callbacks now have equivalents
15966 in the X509_STORE_CTX structure.
15968 X509_STORE_CTX also has a 'flags' field which can be used
15969 to customise the verify behaviour.
15973 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15974 excludes S/MIME capabilities.
15978 * When a certificate request is read in keep a copy of the
15979 original encoding of the signed data and use it when outputting
15980 again. Signatures then use the original encoding rather than
15981 a decoded, encoded version which may cause problems if the
15982 request is improperly encoded.
15986 * For consistency with other BIO_puts implementations, call
15987 buffer_write(b, ...) directly in buffer_puts instead of calling
15990 In BIO_puts, increment b->num_write as in BIO_write.
15992 *Peter.Sylvester@EdelWeb.fr*
15994 * Fix BN_mul_word for the case where the word is 0. (We have to use
15995 BN_zero, we may not return a BIGNUM with an array consisting of
15996 words set to zero.)
16000 * Avoid calling abort() from within the library when problems are
16001 detected, except if preprocessor symbols have been defined
16002 (such as REF_CHECK, BN_DEBUG etc.).
16006 * New openssl application 'rsautl'. This utility can be
16007 used for low-level RSA operations. DER public key
16008 BIO/fp routines also added.
16012 * New Configure entry and patches for compiling on QNX 4.
16014 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16016 * A demo state-machine implementation was sponsored by
16017 Nuron (<http://www.nuron.com/>) and is now available in
16018 demos/state_machine.
16022 * New options added to the 'dgst' utility for signature
16023 generation and verification.
16027 * Unrecognized PKCS#7 content types are now handled via a
16028 catch all ASN1_TYPE structure. This allows unsupported
16029 types to be stored as a "blob" and an application can
16030 encode and decode it manually.
16034 * Fix various signed/unsigned issues to make a_strex.c
16035 compile under VC++.
16037 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16039 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16040 length if passed a buffer. ASN1_INTEGER_to_BN failed
16041 if passed a NULL BN and its argument was negative.
16043 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16045 * Modification to PKCS#7 encoding routines to output definite
16046 length encoding. Since currently the whole structures are in
16047 memory there's not real point in using indefinite length
16048 constructed encoding. However if OpenSSL is compiled with
16049 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16053 * Added BIO_vprintf() and BIO_vsnprintf().
16057 * Added more prefixes to parse for in the strings written
16058 through a logging bio, to cover all the levels that are available
16059 through syslog. The prefixes are now:
16061 PANIC, EMERG, EMR => LOG_EMERG
16062 ALERT, ALR => LOG_ALERT
16063 CRIT, CRI => LOG_CRIT
16064 ERROR, ERR => LOG_ERR
16065 WARNING, WARN, WAR => LOG_WARNING
16066 NOTICE, NOTE, NOT => LOG_NOTICE
16067 INFO, INF => LOG_INFO
16068 DEBUG, DBG => LOG_DEBUG
16070 and as before, if none of those prefixes are present at the
16071 beginning of the string, LOG_ERR is chosen.
16073 On Win32, the `LOG_*` levels are mapped according to this:
16075 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16076 LOG_WARNING => EVENTLOG_WARNING_TYPE
16077 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16081 * Made it possible to reconfigure with just the configuration
16082 argument "reconf" or "reconfigure". The command line arguments
16083 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16084 and are retrieved from there when reconfiguring.
16090 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16092 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16096 * The obj_dat.pl script was messing up the sorting of object
16097 names. The reason was that it compared the quoted version
16098 of strings as a result "OCSP" > "OCSP Signing" because
16099 " > SPACE. Changed script to store unquoted versions of
16100 names and add quotes on output. It was also omitting some
16101 names from the lookup table if they were given a default
16102 value (that is if SN is missing it is given the same
16103 value as LN and vice versa), these are now added on the
16104 grounds that if an object has a name we should be able to
16105 look it up. Finally added warning output when duplicate
16106 short or long names are found.
16110 * Changes needed for Tandem NSK.
16112 *Scott Uroff <scott@xypro.com>*
16114 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16115 RSA_padding_check_SSLv23(), special padding was never detected
16116 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16117 version rollback attacks was not effective.
16119 In s23_clnt.c, don't use special rollback-attack detection padding
16120 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16121 client; similarly, in s23_srvr.c, don't do the rollback check if
16122 SSL 2.0 is the only protocol enabled in the server.
16126 * Make it possible to get hexdumps of unprintable data with 'openssl
16127 asn1parse'. By implication, the functions ASN1_parse_dump() and
16128 BIO_dump_indent() are added.
16132 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16133 these print out strings and name structures based on various
16134 flags including RFC2253 support and proper handling of
16135 multibyte characters. Added options to the 'x509' utility
16136 to allow the various flags to be set.
16140 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16141 Also change the functions X509_cmp_current_time() and
16142 X509_gmtime_adj() work with an ASN1_TIME structure,
16143 this will enable certificates using GeneralizedTime in validity
16144 dates to be checked.
16148 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16149 negative public key encodings) on by default,
16150 NO_NEG_PUBKEY_BUG can be set to disable it.
16154 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16155 content octets. An i2c_ASN1_OBJECT is unnecessary because
16156 the encoding can be trivially obtained from the structure.
16160 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16161 not read locks (`CRYPTO_r_[un]lock`).
16165 * A first attempt at creating official support for shared
16166 libraries through configuration. I've kept it so the
16167 default is static libraries only, and the OpenSSL programs
16168 are always statically linked for now, but there are
16169 preparations for dynamic linking in place.
16170 This has been tested on Linux and Tru64.
16174 * Randomness polling function for Win9x, as described in:
16175 Peter Gutmann, Software Generation of Practically Strong
16180 * Fix so PRNG is seeded in req if using an already existing
16185 * New options to smime application. -inform and -outform
16186 allow alternative formats for the S/MIME message including
16187 PEM and DER. The -content option allows the content to be
16188 specified separately. This should allow things like Netscape
16189 form signing output easier to verify.
16193 * Fix the ASN1 encoding of tags using the 'long form'.
16197 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16198 STRING types. These convert content octets to and from the
16199 underlying type. The actual tag and length octets are
16200 already assumed to have been read in and checked. These
16201 are needed because all other string types have virtually
16202 identical handling apart from the tag. By having versions
16203 of the ASN1 functions that just operate on content octets
16204 IMPLICIT tagging can be handled properly. It also allows
16205 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16206 and ASN1_INTEGER are identical apart from the tag.
16210 * Change the handling of OID objects as follows:
16212 - New object identifiers are inserted in objects.txt, following
16213 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16214 - objects.pl is used to process obj_mac.num and create a new
16216 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16219 This is currently kind of a hack, and the perl code in objects.pl
16220 isn't very elegant, but it works as I intended. The simplest way
16221 to check that it worked correctly is to look in obj_dat.h and
16222 check the array nid_objs and make sure the objects haven't moved
16223 around (this is important!). Additions are OK, as well as
16224 consistent name changes.
16228 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16232 * Addition of the command line parameter '-rand file' to 'openssl req'.
16233 The given file adds to whatever has already been seeded into the
16234 random pool through the RANDFILE configuration file option or
16235 environment variable, or the default random state file.
16239 * mkstack.pl now sorts each macro group into lexical order.
16240 Previously the output order depended on the order the files
16241 appeared in the directory, resulting in needless rewriting
16246 * Patches to make OpenSSL compile under Win32 again. Mostly
16247 work arounds for the VC++ problem that it treats func() as
16248 func(void). Also stripped out the parts of mkdef.pl that
16249 added extra typesafe functions: these no longer exist.
16253 * Reorganisation of the stack code. The macros are now all
16254 collected in safestack.h . Each macro is defined in terms of
16255 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16256 DEBUG_SAFESTACK is now handled in terms of function casts,
16257 this has the advantage of retaining type safety without the
16258 use of additional functions. If DEBUG_SAFESTACK is not defined
16259 then the non typesafe macros are used instead. Also modified the
16260 mkstack.pl script to handle the new form. Needs testing to see
16261 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16262 the default if no major problems. Similar behaviour for ASN1_SET_OF
16263 and PKCS12_STACK_OF.
16267 * When some versions of IIS use the 'NET' form of private key the
16268 key derivation algorithm is different. Normally MD5(password) is
16269 used as a 128 bit RC4 key. In the modified case
16270 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16271 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16272 as the old Netscape_RSA functions except they have an additional
16273 'sgckey' parameter which uses the modified algorithm. Also added
16274 an -sgckey command line option to the rsa utility. Thanks to
16275 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16276 algorithm to openssl-dev.
16280 * The evp_local.h macros were using 'c.##kname' which resulted in
16281 invalid expansion on some systems (SCO 5.0.5 for example).
16282 Corrected to 'c.kname'.
16284 *Phillip Porch <root@theporch.com>*
16286 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16287 a STACK of email addresses from a certificate or request, these look
16288 in the subject name and the subject alternative name extensions and
16289 omit any duplicate addresses.
16293 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16294 This makes DSA verification about 2 % faster.
16298 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16299 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16300 plus overhead for 1024 bit moduli).
16301 This makes exponentiations about 0.5 % faster for 1024 bit
16302 exponents (as measured by "openssl speed rsa2048").
16306 * Rename memory handling macros to avoid conflicts with other
16308 Malloc => OPENSSL_malloc
16309 Malloc_locked => OPENSSL_malloc_locked
16310 Realloc => OPENSSL_realloc
16311 Free => OPENSSL_free
16315 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16316 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16320 * CygWin32 support.
16322 *John Jarvie <jjarvie@newsguy.com>*
16324 * The type-safe stack code has been rejigged. It is now only compiled
16325 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16326 by default all type-specific stack functions are "#define"d back to
16327 standard stack functions. This results in more streamlined output
16328 but retains the type-safety checking possibilities of the original
16333 * The STACK code has been cleaned up, and certain type declarations
16334 that didn't make a lot of sense have been brought in line. This has
16335 also involved a cleanup of sorts in safestack.h to more correctly
16336 map type-safe stack functions onto their plain stack counterparts.
16337 This work has also resulted in a variety of "const"ifications of
16338 lots of the code, especially `_cmp` operations which should normally
16339 be prototyped with "const" parameters anyway.
16343 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16344 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16345 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16346 where all of 'md' is used each time the PRNG is used, but 'state'
16347 is used only indexed by a cyclic counter. As entropy may not be
16348 well distributed from the beginning, 'md' is important as a
16349 chaining variable. However, the output function chains only half
16350 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16351 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16352 in all of 'state' being rewritten, with the new values depending
16353 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16357 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16358 the handshake is continued after ssl_verify_cert_chain();
16359 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16360 can lead to 'unexplainable' connection aborts later.
16362 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16364 * Major EVP API cipher revision.
16365 Add hooks for extra EVP features. This allows various cipher
16366 parameters to be set in the EVP interface. Support added for variable
16367 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16368 setting of RC2 and RC5 parameters.
16370 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16373 Remove lots of duplicated code from the EVP library. For example *every*
16374 cipher init() function handles the 'iv' in the same way according to the
16375 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16376 for CFB and OFB modes they zero ctx->num.
16378 New functionality allows removal of S/MIME code RC2 hack.
16380 Most of the routines have the same form and so can be declared in terms
16383 By shifting this to the top level EVP_CipherInit() it can be removed from
16384 all individual ciphers. If the cipher wants to handle IVs or keys
16385 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16388 Change lots of functions like EVP_EncryptUpdate() to now return a
16389 value: although software versions of the algorithms cannot fail
16390 any installed hardware versions can.
16394 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16395 this option is set, tolerate broken clients that send the negotiated
16396 protocol version number instead of the requested protocol version
16401 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16402 i.e. non-zero for export ciphersuites, zero otherwise.
16403 Previous versions had this flag inverted, inconsistent with
16404 rsa_tmp_cb (..._TMP_RSA_CB).
16406 *Bodo Moeller; problem reported by Amit Chopra*
16408 * Add missing DSA library text string. Work around for some IIS
16409 key files with invalid SEQUENCE encoding.
16413 * Add a document (doc/standards.txt) that list all kinds of standards
16414 and so on that are implemented in OpenSSL.
16418 * Enhance c_rehash script. Old version would mishandle certificates
16419 with the same subject name hash and wouldn't handle CRLs at all.
16420 Added -fingerprint option to crl utility, to support new c_rehash
16425 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16429 * Fix for SSL server purpose checking. Server checking was
16430 rejecting certificates which had extended key usage present
16431 but no ssl client purpose.
16433 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16435 * Make PKCS#12 code work with no password. The PKCS#12 spec
16436 is a little unclear about how a blank password is handled.
16437 Since the password in encoded as a BMPString with terminating
16438 double NULL a zero length password would end up as just the
16439 double NULL. However no password at all is different and is
16440 handled differently in the PKCS#12 key generation code. NS
16441 treats a blank password as zero length. MSIE treats it as no
16442 password on export: but it will try both on import. We now do
16443 the same: PKCS12_parse() tries zero length and no password if
16444 the password is set to "" or NULL (NULL is now a valid password:
16445 it wasn't before) as does the pkcs12 application.
16449 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16450 perror when PEM_read_bio_X509_REQ fails, the error message must
16451 be obtained from the error queue.
16455 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16456 it in ERR_remove_state if appropriate, and change ERR_get_state
16457 accordingly to avoid race conditions (this is necessary because
16458 thread_hash is no longer constant once set).
16462 * Bugfix for linux-elf makefile.one.
16466 * RSA_get_default_method() will now cause a default
16467 RSA_METHOD to be chosen if one doesn't exist already.
16468 Previously this was only set during a call to RSA_new()
16469 or RSA_new_method(NULL) meaning it was possible for
16470 RSA_get_default_method() to return NULL.
16474 * Added native name translation to the existing DSO code
16475 that will convert (if the flag to do so is set) filenames
16476 that are sufficiently small and have no path information
16477 into a canonical native form. Eg. "blah" converted to
16478 "libblah.so" or "blah.dll" etc.
16482 * New function ERR_error_string_n(e, buf, len) which is like
16483 ERR_error_string(e, buf), but writes at most 'len' bytes
16484 including the 0 terminator. For ERR_error_string_n, 'buf'
16487 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16489 * CONF library reworked to become more general. A new CONF
16490 configuration file reader "class" is implemented as well as a
16491 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16492 old `CONF_*` functions are still there, but are reimplemented to
16493 work in terms of the new functions. Also, a set of functions
16494 to handle the internal storage of the configuration data is
16495 provided to make it easier to write new configuration file
16496 reader "classes" (I can definitely see something reading a
16497 configuration file in XML format, for example), called `_CONF_*`,
16498 or "the configuration storage API"...
16500 The new configuration file reading functions are:
16502 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16503 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16505 NCONF_default, NCONF_WIN32
16507 NCONF_dump_fp, NCONF_dump_bio
16509 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16510 NCONF_new creates a new CONF object. This works in the same way
16511 as other interfaces in OpenSSL, like the BIO interface.
16512 `NCONF_dump_*` dump the internal storage of the configuration file,
16513 which is useful for debugging. All other functions take the same
16514 arguments as the old `CONF_*` functions with the exception of the
16515 first that must be a `CONF *` instead of a `LHASH *`.
16517 To make it easier to use the new classes with the old `CONF_*` functions,
16518 the function CONF_set_default_method is provided.
16522 * Add '-tls1' option to 'openssl ciphers', which was already
16523 mentioned in the documentation but had not been implemented.
16524 (This option is not yet really useful because even the additional
16525 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16529 * Initial DSO code added into libcrypto for letting OpenSSL (and
16530 OpenSSL-based applications) load shared libraries and bind to
16531 them in a portable way.
16533 *Geoff Thorpe, with contributions from Richard Levitte*
16535 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16537 * Make sure _lrotl and _lrotr are only used with MSVC.
16539 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16540 (the default implementation of RAND_status).
16542 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16543 to '-clrext' (= clear extensions), as intended and documented.
16544 *Bodo Moeller; inconsistency pointed out by Michael Attili
16545 <attili@amaxo.com>*
16547 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16548 was larger than the MD block size.
16550 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16552 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16553 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16554 using the passed key: if the passed key was a private key the result
16555 of X509_print(), for example, would be to print out all the private key
16560 * des_quad_cksum() byte order bug fix.
16561 *Ulf Möller, using the problem description in krb4-0.9.7, where
16562 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16564 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16567 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16569 * For easily testing in shell scripts whether some command
16570 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16571 returns with exit code 0 iff no command of the given name is available.
16572 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16573 the output goes to stdout and nothing is printed to stderr.
16574 Additional arguments are always ignored.
16576 Since for each cipher there is a command of the same name,
16577 the 'no-cipher' compilation switches can be tested this way.
16579 ('openssl no-XXX' is not able to detect pseudo-commands such
16580 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16584 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16588 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16589 is set; it will be thrown away anyway because each handshake creates
16591 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16592 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16593 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16594 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16598 * New s_client option -ign_eof: EOF at stdin is ignored, and
16599 'Q' and 'R' lose their special meanings (quit/renegotiate).
16600 This is part of what -quiet does; unlike -quiet, -ign_eof
16601 does not suppress any output.
16605 * Add compatibility options to the purpose and trust code. The
16606 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16607 accepts a certificate or CA, this was the previous behaviour,
16608 with all the associated security issues.
16610 X509_TRUST_COMPAT is the old trust behaviour: only and
16611 automatically trust self signed roots in certificate store. A
16612 new trust setting X509_TRUST_DEFAULT is used to specify that
16613 a purpose has no associated trust setting and it should instead
16614 use the value in the default purpose.
16618 * Fix the PKCS#8 DSA private key code so it decodes keys again
16619 and fix a memory leak.
16623 * In util/mkerr.pl (which implements 'make errors'), preserve
16624 reason strings from the previous version of the .c file, as
16625 the default to have only downcase letters (and digits) in
16626 automatically generated reasons codes is not always appropriate.
16630 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16631 using strerror. Previously, ERR_reason_error_string() returned
16632 library names as reason strings for SYSerr; but SYSerr is a special
16633 case where small numbers are errno values, not library numbers.
16637 * Add '-dsaparam' option to 'openssl dhparam' application. This
16638 converts DSA parameters into DH parameters. (When creating parameters,
16639 DSA_generate_parameters is used.)
16643 * Include 'length' (recommended exponent length) in C code generated
16644 by 'openssl dhparam -C'.
16648 * The second argument to set_label in perlasm was already being used
16649 so couldn't be used as a "file scope" flag. Moved to third argument
16654 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16655 instead of RAND_bytes for encryption IVs and salts.
16659 * Include RAND_status() into RAND_METHOD instead of implementing
16660 it only for md_rand.c Otherwise replacing the PRNG by calling
16661 RAND_set_rand_method would be impossible.
16665 * Don't let DSA_generate_key() enter an infinite loop if the random
16666 number generation fails.
16670 * New 'rand' application for creating pseudo-random output.
16674 * Added configuration support for Linux/IA64
16676 *Rolf Haberrecker <rolf@suse.de>*
16678 * Assembler module support for Mingw32.
16682 * Shared library support for HPUX (in shlib/).
16684 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16686 * Shared library support for Solaris gcc.
16688 *Lutz Behnke <behnke@trustcenter.de>*
16690 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16692 * PKCS7_encrypt() was adding text MIME headers twice because they
16693 were added manually and by SMIME_crlf_copy().
16697 * In bntest.c don't call BN_rand with zero bits argument.
16699 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16701 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16702 case was implemented. This caused BN_div_recp() to fail occasionally.
16706 * Add an optional second argument to the set_label() in the perl
16707 assembly language builder. If this argument exists and is set
16708 to 1 it signals that the assembler should use a symbol whose
16709 scope is the entire file, not just the current function. This
16710 is needed with MASM which uses the format label:: for this scope.
16712 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16714 * Change the ASN1 types so they are typedefs by default. Before
16715 almost all types were #define'd to ASN1_STRING which was causing
16716 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16721 * Change names of new functions to the new get1/get0 naming
16722 convention: After 'get1', the caller owns a reference count
16723 and has to call `..._free`; 'get0' returns a pointer to some
16724 data structure without incrementing reference counters.
16725 (Some of the existing 'get' functions increment a reference
16726 counter, some don't.)
16727 Similarly, 'set1' and 'add1' functions increase reference
16728 counters or duplicate objects.
16732 * Allow for the possibility of temp RSA key generation failure:
16733 the code used to assume it always worked and crashed on failure.
16737 * Fix potential buffer overrun problem in BIO_printf().
16738 *Ulf Möller, using public domain code by Patrick Powell; problem
16739 pointed out by David Sacerdote <das33@cornell.edu>*
16741 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16742 RAND_egd() and RAND_status(). In the command line application,
16743 the EGD socket can be specified like a seed file using RANDFILE
16748 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16749 Some CAs (e.g. Verisign) distribute certificates in this form.
16753 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16754 list to exclude them. This means that no special compilation option
16755 is needed to use anonymous DH: it just needs to be included in the
16760 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16761 EVP_MD_type. The old functionality is available in a new macro called
16762 EVP_MD_md(). Change code that uses it and update docs.
16766 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16767 where the `void *` argument is replaced by a function pointer argument.
16768 Previously `void *` was abused to point to functions, which works on
16769 many platforms, but is not correct. As these functions are usually
16770 called by macros defined in OpenSSL header files, most source code
16771 should work without changes.
16775 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16776 sections with information on -D... compiler switches used for
16777 compiling the library so that applications can see them. To enable
16778 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16779 must be defined. E.g.,
16780 #define OPENSSL_ALGORITHM_DEFINES
16781 #include <openssl/opensslconf.h>
16782 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16784 *Richard Levitte, Ulf and Bodo Möller*
16786 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16791 * Change the 'other' type in certificate aux info to a STACK_OF
16792 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16793 the required ASN1 format: arbitrary types determined by an OID.
16797 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16798 argument to 'req'. This is not because the function is newer or
16799 better than others it just uses the work 'NEW' in the certificate
16800 request header lines. Some software needs this.
16804 * Reorganise password command line arguments: now passwords can be
16805 obtained from various sources. Delete the PEM_cb function and make
16806 it the default behaviour: i.e. if the callback is NULL and the
16807 usrdata argument is not NULL interpret it as a null terminated pass
16808 phrase. If usrdata and the callback are NULL then the pass phrase
16809 is prompted for as usual.
16813 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16814 the support is automatically enabled. The resulting binaries will
16815 autodetect the card and use it if present.
16817 *Ben Laurie and Compaq Inc.*
16819 * Work around for Netscape hang bug. This sends certificate request
16820 and server done in one record. Since this is perfectly legal in the
16821 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16822 the bugs/SSLv3 entry for more info.
16826 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16830 * Add -rand argument to smime and pkcs12 applications and read/write
16835 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16839 * Add command line password options to the remaining applications.
16843 * Bug fix for BN_div_recp() for numerators with an even number of
16848 * More tests in bntest.c, and changed test_bn output.
16852 * ./config recognizes MacOS X now.
16856 * Bug fix for BN_div() when the first words of num and divisor are
16857 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16861 * Add support for various broken PKCS#8 formats, and command line
16862 options to produce them.
16866 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16867 get temporary BIGNUMs from a BN_CTX.
16871 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16876 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16877 include a #define from the old name to the new. The original intent
16878 was that statically linked binaries could for example just call
16879 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16880 link with digests. This never worked because SSLeay_add_all_digests()
16881 and SSLeay_add_all_ciphers() were in the same source file so calling
16882 one would link with the other. They are now in separate source files.
16886 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16890 * Use a less unusual form of the Miller-Rabin primality test (it used
16891 a binary algorithm for exponentiation integrated into the Miller-Rabin
16892 loop, our standard modexp algorithms are faster).
16896 * Support for the EBCDIC character set completed.
16898 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16900 * Source code cleanups: use const where appropriate, eliminate casts,
16901 use `void *` instead of `char *` in lhash.
16905 * Bugfix: ssl3_send_server_key_exchange was not restartable
16906 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16907 this the server could overwrite ephemeral keys that the client
16912 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16913 using 50 iterations of the Rabin-Miller test.
16915 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16916 iterations of the Rabin-Miller test as required by the appendix
16917 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16918 As BN_is_prime_fasttest includes trial division, DSA parameter
16919 generation becomes much faster.
16921 This implies a change for the callback functions in DSA_is_prime
16922 and DSA_generate_parameters: The callback function is called once
16923 for each positive witness in the Rabin-Miller test, not just
16924 occasionally in the inner loop; and the parameters to the
16925 callback function now provide an iteration count for the outer
16926 loop rather than for the current invocation of the inner loop.
16927 DSA_generate_parameters additionally can call the callback
16928 function with an 'iteration count' of -1, meaning that a
16929 candidate has passed the trial division test (when q is generated
16930 from an application-provided seed, trial division is skipped).
16934 * New function BN_is_prime_fasttest that optionally does trial
16935 division before starting the Rabin-Miller test and has
16936 an additional BN_CTX * argument (whereas BN_is_prime always
16937 has to allocate at least one BN_CTX).
16938 'callback(1, -1, cb_arg)' is called when a number has passed the
16939 trial division stage.
16943 * Fix for bug in CRL encoding. The validity dates weren't being handled
16948 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16952 * New function BN_pseudo_rand().
16956 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16957 bignum version of BN_from_montgomery() with the working code from
16958 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16963 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16964 made it impossible to use the same SSL_SESSION data structure in
16965 SSL2 clients in multiple threads.
16969 * The return value of RAND_load_file() no longer counts bytes obtained
16970 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16971 to seed the PRNG (previously an explicit byte count was required).
16973 *Ulf Möller, Bodo Möller*
16975 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16976 used `char *` instead of `void *` and had casts all over the place.
16980 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16984 * Retain source code compatibility for BN_prime_checks macro:
16985 BN_is_prime(..., BN_prime_checks, ...) now uses
16986 BN_prime_checks_for_size to determine the appropriate number of
16987 Rabin-Miller iterations.
16991 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16992 DH_CHECK_P_NOT_SAFE_PRIME.
16993 (Check if this is true? OpenPGP calls them "strong".)
16997 * Merge the functionality of "dh" and "gendh" programs into a new program
16998 "dhparam". The old programs are retained for now but will handle DH keys
16999 (instead of parameters) in future.
17003 * Make the ciphers, s_server and s_client programs check the return values
17004 when a new cipher list is set.
17008 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17009 ciphers. Before when the 56bit ciphers were enabled the sorting was
17012 The syntax for the cipher sorting has been extended to support sorting by
17013 cipher-strength (using the strength_bits hard coded in the tables).
17014 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17016 Fix a bug in the cipher-command parser: when supplying a cipher command
17017 string with an "undefined" symbol (neither command nor alphanumeric
17018 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17019 an error is flagged.
17021 Due to the strength-sorting extension, the code of the
17022 ssl_create_cipher_list() function was completely rearranged. I hope that
17023 the readability was also increased :-)
17025 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17027 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17028 for the first serial number and places 2 in the serial number file. This
17029 avoids problems when the root CA is created with serial number zero and
17030 the first user certificate has the same issuer name and serial number
17035 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17036 the new code. Add documentation for this stuff.
17040 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17041 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17042 structures and behave in an analogous way to the X509v3 functions:
17043 they shouldn't be called directly but wrapper functions should be used
17046 So we also now have some wrapper functions that call the X509at functions
17047 when passed certificate requests. (TO DO: similar things can be done with
17048 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17049 things. Some of these need some d2i or i2d and print functionality
17050 because they handle more complex structures.)
17054 * Add missing #ifndefs that caused missing symbols when building libssl
17055 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17056 NO_RSA in `ssl/s2*.c`.
17058 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17060 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17061 has a return value which indicates the quality of the random data
17062 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17063 error queue. New function RAND_pseudo_bytes() generates output that is
17064 guaranteed to be unique but not unpredictable. RAND_add is like
17065 RAND_seed, but takes an extra argument for an entropy estimate
17066 (RAND_seed always assumes full entropy).
17070 * Do more iterations of Rabin-Miller probable prime test (specifically,
17071 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17072 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17073 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17074 false-positive rate of at most 2^-80 for random input.
17078 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17082 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17083 in the 0.9.5 release), this returns the chain
17084 from an X509_CTX structure with a dup of the stack and all
17085 the X509 reference counts upped: so the stack will exist
17086 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17089 Also make SSL_SESSION_print() print out the verify return
17094 * Add manpage for the pkcs12 command. Also change the default
17095 behaviour so MAC iteration counts are used unless the new
17096 -nomaciter option is used. This improves file security and
17097 only older versions of MSIE (4.0 for example) need it.
17101 * Honor the no-xxx Configure options when creating .DEF files.
17105 * Add PKCS#10 attributes to field table: challengePassword,
17106 unstructuredName and unstructuredAddress. These are taken from
17107 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17108 international characters are used.
17110 More changes to X509_ATTRIBUTE code: allow the setting of types
17111 based on strings. Remove the 'loc' parameter when adding
17112 attributes because these will be a SET OF encoding which is sorted
17117 * Initial changes to the 'req' utility to allow request generation
17118 automation. This will allow an application to just generate a template
17119 file containing all the field values and have req construct the
17122 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17123 used all over the place including certificate requests and PKCS#7
17124 structures. They are currently handled manually where necessary with
17125 some primitive wrappers for PKCS#7. The new functions behave in a
17126 manner analogous to the X509 extension functions: they allow
17127 attributes to be looked up by NID and added.
17129 Later something similar to the X509V3 code would be desirable to
17130 automatically handle the encoding, decoding and printing of the
17131 more complex types. The string types like challengePassword can
17132 be handled by the string table functions.
17134 Also modified the multi byte string table handling. Now there is
17135 a 'global mask' which masks out certain types. The table itself
17136 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17137 is useful when for example there is only one permissible type
17138 (as in countryName) and using the mask might result in no valid
17143 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17144 SSL_get_peer_finished to allow applications to obtain the latest
17145 Finished messages sent to the peer or expected from the peer,
17146 respectively. (SSL_get_peer_finished is usually the Finished message
17147 actually received from the peer, otherwise the protocol will be aborted.)
17149 As the Finished message are message digests of the complete handshake
17150 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17151 be used for external authentication procedures when the authentication
17152 provided by SSL/TLS is not desired or is not enough.
17156 * Enhanced support for Alpha Linux is added. Now ./config checks if
17157 the host supports BWX extension and if Compaq C is present on the
17158 $PATH. Just exploiting of the BWX extension results in 20-30%
17159 performance kick for some algorithms, e.g. DES and RC4 to mention
17160 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17165 * Add support for MS "fast SGC". This is arguably a violation of the
17166 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17167 weak crypto and after checking the certificate is SGC a second one
17168 with strong crypto. MS SGC stops the first handshake after receiving
17169 the server certificate message and sends a second client hello. Since
17170 a server will typically do all the time consuming operations before
17171 expecting any further messages from the client (server key exchange
17172 is the most expensive) there is little difference between the two.
17174 To get OpenSSL to support MS SGC we have to permit a second client
17175 hello message after we have sent server done. In addition we have to
17176 reset the MAC if we do get this second client hello.
17180 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17181 if a DER encoded private key is RSA or DSA traditional format. Changed
17182 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17183 format DER encoded private key. Newer code should use PKCS#8 format which
17184 has the key type encoded in the ASN1 structure. Added DER private key
17185 support to pkcs8 application.
17189 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17190 ciphersuites has been selected (as required by the SSL 3/TLS 1
17191 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17192 is set, we interpret this as a request to violate the specification
17193 (the worst that can happen is a handshake failure, and 'correct'
17194 behaviour would result in a handshake failure anyway).
17198 * In SSL_CTX_add_session, take into account that there might be multiple
17199 SSL_SESSION structures with the same session ID (e.g. when two threads
17200 concurrently obtain them from an external cache).
17201 The internal cache can handle only one SSL_SESSION with a given ID,
17202 so if there's a conflict, we now throw out the old one to achieve
17207 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17208 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17209 some routines that use cipher OIDs: some ciphers do not have OIDs
17210 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17215 * Simplify the trust setting structure and code. Now we just have
17216 two sequences of OIDs for trusted and rejected settings. These will
17217 typically have values the same as the extended key usage extension
17218 and any application specific purposes.
17220 The trust checking code now has a default behaviour: it will just
17221 check for an object with the same NID as the passed id. Functions can
17222 be provided to override either the default behaviour or the behaviour
17223 for a given id. SSL client, server and email already have functions
17224 in place for compatibility: they check the NID and also return "trusted"
17225 if the certificate is self signed.
17229 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17230 traditional format into an EVP_PKEY structure.
17234 * Add a password callback function PEM_cb() which either prompts for
17235 a password if usr_data is NULL or otherwise assumes it is a null
17236 terminated password. Allow passwords to be passed on command line
17237 environment or config files in a few more utilities.
17241 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17242 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17243 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17244 Update documentation.
17248 * Support for ASN1 "NULL" type. This could be handled before by using
17249 ASN1_TYPE but there wasn't any function that would try to read a NULL
17250 and produce an error if it couldn't. For compatibility we also have
17251 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17252 don't allocate anything because they don't need to.
17256 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17259 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17261 * Rebuild of the memory allocation routines used by OpenSSL code and
17262 possibly others as well. The purpose is to make an interface that
17263 provide hooks so anyone can build a separate set of allocation and
17264 deallocation routines to be used by OpenSSL, for example memory
17265 pool implementations, or something else, which was previously hard
17266 since Malloc(), Realloc() and Free() were defined as macros having
17267 the values malloc, realloc and free, respectively (except for Win32
17268 compilations). The same is provided for memory debugging code.
17269 OpenSSL already comes with functionality to find memory leaks, but
17270 this gives people a chance to debug other memory problems.
17272 With these changes, a new set of functions and macros have appeared:
17274 CRYPTO_set_mem_debug_functions() [F]
17275 CRYPTO_get_mem_debug_functions() [F]
17276 CRYPTO_dbg_set_options() [F]
17277 CRYPTO_dbg_get_options() [F]
17278 CRYPTO_malloc_debug_init() [M]
17280 The memory debug functions are NULL by default, unless the library
17281 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17282 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17283 gives the standard debugging functions that come with OpenSSL) or
17284 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17285 provided by the library user) must be used. When the standard
17286 debugging functions are used, CRYPTO_dbg_set_options can be used to
17287 request additional information:
17288 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17289 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17291 Also, things like CRYPTO_set_mem_functions will always give the
17292 expected result (the new set of functions is used for allocation
17293 and deallocation) at all times, regardless of platform and compiler
17296 To finish it up, some functions that were never use in any other
17297 way than through macros have a new API and new semantic:
17299 CRYPTO_dbg_malloc()
17300 CRYPTO_dbg_realloc()
17303 All macros of value have retained their old syntax.
17305 *Richard Levitte and Bodo Moeller*
17307 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17308 ordering of SMIMECapabilities wasn't in "strength order" and there
17309 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17314 * Some ASN1 types with illegal zero length encoding (INTEGER,
17315 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17317 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17319 * Merge in my S/MIME library for OpenSSL. This provides a simple
17320 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17321 functionality to handle multipart/signed properly) and a utility
17322 called 'smime' to call all this stuff. This is based on code I
17323 originally wrote for Celo who have kindly allowed it to be
17324 included in OpenSSL.
17328 * Add variants des_set_key_checked and des_set_key_unchecked of
17329 des_set_key (aka des_key_sched). Global variable des_check_key
17330 decides which of these is called by des_set_key; this way
17331 des_check_key behaves as it always did, but applications and
17332 the library itself, which was buggy for des_check_key == 1,
17333 have a cleaner way to pick the version they need.
17337 * New function PKCS12_newpass() which changes the password of a
17342 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17343 dynamic mix. In both cases the ids can be used as an index into the
17344 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17345 functions so they accept a list of the field values and the
17346 application doesn't need to directly manipulate the X509_TRUST
17351 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17356 * Modify the way the V3 extension code looks up extensions. This now
17357 works in a similar way to the object code: we have some "standard"
17358 extensions in a static table which is searched with OBJ_bsearch()
17359 and the application can add dynamic ones if needed. The file
17360 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17361 updated whenever a new extension is added to the core code and kept
17362 in ext_nid order. There is a simple program 'tabtest.c' which checks
17363 this. New extensions are not added too often so this file can readily
17364 be maintained manually.
17366 There are two big advantages in doing things this way. The extensions
17367 can be looked up immediately and no longer need to be "added" using
17368 X509V3_add_standard_extensions(): this function now does nothing.
17369 Side note: I get *lots* of email saying the extension code doesn't
17370 work because people forget to call this function.
17371 Also no dynamic allocation is done unless new extensions are added:
17372 so if we don't add custom extensions there is no need to call
17373 X509V3_EXT_cleanup().
17377 * Modify enc utility's salting as follows: make salting the default. Add a
17378 magic header, so unsalted files fail gracefully instead of just decrypting
17379 to garbage. This is because not salting is a big security hole, so people
17380 should be discouraged from doing it.
17384 * Fixes and enhancements to the 'x509' utility. It allowed a message
17385 digest to be passed on the command line but it only used this
17386 parameter when signing a certificate. Modified so all relevant
17387 operations are affected by the digest parameter including the
17388 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17389 DSA key was used because it didn't fix the digest.
17393 * Initial certificate chain verify code. Currently tests the untrusted
17394 certificates for consistency with the verify purpose (which is set
17395 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17397 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17398 this is because it will reject chains with invalid extensions whereas
17399 every previous version of OpenSSL and SSLeay made no checks at all.
17401 Trust code: checks the root CA for the relevant trust settings. Trust
17402 settings have an initial value consistent with the verify purpose: e.g.
17403 if the verify purpose is for SSL client use it expects the CA to be
17404 trusted for SSL client use. However the default value can be changed to
17405 permit custom trust settings: one example of this would be to only trust
17406 certificates from a specific "secure" set of CAs.
17408 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17409 which should be used for version portability: especially since the
17410 verify structure is likely to change more often now.
17412 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17413 to set them. If not set then assume SSL clients will verify SSL servers
17416 Two new options to the verify program: -untrusted allows a set of
17417 untrusted certificates to be passed in and -purpose which sets the
17418 intended purpose of the certificate. If a purpose is set then the
17419 new chain verify code is used to check extension consistency.
17423 * Support for the authority information access extension.
17427 * Modify RSA and DSA PEM read routines to transparently handle
17428 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17429 public keys in a format compatible with certificate
17430 SubjectPublicKeyInfo structures. Unfortunately there were already
17431 functions called *_PublicKey_* which used various odd formats so
17432 these are retained for compatibility: however the DSA variants were
17433 never in a public release so they have been deleted. Changed dsa/rsa
17434 utilities to handle the new format: note no releases ever handled public
17435 keys so we should be OK.
17437 The primary motivation for this change is to avoid the same fiasco
17438 that dogs private keys: there are several incompatible private key
17439 formats some of which are standard and some OpenSSL specific and
17440 require various evil hacks to allow partial transparent handling and
17441 even then it doesn't work with DER formats. Given the option anything
17442 other than PKCS#8 should be dumped: but the other formats have to
17443 stay in the name of compatibility.
17445 With public keys and the benefit of hindsight one standard format
17446 is used which works with EVP_PKEY, RSA or DSA structures: though
17447 it clearly returns an error if you try to read the wrong kind of key.
17449 Added a -pubkey option to the 'x509' utility to output the public key.
17450 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17451 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17452 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17453 that do the same as the `EVP_PKEY_assign_*()` except they up the
17454 reference count of the added key (they don't "swallow" the
17459 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17460 CRLs would fail if the file contained no certificates or no CRLs:
17461 added a new function to read in both types and return the number
17462 read: this means that if none are read it will be an error. The
17463 DER versions of the certificate and CRL reader would always fail
17464 because it isn't possible to mix certificates and CRLs in DER format
17465 without choking one or the other routine. Changed this to just read
17466 a certificate: this is the best we can do. Also modified the code
17467 in `apps/verify.c` to take notice of return codes: it was previously
17468 attempting to read in certificates from NULL pointers and ignoring
17469 any errors: this is one reason why the cert and CRL reader seemed
17470 to work. It doesn't check return codes from the default certificate
17471 routines: these may well fail if the certificates aren't installed.
17475 * Code to support otherName option in GeneralName.
17479 * First update to verify code. Change the verify utility
17480 so it warns if it is passed a self signed certificate:
17481 for consistency with the normal behaviour. X509_verify
17482 has been modified to it will now verify a self signed
17483 certificate if *exactly* the same certificate appears
17484 in the store: it was previously impossible to trust a
17485 single self signed certificate. This means that:
17486 openssl verify ss.pem
17487 now gives a warning about a self signed certificate but
17488 openssl verify -CAfile ss.pem ss.pem
17493 * For servers, store verify_result in SSL_SESSION data structure
17494 (and add it to external session representation).
17495 This is needed when client certificate verifications fails,
17496 but an application-provided verification callback (set by
17497 SSL_CTX_set_cert_verify_callback) allows accepting the session
17498 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17499 but returns 1): When the session is reused, we have to set
17500 ssl->verify_result to the appropriate error code to avoid
17503 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17505 * Fix a bug in the new PKCS#7 code: it didn't consider the
17506 case in PKCS7_dataInit() where the signed PKCS7 structure
17507 didn't contain any existing data because it was being created.
17509 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17511 * Add a salt to the key derivation routines in enc.c. This
17512 forms the first 8 bytes of the encrypted file. Also add a
17513 -S option to allow a salt to be input on the command line.
17517 * New function X509_cmp(). Oddly enough there wasn't a function
17518 to compare two certificates. We do this by working out the SHA1
17519 hash and comparing that. X509_cmp() will be needed by the trust
17524 * SSL_get1_session() is like SSL_get_session(), but increments
17525 the reference count in the SSL_SESSION returned.
17527 *Geoff Thorpe <geoff@eu.c2.net>*
17529 * Fix for 'req': it was adding a null to request attributes.
17530 Also change the X509_LOOKUP and X509_INFO code to handle
17531 certificate auxiliary information.
17535 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17540 * Add the possibility to add extra information to the memory leak
17541 detecting output, to form tracebacks, showing from where each
17542 allocation was originated: CRYPTO_push_info("constant string") adds
17543 the string plus current file name and line number to a per-thread
17544 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17545 is like calling CYRPTO_pop_info() until the stack is empty.
17546 Also updated memory leak detection code to be multi-thread-safe.
17550 * Add options -text and -noout to pkcs7 utility and delete the
17551 encryption options which never did anything. Update docs.
17555 * Add options to some of the utilities to allow the pass phrase
17556 to be included on either the command line (not recommended on
17557 OSes like Unix) or read from the environment. Update the
17558 manpages and fix a few bugs.
17562 * Add a few manpages for some of the openssl commands.
17566 * Fix the -revoke option in ca. It was freeing up memory twice,
17567 leaking and not finding already revoked certificates.
17571 * Extensive changes to support certificate auxiliary information.
17572 This involves the use of X509_CERT_AUX structure and X509_AUX
17573 functions. An X509_AUX function such as PEM_read_X509_AUX()
17574 can still read in a certificate file in the usual way but it
17575 will also read in any additional "auxiliary information". By
17576 doing things this way a fair degree of compatibility can be
17577 retained: existing certificates can have this information added
17578 using the new 'x509' options.
17580 Current auxiliary information includes an "alias" and some trust
17581 settings. The trust settings will ultimately be used in enhanced
17582 certificate chain verification routines: currently a certificate
17583 can only be trusted if it is self signed and then it is trusted
17588 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17589 The problem was that one of the replacement routines had not been working
17590 since SSLeay releases. For now the offending routine has been replaced
17591 with non-optimised assembler. Even so, this now gives around 95%
17592 performance improvement for 1024 bit RSA signs.
17596 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17597 handling. Most clients have the effective key size in bits equal to
17598 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17599 A few however don't do this and instead use the size of the decrypted key
17600 to determine the RC2 key length and the AlgorithmIdentifier to determine
17601 the effective key length. In this case the effective key length can still
17602 be 40 bits but the key length can be 168 bits for example. This is fixed
17603 by manually forcing an RC2 key into the EVP_PKEY structure because the
17604 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17605 the key length and effective key length are equal.
17609 * Add a bunch of functions that should simplify the creation of
17610 X509_NAME structures. Now you should be able to do:
17611 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17612 and have it automatically work out the correct field type and fill in
17613 the structures. The more adventurous can try:
17614 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17615 and it will (hopefully) work out the correct multibyte encoding.
17619 * Change the 'req' utility to use the new field handling and multibyte
17620 copy routines. Before the DN field creation was handled in an ad hoc
17621 way in req, ca, and x509 which was rather broken and didn't support
17622 BMPStrings or UTF8Strings. Since some software doesn't implement
17623 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17624 using the dirstring_type option. See the new comment in the default
17625 openssl.cnf for more info.
17629 * Make crypto/rand/md_rand.c more robust:
17630 - Assure unique random numbers after fork().
17631 - Make sure that concurrent threads access the global counter and
17632 md serializably so that we never lose entropy in them
17633 or use exactly the same state in multiple threads.
17634 Access to the large state is not always serializable because
17635 the additional locking could be a performance killer, and
17636 md should be large enough anyway.
17640 * New file `apps/app_rand.c` with commonly needed functionality
17641 for handling the random seed file.
17643 Use the random seed file in some applications that previously did not:
17645 dsaparam -genkey (which also ignored its '-rand' option),
17648 x509 (when signing).
17649 Except on systems with /dev/urandom, it is crucial to have a random
17650 seed file at least for key creation, DSA signing, and for DH exchanges;
17651 for RSA signatures we could do without one.
17653 gendh and gendsa (unlike genrsa) used to read only the first byte
17654 of each file listed in the '-rand' option. The function as previously
17655 found in genrsa is now in app_rand.c and is used by all programs
17656 that support '-rand'.
17660 * In RAND_write_file, use mode 0600 for creating files;
17661 don't just chmod when it may be too late.
17665 * Report an error from X509_STORE_load_locations
17666 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17670 * New function ASN1_mbstring_copy() this copies a string in either
17671 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17672 into an ASN1_STRING type. A mask of permissible types is passed
17673 and it chooses the "minimal" type to use or an error if not type
17678 * Add function equivalents to the various macros in asn1.h. The old
17679 macros are retained with an `M_` prefix. Code inside the library can
17680 use the `M_` macros. External code (including the openssl utility)
17681 should *NOT* in order to be "shared library friendly".
17685 * Add various functions that can check a certificate's extensions
17686 to see if it usable for various purposes such as SSL client,
17687 server or S/MIME and CAs of these types. This is currently
17688 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17689 verification. Also added a -purpose flag to x509 utility to
17690 print out all the purposes.
17694 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17699 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17700 for, obtain and decode and extension and obtain its critical flag.
17701 This allows all the necessary extension code to be handled in a
17702 single function call.
17706 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17707 platforms. See crypto/rc4/rc4_enc.c for further details.
17711 * New -noout option to asn1parse. This causes no output to be produced
17712 its main use is when combined with -strparse and -out to extract data
17713 from a file (which may not be in ASN.1 format).
17717 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17718 when producing the local key id.
17720 *Richard Levitte <levitte@stacken.kth.se>*
17722 * New option -dhparam in s_server. This allows a DH parameter file to be
17723 stated explicitly. If it is not stated then it tries the first server
17724 certificate file. The previous behaviour hard coded the filename
17729 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17730 a public key to be input or output. For example:
17731 openssl rsa -in key.pem -pubout -out pubkey.pem
17732 Also added necessary DSA public key functions to handle this.
17736 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17737 in the message. This was handled by allowing
17738 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17740 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17742 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17743 to the end of the strings whereas this didn't. This would cause problems
17744 if strings read with d2i_ASN1_bytes() were later modified.
17746 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17748 * Fix for base64 decode bug. When a base64 bio reads only one line of
17749 data and it contains EOF it will end up returning an error. This is
17750 caused by input 46 bytes long. The cause is due to the way base64
17751 BIOs find the start of base64 encoded data. They do this by trying a
17752 trial decode on each line until they find one that works. When they
17753 do a flag is set and it starts again knowing it can pass all the
17754 data directly through the decoder. Unfortunately it doesn't reset
17755 the context it uses. This means that if EOF is reached an attempt
17756 is made to pass two EOFs through the context and this causes the
17757 resulting error. This can also cause other problems as well. As is
17758 usual with these problems it takes *ages* to find and the fix is
17759 trivial: move one line.
17761 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17763 * Ugly workaround to get s_client and s_server working under Windows. The
17764 old code wouldn't work because it needed to select() on sockets and the
17765 tty (for keypresses and to see if data could be written). Win32 only
17766 supports select() on sockets so we select() with a 1s timeout on the
17767 sockets and then see if any characters are waiting to be read, if none
17768 are present then we retry, we also assume we can always write data to
17769 the tty. This isn't nice because the code then blocks until we've
17770 received a complete line of data and it is effectively polling the
17771 keyboard at 1s intervals: however it's quite a bit better than not
17772 working at all :-) A dedicated Windows application might handle this
17773 with an event loop for example.
17777 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17778 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17779 will be called when RSA_sign() and RSA_verify() are used. This is useful
17780 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17781 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17782 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17783 This necessitated the support of an extra signature type NID_md5_sha1
17784 for SSL signatures and modifications to the SSL library to use it instead
17785 of calling RSA_public_decrypt() and RSA_private_encrypt().
17789 * Add new -verify -CAfile and -CApath options to the crl program, these
17790 will lookup a CRL issuers certificate and verify the signature in a
17791 similar way to the verify program. Tidy up the crl program so it
17792 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17793 less strict. It will now permit CRL extensions even if it is not
17794 a V2 CRL: this will allow it to tolerate some broken CRLs.
17798 * Initialize all non-automatic variables each time one of the openssl
17799 sub-programs is started (this is necessary as they may be started
17800 multiple times from the "OpenSSL>" prompt).
17802 *Lennart Bang, Bodo Moeller*
17804 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17805 removing all other RSA functionality (this is what NO_RSA does). This
17806 is so (for example) those in the US can disable those operations covered
17807 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17812 * Non-copying interface to BIO pairs.
17813 (still largely untested)
17817 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17818 ASCII string. This was handled independently in various places before.
17822 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17823 UTF8 strings a character at a time.
17827 * Use client_version from client hello to select the protocol
17828 (s23_srvr.c) and for RSA client key exchange verification
17829 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17833 * Add various utility functions to handle SPKACs, these were previously
17834 handled by poking round in the structure internals. Added new function
17835 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17836 print, verify and generate SPKACs. Based on an original idea from
17837 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17841 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17845 * Allow the config file extension section to be overwritten on the
17846 command line. Based on an original idea from Massimiliano Pala
17847 <madwolf@comune.modena.it>. The new option is called -extensions
17848 and can be applied to ca, req and x509. Also -reqexts to override
17849 the request extensions in req and -crlexts to override the crl extensions
17854 * Add new feature to the SPKAC handling in ca. Now you can include
17855 the same field multiple times by preceding it by "XXXX." for example:
17858 this is the same syntax as used in the req config file.
17862 * Allow certificate extensions to be added to certificate requests. These
17863 are specified in a 'req_extensions' option of the req section of the
17864 config file. They can be printed out with the -text option to req but
17865 are otherwise ignored at present.
17869 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17870 data read consists of only the final block it would not decrypted because
17871 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17872 A misplaced 'break' also meant the decrypted final block might not be
17873 copied until the next read.
17877 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17878 a few extra parameters to the DH structure: these will be useful if
17879 for example we want the value of 'q' or implement X9.42 DH.
17883 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17884 provides hooks that allow the default DSA functions or functions on a
17885 "per key" basis to be replaced. This allows hardware acceleration and
17886 hardware key storage to be handled without major modification to the
17887 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17888 associated functions.
17892 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17893 as "read only": it can't be written to and the buffer it points to will
17894 not be freed. Reading from a read only BIO is much more efficient than
17895 a normal memory BIO. This was added because there are several times when
17896 an area of memory needs to be read from a BIO. The previous method was
17897 to create a memory BIO and write the data to it, this results in two
17898 copies of the data and an O(n^2) reading algorithm. There is a new
17899 function BIO_new_mem_buf() which creates a read only memory BIO from
17900 an area of memory. Also modified the PKCS#7 routines to use read only
17905 * Bugfix: ssl23_get_client_hello did not work properly when called in
17906 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17907 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17908 but a retry condition occurred while trying to read the rest.
17912 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17913 NID_pkcs7_encrypted by default: this was wrong since this should almost
17914 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17915 the encrypted data type: this is a more sensible place to put it and it
17916 allows the PKCS#12 code to be tidied up that duplicated this
17921 * Changed obj_dat.pl script so it takes its input and output files on
17922 the command line. This should avoid shell escape redirection problems
17927 * Initial support for certificate extension requests, these are included
17928 in things like Xenroll certificate requests. Included functions to allow
17929 extensions to be obtained and added.
17933 * -crlf option to s_client and s_server for sending newlines as
17934 CRLF (as required by many protocols).
17938 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17940 * Install libRSAglue.a when OpenSSL is built with RSAref.
17942 *Ralf S. Engelschall*
17944 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17946 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17948 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17953 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17954 DH parameters/keys (q is lost during that conversion, but the resulting
17955 DH parameters contain its length).
17957 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17958 much faster than DH_generate_parameters (which creates parameters
17959 where `p = 2*q + 1`), and also the smaller q makes DH computations
17960 much more efficient (160-bit exponentiation instead of 1024-bit
17961 exponentiation); so this provides a convenient way to support DHE
17962 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17963 utter importance to use
17964 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17966 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17967 when such DH parameters are used, because otherwise small subgroup
17968 attacks may become possible!
17972 * Avoid memory leak in i2d_DHparams.
17976 * Allow the -k option to be used more than once in the enc program:
17977 this allows the same encrypted message to be read by multiple recipients.
17981 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17982 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17983 it will always use the numerical form of the OID, even if it has a short
17988 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17989 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17990 otherwise bn_mod_exp was called. In the case of hardware keys for example
17991 no private key components need be present and it might store extra data
17992 in the RSA structure, which cannot be accessed from bn_mod_exp.
17993 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17994 private key operations.
17998 * Added support for SPARC Linux.
18002 * pem_password_cb function type incompatibly changed from
18003 typedef int pem_password_cb(char *buf, int size, int rwflag);
18005 ....(char *buf, int size, int rwflag, void *userdata);
18006 so that applications can pass data to their callbacks:
18007 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18008 additional void * argument, which is just handed through whenever
18009 the password callback is called.
18011 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18013 New function SSL_CTX_set_default_passwd_cb_userdata.
18015 Compatibility note: As many C implementations push function arguments
18016 onto the stack in reverse order, the new library version is likely to
18017 interoperate with programs that have been compiled with the old
18018 pem_password_cb definition (PEM_whatever takes some data that
18019 happens to be on the stack as its last argument, and the callback
18020 just ignores this garbage); but there is no guarantee whatsoever that
18023 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18024 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18025 problems not only on Windows, but also on some Unix platforms.
18026 To avoid problematic command lines, these definitions are now in an
18027 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18028 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18032 * MIPS III/IV assembler module is reimplemented.
18036 * More DES library cleanups: remove references to srand/rand and
18037 delete an unused file.
18041 * Add support for the free Netwide assembler (NASM) under Win32,
18042 since not many people have MASM (ml) and it can be hard to obtain.
18043 This is currently experimental but it seems to work OK and pass all
18044 the tests. Check out INSTALL.W32 for info.
18048 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18049 without temporary keys kept an extra copy of the server key,
18050 and connections with temporary keys did not free everything in case
18055 * New function RSA_check_key and new openssl rsa option -check
18056 for verifying the consistency of RSA keys.
18058 *Ulf Moeller, Bodo Moeller*
18060 * Various changes to make Win32 compile work:
18061 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18062 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18063 comparison" warnings.
18064 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18068 * Add a debugging option to PKCS#5 v2 key generation function: when
18069 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18070 derived keys are printed to stderr.
18074 * Copy the flags in ASN1_STRING_dup().
18076 *Roman E. Pavlov <pre@mo.msk.ru>*
18078 * The x509 application mishandled signing requests containing DSA
18079 keys when the signing key was also DSA and the parameters didn't match.
18081 It was supposed to omit the parameters when they matched the signing key:
18082 the verifying software was then supposed to automatically use the CA's
18083 parameters if they were absent from the end user certificate.
18085 Omitting parameters is no longer recommended. The test was also
18086 the wrong way round! This was probably due to unusual behaviour in
18087 EVP_cmp_parameters() which returns 1 if the parameters match.
18088 This meant that parameters were omitted when they *didn't* match and
18089 the certificate was useless. Certificates signed with 'ca' didn't have
18092 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18094 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18095 The interface is as follows:
18096 Applications can use
18097 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18098 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18099 "off" is now the default.
18100 The library internally uses
18101 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18102 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18103 to disable memory-checking temporarily.
18105 Some inconsistent states that previously were possible (and were
18106 even the default) are now avoided.
18108 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18109 with each memory chunk allocated; this is occasionally more helpful
18110 than just having a counter.
18112 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18114 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18119 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18120 which largely parallels "options", but is for changing API behaviour,
18121 whereas "options" are about protocol behaviour.
18122 Initial "mode" flags are:
18124 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18125 a single record has been written.
18126 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18127 retries use the same buffer location.
18128 (But all of the contents must be
18133 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18136 * Fix problems with no-hmac etc.
18138 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18140 * New functions RSA_get_default_method(), RSA_set_method() and
18141 RSA_get_method(). These allows replacement of RSA_METHODs without having
18142 to mess around with the internals of an RSA structure.
18146 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18147 Also really enable memory leak checks in openssl.c and in some
18150 *Chad C. Mulligan, Bodo Moeller*
18152 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18153 up the length of negative integers. This has now been simplified to just
18154 store the length when it is first determined and use it later, rather
18155 than trying to keep track of where data is copied and updating it to
18157 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18159 * Add a new function PKCS7_signatureVerify. This allows the verification
18160 of a PKCS#7 signature but with the signing certificate passed to the
18161 function itself. This contrasts with PKCS7_dataVerify which assumes the
18162 certificate is present in the PKCS#7 structure. This isn't always the
18163 case: certificates can be omitted from a PKCS#7 structure and be
18164 distributed by "out of band" means (such as a certificate database).
18168 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18169 function prototypes in pem.h, also change util/mkdef.pl to add the
18170 necessary function names.
18174 * mk1mf.pl (used by Windows builds) did not properly read the
18175 options set by Configure in the top level Makefile, and Configure
18176 was not even able to write more than one option correctly.
18177 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18181 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18182 file to be loaded from a BIO or FILE pointer. The BIO version will
18183 for example allow memory BIOs to contain config info.
18187 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18188 Whoever hopes to achieve shared-library compatibility across versions
18189 must use this, not the compile-time macro.
18190 (Exercise 0.9.4: Which is the minimum library version required by
18192 Note: All this applies only to multi-threaded programs, others don't
18197 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18198 through a BIO pair triggered the default case, i.e.
18199 SSLerr(...,SSL_R_UNKNOWN_STATE).
18203 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18204 can use the SSL library even if none of the specific BIOs is
18209 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18210 for the encoded length.
18212 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18214 * Add initial documentation of the X509V3 functions.
18218 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18219 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18220 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18221 secure PKCS#8 private key format with a high iteration count.
18225 * Fix determination of Perl interpreter: A perl or perl5
18226 *directory* in $PATH was also accepted as the interpreter.
18228 *Ralf S. Engelschall*
18230 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18231 wrong with it but it was very old and did things like calling
18232 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18233 unusual formatting.
18237 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18238 to use the new extension code.
18242 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18243 with macros. This should make it easier to change their form, add extra
18244 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18249 * Add to configuration table a new entry that can specify an alternative
18250 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18251 according to Mark Crispin <MRC@Panda.COM>.
18255 * DES CBC did not update the IV. Weird.
18259 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18260 Changing the behaviour of the former might break existing programs --
18261 where IV updating is needed, des_ncbc_encrypt can be used.
18264 * When bntest is run from "make test" it drives bc to check its
18265 calculations, as well as internally checking them. If an internal check
18266 fails, it needs to cause bc to give a non-zero result or make test carries
18267 on without noticing the failure. Fixed.
18271 * DES library cleanups.
18275 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18276 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18277 ciphers. NOTE: although the key derivation function has been verified
18278 against some published test vectors it has not been extensively tested
18279 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18284 * Instead of "mkdir -p", which is not fully portable, use new
18285 Perl script "util/mkdir-p.pl".
18289 * Rewrite the way password based encryption (PBE) is handled. It used to
18290 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18291 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18292 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18293 the 'parameter' field of the AlgorithmIdentifier is passed to the
18294 underlying key generation function so it must do its own ASN1 parsing.
18295 This has also changed the EVP_PBE_CipherInit() function which now has a
18296 'parameter' argument instead of literal salt and iteration count values
18297 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18301 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18302 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18303 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18304 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18305 value was just used as a "magic string" and not used directly its
18306 value doesn't matter.
18310 * Introduce some semblance of const correctness to BN. Shame C doesn't
18315 * "linux-sparc64" configuration (ultrapenguin).
18317 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18318 "linux-sparc" configuration.
18320 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18322 * config now generates no-xxx options for missing ciphers.
18326 * Support the EBCDIC character set (work in progress).
18327 File ebcdic.c not yet included because it has a different license.
18329 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18331 * Support BS2000/OSD-POSIX.
18333 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18335 * Make callbacks for key generation use `void *` instead of `char *`.
18339 * Make S/MIME samples compile (not yet tested).
18343 * Additional typesafe stacks.
18347 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18351 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18353 * New configuration variant "sco5-gcc".
18355 * Updated some demos.
18357 *Sean O Riordain, Wade Scholine*
18359 * Add missing BIO_free at exit of pkcs12 application.
18363 * Fix memory leak in conf.c.
18367 * Updates for Win32 to assembler version of MD5.
18371 * Set #! path to perl in `apps/der_chop` to where we found it
18372 instead of using a fixed path.
18376 * SHA library changes for irix64-mips4-cc.
18380 * Improvements for VMS support.
18384 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18386 * Bignum library bug fix. IRIX 6 passes "make test" now!
18387 This also avoids the problems with SC4.2 and unpatched SC5.
18389 *Andy Polyakov <appro@fy.chalmers.se>*
18391 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18392 These are required because of the typesafe stack would otherwise break
18393 existing code. If old code used a structure member which used to be STACK
18394 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18395 sk_num or sk_value it would produce an error because the num, data members
18396 are not present in STACK_OF. Now it just produces a warning. sk_set
18397 replaces the old method of assigning a value to sk_value
18398 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18399 that does this will no longer work (and should use sk_set instead) but
18400 this could be regarded as a "questionable" behaviour anyway.
18404 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18405 correctly handle encrypted S/MIME data.
18409 * Change type of various DES function arguments from des_cblock
18410 (which means, in function argument declarations, pointer to char)
18411 to des_cblock * (meaning pointer to array with 8 char elements),
18412 which allows the compiler to do more typechecking; it was like
18413 that back in SSLeay, but with lots of ugly casts.
18415 Introduce new type const_des_cblock.
18419 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18420 problems: find RecipientInfo structure that matches recipient certificate
18421 and initialise the ASN1 structures properly based on passed cipher.
18425 * Belatedly make the BN tests actually check the results.
18429 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18430 to and from BNs: it was completely broken. New compilation option
18431 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18432 key elements as negative integers.
18436 * Reorganize and speed up MD5.
18438 *Andy Polyakov <appro@fy.chalmers.se>*
18442 *Richard Levitte <richard@levitte.org>*
18444 * New option -out to asn1parse to allow the parsed structure to be
18445 output to a file. This is most useful when combined with the -strparse
18446 option to examine the output of things like OCTET STRINGS.
18450 * Make SSL library a little more fool-proof by not requiring any longer
18451 that `SSL_set_{accept,connect}_state` be called before
18452 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18453 in many applications because usually everything *appeared* to work as
18454 intended anyway -- now it really works as intended).
18458 * Move openssl.cnf out of lib/.
18462 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18463 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18464 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18466 *Ralf S. Engelschall*
18468 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18469 handle PKCS#7 enveloped data properly.
18471 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18473 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18474 copying pointers. The cert_st handling is changed by this in
18475 various ways (and thus what used to be known as ctx->default_cert
18476 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18477 any longer when s->cert does not give us what we need).
18478 ssl_cert_instantiate becomes obsolete by this change.
18479 As soon as we've got the new code right (possibly it already is?),
18480 we have solved a couple of bugs of the earlier code where s->cert
18481 was used as if it could not have been shared with other SSL structures.
18483 Note that using the SSL API in certain dirty ways now will result
18484 in different behaviour than observed with earlier library versions:
18485 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18486 does not influence s as it used to.
18488 In order to clean up things more thoroughly, inside SSL_SESSION
18489 we don't use CERT any longer, but a new structure SESS_CERT
18490 that holds per-session data (if available); currently, this is
18491 the peer's certificate chain and, for clients, the server's certificate
18492 and temporary key. CERT holds only those values that can have
18493 meaningful defaults in an SSL_CTX.
18497 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18498 from the internal representation. Various PKCS#7 fixes: remove some
18499 evil casts and set the enc_dig_alg field properly based on the signing
18504 * Allow PKCS#12 password to be set from the command line or the
18505 environment. Let 'ca' get its config file name from the environment
18506 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18511 * Allow certificate policies extension to use an IA5STRING for the
18512 organization field. This is contrary to the PKIX definition but
18513 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18518 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18519 without disallowing inline assembler and the like for non-pedantic builds.
18523 * Support Borland C++ builder.
18525 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18531 * SHA-1 cleanups and performance enhancements.
18533 *Andy Polyakov <appro@fy.chalmers.se>*
18535 * Sparc v8plus assembler for the bignum library.
18537 *Andy Polyakov <appro@fy.chalmers.se>*
18539 * Accept any -xxx and +xxx compiler options in Configure.
18543 * Update HPUX configuration.
18547 * Add missing `sk_<type>_unshift()` function to safestack.h
18549 *Ralf S. Engelschall*
18551 * New function SSL_CTX_use_certificate_chain_file that sets the
18552 "extra_cert"s in addition to the certificate. (This makes sense
18553 only for "PEM" format files, as chains as a whole are not
18558 * Support verify_depth from the SSL API.
18559 x509_vfy.c had what can be considered an off-by-one-error:
18560 Its depth (which was not part of the external interface)
18561 was actually counting the number of certificates in a chain;
18562 now it really counts the depth.
18566 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18567 instead of X509err, which often resulted in confusing error
18568 messages since the error codes are not globally unique
18569 (e.g. an alleged error in ssl3_accept when a certificate
18570 didn't match the private key).
18572 * New function SSL_CTX_set_session_id_context that allows to set a default
18573 value (so that you don't need SSL_set_session_id_context for each
18574 connection using the SSL_CTX).
18578 * OAEP decoding bug fix.
18582 * Support INSTALL_PREFIX for package builders, as proposed by
18587 * New Configure options "threads" and "no-threads". For systems
18588 where the proper compiler options are known (currently Solaris
18589 and Linux), "threads" is the default.
18593 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18597 * Install various scripts to $(OPENSSLDIR)/misc, not to
18598 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18599 such as /usr/local/bin.
18603 * "make linux-shared" to build shared libraries.
18605 *Niels Poppe <niels@netbox.org>*
18607 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18611 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18612 extension adding in x509 utility.
18616 * Remove NOPROTO sections and error code comments.
18620 * Partial rewrite of the DEF file generator to now parse the ANSI
18625 * New Configure options --prefix=DIR and --openssldir=DIR.
18629 * Complete rewrite of the error code script(s). It is all now handled
18630 by one script at the top level which handles error code gathering,
18631 header rewriting and C source file generation. It should be much better
18632 than the old method: it now uses a modified version of Ulf's parser to
18633 read the ANSI prototypes in all header files (thus the old K&R definitions
18634 aren't needed for error creation any more) and do a better job of
18635 translating function codes into names. The old 'ASN1 error code embedded
18636 in a comment' is no longer necessary and it doesn't use .err files which
18637 have now been deleted. Also the error code call doesn't have to appear all
18638 on one line (which resulted in some large lines...).
18642 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18646 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18647 0 (which usually indicates a closed connection), but continue reading.
18651 * Fix some race conditions.
18655 * Add support for CRL distribution points extension. Add Certificate
18656 Policies and CRL distribution points documentation.
18660 * Move the autogenerated header file parts to crypto/opensslconf.h.
18664 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18665 8 of keying material. Merlin has also confirmed interop with this fix
18666 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18668 *Merlin Hughes <merlin@baltimore.ie>*
18670 * Fix lots of warnings.
18672 *Richard Levitte <levitte@stacken.kth.se>*
18674 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18675 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18677 *Richard Levitte <levitte@stacken.kth.se>*
18679 * Fix problems with sizeof(long) == 8.
18681 *Andy Polyakov <appro@fy.chalmers.se>*
18683 * Change functions to ANSI C.
18687 * Fix typos in error codes.
18689 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18691 * Remove defunct assembler files from Configure.
18695 * SPARC v8 assembler BIGNUM implementation.
18697 *Andy Polyakov <appro@fy.chalmers.se>*
18699 * Support for Certificate Policies extension: both print and set.
18700 Various additions to support the r2i method this uses.
18704 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18705 return a const string when you are expecting an allocated buffer.
18709 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18710 types DirectoryString and DisplayText.
18714 * Add code to allow r2i extensions to access the configuration database,
18715 add an LHASH database driver and add several ctx helper functions.
18719 * Fix an evil bug in bn_expand2() which caused various BN functions to
18720 fail when they extended the size of a BIGNUM.
18724 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18725 support typesafe stack.
18729 * Fix typo in SSL_[gs]et_options().
18731 *Nils Frostberg <nils@medcom.se>*
18733 * Delete various functions and files that belonged to the (now obsolete)
18734 old X509V3 handling code.
18738 * New Configure option "rsaref".
18742 * Don't auto-generate pem.h.
18746 * Introduce type-safe ASN.1 SETs.
18750 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18752 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18754 * Introduce type-safe STACKs. This will almost certainly break lots of code
18755 that links with OpenSSL (well at least cause lots of warnings), but fear
18756 not: the conversion is trivial, and it eliminates loads of evil casts. A
18757 few STACKed things have been converted already. Feel free to convert more.
18758 In the fullness of time, I'll do away with the STACK type altogether.
18762 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18763 specified in `<certfile>` by updating the entry in the index.txt file.
18764 This way one no longer has to edit the index.txt file manually for
18765 revoking a certificate. The -revoke option does the gory details now.
18767 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18769 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18770 `-text` option at all and this way the `-noout -text` combination was
18771 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18773 *Ralf S. Engelschall*
18775 * Make sure a corresponding plain text error message exists for the
18776 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18777 verify callback function determined that a certificate was revoked.
18779 *Ralf S. Engelschall*
18781 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18782 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18783 all available ciphers including rc5, which was forgotten until now.
18784 In order to let the testing shell script know which algorithms
18785 are available, a new (up to now undocumented) command
18786 `openssl list-cipher-commands` is used.
18790 * Bugfix: s_client occasionally would sleep in select() when
18791 it should have checked SSL_pending() first.
18795 * New functions DSA_do_sign and DSA_do_verify to provide access to
18796 the raw DSA values prior to ASN.1 encoding.
18800 * Tweaks to Configure
18802 *Niels Poppe <niels@netbox.org>*
18804 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18809 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18813 * New config option to avoid instructions that are illegal on the 80386.
18814 The default code is faster, but requires at least a 486.
18818 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18819 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18820 same as SSL2_VERSION anyway.
18824 * New "-showcerts" option for s_client.
18828 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18829 application. Various cleanups and fixes.
18833 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18834 modify error routines to work internally. Add error codes and PBE init
18835 to library startup routines.
18839 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18840 packing functions to asn1 and evp. Changed function names and error
18841 codes along the way.
18845 * PKCS12 integration: and so it begins... First of several patches to
18846 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18847 objects to objects.h
18851 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18852 and display support for Thawte strong extranet extension.
18856 * Add LinuxPPC support.
18858 *Jeff Dubrule <igor@pobox.org>*
18860 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18861 bn_div_words in alpha.s.
18863 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18865 * Make sure the RSA OAEP test is skipped under -DRSAref because
18866 OAEP isn't supported when OpenSSL is built with RSAref.
18868 *Ulf Moeller <ulf@fitug.de>*
18870 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18871 so they no longer are missing under -DNOPROTO.
18873 *Soren S. Jorvang <soren@t.dk>*
18875 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18877 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18878 doesn't work when the session is reused. Coming soon!
18882 * Fix a security hole, that allows sessions to be reused in the wrong
18883 context thus bypassing client cert protection! All software that uses
18884 client certs and session caches in multiple contexts NEEDS PATCHING to
18885 allow session reuse! A fuller solution is in the works.
18887 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18889 * Some more source tree cleanups (removed obsolete files
18890 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18891 permission on "config" script to be executable) and a fix for the INSTALL
18894 *Ulf Moeller <ulf@fitug.de>*
18896 * Remove some legacy and erroneous uses of malloc, free instead of
18899 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18901 * Make rsa_oaep_test return non-zero on error.
18903 *Ulf Moeller <ulf@fitug.de>*
18905 * Add support for native Solaris shared libraries. Configure
18906 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18907 if someone would make that last step automatic.
18909 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18911 * ctx_size was not built with the right compiler during "make links". Fixed.
18915 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18916 except NULL ciphers". This means the default cipher list will no longer
18917 enable NULL ciphers. They need to be specifically enabled e.g. with
18918 the string "DEFAULT:eNULL".
18922 * Fix to RSA private encryption routines: if p < q then it would
18923 occasionally produce an invalid result. This will only happen with
18924 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18928 * Be less restrictive and allow also `perl util/perlpath.pl
18929 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18930 because this way one can also use an interpreter named `perl5` (which is
18931 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18932 installed as `perl`).
18934 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18936 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18938 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18940 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18941 advapi32.lib to Win32 build and change the pem test comparison
18942 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18943 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18944 and crypto/des/ede_cbcm_enc.c.
18948 * DES quad checksum was broken on big-endian architectures. Fixed.
18952 * Comment out two functions in bio.h that aren't implemented. Fix up the
18953 Win32 test batch file so it (might) work again. The Win32 test batch file
18954 is horrible: I feel ill....
18958 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18959 in e_os.h. Audit of header files to check ANSI and non ANSI
18960 sections: 10 functions were absent from non ANSI section and not exported
18961 from Windows DLLs. Fixed up libeay.num for new functions.
18965 * Make `openssl version` output lines consistent.
18967 *Ralf S. Engelschall*
18969 * Fix Win32 symbol export lists for BIO functions: Added
18970 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18971 to ms/libeay{16,32}.def.
18973 *Ralf S. Engelschall*
18975 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18976 fine under Unix and passes some trivial tests I've now added. But the
18977 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18978 added to make sure no one expects that this stuff really works in the
18979 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18980 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18983 *Ralf S. Engelschall*
18985 * Fix the generation of two part addresses in perl.
18987 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18989 * Add config entry for Linux on MIPS.
18991 *John Tobey <jtobey@channel1.com>*
18993 * Make links whenever Configure is run, unless we are on Windoze.
18997 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18998 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19003 * Add a useful kludge to allow package maintainers to specify compiler and
19004 other platforms details on the command line without having to patch the
19005 Configure script every time: One now can use
19006 `perl Configure <id>:<details>`,
19007 i.e. platform ids are allowed to have details appended
19008 to them (separated by colons). This is treated as there would be a static
19009 pre-configured entry in Configure's %table under key `<id>` with value
19010 `<details>` and `perl Configure <id>` is called. So, when you want to
19011 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19012 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19013 now, which overrides the FreeBSD-elf entry on-the-fly.
19015 *Ralf S. Engelschall*
19017 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19021 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19022 on the `perl Configure ...` command line. This way one can compile
19023 OpenSSL libraries with Position Independent Code (PIC) which is needed
19024 for linking it into DSOs.
19026 *Ralf S. Engelschall*
19028 * Remarkably, export ciphers were totally broken and no-one had noticed!
19033 * Cleaned up the LICENSE document: The official contact for any license
19034 questions now is the OpenSSL core team under openssl-core@openssl.org.
19035 And add a paragraph about the dual-license situation to make sure people
19036 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19037 to the OpenSSL toolkit.
19039 *Ralf S. Engelschall*
19041 * General source tree makefile cleanups: Made `making xxx in yyy...`
19042 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19043 Additionally cleaned up the `make links` target: Remove unnecessary
19044 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19045 to speed processing and no longer clutter the display with confusing
19046 stuff. Instead only the actually done links are displayed.
19048 *Ralf S. Engelschall*
19050 * Permit null encryption ciphersuites, used for authentication only. It used
19051 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19052 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19057 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19058 signed attributes when verifying signatures (this would break them),
19059 the detached data encoding was wrong and public keys obtained using
19060 X509_get_pubkey() weren't freed.
19064 * Add text documentation for the BUFFER functions. Also added a work around
19065 to a Win95 console bug. This was triggered by the password read stuff: the
19066 last character typed gets carried over to the next fread(). If you were
19067 generating a new cert request using 'req' for example then the last
19068 character of the passphrase would be CR which would then enter the first
19073 * Added the new 'Includes OpenSSL Cryptography Software' button as
19074 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19075 button and can be used by applications based on OpenSSL to show the
19076 relationship to the OpenSSL project.
19078 *Ralf S. Engelschall*
19080 * Remove confusing variables in function signatures in files
19081 ssl/ssl_lib.c and ssl/ssl.h.
19083 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19085 * Don't install bss_file.c under PREFIX/include/
19087 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19089 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19090 functions that return function pointers and has support for NT specific
19091 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19092 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19093 unsigned to signed types: this was killing the Win32 compile.
19097 * Add new certificate file to stack functions,
19098 SSL_add_dir_cert_subjects_to_stack() and
19099 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19100 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19101 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19102 This means that Apache-SSL and similar packages don't have to mess around
19103 to add as many CAs as they want to the preferred list.
19107 * Experiment with doxygen documentation. Currently only partially applied to
19109 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19110 openssl.doxy as the configuration file.
19114 * Get rid of remaining C++-style comments which strict C compilers hate.
19116 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19118 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19119 compiled in by default: it has problems with large keys.
19123 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19124 DH private keys and/or callback functions which directly correspond to
19125 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19126 is needed for applications which have to configure certificates on a
19127 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19129 For the RSA certificate situation is makes no difference, but
19130 for the DSA certificate situation this fixes the "no shared cipher"
19131 problem where the OpenSSL cipher selection procedure failed because the
19132 temporary keys were not overtaken from the context and the API provided
19133 no way to reconfigure them.
19134 The new functions now let applications reconfigure the stuff and they
19135 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19136 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19137 non-public-API function ssl_cert_instantiate() is used as a helper
19138 function and also to reduce code redundancy inside ssl_rsa.c.
19140 *Ralf S. Engelschall*
19142 * Move s_server -dcert and -dkey options out of the undocumented feature
19143 area because they are useful for the DSA situation and should be
19144 recognized by the users.
19146 *Ralf S. Engelschall*
19148 * Fix the cipher decision scheme for export ciphers: the export bits are
19149 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19150 SSL_EXP_MASK. So, the original variable has to be used instead of the
19151 already masked variable.
19153 *Richard Levitte <levitte@stacken.kth.se>*
19155 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19157 *Richard Levitte <levitte@stacken.kth.se>*
19159 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19160 from `int` to `unsigned int` because it is a length and initialized by
19161 EVP_DigestFinal() which expects an `unsigned int *`.
19163 *Richard Levitte <levitte@stacken.kth.se>*
19165 * Don't hard-code path to Perl interpreter on shebang line of Configure
19166 script. Instead use the usual Shell->Perl transition trick.
19168 *Ralf S. Engelschall*
19170 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19171 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19172 -noout -modulus` as it's already the case for `openssl rsa -noout
19173 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19174 currently the public key is printed (a decision which was already done by
19175 `openssl dsa -modulus` in the past) which serves a similar purpose.
19176 Additionally the NO_RSA no longer completely removes the whole -modulus
19177 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19180 *Ralf S. Engelschall*
19182 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19183 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19185 *Arne Ansper <arne@ats.cyber.ee>*
19187 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19188 to be added. Now both 'req' and 'ca' can use new objects defined in the
19193 * Add cool BIO that does syslog (or event log on NT).
19195 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19197 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19198 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19199 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19200 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19204 * Add preliminary config info for new extension code.
19208 * Make RSA_NO_PADDING really use no padding.
19210 *Ulf Moeller <ulf@fitug.de>*
19212 * Generate errors when private/public key check is done.
19216 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19217 for some CRL extensions and new objects added.
19221 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19222 key usage extension and fuller support for authority key id.
19226 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19227 padding method for RSA, which is recommended for new applications in PKCS
19228 #1 v2.0 (RFC 2437, October 1998).
19229 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19230 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19231 against Bleichbacher's attack on RSA.
19232 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19235 * Updates to the new SSL compression code
19237 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19239 * Fix so that the version number in the master secret, when passed
19240 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19241 (because the server will not accept higher), that the version number
19242 is 0x03,0x01, not 0x03,0x00
19244 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19246 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19247 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19248 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19252 * Support for RAW extensions where an arbitrary extension can be
19253 created by including its DER encoding. See `apps/openssl.cnf` for
19258 * Make sure latest Perl versions don't interpret some generated C array
19259 code as Perl array code in the crypto/err/err_genc.pl script.
19261 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19263 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19264 not many people have the assembler. Various Win32 compilation fixes and
19265 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19266 build instructions.
19270 * Modify configure script 'Configure' to automatically create crypto/date.h
19271 file under Win32 and also build pem.h from pem.org. New script
19272 util/mkfiles.pl to create the MINFO file on environments that can't do a
19273 'make files': perl util/mkfiles.pl >MINFO should work.
19277 * Major rework of DES function declarations, in the pursuit of correctness
19278 and purity. As a result, many evil casts evaporated, and some weirdness,
19279 too. You may find this causes warnings in your code. Zapping your evil
19280 casts will probably fix them. Mostly.
19284 * Fix for a typo in asn1.h. Bug fix to object creation script
19285 obj_dat.pl. It considered a zero in an object definition to mean
19286 "end of object": none of the objects in objects.h have any zeros
19287 so it wasn't spotted.
19289 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19291 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19292 Masking (CBCM). In the absence of test vectors, the best I have been able
19293 to do is check that the decrypt undoes the encrypt, so far. Send me test
19294 vectors if you have them.
19298 * Correct calculation of key length for export ciphers (too much space was
19299 allocated for null ciphers). This has not been tested!
19303 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19304 message is now correct (it understands "crypto" and "ssl" on its
19305 command line). There is also now an "update" option. This will update
19306 the util/ssleay.num and util/libeay.num files with any new functions.
19308 perl util/mkdef.pl crypto ssl update
19309 it will update them.
19313 * Overhauled the Perl interface:
19314 - ported BN stuff to OpenSSL's different BN library
19315 - made the perl/ source tree CVS-aware
19316 - renamed the package from SSLeay to OpenSSL (the files still contain
19317 their history because I've copied them in the repository)
19318 - removed obsolete files (the test scripts will be replaced
19319 by better Test::Harness variants in the future)
19321 *Ralf S. Engelschall*
19323 * First cut for a very conservative source tree cleanup:
19324 1. merge various obsolete readme texts into doc/ssleay.txt
19325 where we collect the old documents and readme texts.
19326 2. remove the first part of files where I'm already sure that we no
19327 longer need them because of three reasons: either they are just temporary
19328 files which were left by Eric or they are preserved original files where
19329 I've verified that the diff is also available in the CVS via "cvs diff
19330 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19331 the crypto/md/ stuff).
19333 *Ralf S. Engelschall*
19335 * More extension code. Incomplete support for subject and issuer alt
19336 name, issuer and authority key id. Change the i2v function parameters
19337 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19338 what that's for :-) Fix to ASN1 macro which messed up
19339 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19343 * Preliminary support for ENUMERATED type. This is largely copied from the
19348 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19350 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19352 * Make sure `make rehash` target really finds the `openssl` program.
19354 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19356 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19357 like to hear about it if this slows down other processors.
19361 * Add CygWin32 platform information to Configure script.
19363 *Alan Batie <batie@aahz.jf.intel.com>*
19365 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19367 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19369 * New program nseq to manipulate netscape certificate sequences
19373 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19378 * Fixes to BN code. Previously the default was to define BN_RECURSION
19379 but the BN code had some problems that would cause failures when
19380 doing certificate verification and some other functions.
19382 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19384 * Add ASN1 and PEM code to support netscape certificate sequences.
19388 * Add ASN1 and PEM code to support netscape certificate sequences.
19392 * Add several PKIX and private extended key usage OIDs.
19396 * Modify the 'ca' program to handle the new extension code. Modify
19397 openssl.cnf for new extension format, add comments.
19401 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19402 and add a sample to openssl.cnf so req -x509 now adds appropriate
19407 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19408 error code, add initial support to X509_print() and x509 application.
19412 * Takes a deep breath and start adding X509 V3 extension support code. Add
19413 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19414 stuff is currently isolated and isn't even compiled yet.
19418 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19419 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19420 Removed the versions check from X509 routines when loading extensions:
19421 this allows certain broken certificates that don't set the version
19422 properly to be processed.
19426 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19427 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19428 can still be regenerated with "make depend".
19432 * Spelling mistake in C version of CAST-128.
19434 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19436 * Changes to the error generation code. The perl script err-code.pl
19437 now reads in the old error codes and retains the old numbers, only
19438 adding new ones if necessary. It also only changes the .err files if new
19439 codes are added. The makefiles have been modified to only insert errors
19440 when needed (to avoid needlessly modifying header files). This is done
19441 by only inserting errors if the .err file is newer than the auto generated
19442 C file. To rebuild all the error codes from scratch (the old behaviour)
19443 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19444 or delete all the .err files.
19448 * CAST-128 was incorrectly implemented for short keys. The C version has
19449 been fixed, but is untested. The assembler versions are also fixed, but
19450 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19451 to regenerate it if needed.
19452 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19453 Hagino <itojun@kame.net>*
19455 * File was opened incorrectly in randfile.c.
19457 *Ulf Möller <ulf@fitug.de>*
19459 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19460 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19461 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19462 al: it's just almost always a UTCTime. Note this patch adds new error
19463 codes so do a "make errors" if there are problems.
19467 * Correct Linux 1 recognition in config.
19469 *Ulf Möller <ulf@fitug.de>*
19471 * Remove pointless MD5 hash when using DSA keys in ca.
19473 *Anonymous <nobody@replay.com>*
19475 * Generate an error if given an empty string as a cert directory. Also
19476 generate an error if handed NULL (previously returned 0 to indicate an
19477 error, but didn't set one).
19479 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19481 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19485 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19486 parameters. This was causing a warning which killed off the Win32 compile.
19490 * Remove C++ style comments from crypto/bn/bn_local.h.
19492 *Neil Costigan <neil.costigan@celocom.com>*
19494 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19495 based on a text string, looking up short and long names and finally
19496 "dot" format. The "dot" format stuff didn't work. Added new function
19497 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19498 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19499 OID is not part of the table.
19503 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19504 X509_LOOKUP_by_alias().
19508 * Sort openssl functions by name.
19512 * Get the `gendsa` command working and add it to the `list` command. Remove
19513 encryption from sample DSA keys (in case anyone is interested the password
19518 * Make *all* `*_free` functions accept a NULL pointer.
19520 *Frans Heymans <fheymans@isaserver.be>*
19522 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19525 *Anonymous <nobody@replay.com>*
19527 * s_server should send the CAfile as acceptable CAs, not its own cert.
19529 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19531 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19533 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19535 * Temp key "for export" tests were wrong in s3_srvr.c.
19537 *Anonymous <nobody@replay.com>*
19539 * Add prototype for temp key callback functions
19540 SSL_CTX_set_tmp_{rsa,dh}_callback().
19544 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19545 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19549 * X509_name_add_entry() freed the wrong thing after an error.
19551 *Arne Ansper <arne@ats.cyber.ee>*
19553 * rsa_eay.c would attempt to free a NULL context.
19555 *Arne Ansper <arne@ats.cyber.ee>*
19557 * BIO_s_socket() had a broken should_retry() on Windoze.
19559 *Arne Ansper <arne@ats.cyber.ee>*
19561 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19563 *Arne Ansper <arne@ats.cyber.ee>*
19565 * Make sure the already existing X509_STORE->depth variable is initialized
19566 in X509_STORE_new(), but document the fact that this variable is still
19567 unused in the certificate verification process.
19569 *Ralf S. Engelschall*
19571 * Fix the various library and `apps/` files to free up pkeys obtained from
19572 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19576 * Fix reference counting in X509_PUBKEY_get(). This makes
19577 demos/maurice/example2.c work, amongst others, probably.
19579 *Steve Henson and Ben Laurie*
19581 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19582 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19583 are no longer created. This way we have a single and consistent command
19584 line interface `openssl <command>`, similar to `cvs <command>`.
19586 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19588 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19589 BIT STRING wrapper always have zero unused bits.
19593 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19597 * Make the top-level INSTALL documentation easier to understand.
19601 * Makefiles updated to exit if an error occurs in a sub-directory
19602 make (including if user presses ^C) [Paul Sutton]
19604 * Make Montgomery context stuff explicit in RSA data structure.
19608 * Fix build order of pem and err to allow for generated pem.h.
19612 * Fix renumbering bug in X509_NAME_delete_entry().
19616 * Enhanced the err-ins.pl script so it makes the error library number
19617 global and can add a library name. This is needed for external ASN1 and
19618 other error libraries.
19622 * Fixed sk_insert which never worked properly.
19626 * Fix ASN1 macros so they can handle indefinite length constructed
19627 EXPLICIT tags. Some non standard certificates use these: they can now
19632 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19633 into a single doc/ssleay.txt bundle. This way the information is still
19634 preserved but no longer messes up this directory. Now it's new room for
19635 the new set of documentation files.
19637 *Ralf S. Engelschall*
19639 * SETs were incorrectly DER encoded. This was a major pain, because they
19640 shared code with SEQUENCEs, which aren't coded the same. This means that
19641 almost everything to do with SETs or SEQUENCEs has either changed name or
19642 number of arguments.
19644 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19646 * Fix test data to work with the above.
19650 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19651 was already fixed by Eric for 0.9.1 it seems.
19653 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19655 * Autodetect FreeBSD3.
19659 * Fix various bugs in Configure. This affects the following platforms:
19663 unixware-2.0-pentium
19668 * Eliminate generated files from CVS. Reorder tests to regenerate files
19669 before they are needed.
19673 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19677 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19679 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19680 changed SSLeay to OpenSSL in version strings.
19682 *Ralf S. Engelschall*
19684 * Some fixups to the top-level documents.
19688 * Fixed the nasty bug where rsaref.h was not found under compile-time
19689 because the symlink to include/ was missing.
19691 *Ralf S. Engelschall*
19693 * Incorporated the popular no-RSA/DSA-only patches
19694 which allow to compile an RSA-free SSLeay.
19696 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19698 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19699 when "ssleay" is still not found.
19701 *Ralf S. Engelschall*
19703 * Added more platforms to Configure: Cray T3E, HPUX 11,
19705 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19707 * Updated the README file.
19709 *Ralf S. Engelschall*
19711 * Added various .cvsignore files in the CVS repository subdirs
19712 to make a "cvs update" really silent.
19714 *Ralf S. Engelschall*
19716 * Recompiled the error-definition header files and added
19717 missing symbols to the Win32 linker tables.
19719 *Ralf S. Engelschall*
19721 * Cleaned up the top-level documents;
19722 o new files: CHANGES and LICENSE
19723 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19724 o merged COPYRIGHT into LICENSE
19725 o removed obsolete TODO file
19726 o renamed MICROSOFT to INSTALL.W32
19728 *Ralf S. Engelschall*
19730 * Removed dummy files from the 0.9.1b source tree:
19731 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19732 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19733 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19734 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19735 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19737 *Ralf S. Engelschall*
19739 * Added various platform portability fixes.
19743 * The Genesis of the OpenSSL rpject:
19744 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19745 Young and Tim J. Hudson created while they were working for C2Net until
19748 *The OpenSSL Project*
19750 ### Changes between 0.9.0b and 0.9.1b [not released]
19752 * Updated a few CA certificates under certs/
19756 * Changed some BIGNUM api stuff.
19760 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19761 DGUX x86, Linux Alpha, etc.
19765 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19766 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19771 * Add -strparse option to asn1pars program which parses nested
19774 *Dr Stephen Henson <shenson@bigfoot.com>*
19776 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19780 * DSA fix for "ca" program.
19784 * Added "-genkey" option to "dsaparam" program.
19788 * Added RIPE MD160 (rmd160) message digest.
19792 * Added -a (all) option to "ssleay version" command.
19796 * Added PLATFORM define which is the id given to Configure.
19800 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19804 * Extended the ASN.1 parser routines.
19808 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19812 * Added a BN_CTX to the BN library.
19816 * Fixed the weak key values in DES library
19820 * Changed API in EVP library for cipher aliases.
19824 * Added support for RC2/64bit cipher.
19828 * Converted the lhash library to the crypto/mem.c functions.
19832 * Added more recognized ASN.1 object ids.
19836 * Added more RSA padding checks for SSL/TLS.
19840 * Added BIO proxy/filter functionality.
19844 * Added extra_certs to SSL_CTX which can be used
19845 send extra CA certificates to the client in the CA cert chain sending
19846 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19850 * Now Fortezza is denied in the authentication phase because
19851 this is key exchange mechanism is not supported by SSLeay at all.
19855 * Additional PKCS1 checks.
19859 * Support the string "TLSv1" for all TLS v1 ciphers.
19863 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19864 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19868 * Fixed a few memory leaks.
19872 * Fixed various code and comment typos.
19876 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19877 bytes sent in the client random.
19879 *Edward Bishop <ebishop@spyglass.com>*
19883 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
19884 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
19885 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
19886 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19887 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19888 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19889 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19890 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19891 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19892 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19893 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19894 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19895 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19896 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19897 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19898 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19899 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19900 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19901 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19902 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19903 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19904 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19905 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19906 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19907 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19908 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19909 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19910 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19911 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19912 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19913 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19914 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19915 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19916 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19917 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19918 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19919 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19920 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19921 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19922 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19923 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19924 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19925 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19926 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19927 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19928 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19929 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19930 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19931 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19932 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19933 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19934 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19935 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19936 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19937 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19938 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19939 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19940 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19941 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19942 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19943 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19944 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19945 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19946 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19947 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19948 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19949 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19950 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19951 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19952 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19953 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19954 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19955 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19956 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19957 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19958 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19959 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19960 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19961 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19962 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19963 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19964 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19965 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19966 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19967 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19968 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19969 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19970 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19971 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19972 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19973 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19974 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19975 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19976 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19977 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19978 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19979 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19980 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19981 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19982 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19983 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19984 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19985 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19986 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19987 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19988 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19989 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19990 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19991 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19992 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19993 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19994 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19995 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19996 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19997 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19998 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19999 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20000 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20001 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20002 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20003 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20004 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20005 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20006 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20007 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20008 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20009 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20010 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20011 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20012 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20013 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20014 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20015 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20016 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20017 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20018 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20019 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20020 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20021 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20022 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20023 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20024 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20025 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20026 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20027 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20028 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20029 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20030 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20031 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20032 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20033 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20034 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20035 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20036 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20037 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20038 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20039 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20040 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20041 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20042 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20043 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20044 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20045 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20046 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20047 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20048 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20049 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20050 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20051 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20052 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20053 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20054 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20055 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20056 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20057 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20058 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20059 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20060 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20061 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20062 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20063 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20064 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20065 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20066 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20067 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655